<%BANNER%>

Improving Antiretroviral Medication Adherence among Adolescents with Human Immunodeficiency Virus (HIV)

Permanent Link: http://ufdc.ufl.edu/UFE0024717/00001

Material Information

Title: Improving Antiretroviral Medication Adherence among Adolescents with Human Immunodeficiency Virus (HIV) A Case Series Pilot Intervention Study
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Gray, Wendy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Maintaining an adherence level of 95% or more is critical in optimizing treatment effectiveness, suppressing viral load, and preventing medication resistance in patients with HIV. Although adolescents are least likely to be adherent, they have been largely ignored by the intervention literature. The current study is the first-known prospective intervention designed to target adherence to antiretroviral medications among adolescents with HIV. Participants were four perinatally-infected adolescents who received a seven-week intervention focused on problem solving and family communication to improve adherence to their HIV regimen. The intervention was delivered through alternating home and telephone sessions and a multi-method adherence assessment approach was used to track adherence at baseline, during treatment, and at three-month follow-up. Viral load was obtained at all assessment points. All adolescents experienced an improvement in adherence levels following implementation of problem solving but, for the most part, these improvements were not maintained over time. Changes in the adolescent?s routine were associated with declines in adherence. Seventy-five percent of the sample experienced a decrease in viral load by the end of treatment. Medication resistance and immune suppression affected the degree to which improvements in adherence affected viral load in one participant. All participants reported a decrease in the severity of barriers experienced from pre- to post-treatment. Participant knowledge of their medication regimen was high at pre-treatment and did not change over time. Contrary to the initial hypothesis, parent perception of family conflict increased from pre- to post-treatment for some dyads. Adopting an individualized approach to help adolescents problem solve and overcome barriers may be promising strategy to improve their adherence. Interventions focused on improving adolescent adherence may benefit from helping the adolescent find ways to incorporate their regimen into their regular routine as opposed to encouraging the adolescent to bend their routine to the demands of their regimen. Adopting a proactive problem solving approach may help minimize the extent to which changes in routine result in declines in adherence. Further research, characterized by randomization to treatment, multi-site collaboration, and long-term monitoring of adherence are greatly needed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wendy Gray.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Janicke, David M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0024717:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024717/00001

Material Information

Title: Improving Antiretroviral Medication Adherence among Adolescents with Human Immunodeficiency Virus (HIV) A Case Series Pilot Intervention Study
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Gray, Wendy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Maintaining an adherence level of 95% or more is critical in optimizing treatment effectiveness, suppressing viral load, and preventing medication resistance in patients with HIV. Although adolescents are least likely to be adherent, they have been largely ignored by the intervention literature. The current study is the first-known prospective intervention designed to target adherence to antiretroviral medications among adolescents with HIV. Participants were four perinatally-infected adolescents who received a seven-week intervention focused on problem solving and family communication to improve adherence to their HIV regimen. The intervention was delivered through alternating home and telephone sessions and a multi-method adherence assessment approach was used to track adherence at baseline, during treatment, and at three-month follow-up. Viral load was obtained at all assessment points. All adolescents experienced an improvement in adherence levels following implementation of problem solving but, for the most part, these improvements were not maintained over time. Changes in the adolescent?s routine were associated with declines in adherence. Seventy-five percent of the sample experienced a decrease in viral load by the end of treatment. Medication resistance and immune suppression affected the degree to which improvements in adherence affected viral load in one participant. All participants reported a decrease in the severity of barriers experienced from pre- to post-treatment. Participant knowledge of their medication regimen was high at pre-treatment and did not change over time. Contrary to the initial hypothesis, parent perception of family conflict increased from pre- to post-treatment for some dyads. Adopting an individualized approach to help adolescents problem solve and overcome barriers may be promising strategy to improve their adherence. Interventions focused on improving adolescent adherence may benefit from helping the adolescent find ways to incorporate their regimen into their regular routine as opposed to encouraging the adolescent to bend their routine to the demands of their regimen. Adopting a proactive problem solving approach may help minimize the extent to which changes in routine result in declines in adherence. Further research, characterized by randomization to treatment, multi-site collaboration, and long-term monitoring of adherence are greatly needed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Wendy Gray.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Janicke, David M.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0024717:00001


This item has the following downloads:


Full Text

PAGE 1

IMPROVING ANTIR ETROVIRAL MEDICA TION ADHERENCE AMONG ADOLESCENTS WITH HUMAN IMMUNODEFICIENCY VIRU S (HIV): A CASE SERIES PILOT INTERVENTION STUDY By WENDY NOVOA GRAY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010 1

PAGE 2

2010 W endy Novoa Gray 2

PAGE 3

To m y husband, who has supported me through the years as Ive pursued my goals 3

PAGE 4

ACKNOWL EDGMENTS I would like to thank the fo llowing individuals: Dr. David Janicke, for his support, dedication to my professional development, and willingness to embark on a research study outside of his area of expertise in support of my clinical and research interests, Dr. Eileen Fennell, for providing me with the clinical oppor tunity that sparked a lifelong commitment to working with adolescents with Human Immunodefi ciency Virus (HIV). Dr. Robert Lawrence, Vicky Campbell, Peggy Borum, Bill Harbilas, Martha Buffington, and other members of the University of Florida (UF) Pediatric Infectious Disease Clinic, for their critical support in integrating this research proj ect within a medical setting. Da nielle Driscoll, for her strong attention to detail and commitment to the journe y of completing this project. The UF Center for Pediatric and Family Studies, the Geoffrey Cl ark Ryan Memorial Fund, and the College of Public Health and Health Professions, for their financial support of this project. And most importantly, I would like to thank the adolescent s and their families who invited me into their homes and were open to trying new strategies to improve their adherence by participating in this research study. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................10 INTRODUCTION.........................................................................................................................12 Treatment of HIV in Children and Adolescents.....................................................................13 Adherence in Children and Adolescents with HIV................................................................15 Correlates of Adherence..................................................................................................16 Limitations of Adherence Literature...............................................................................19 Interventions to Improve A dherence in Pediatric HIV...........................................................25 Limitations of the Intervention Literature..............................................................................29 Current Study Aims and Hypotheses......................................................................................32 METHOD......................................................................................................................... .............34 Participants.............................................................................................................................34 Inclusion Criteria.............................................................................................................34 Exclusion Criteria............................................................................................................34 Criteria for Identifying Nonadherent Patients.................................................................35 Experimental Design............................................................................................................ ..35 Procedure................................................................................................................................36 Recruitment.................................................................................................................... .36 Compensation..................................................................................................................36 Initial Screening and Assessment....................................................................................36 Baseline Monitoring........................................................................................................37 Schedule for Assessment.................................................................................................37 Measures.................................................................................................................................38 Demographic Questionnaire............................................................................................38 Treatment Interview Protocol..........................................................................................38 Conflict Behavior Questionnaire.....................................................................................38 Pediatric AIDS Clinical Tria ls Group Adherence Module 2...........................................39 Client Satisfaction Questionnaire....................................................................................39 Medication Use................................................................................................................3 9 Electronic monitoring...............................................................................................40 Pill counts.................................................................................................................41 Self-reported adherence............................................................................................41 Medication Adherence.....................................................................................................41 Viral Load..................................................................................................................... ...42 Intervention Program..............................................................................................................42 5

PAGE 6

Session Form at................................................................................................................43 Treatment sessions...........................................................................................................43 Data Analyses.........................................................................................................................48 RESULTS......................................................................................................................................53 Participant Data............................................................................................................... .......53 Participant 1.....................................................................................................................53 Background information..........................................................................................53 Construction of composite adherence score.............................................................53 Treatment and adherence.........................................................................................54 Visual inspection......................................................................................................55 Statistical analyses....................................................................................................56 Virologic functioning...............................................................................................57 Knowledge of medication regimen..........................................................................57 Barriers to adherence................................................................................................58 Parent-child conflict.................................................................................................58 Participant 2.....................................................................................................................58 Background information..........................................................................................58 Construction of composite adherence score.............................................................59 Treatment and adherence.........................................................................................59 Visual inspection......................................................................................................61 Statistical analyses....................................................................................................61 Virologic functioning...............................................................................................62 Knowledge of medication regimen..........................................................................62 Barriers to adherence................................................................................................63 Parent-child conflict.................................................................................................63 Participant 3.....................................................................................................................63 Background information..........................................................................................63 Construction of composite adherence score.............................................................64 Treatment and adherence.........................................................................................64 Visual inspection......................................................................................................66 Statistical analyses....................................................................................................67 Virologic functioning...............................................................................................67 Knowledge of medication regimen..........................................................................68 Barriers to adherence................................................................................................68 Parent-child conflict.................................................................................................68 Participant 4.....................................................................................................................69 Background information..........................................................................................69 Construction of composite adherence score.............................................................69 Treatment and adherence.........................................................................................70 Visual inspection......................................................................................................72 Statistical analyses....................................................................................................72 Virologic functioning...............................................................................................73 Knowledge of medication regimen..........................................................................73 Barriers to adherence................................................................................................74 Parent-child conflict.................................................................................................74 6

PAGE 7

Treatm ent Satisfaction......................................................................................................... ...74 DISCUSSION................................................................................................................................88 Possible Active Treatment Variables......................................................................................88 Changes in Viral Load, Barriers to Adhe rence, Medication Knowledge, and Family Conflict................................................................................................................................93 Utility of Multi-Method Adherence Assessment....................................................................96 Limitations.................................................................................................................... ..........99 Strengths...................................................................................................................... .........101 Conclusions and Future Directions.......................................................................................103 APPENDIX A SAMPLE GRAPH OF PATIENT ADHERE NCE FROM POWERVIEW SOFTWARE...104 B HIV MYTHS HANDOUT...................................................................................................105 C PICTORIAL REPRESENTATION OF MEDICATION REGIMEN..................................107 D PROBLEM SOLVING HANDOUT....................................................................................108 E COMMUNICATION HANDOUT.......................................................................................115 F FAMILY ROLES HANDOUT.............................................................................................120 G DEVELOPMENTAL CHANGES AND ADHERENCE HANDOUT................................122 LIST OF REFERENCES.............................................................................................................127 BIOGRAPHICAL SKETCH.......................................................................................................135 7

PAGE 8

LIST OF TABLES Table page 2-1 Intervention schedule of assessment..................................................................................51 2-2 Intervention sessions by format and topic.........................................................................52 3-1 Descriptive statistics of CAS data for all participants.......................................................75 8

PAGE 9

LIST OF FI GURES Figure page 3-1 Comparison of individual adherence measures for Participant 1......................................76 3-2 Composite Adherence for Participant 1.............................................................................77 3-3 Participant 1s correlogram as a function of time lag indicating independence of adherence data points.........................................................................................................78 3-4 Comparison of individual adherence measures for Participant 2......................................79 3-5 Composite adherence for Participant 2..............................................................................80 3-6 Participant 2s correlogram as a function of time lag indicating independence of adherence data points.........................................................................................................81 3-7 Comparison of individual adherence measures for Participant 3......................................82 3-8 Composite adherence for Participant 3..............................................................................83 3-9 Participant 3s correlogram as a function of time lag indicating independence of adherence data points.........................................................................................................84 3-10 Comparison of individual adhere nce measures for Participant 4......................................85 3-11 Composite adherence for Participant 4..............................................................................86 3-12 Participant 4s correlogram as a func tion of time lag indicating independence of adherence data points.........................................................................................................87 9

PAGE 10

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy IMPROVING ANTIRETROVIRAL MEDICA TION ADHERENCE AMONG ADOLESCENTS WITH HIV: A PILOT STUDY By Wendy Novoa Gray August 2010 Chair: David M. Janicke Cochair: Eileen B. Fennell Major: Psychology Maintaining an adherence level of 95% or more is critical in optimizing treatment effectiveness, suppressing viral load, and preventing medication resi stance in patients with HIV. Although adolescents are least like ly to be adherent, they have been largely ignored by the intervention literature. The curr ent study is the first-known prosp ective intervention designed to target adherence to antiretrovir al medications among adolescents with HIV. Participants were four perinatally-infected adolescents who r eceived a seven-week intervention focused on problem solving and family communication to improve adherence to their HIV regimen. The intervention was delivered through alternating home and telephone sessions and a multi-method adherence assessment approach was used to track adherence at baseline, dur ing treatment, and at three-month follow-up. Viral load was obtained at all assessment points. All adolescents experienced an improvement in adherence leve ls following implementation of problem solving but, for the most part, these improvements were not maintained over time. Changes in the adolescents routine were associated with declin es in adherence. Sevent y-five percent of the sample experienced a decrease in viral load by th e end of treatment. Medication resistance and immune suppression affected the degree to which improvements in adherence affected viral load 10

PAGE 11

11 in one participant. All participants reported a decrease in the severity of barriers experienced from preto post-treatment. Participant knowledge of their me dication regimen was high at pretreatment and did not change over time. Contrary to the initial hypothesis, parent perception of family conflict increased from preto post-trea tment for some dyads. Adopting an individualized approach to help adolescents problem solve a nd overcome barriers may be promising strategy to improve their adherence. Interventions focuse d on improving adolescent adherence may benefit from helping the adolescent find wa ys to incorporate th eir regimen into their regular routine as opposed to encouraging the adolescent to bend th eir routine to the demands of their regimen. Adopting a proactive problem solv ing approach may help minimize the extent to which changes in routine result in declines in adherence. Fu rther research, characte rized by randomization to treatment, multi-site coll aboration, and long-term monitoring of adherence are greatly needed.

PAGE 12

CHAP TER 1 INTRODUCTION Acquired immunodeficiency syndrome (AIDS) is a life-threatening chronic illness that has reached epidemic proportions. Worldwide, it is estimated that between 34.1-47.1 million individuals have AIDS or are in fected with the human immunodefi ciency virus (HIV), the cause of AIDS (Joint United Nations Programme on HIV/AIDS, 2006). In 2005, up to 2.3 million children under the age of 15 were infected with HIV, and it is estimated that 1,500 children are infected with the virus daily (World Health Or ganization, 2006). In the United States, there are an estimated 45,669 individuals living with HI V/AIDS. Of these, 10,000 are children under the age of 13 (Centers for Disease Control and Pr evention, 2005). A significantly large portion of those infected with HIV are children from ethni c minority groups (58% African American, 23% Hispanic), those living in poverty (Armstrong, Willen, & Sorgen, 2003), and in single-parent homes (Brown, Lourie, & Pao, 2000). In 1982, the Centers for Disease Control and Prevention reported the first case of AIDS in children (Chadwick & Yogev, 1995); just one year af ter the illness was firs t identified in adults (National Institute of Allergy and Infectious Diseases, 2005). As both the adult and pediatric HIV populations have grown over the past two decades, much attention has been given to studying the viral structure and tr ansmission of HIV in order to develop medications to slow down the diseases progression and find a cure. HIV is a single-stranded ribonucleic acid (R NA) retrovirus that weakens the immune system through a process that begins by targetin g T-cells, which help the immune system fight off infections (Armstrong et al., 2003). Once attach ed to a T-cell, the vi rus injects its genetic material and uses the cells internal structur e to replicate HIV-specific RNA proteins. This process leads to the eventual destruction of the T-cell and the release of additional viruses into 12

PAGE 13

the bloodstream where they can at tack other T-cells, replicate, and proliferate throughout the body. As the virus continues attacking the immune system by destroying T-cells, it weakens the bodys ability to fight off opportunistic inf ections such as pneumonia, disseminated cytomegalovirus, and various cancers (National Institute of Allergy and Infectious Diseases, 2005). HIV is transmitted through contact with blood and other bodily fluids via two methods of transmission. Vertical, or perina tal, transmission refers to mother-to-child viral transmission during pregnancy, birth, or breast feeding (National Inst itute of Allergy and Infectious Diseases, 2005). Approximately one-third to one-quarter of women infected with HIV who do not receive antiretroviral treatment during their pregnancy wi ll pass on the virus to their infant (National Institute of Allergy and Infecti ous Diseases, 2005). Prior to the advent of preventative treatment for pregnant women who were HIV positive, pe rinatal transmission of HIV was the most common method of infection am ong children, accounting for 91% of all documented HIV cases (Armstrong et al. 2003). Fortunatel y, with the increased awareness of the importance of pre-natal preventative HIV testing and treatment, significantly fewer children are born with HIV (Centers for Disease Control and Prevention, 2005; Lindegren et al., 1999). Ho rizontal transmission refers to viral transmission through other modalities such as sexual activity, drug use, or blood transfusion (Armstrong et al., 2003). Currently, through risky sexual behavior, drug use, and limited access to HIV prevention services, adolescen ts and young adults are one of the most atrisk groups for contracting HIV, comprising ha lf of all new infections (Secord & CotroneiCascardo, 2007; World Health Organization 2005). Treatment of HIV in Children and Adolescents Prior to the advent of antir etroviral medication, HIV was th e leading cause of infant mortality in major American cities (S ecord & Cotronei-Cascardo, 2007). Fortunately, 13

PAGE 14

advancem ents in treatment over the past two de cades have enabled a greater number of children born with HIV to survive into adolescence and adulthood (McConnell et al., 2005). Treatment for HIV first emerged in the late 1980s and consisted of single and dual agent nucleoside reverse transcriptase inhibitors (NRTI; Br ogly, Williams, Seage, Oleske, Van Dyke, & McIntosh, 2005). With guidance from numerous c linical trials, treatment of HIV has evolved from a single agent approach to more effective multi-drug regimen (Borgly et al., 2005). These multi-drug approaches, also known as highly acti ve antiretroviral th erapy (HAART), suppress viral proliferation by targeting the various stages of HIV replication (Dong, 2007). Liquid formulations of many antiretroviral medications have helped increase the accessibility of these drugs to younger children and those with di fficulty swallowing pills (Secord & CotroneiCascardo, 2006). Although there is no cure for HIV, existing treatments have helped reduce HIV-related mortality and morbidity, improve quality of life, and restore and preserve immunologic function by suppressi ng viral load (Panel on Antiretroviral Guidelines for Adult and Adolescents, 2006). Currently, expert recommendations for antiretroviral treatment in adolescents and adults consist of either two NRTIs and one non-nucleos ide reverse transcriptas e inhibitor (NNRTI) or one protease inhibitor (PI) with two NRTIs (Panel on Antiretrov iral Guidelines for Adult and Adolescents, 2006). It is importa nt to note that this reco mmended regimen may not be appropriate for all patients and may vary based on other factors such as clinician judgment, patient age, and degree of viral suppression or medication resistance (Bro gly et al., 2005; Panel on Antiretroviral Guidelines fo r Adult and Adolescents, 2006). Most children and adolescents with HIV are prescribed several antiretroviral medications, often with multiple dosings per day (Steele & Grauer, 2003). In addition to being potentially 14

PAGE 15

tim e-intensive, some medicines have unpleasant tastes and side-e ffects such as liver toxicity, hyperglycemia, hyperlipidemia, osteoporosis, and lactic acidosis (Pan el on Antiretroviral Guidelines for Adult and Adolescents, 2006). B ecause of these and other factors discussed below, optimal adherence to these treatment regimens is a challenge for many children and their families. Adherence in Children and Adolescents with HIV Adherence refers to the degree to which a pe rsons behavior coincides with medical or health advice (Haynes, 1979, pp. 1-2). Although being adherent does not guarantee a positive health status, adherence tends to predict health in pediatric HI V (Steele & Grauer, 2003). Indeed, children who report being less adherent tend to have higher viral loads (Puga, 2006; Williams et al., 2006). Current recommendations suggest that th e greatest chance of a person benefiting from treatment with HAART results from adherence of 95% or more (Paterson et al., 2001; Puga, 2006). Unfortunately, existing esti mates of adherence to antiret roviral medications among the pediatric population are much lower. Only 34% of all families are over 90% adherent, with most adherence rates ranging from less than 50% up to 75% (Katko, Johnson, Fowler, & Turner, 2001; Steele et al., 2001). Half of all patients fail to take their medications as prescribed (Puga, 2006) and between 26-59% of patients report missing doses during a given week (Feingold, Rutstein, Meislich, Brown & Rudy, 2000; Murp hy et al., 2001; Reddington et al., 2000). Alarmingly, a study of adolescents found that on ly 41% reported full adherence to their prescribed regimen (Murphy et al., 2001). Although nonadherence has significant conseque nces in most pediatric conditions, the impact of nonadherence in HIV can be severe and possibly irreversible. Subt herapeutic levels of HAART medication may be ineffective at preven ting viral proliferation which may lead to symptomatic HIV, replication of the HIV virus, and the development of medication resistance 15

PAGE 16

(Gavin & Y ogev, 2002; Sethi, Celentano, Ga nge, Moore, & Gallant, 2003). Children who become resistant to a particular medication may also develop resistance to other medications within the same therapeutic class, further lim iting their treatment options (Altice & Friedland, 1998). Over time, those with chronic nonadherence may develop resistance to several, if not all, antiretroviral medications (Sethi et al., 2003). Correlates of Adherence Given the severity of nonadherence in HIV, re search examining predictors of, and potential factors affecting, adherence is cri tically needed. Compared to the extensive adherence research in other pediatric medical conditions such as di abetes (De Civita & D obkin, 2004), research on correlates and predictors of adhe rence in pediatric HIV is in its infancy (Steele & Grauer, 2003). Although the more extensive adult HIV literature on adherence (e.g., Adam, Maticke-Tynedale, & Cohen, 2003, Remien et al., 2003) has the poten tial to inform the pediatric knowledge base, various aspects of living with HIV, such as lo sing a parent from AIDS and growing up with a chronic, life-threatening illness, are unique to children and can potentially impact adherence (Steele & Nelson, 2007). Similar to research in the general pediatri c literature, certain ch aracteristics of the treatment regimen have been found to negativel y impact adherence in HIV. In general, medication adherence is a greater problem in th e treatment of chronic medical conditions, such as HIV, than acute conditions (Rapoff, 1999). Le ngth of treatment with an tiretroviral therapy is negatively correlated with adherence, with childr en on lengthier courses of treatment being less likely to adhere (Martinez et al., 2000). Child ren who consider their regimen burdensome (e.g., too many pills, interferes with my schedule) tend to be less adherent than children who have less complex regimens or view their regimen as le ss intrusive (Belzer, Fuchs, Luftman, & Tucker, 16

PAGE 17

1999; Goode, McMaugh, Crisp, W ales, & Ziegler, 2003; Rapoff, 1999). Finally, children who experience greater medication side-effects are less likely to be adherent (Belzer et al., 1999). Child factors have also been associated w ith adherence. Specifically, children with a greater number of depressive symptoms tend to have poorer levels of medication adherence (Murphy et al., 2001). Children who view their HIV medication as a reminder that they are HIV positive are also less likely to be adherent (Belzer et al., 1999). Although medical providers believe that children who have been informed of their HIV status ar e less likely to be resistant to their treatment regimen (Brackis-Cott, Mellins, Abrams, Reval, & Dolezal, 2003), the data are mixed and do not allow for any definitive conclusi ons to be made (Wiener, Mellins, Marhefka, & Battles, 2007). Child age has been found to be a significant pred ictor of adherence, w ith adolescents being less likely to adhere than younger children (M ellins, Brackis-Cott, Dolezal, & Abrams, 2004; Williams et al., 2006). This may be due to a premature shifting of responsibility for the medication regimen from the parent to the child as the child approaches adolescence (Steele & Grauer, 2003). The developmental and social cha llenges children face as they enter adolescence further contribute to non-adhere nce (Mellins et al., 2004; Stee le & Grauer, 2003) as children experience tremendous changes in their biologi cal and psychological functioning. During this time, adolescents attempt to es tablish their personal identi ty, increase their autonomy, and explore their sexuality through ro mantic relationships (Holmbec k, 2002). These processes may at times be at odds with adherence to the HIV medication regimen. Adolescents in close intimate relationships often struggle with disclosing their illness out of the fear of rejection (Wiener, Battles, & Heilman, 2000). This may lead to so me adolescents hiding their HIV status by skipping doses when around others so as to avoid drawing unwanted at tention to themselves 17

PAGE 18

(Rao, Kekwaletswe, Hosek, Martinez, & Rodrig u ez, 2007). Because adolescence is a critical time in which lifelong positive and risky health behaviors are established, intervening with adolescents who have poor adherence to their HIV regimen is extremely important (Holmbeck, 2002). Certain caregiver characteristics have also been associated with nonadherence. Parents with less ability to name their childs medications, a reduced se nse of self-efficacy with regard to their ability to correctly administer their childs medications, lower perceptions of medication efficacy, and a greater concern of others disc overing their childs medi cal condition are more likely to have children with poor adherence (Red dington et al., 2000). It may be that parents who are afraid of others discovering their childs HIV status limit their ability to take advantage of adherence-enhancing social suppor t networks such as their child s school nurse who can assist with administering medications within th e school setting (Steele & Grauer, 2003). Specific family characteristics have been a ssociated with non-adherence in children and adolescents with HIV. Parent-chi ld communication is positively associated with adherence, with dyads reporting poorer parent-chi ld communication being more likely to be non-adherent (Mellins et al., 2004). Other family factors such as parent-child re lationship quality and caregiver-perceived quality of life and stress are also strongly asso ciated with adherence (Mellins et al., 2004; Miller, Bishop, Herman & Stein, 2007). These relationships make intuitive sense as so much of a childs functioning occurs within and is affected by the larger family system (Cunningham, Naar-King, Ellis, Peju an, & Secord, 2006). Given that parent-child conflict tends to increase during adolescence (Robin & Foster, 1989), it is possible that targeting family factors, such as parent-child communication a nd joint problem solving, may help to improve adolescent adherence to antiretroviral medicatio ns. Indeed, these family factors are highly 18

PAGE 19

am enable to intervention and should be targeted when addressing adherence in children and adolescents with HIV (Mellins et al., 2004). Limitations of Adherence Literature In a recent review of the literature, Steele and Grauer (2003 ) noted three major challenges facing the integration of existi ng research on adherence in pedi atric HIV: the cross-sectional nature of the current literature a lack of coherence in adhere nce conceptualiza tion, and the high variability in adherence assessment methodologies. These issues have significantly impacted the design of intervention programs to improve adhere nce as no coherent pi cture of adherence to antiretroviral medications can be drawn to guide research (Steele & Grauer, 2003). As previously mentioned, treatment for HIV requires multiple drug regimens over extended periods of time. Because adherence is a dynamic process that is multiply determined (Brown, 2002), cross-sectional st udies may be inadequate in assessing the complexity of adhering to a long-term medication regime n (La Greca & Bearman, 2003). Prospective longitudinal assessment is greatly needed to be tter understand the proce ss of adherence in youth with HIV (Liu et al., 2001; Steele & Grauer, 2003). Adherence varies greatly by how it is c onceptualized (Rapoff, 1999). Categorical conceptualizations of adherence, often used in initial studies, use specified criteria and labels (e.g., poor, fair, good) to classify patient adhere nce (La Greca & Bearman, 2003). In the pediatric HIV literature, categorical conceptualizations have made it difficult to compare results across studies as arbitrary cutoffs w ith no physiological basis have been used (Steele & Grauer, 2003). Because of this, studies which conceptualized adhe rence as a categorical variable, such as that conducted by Feingold and colleagues (Feingold et al ., 2001), have a limited ability to inform the literature. Recently, a move toward conceptua lizing adherence as a continuous, multidimensional variable has emerged in the pediatric literat ure (La Greca & Bearman, 2003). A greater number 19

PAGE 20

of studies in pediatric HIV now report adherence as a continuous variab le in the f orm of a percent by dividing the number of adherence behaviors complete d by the number of behaviors prescribed (La Greca & Bearman, 2003). Despit e the move toward viewing adherence as continuous variable, there remains high variabilit y in what constitutes adherence, with no two studies in Steele and Grauers (2003) comprehensive review con ceptualizing adherence in the same way. The high variability in how adherence is measured serves as another barrier in integrating the existing literature. Adherence in pediatric po pulations has been assessed in various ways. Drug assays, self-report, pill counts, pharmacy refill history, provider rating, and electronic monitoring are among the most commonly used assessment methods in pediatric HIV (Farley, Hines, Musk, Ferrus, & Tepper, 2003; Simoni Montgomery, Martin, New, Demas, & Rana, 2007). Although each method is able to provide some measurement of adherence, assessment methods vary in the accuracy and quality of inform ation they can provide as well as in the costs of using them. Blood assays measure a drugs concentration or presence in the bloodstream (La Greca & Bearman, 2003). They provide a direct and object ive measure of drug exposure and can be used to reliably determine adherence over short pe riods of time (La Greca & Bearman, 2003; Liu et al., 2001). Although drug assays are informative th ey can also be misleading. They are highly influenced by short-term adherence and are thus only a snapshot of behavior (Liu et al., 2003). Thus, a patient who is generally nonadherent but ta kes all of their medicati on just prior to having a drug assay will appear highly adherent (La Greca & Bearman, 2003). Additionally, because the therapeutic drug concentration levels for antiretr oviral medications are ill -defined in pediatric populations, blood assays are somewhat limited in their ability to measure adherence in HIV 20

PAGE 21

(Puga, 2006). This, in ad dition to the high cost of this assessment method and limited insurance reimbursement rates, make drug assays an impr actical method for assessing adherence in this population (Puga, 2006). Self-report is able to span a greater period of time than drug assays and serves as an inexpensive method of measuring adherence (L a Greca & Bearman, 2003). It comes in various forms and can be collected through interviews, questionnaires, and self-monitoring (La Greca & Bearman, 2003, Marhefka, Farley, Rodrigue, Sandrik, Sleasman, & Tepper, 2004; Wiener, Riekert, Ryder, & Wood, 2004). Self-report sp anning a shorter amount of time (e.g., 24-hour recall) resulting from direct and objective questioning tends to pr ovide more reliable reports than recalls of greater time spans (La Greca & B earman, 2003; Marhefka, Tepper, Farley, Sleasman, & Mellins, 2006). Althou gh relatively easy to obtain, questions about the validity of self-report have arisen as several studies have found th at self-report is highly influenced by social desirability, tends to overestimate adherence, and has an inconsis tent association with virologic response (Farley et al., 2003; Frey & Naar-King, 2000; Melbourne, Geletko, Brown, WilleyLessne, Chase, & Fisher, 1999; Naar-King, Fr ey, Harris, & Arfken, 2005). Though few studies have been able to demonstrate the external vali dity of self-re port data (Steele & Grauer, 2003), self-report remains one of the mo st utilized methods for assessi ng adherence in HIV. Self-report can provide information about vari ous adherence-related behaviors such as dietary intake, which may enhance medication absorbency, sleep patterns, and division of family responsibility for disease management (La Greca & Bearman, 2003; Naar-King et al., 2005, Simoni et al., 2007; Wiener et al., 2004). Several measures, such as the 24-hour recall interv iew (Marhefka et al., 2006) and the Treatment Interview Protocol (Marhefka et al., 2004), have been developed to 21

PAGE 22

im prove the accuracy of self-report data and have been found to correlate with data from electronic monitoring and pharmacy records (Far ley et al., 2003; Watson & Farley, 1999). Two studies in the pediatric HIV literature s upport the use of medical provider rating as a form of assessing medication adherence (Farley et al., 2003; Naar-King et al., 2005). Health providers in pediatric HIV are in a good position to assess adherence as they tend to have connected and long-lasting relations hips with their patients and th eir families (Naar-King et al., 2005). Provider ratings of adherence are often lo wer than self-report ratings and have been associated with viral load (F arley et al., 2003, Naar-King et al ., 2005). Though a potential source of adherence assessment, provider reports still overestimate adherence (Wiener et al., 2004). Ratings can be influenced by other factors such as patient report, or j udgments about adherence based on history or family knowledge (La Gr eca & Bearman, 2003; Naar-King et al., 2005). Pill counts measure adherence by comparing the amount of medication remaining in a container with the amount that would be expected to be remaining if the patient had taken the medication as prescribed (La Greca & Bear man, 2003). Though they are considered more reliable than self-report (La Greca & Bearman, 2003) pill counts have significant disadvantages. As with self-report, pill counts tend to overestimate adherence (Steele et al., 2001). They are difficult to accurately obtain when patients forg et to bring in their pill bottles to clinic appointments (Naar-King et al., 2005) and do no t guarantee that the pills missing have been ingested (La Greca & Bearman, 2003). Because they are computationally complex and timeconsuming, pill counts are better su ited for research purposes and short-term regimens (La Greca & Bearman, 2003; Liu et al., 2001). Pharmacy refill history measures adherence by examining the extent to which prescribed medications are ordered on time (La Greca & Bear man, 2003). They are less cumbersome than 22

PAGE 23

pill counts and are relatively easy an d inexpensive to obtain (Liu et al., 2001). Recent studies show a significant association between pharmacy refill data and viral load among children (Farley et al., 2003; Katko et al ., 2001; Marhefka et al., 2004; Ma rhefka et al., 2006). Despite the superiority of pharmacy refill data over other forms of adherence assessment, data can easily be invalidated if a patient frequen tly uses different pharmacies or stores medication at home (Marhefka et al., 2004). Additionally, pharmacy data is only able to provide a general picture of adherence. As such, it cannot pr ovide detailed in formation on day-to-day adherence behaviors or whether missing medications have been actually consumed by the patient (Marhefka et al., 2004; Melbourne et al., 1999). Electronic monitoring is the most recent innovation in adherence measurement. Through the use of a microchip embedded in prescription pill caps, medication dosing events are recorded each time the patient opens thei r bottle to retrieve medication (APREX, 1998). In addition to providing information on overall dosing behavior, elec tronic monitoring can provide real-time information about the time and date of each dosing event (Melbourne et al., 1999). This may help identify patterns and devi ations of medication use (e .g., patient is non-adherent on weekends, patient adherence increas es just prior to their appointment) that self-report and pill counts are not able to detect (Farmer, 1999; Schwed et al., 1999). Despite the abundance of information that can be obtained from electroni c monitoring, several disadvantages have limited its use. The biggest barrier to implementing el ectronic monitoring is cost (Wiener et al., 2004). The high price of each monitoring cap combined w ith the numerous medications prescribed for children with HIV limits the clin ical utility of electronic m onitoring (Naar-King et al., 2005). Some researchers have made methodological ac commodations such as monitoring adherence to only one antiretroviral medication in order to incorporate the adva ntages of electronic monitoring 23

PAGE 24

(Farley et al., 2003; Liu et al ., 2001; Melbourne et al., 1999). T hough this strategy is able to provide som e information on patient adherence pa tterns, this information is limited to only the medication being monitored as patients may be more adherent to some medications than others (De Civita & Dobkin, 2004). A second limitation of electronic monitoring is its inability to monitor other common medication forms in childr en, such as liquid and powder formulations (Bova et al., 2005; Farley et al., 2003; Simoni et al., 2007). Finally, it has been suggested that electronic monitoring may actually underestimate adherence as it is only able to document when pill bottles are opened (LaFleur & Oderda, 2004; Naar-King, Frey, Harris, & Arfken, 2005). Thus, a patients who opens their pill bottle once but retrieves seve ral dosings for use in a pill box or to lay out pills in advance will be seen as less adherent despite actual medication taking behaviors (Bova, Fennie, Knafl, Dieckhaus, Watrous, & Williams, 2005; La Greca & Bearman, 2003, Liu et al., 2001). As with pill counts and ph armacy refill history, electronic monitoring cannot guarantee that patients actu ally ingest the medication pres cribed (Liu et al., 2001). Despite these limitations, electr onic monitoring has been found to have higher specificity and predictive value than other form s of assessing adherence and is associated with virologic response (Farley et al., 2003). As previously mentioned, all of the a bove methods of assessing adherence have disadvantages in terms of accuracy and/or co st. Currently, there is no gold standard for measuring adherence in this population (Marhefk a et al., 2004). As such, no method should be used in isolation (Dolezal, Mellins, Brackis-Cott, & Abrams, 2003; Liu et al., 2001). Instead, assessment methods should be combined to provide a more accurate estimate of adherence (Puga, 2006). Research by Liu and colleagues (Liu et al., 2001) has a dvocated the use of a composite adherence score based on the combina tion of information from electronic monitoring, 24

PAGE 25

pill counts, and self-rep ort. Pre liminary data support the superi ority of a composite score over individual assessment methods su ch as self-report a nd pill count (Liu et al., 2001). However, given that Lius composite score calculations rely upon repeated measures longitudinal growth modeling, which requires a far greater number of participants than those commonly found in pediatric research studies, the use of this model in the pediatric literature is somewhat limited. Interventions to Improve Adherence in Pediatric HIV Difficulty integrating the extant adherence liter ature in HIV has limited its ability to inform adherence-improving intervention studies. To ou r knowledge, the first pu blished intervention study was conducted by Rogers and colleagues (Rogers, Miller, Murphey, Tanney, & Fortune, 2001) to improve future adherence to medica tions by targeting the acceptability of HAART treatment among a group of HIV positive adoles cents who had never been treated with antiretroviral medications. Guided by the Transt heoretical Model of Change (Prochaska & DiClemente, 1983), Rogers and colleagues de signed a 6-8 session intervention program delivered on a weekly basis during each pa tients medical appointment. The intervention consisted of presenting the patient with audio and videotapes depicti ng a newly diagnosed HIV positive adolescent female as she joined a support gr oup to help her cope with the impact of her illness and adjusted to the demands of her medi cal treatment. Despite th e innovative approach of this study, several methodological weaknesses must be noted. Of the 65 adolescents who agreed to participate in the study, only 18 (28%) completed the pr ogram. Of these 18, only seven (38%) participants received the progr am as it was intended to be delivered. At times, study sessions were not administered because they interfered with the patients health care delivery, as all sessions were delivered during patient medical appointments. Because the study treatment schedule was more intensive than what patients typically received through their standard medical 25

PAGE 26

care (i.e., w eekly clinic-based appointments), tr eatment burden was an issue and contributed to the studys high attrition rate. The methodological weaknesses of the Rogers et al. (2001) study refl ect the difficulty of conducting clinic-based interventi ons with this population. Since this time, researchers have shifted to the delivery of home-based interv entions. Home-based interventions minimize treatment burden in multiple ways. First, by bringing the treatment to the patient, home-based interventions limit the need for superfluous clinic visits. Thr ough the elimination of unnecessary clinic-related expenses, home-based interventions may serve as a more cost-effective approach to the delivery of services. Home-bas ed visits may also be helpful in reducing study attrition as they increase participant convenien ce by minimizing participant travel burden. Finally, home-based treatments can be delivered wit hout interfering with the delivery of standard medical care, an issue that repeatedly came up in the Rogers et al. (2001) study. Because of the multiple advantages associated with home-based treatments interventions since Roge rs et al. (2001) have primarily been home-based. Using the well-researched Health Belief Model (Becker, Drachman, & Kirscht, 1972; Becker, Maiman, Kirscht, Haefner, & Drasch man, 1977) as a guide, Berrien and colleagues (Berrien, Salazar, Reynolds, & McKay, 2004) co nducted a randomized, home-based, non-blind pilot study to improve the adherence of 18 ch ildren (ages 1.5-19 years) with HIV. The study focused on improving patient/careg iver knowledge and adherence by providing education for the HIV regimen through eight struct ured home visits over a three month period. The intervention was guided by a medication knowledge questionnai re completed at the beginning of the study which identified barriers to adherence and misc onceptions or lack of knowledge regarding the childs HIV regimen. When compared to the cont rol group, significant differences were noted in 26

PAGE 27

participant knowledge of their regim en at post-treatment, with the treatment group reporting greater knowledge and understandi ng of their medications. No signi ficant differences were noted in self-reported adherence or pharmacy refill hi story. No significant differences were noted in either viral load or CD4 count. Six to eleven month follow-up assessments suggested that those few participants who experienced significant improvements in their viral load were able to maintain their health status over time. Although this study had several methodological improvements over the previously described Ro gers and colleagues (2001) study, such as the inclusion of methods to assess a dherence, laboratory results, and a follow-up assessment, results suggest that the study was unable to demonstrate clinically significant change in health status. No criteria for poor adherence were specified and only one child over the age of 16 completed the protocol, limiting the studys generalizability to adolescents with chronic non-adherence. Although the above two studies se rved as important stepping stones for the development of future intervention studies, they failed to accoun t for many of the complex behaviors and systemlevel factors that comprise adherence such as child characteristics and family functioning (Cunningham et al., 2006). With the belief that child nonadherence is multiply determined, Cunningham and colleagues (Cunningham et al ., 2006) reported on the adaptation of Multisystemic Family Therapy (MST; He nggeler, Schoenwald, Borduin, Rowland, & Cunningham, 1998) to improve antiretroviral me dication adherence in P.D.S., a 12 year-old African American male. Drawing from evidence-based treatments such as cognitive-behavioral therapy, parent training, and stru ctural family therapy, MST was used to change the systems (e.g., individual, family, community) affecting or maintaining P.D.S.s nonadherence. After five months of intensive MST home and community-bas ed visits 2-3 times per week, P.D.S.s viral load declined, and remained at undetectable levels for 12 mont hs. MST was reinstituted when 27

PAGE 28

P.D.S.s viral load increased and sim ilar reducti ons were noted at the end of treatment. Given the single-case nature of this intervention, the authors noted their inab ility to determine if changes in P.D.S.s viral load were the result of the intervention or other non-specific factors such as the mere presence of the research team. Adherence was not measured in the study. Multisystemic therapy was also applied in a larger study by Ellis, Naar-King, Cunningham, and Secord (2006) with 19 children and their families. Participants received home-based MST sessions 2-3 times per week. Because MST is goal-driven, the intervention was not time-limited. Instead, treatment was only disc ontinued when treatment goals were met. On average, participants received MST for almost seven m onths (46 sessions). No changes in caregiverreported adherence were seen upon completion of the intervention but statistically significant improvements were noted in caregiver knowledge. Clinically significant re ductions were seen in viral load, with these changes maintained at three-month follow-up. Although Elliss intervention shows promise, seve ral methodological weaknesses must be noted. First, this study was conducted via retrospective chart review. The authors were limited in their ability to examine the process of adherence as all da ta were archival. Second, only one method of measuring adherence was used. The authors noted a ceiling effect with regard to self-report adherence as caregivers reported high levels of adherence at the beginning of the study despite contradictory laboratory data. Finally, the authors acknowledged the cost of MST ($5,500 to $6,000 per patient) as a significant barrier to the implementation of treatment. Preliminary studies of MST have supported its use as a viable treatm ent option to improve adherence in the pediatric HIV population. Adopting a socioecologica l framework such that used in MST provides significant advantages ove r the cognitive or e ducation-only programs previously used in this population. By acknowledging adherence to be a complex and multiply28

PAGE 29

determ ined behavior, a socioecological approach is able to target severa l system-level domains of adherence, such as child and family-syste m factors previously discussed (Cunningham et al., 2006). Another advantage over prio r interventions is MSTs ab ility to maintain positive adherence behaviors by helping patients develo p life-long skills such as improved problemsolving. These skills may enable the maintenance of good adherence long af ter the withdrawal of an intervention. Limitations of the Intervention Literature A general criticism of the existing interventi on literature, including MS T, is the lack of prospective studies employing multiple methods of assessing adherence (Simoni et al., 2007). At best, existing interventions have adopted sing le-method approaches. Some interventions have failed to include even one measure of adherence. When adherence is measured, self-report tends to be the most commonly used method, with self -reports ranging from 24-hour recalls to recall periods of over a month or more. Given the consistent finding that self-report tends to overestimate adherence (Farley et al., 2003; Frey & Naar-King, 2000; Melbourne et al., 1999; Naar-King et al., 2005), future studies employing multiple methods of measuring adherence would have several methodological advantages ov er the existing literature (Steele & Grauer, 2003). The extant intervention literature has also b een criticized for its uni versal application of one specific treatment to children from very diffe rent age ranges without consideration of the unique needs of participants from various de velopmental stag es (Simoni et al., 2007, Steele & Grauer, 2003). A consequence of th is approach has been that littl e to no attention has been given to the high variability in normative developmen tal tasks across age groups (Simoni et al., 2007; Steele & Grauer, 2003). Thus, issues that are salient to one age group, such as adolescence, may not be adequately targeted in an interventi on with participants ra nging from infancy through 29

PAGE 30

young adulthood (Steele & Grauer, 2003). Interventions targeting a specific age range m ay be an improvement over existing interventions as th ey would allow for targeting of normative developmental tasks (Holmbeck, 2002). Such in terventions have been limited by the small number of children with HIV who are of the sa me developmental stage and living in the same geographical area (Simoni et al., 2007). Single-ca se experimental designs may be an effective approach to dealing with small participant samp le sizes and may prove to be an informative precursor to larger, multi-site studies. As a group, adolescents have been largely ignored by the inte rvention literature. This is surprising given research showing that adolescen ts are at greatest risk for non-adherence (Mellins et al., 2004; Williams et al., 2006). Adol escents and young adults comprise the largest group of newly infected individuals with HIV (Secord & Cotronei-Cascardo, 2007; World Health Organization 2005). Additio nally, the relative aging of th e pediatric HIV population, with many children exposed to the virus in the late 1980s and early 1990s surviving into adolescence and young adulthood, fu rther contributes to the la rge numbers of adolescents currently living with HIV (M cConnell et al., 2005). The increased number of adolescents currently infected, coupled with greater rate s of non-adherence among this age group (Murphy et al. 2001, Williams et al., 2006), implies a critical need for interventions targeting adherence among adolescents. Despite being the most promising approach to improving adherence in children with HIV, MST is limited in its applicab ility to the greater pediatri c HIV population. Although several researchers have supported the ut ility of home visits among this population (e.g., Berrien et al., 2004; Cunningham et al., 2006; Ellis et al., 2006) the time-intensive nature of MST (2-3 home visits per week) contribute to high treatment cost and limited feasibility outside of the research 30

PAGE 31

setting. Thus, there is a critical need for the developm ent of a more cost-effective and less timeintensive home-based intervention to improve adherence among adolescents with HIV. A behavioral-family systems model (BFST; R obin & Foster, 1989) may be an alternative, yet efficacious, approach to targeting adherenc e in adolescents with HIV. Behavioral-family systems therapy targets parent-adolescent commun ication and problem-solving skills to improve overall functioning in distressed families (Robi n & Foster, 1989). Trad itionally used with families experiencing high levels of conflict, successful applications of BFST have been reported in other problematic areas of adolescent functio ning such as eating disorders (Robin, 2003), and Attention-Deficit/Hyperac tivity Disorder (Robin, 1998). With notable exceptions (e.g., Harris, & Mertli ck, 2003; Quittner, Drot ar, Iveres-Landis, Slocum, Seidner, & Jacobson, 2000; Wysocki, Harris, & Buckloh, 2006; Wysocki, Greco, & Harris, 2000), applications of BFST to children and adolescents w ith chronic illnesses have been limited. Adapting a behavioral-family systems appro ach to target adherence in adolescents with HIV may be an efficacious treatment. It has been suggested that focusing specifically on improving family functioning by helping dyads build positive and effectiv e communication skills may positively impact adherence to antiretrovi ral medications (Mellin s et al., 2004). BFST places specific emphasis on this aspect of fam ily functioning and also incorporates problemsolving skills training to help reduce the frequency and severity of conflict within the family system. Though research finding poorer parent-child co mmunication and relationship quality to be associated with non-adherence among adolescents with HIV supports the a pplication of a BFSToriented intervention to this population, no known research has examined the extent to which adolescent problem-solving abilities may imp act adherence. Support for incorporating a 31

PAGE 32

problem -solving approach comes from the a dult HIV literature. Problem-solving has been consistently and successfully used in several intervention studies aimed at improving adherence to antiretroviral medications in adults (e.g., Davies et al., 2006; Johnson, Gamarel, & Rose, 2006; Remien et al., 2005). Incorporating problem-solving skills training in an adherenceimproving intervention for adolescents may have similar benefits. An additional benefit of including problem-solving skills tr aining is that it provides adol escents with lifelong conflictresolution skills that can be applied to the many challenges they will face in life (Robin & Foster, 1989). The two principal components of a beha vioral-family systems-oriented approach (problem-solving and family co mmunication) are well suited for an intervention to improve adherence among adolescents with HIV. Current Study Aims and Hypotheses The current study expands upon the existing interven tion literature in several ways. It is the first-known prospective, behavioral-family system s-oriented intervention approach designed to target adherence to antiretrov iral medications among adolescents with HIV. Through alternating weekly home and telephone session s, the current intervention targ ets both the adolescent and the family system to improve adherence through th e development of effective problem-solving and communication skills as they relate to the HIV regimen. A second advantage over the extant literature is the narrower focus on the adolescent age group. By focusing solely on adolescents with HIV, this study targets an important yet greatly understudied population. Finally, the use of multiple methods for assessing adherence (ex., el ectronic monitoring, self -report, pill count, and laboratory results) serves as a significant improvement over previous research which has predominantly utilized single-method approaches. 32

PAGE 33

Study aim s and hypotheses are as follows: Aim 1: To conduct a family-oriented interven tion to improve family communication and problem-solving skills to increase adherence to prescribed antiretroviral medications among adolescents with HIV. Hypothesis 1.1: Adolescents wi ll have improved rates of adherence from pre-to posttreatment as determined by an adherence composite score. Hypothesis 1.2: Adherence rates will be maintained at three-month follow-up. Aim 2. To examine the effect of the in tervention on patient virologic functioning. Hypothesis 2.1: Adolescents completing th e intervention will experience a 1 log10 reduction in viral load from preto post-treatment. Hypothesis 2.2: Viral load at post-treatment will either be maintained or decreased at threemonth follow-up. Aim 3. To improve caregiver and patient knowledge of the medication regimen. Hypothesis 3: Participants co mpleting the intervention will have improved knowledge of their medication regimen at post-treatment comp ared to their knowledge at pre-treatment. Aim 4. To identify and reduce barriers to adhere nce to the treatment regimen as a result of the intervention. Hypothesis 4: Participants will report a signi ficant decline in the number of barriers to adherence from preto post-treatment. Aim 5. To identify and remediate fam ily conflict around tr eatment tasks. Hypothesis 5: Participants will report a signi ficant decline in their degree of family conflict, as determined by the Conflict Behavi or Questionnaire (Robin & Foster, 1989), from preto post-treatment. 33

PAGE 34

CHAP TER 2 METHOD Participants Participants were four adol escents (ages 13, 15, 17, and 17 year s) with a diagnosis of HIV and their parent/legal guardian attending a regularly scheduled appointment at the Pediatric Infectious Disease Clinic directed by Robert Lawrence, M.D. at University of Florida. Inclusion Criteria Child eligibility for participation included: (1) currently receiving medication for treatment of a diagnosis of HIV with no planned change in medication type (dosing changes were allowed), (2) being aware of their HIV diagnosis, (3) between the ages of 11-18, (4) accompanied by a parent/legal guardian, (5) able to speak, read, and understand English, (6) having at least one medication amenable to el ectronic monitoring, and (7) referred to the study by their physician for problems w ith adherence to their current medication regimen. Parent eligibility for participation incl uded: (1) having legal guardianship of the child, (2) living in the same household as the participating child for the duration of the study, (3) living within a 90minute driving distance of Gainesvi lle, Florida, (4) not planning to move out of the area within the next year, and (5) able to be c ontacted by telephone on a weekly basis. Exclusion Criteria Children and/or parents were excluded from the study for any factors that could negatively impact their ability to successf ully complete a telephone-based intervention designed to improve adherence to antiretroviral medications. Excl usion criteria included the presence of: (1) significant cognitive or developmental delay, (2) an inability to communicate via telephone, (3) the presence of a major psychiatric illness or medi cal condition in either the child or parent that impaired judgment, (4) current participation in another intervention designed to improve 34

PAGE 35

adherence, (5) any other m edical or behavioral condition that, in the op inion of staff, would adversely affect participa tion in the intervention. Criteria for Identifying Nonadherent Patients. All potential participants were considered nonadherent based upon physician referral. To identify nonadherent patients, physicians had access to patient self-reported adherence, laboratory data, and pharmacy refill records. Experimental Design Although randomized clinical trials provide the strongest evid ence for the causal relationship between an intervention and pati ent outcome (Clingempeel & Henggeler, 2002), the small number of adolescents with HIV living in the north-central Florida area limited the feasibility of this approach. A single-case design was used. Single-case experimental designs are well-suited to small sample sizes as they involve the intensive study of individual subj ects with the purpose of learni ng and measuring the effects of treatment on different individuals (Barlow & Hersen, 1984). Through their controlled and reliable measurement of clinical change, singlecase designs contribute significantly to our understanding of the effec tiveness of interventions and can se rve as informative precursors to larger group-based interventions (Barlow & Hersen, 1984; Tervo, Estrem, Bryson-Brockmann, & Symons, 2003). Due to the longitudinal nature of this study, the following steps were taken to minimize attrition: (1) increasing compensation across time to encourage continued study enrollment, (2) alternating weekly home and phone sessions to minimize participant burden, and (3) coordinating regular clinic appoint ments into the study protocol to reduce participant travel for study-related assessments. 35

PAGE 36

Procedure Recruitment Eligib le study participants were approached while atte nding a regularly scheduled appointment at the Pediatric Infectious Diseases Clinic by a trained member of the research team. Interested participants were provided with detailed study information. Parent informed consent and child assent were obtai ned prior to all data collection. All potential participants were informed that study refusal would not adversely affect th eir medical care. Compensation Participants were given $15 in recompense for completion of the pre-treatment assessment, $20 for completion of the treatment midpoint (s ession four), $25 for completion of the posttreatment assessment, and $45 for completion of the three-month follow-up. Participants were given an additional $20 for returning functioning electronic monitoring caps at the end of the study. Thus, participants were eligible to re ceive up to $125, but no less than $105, for successful completion of the entire study. Initial Screening and Assessment Dyads meeting initial eligibility criteria comp leted a semi-structured interview designed to assess current adherence beliefs and practices. Pa rents and children also completed self-report questionnaires designed to obtain basic demographic and contact information. Adolescents were given one Medication Event Monitoring System (MEM S) cap for their selected medication to be monitored and received instruction on proper MEMS cap use. Monitoring was limited to an antiretroviral medication that was in pill form and was currently contained in standard prescription plastic vial. Medicatio ns not housed in plastic vials (e.g., pill box), liquid, and powder medications were not eligible for MEMS cap use. Physician 36

PAGE 37

rating of adherence was obtained via a visual analog scale. Infor mation regarding patient viral load and medication regimen were obtained from medical records. Baseline Monitoring During the initial assessment, participants iden tified a day and time in which they would be available each week for contact by the research t eam in order to obtain baseline adherence data. MEMS cap and pill count data were collected du ring weekly scheduled home visits by research team members trained in the proper MEMS cap da ta collection and pill count procedures who were not directly involved with the interven tion and were blind to study aims. Self-report adherence data were obtained via a participant-comple ted questionnaire. Schedule for Assessment Participants completed four major assessments throughout their participation in the study (See Table 1). The first assessment occurred i mmediately following informed consent and was used to determine study eligibilit y. Participants who met initial study criteria were provided with a MEMS cap and were informed that their a dherence would be monitored on a weekly basis throughout the entire intervention via self-report, pill counts, a nd MEMS cap data. Pre-arranged weekly home visits by a member of the research team were re quired to obtain MEMS cap and pill count data. An adherence composite score based on MEMS cap, pill count, and self-report data were used to determine the starting time fo r the pre-treatment assessment. The pre-treatment assessment occurred at the participants next scheduled clinic appointment after a stable adherence pattern (i.e., two data points in wh ich adherence was not improving), was obtained. Thus, participants had varying lengths of baselin e monitoring prior to in itiating the intervention. A post-treatment assessment occurred at the end of the intervention program. Participants also completed a follow-up assessment approximat ely three months afte r their post-treatment assessment. Viral load data were obtained at each clinic visit as part of th e patients regular care. 37

PAGE 38

Measures Demographic Ques tionnaire A demographic questionnaire designed for this study was used to obtain information regarding family background information such as: child age, gender, race, age at diagnosis, age of disclosure, and parent marital status, educat ion, and family income. The participating parent completed the demographic questionna ire during the initial visit. Treatment Interview Protocol The Treatment Interview Protocol (TIP; Marhef ka et al., 2004) is a qualitative structured interview designed to assess a pa tients typical adherence to th eir prescribed medical regimen. The TIP assesses adherence behaviors by asking caregivers a bout their childs actual regimen behaviors prior to soliciting careg iver knowledge of their childs prescribed regimen. Caregiver knowledge of the prescribed regimen, as measured by the TIP, is associated with pharmacy refill history, a proxy of medication adherence. For th e purpose of this study, participants completed the portion of the TIP designed to assess knowledg e of the medication regimen. This information was used to inform and guide the current in tervention program and took approximately five minutes to complete. The TIP was also administered at post-treatment to examine changes in medication regimen knowledge as a result of the intervention. Conflict Behavior Questionnaire The Conflict Behavior Questionnaire (CBQ, Robin & Foster, 1989) assesses parent-child conflict, arguments, and disagree ments over the past two weeks via a 20-item true or false scale. Parent and child-report versions of the CBQ we re used in this study. The CBQ has been used extensively throughout the literature, has adequate internal consistency, and has been found to discriminate between distressed and nondistre ssed families (Robin & Foster, 1989). The CBQ 38

PAGE 39

was adm inistered at both preand post-treatment to examine changes in parent-adolescent conflict. Pediatric AIDS Clinical Trials Group Adherence Module 2 The Pediatric AIDS Clinical Trials Group (PACTG) Adherence Module 2 is designed to assess general reasons for non-adherence to prescribed medications. Because a patients level of adherence may vary across medica tions (De Civita & Dobkin, 2004) participants are asked to identify reasons for non-adherence for each indi vidual medication. Patients are asked to endorse their reasons for non-adherence from a list provided in Module 2 a nd are given the opportunity to identify any reasons not listed in the module. Th is information was used to inform and guide the current intervention program and was administered at pre-treatment, post-treatment, and at threemonth follow-up. Client Satisfaction Questionnaire The Client Satisfaction Questionnaire (CSQ ; Larsen, Attkisson, Hargreaves, & Nguyen, 1979) is an eight-item measure designed to measur e client satisfaction w ith services. Items for the CSQ were selected on the basis of ratings by mental health profe ssionals of a number of items that could be related to client satisfaction as well and by subsequent factor analysis. The CSQ-8 is uni-dimensional, yielding a homogeneous estimate of general satisfaction with services. Adolescent and caregiver treatment sati sfaction were measured by the CSQ-8 immediately after completing the intervention. Pa rticipating caregivers and children completed this measure separately. Medication Use Medication use was measured through electroni c monitoring, pill counts, and self-report. Because the cost of electronic m onitoring of all medications exceed ed the financial resources of this study, only one medication was selected for electronic monitori ng at the physicians 39

PAGE 40

discretion. However, all patient antiretroviral medica tions w ere measured via pill counts and self-report. Electronic monitoring Medication Event Monitoring System (MEMS) TrackCaps (APREX, 1998) are considered a reliable and innovative method for assessi ng patient adherence (La Greca & Bearman, 2003). MEMS caps fit most pharmacy bottles and monitor vial openings (events) via a microprocessor imbedded in the bottle cap. Each event is date and time-stamped with an accuracy of 30 seconds. Each MEMS cap is waterproof, can stor e up to 3800 medication ev ents, and is capable of retaining stored data for y ears after loss of battery power. MEMS caps were placed on one plastic medicatio n vial prescribed to the patient for the management of their HIV. The medication select ed for monitoring was at the discretion of the medical team. When the patients medical regime n consisted of medications of different dosing frequencies, the medication with the highest dosing frequency was selected. When dosing frequencies were the same among two or more me dications, the medication perceived to have the lowest adherence rate was select ed for electronic monitoring. MEMS cap data were collected on a weekly basis during baseline monitoring. Encrypted MEMS cap data were downloaded onto a battery-powered portable co mmunication device while in the participants home. Data were later uploaded onto the Aardex Ltd. HIPPA-compliant PowerView Version 3 software on a personal co mputer to obtain numerical and graphical representation of the data. A sample graph from the PowerView software is included in Appendix A. Participants were instructed on MEMS cap use at the initial screening visit and additional MEMS caps were available during each home visit to replace any malfunctioning caps. All caps were collected at the post-treatment assessment and were returned to participants one month 40

PAGE 41

prior to their three-m onth follow-up appointment to complete the one month of monitoring required for the follow-up assessment. Pill counts Pill counts were obtained by a member of the research team who was trained by the pharmacist of the UF/Shands Infect ious Disease Clinic or the Prin cipal Investigator in the proper procedure for counting pills. Pill counts were obtained on a weekly basis during baseline monitoring and on a bi-weekly basis during treatment to coincide with scheduled home visits. Pill counts were also obtained one month prior to, and during, the patients three-month in-clinic follow-up visit. Patients received a reminder call the day before their clinic visit to help minimize the number of patients who forgot to bring their pill bottles. A home visit was scheduled for those patients forgetting to bring in their medications to the clinic. Self-reported adherence Self-report medication use was collected on a weekly basis. Participants completed a paper-based monitoring log once a week that was customized for their specific regimen (e.g., number of pills, and frequency of dosing). For e ach expected dose, participants were asked to indicate: (1) yes or no if th ey took the dose as prescribed a nd (2) if no, participants were asked to write down why they did not take the do se. Asking participants to provide a reason for missed doses helped inform treatment by identifying any additional barriers that needed to be addressed by the intervention. Medication Adherence Information obtained from participant MEMS cap data, pill counts, and self-report were used to generate a composite medication adherence score guided by the theoretical underpinnings of Liu et al.s ( 2000) composite adherence score. In the event that MEMS cap data were missing or invalid, CAS values were ob tained from pill count adherence data. In the 41

PAGE 42

event th at MEMS cap and pill count data were both missing or invalid, self-report data were used. MEMS, pill count, and self-report data were considered invalid if: 1) the data were missing, 2) visual inspection suggested that the data were clearly discrepant from all other adherence values obtained. Viral Load A patients viral load refers to the estimated amount of HIV RNA copies per milliliter of blood plasma (AIDSinfo, 2005). Viral load serves as an indicator of disease progression and to what extent treatment is working (AIDSinfo, 2005). The patients estimated viral load was obtained as a part of the patie nts usual medical care and was taken from medical records corresponding to the patients in itial, pre-treatment, post-treatment, and three-month follow-up assessment visits. Intervention Program The current intervention program used a be havioral-family systems-oriented approach targeting family communication and problem-s olving skills to improve adherence to antiretroviral medications among adolescents with HIV. Dyads received seven sessions focusing on improving parent-adolescent co mmunication and using a problem-s olving approach to reduce barriers to adherence. Throughout the intervention, participants were reinforced for positive selfcare behaviors. All sessions occurred during an agreed upon time by the in terventionist and the participants and lasted approximately 40-50 minu tes. Both the particip ating parent and child were required to be present for each session. The intervention was delivere d by an advanced graduate student with a background in pediatric psychology who also had over two years of experience working in a pediatric HIV clinic. Beyond the possibility of having briefly encountered families through the delivery of 42

PAGE 43

consultativ e services in the HIV clinic, the interv entionist did not have a prior relationship with any of the participating a dolescents or caregivers. Session Format During the intervention, children and parents part icipated in alternating weekly phone calls and in-person sessions with a behavioral family systems therapy orient ation (Robin & Foster, 1989) to assess their adherence and address any barriers to following their treatment regimen. Participants received alternating in-person and telephone sessi ons. For a listing of session format and content, please see Table 2. Treatment sessions The first treatment session occurred in person at the participants regularly scheduled clinic appointment after at least two data points in which adherence was not improving were obtained through baseline monitoring. The main purpose of this session was to provide the family with a constructive experience to discuss their difficulties, reframe any at tributions or beliefs that may interfere with future treatment, gather information related to current adherence, parent-adolescent conflict, and family processes, and establish appropr iate expectations for treatment. At this visit, participants completed a structured adherence interview (TIP) and self-report questionnaires. Participants were provided with educational information about HIV through an interactive quiz format. All information was presented at a deve lopmentally-appropriate level and participants were given written information of the session content (Appendix B) Participants were provided with a written list of their medication regimen which included each medication name (brand name and generic), dosage, frequency of dosage, time frame when medication should be taken, and any special instructions (e.g., take with f ood). A pictorial adherence sheet (Appendix C) was also provided to assist patient s who preferred a visual represen tation of their medication regimen. 43

PAGE 44

At the end of the session, participants were provi ded with su mmary feedback from their therapist and information about their future treatment. The second session occurred in the familys home approximately one week after the first treatment session. This session incorporated in formation obtained from the initial and pretreatment assessment to help families identify regimen-specific barriers to adherence. Families reviewed all of the barriers listed from th eir completion of the PACTG Adherence Module 2 measure and rated each barrier in terms of: (1) how difficult they believe each barrier would be to overcome, and (2) the importance of overcoming each barrier. Participants then targeted a barrier believed to be important, but of lesser difficulty to overcome using the problem-solving framework ou tlined by Robin and Foster (1989). Participant materials for session two are presented in Appendi x D. Instructions, modeling, role playing, and corrective feedback were provi ded to guide the family through the following problem-solving process targeting their select ed barrier to adherence: Problem Definition: During this phase, both the adolescent and their parent were allowed to express their perspective on the selected ba rrier to adherence. Each family member was guided by their therapist in sta ting their view of the problem in a precise, concise, and nonaccusatory manner. Each family member was then asked to verify their understanding of each others perspective through reflection. Inaccurate re flections were clarified by the person stating the problem so that family members would be able to understand each others perspectives and have a clear understanding of the problem to be addressed in the following phases. Generation of alternative solu tions: This phase provided family members the opportunity to suggest a variety of strategies to overcome their identified barrier to adherence. During this phase, family members were instructed to list as many ideas as possible, no matter how 44

PAGE 45

outlandish or ridiculous they m ay seem. The adol escent was asked to record all ideas generate on a problem-solving worksheet (Appendix D) in order to maintain their engage ment in the process. Family members were specifically instructed to defer evaluation of any ideas until a later phase, as evaluation could interfere with the generati on of high quality and novel solutions. Thus, the therapist interrupted any pre-mature evaluations during this phase in an effort to keep the conversation on task and maintain a non-evaluativ e atmosphere with ba lanced participation among family members. Participants progresse d to the decision making phase once they generated a workable list of potential solutio ns. A minimum of 3-4 solutions that went beyond each family members initial position on the problem were required to move to the next phase. The therapist occasionally interjected a suggestion to help diffuse any tense situations and encourage the generation of creative alternatives. Decision making: The therapist began this pha se by providing the pa rent and adolescent with a brief discussion of the rationale for the decision-making process. During this phase, the parent and adolescent took turns evaluating each idea generated from the previous phase. Participants were asked to evaluate and cl early state the perceive d negative and positive consequences of each idea for themselves and fo r the rest of their family. Participants were prompted to comment on the feasibility of each idea in addition to the shortand long-term consequences of implementing each idea. Parents and adolescents then independently rated each idea with either a plus or a minus. After all solutions were evaluate d, dyads reviewed their ratings to identify any solutions that were rated positively by ev eryone. Participants who reached a consensus on one or more ideas either select ed one idea, or combined several ideas, to implement. Participants who were not able to re ach a consensus were assisted by the therapist in negotiating a compromise. A solu tion rated positively by either the adoles cent or parent was 45

PAGE 46

chosen for discussion. Participants were asked to restate their evaluation of the solution and note areas of dis agreement between them. The pers on who was viewed as making the greatest compromise through the implementa tion of the solution was prompted to provide a variation of the current solution for further consideration. Family members evaluated the new proposed solution while the therapist played an active medi ating role to help the family arrive at a consensus. Planning implementation: During this phase, ad olescents and parents generated a step-bystep plan for implementing the chosen solution. A thorough plan outlined specific responsibilities/behaviors for each family member (e.g., parent will check to make sure child takes medicine), identified any resources that may be needed (e.g., use of a wrist watch or cell phone alarm to remind child to take medicine), a nd anticipated any difficulties that may arise in implementing the agreed upon solution. Family me mbers were asked to monitor each others compliance to the terms outlined in the plan. Renegotiation: The renegotiati on phase of problem-solving was only invoked when the family had been unsuccessful at using their ini tial solution to reach th eir goal. The therapist engaged in a thorough review of each persons at tempts at implementation. Information obtained form this review helped place the failure in the context of the problem-solving framework and assisted in revising th e current plan. Problem-solving played a role throughout the entire interven tion as additional barriers arose. Negative communication habits which inte rfered with productive ve rbal interchanges were identified and modified by the therapist during all problem-solving sessions. The third session occurred over the telephone. This session focused on evaluating the success of the problem-solving approach implemente d in session two. Participants evaluated the 46

PAGE 47

success o r failure of their previously established goal. Those reporting problem resolution/goal attainment were asked to identify and target a ba rrier of slightly greater difficulty and importance to overcome using the problem-solving framewor k described above. Participants who did not reach their goal re-evaluated their current plan, identified any barriers they experienced, and implemented the problem-solving framework to either revise their current goal or revise their plan. Problem-solving exercises were used to assist participants in need of additional experience with problem orientation and the problem-solving process. The fourth session was conducted in-person. Pa rticipants reported on their degree of success with the problem-solving approach outlined in the previ ous session. If the previously agreed upon plan was unsuccessful, the problem-solving process was conducted to either revise the plan or the goal or to break down the proble m into smaller, more manageable components. Participants reporting successful re solution of their previously ta rgeted problem were asked to target the next barrier on their list in terms of difficulty and importance. The main goal of this session was to provide the family with communicati on skills training. Famili es participated in a didactic session that provided rationale for communication training and discussed common parent/child communication errors. Various family communication patterns (ex. verbal, nonverbal, and mixed) were discussed and positive family communication strategies were emphasized. Participants were provided a writte n copy of the session materials (Appendix E). With guidance from their therapist, participan ts identified their own problematic communication patterns and practiced positive communication alternat ives via role-playing. Participants received feedback from their therapist and agreed to target these specific communication strategies throughout the week. 47

PAGE 48

The fifth session occurred over the telephone and focused on reviewing the previous weeks homework on problem atic communication patterns and the problem-solving process. Each participant was allowed to express their opinion on what went we ll and needed additional work with regard to their targeted communication patterns. Participants we re asked to reflect on their own progress as well as problem-solving any additional issues the family wanted to address. The sixth session occurred in person and focu sed on defining family roles. Participants identified their specific roles within the family in general, as well as in regard to the teens HIV regimen (e.g., who is responsible for refilling the medication) and used positive communication strategies to resolve ambiguous roles pertaining to the HIV regimen (Appendix F). Parents were given information about the developmental ch anges from childhood to adolescence and the potential impact of these changes on their teens adherence (Appendix G). The seventh session occurred over the phone a nd was led by the dyad as they reviewed their recent problem solving efforts. Dyads we re encouraged to provide feedback and model appropriate problem-solving and co mmunication skills. The therapist served as a passive listener throughout most of the telephone session. The session concluded with feedback from the therapist. Data Analyses For Hypothesis 1.1, that adolesce nts would have improved rates of adherence from pre-to post-treatment as determined by an adherence composite score, da ta were graphically displayed in percent adherence form and visually inspec ted according to guidelines set forth by Kazdin (1982). Autocorrelations were run to identify serial de pendency in the data before all statistical analyses. Significant autocorrelations indicate th e presence of serial dependency and dictate the 48

PAGE 49

need for specialized s tatistical techniques such as time series analysis.An autocorrelation is computed by correlating each data point with another data point later on in the series. Correlating a data point with the po int immediately after it is referred to a lag-1 co rrelation. The correlation of a data point with another point two positions after it in the series is known as a lag-2 correlation. Depending upon the number of data point s available, autocorrelations of numerous lags can be calculated. Though a la g-1 correlation is usually suffi cient in identifying serial dependency in the data, autocorrel ations of greater than lag-1 a llow for a finer grained analysis of serial dependency (Barlow & Hersen, 1984). In the current study, autocorrelations of numerous lags, as allowable by th e data, were calculated. Data not serially dependent were subject to Mann-Whitney U-tests. For Hypothesis 1.2, that adherence treatment ga ins would be maintained at three-month follow-up, data were graphically displayed and vi sually inspected. Changes between phases were evaluated using Mann-Whitney U-tests. For Hypothesis 2.1, that adolescents completing the intervention would experience a reduction in viral load from preto post-tre atment, a 1 log10 reduction in viral load was considered to be a clinically significant change. For Hypothesis 2.2, that viral lo ad at post-treatment would either be maintained or decreased at three-month follow-up, a 1 log10 increas e in viral load was considered a clinically significant increase in viral load. For Hypothesis 3, that participants comple ting the intervention would have improved knowledge of their medication regimen at post-tr eatment compared to their knowledge at pretreatment, data were compared across phase s. An improvement in regimen knowledge was 49

PAGE 50

identified by an increase in percent of medications correctly na med, percent of doses correctly identified, and percent of dosing frequencies correctly identified. For Hypothesis 4, that participants would re port a significant declin e in the number of barriers to adherence from preto post-treatment, barrier severity scores were compared across phases. Decreases in barriers were identified by a d ecline in barrier severity score from preto post-treatment. For Hypothesis 5, that participants would re port a significant declin e in their degree of family conflict, as determined by the Conflict Behavior Questionnaire (CBQ; Robin & Foster, 1989), from preto posttreatment, CBQ scores were compared across phases. Decreases in family conflict were identified by a decline in CBQ scores from preto post-treatment. For those families reporting clinically elevated scores on the Conflict Behavior Questionnaire (Robin & Foster, 1989) at pre-treatment, a reduction in reported conflict to non-clinical levels was considered a clinically significant change. 50

PAGE 51

Table 2-1. Intervention schedule of assessm ent Study Phase Length of Phase Measures Completed Initial Screening & Assessment 1 hour in-clinic assessment Informed Consent Demographic questionnaire Medication use (physician rating) Viral load Baseline Monitoring Until a consistent pattern of adherence is obtained Medication use (self-report, MEMS, pill counts) Pre-Treatment Assessment/Session 1 50-minute in-clinic assessment Adherence Interview (TIP) PACTG Adherence Module 2 Conflict Behavior Questionnaire Viral load Medication use (self-report, MEMS, pill counts, physician rating) Treatment Assessment 8 weekly sessions Note: Session 1 occurs during pretreatment assessment Medication use (self-report, MEMS, pill counts) Post-Treatment Assessment 30-minute in-clinic assessment Medication use (self-report, MEMS, pill counts, physician rating) Conflict Behavior Questionnaire PACTG Adherence Module 2 Viral load Follow-up Assessment Approximately 3 months PACTG Adherence Module 2 Viral load Medication use (self-report, MEMS, pill counts, physician rating) 51

PAGE 52

Table 2-2. Intervention sessions by f ormat and topic Session # Session Format/ Location Topic 1 In-clinic visit Introduction to treatment program, HIV education, medication regimen review 2 In-person home visit Introduction to problem-solving 3 Telephone Review of problem-solving progress 4 In-person home visit Family communication skills training 5 Telephone Review of problem-s olving and communication progress 6 In-person visit Defining family roles and expectations 7 Telephone Family-led problem-solving se ssion, self-evaluation of progress 52

PAGE 53

CHAP TER 3 RESULTS Participant Data Participant 1 Background information Participant 1 (P1) was a 17-year-old Afri can-American female in the custody of her maternal aunt. Caregiver reported annual family income fell between $20,000 and $29,999. Participant 1 was diagnosed with HIV at birth an d was informed of her diagnosis at age 14. Her prescribed treatment regimen consisted of two NRTIs (Viread 300 mg once a day; Epzicom 1 tablet once a day) and two PIs (Norvir 100mg once a day; Reyataz 300 mg once a day), all of which were taken at night. According to the me dical team, P1 had a long standing history of poor adherence to her treatment regimen. Pharmacy reco rds indicated that prior to enrolling in the current study, P1 had not refi lled her medications in over three months. At the time of enrollment, physician estimated adherence for the past month was 36%. Construction of composite adherence score Per the suggestion of her physician, Reyataz was selected as the medication to be monitored electronically throughout the entire study. Adherence data as measured by MEMS, pill count, and self-report for Reyataz are presen ted in Figure 1. As seen in Figure 1, MEMS cap estimates were significantly below pill count and self-report estimates during the treatment and follow-up phases of the study. Midway through treatme nt the researchers became aware that the MEMS had been dropped several times on a hard surface. Participant 1s MEMS was then replaced with a new MEMS and additional education on proper MEMS cap use was provided. Despite this, MEMS estimates continued to be si gnificantly below pill count and self-report data. Though pill count and self-report data were highly correlated with one another (95.2%), MEMS 53

PAGE 54

cap associations with pill count data (-0.7%) we re trivial. Correlations between MEMS and selfreport d ata were low (19.2%). Us ing previously described guidelines for identifying missing or invalid data, MEMS cap data were only used for baseline week two and week four in constructing the composite adherence score. Pill counts, which were more consistent with selfreport as well as with viral load data, were used for the majority of composite adherence score calculations (baseline weeks one & three, treatment sessions tw o, four, & six). During the weeks in which phone sessions were conducted and pill counts could not be obtained (sessions three, five, & seven), self-report data were used as the foundation for CAS. As MEMS cap data were deemed invalid for the follow-up period and pill count data did not capture week-to-week variability in adherence rates, se lf-report data were used for all follow-up CAS. It must be noted, however, that average self-reported adherence during the follow-up period (95%) was similar to estimated monthly pill count adherence (92.6%). Treatment and adherence Composite adherence data for P1 are presen ted in Figure 2. During the baseline monitoring period, her adherence ranged from 0% to 100%, with an average adherence rate of 64.7%. Physician estimate of adherence for the baseline period was 86%. According to self-report adherence sheets co mpleted throughout the baseline monitoring period, forgetting was the most commonly cite d reason for missing a dose. This barrier was specifically targeted during treatment. In sessi on two, P1 and her guardian agreed to implement a new incentive system to help improve her adheren ce. According to this system, P1 would earn points toward being allowed to redecorate he r room each time she remembered to take her medication. Unfortunately, this system was not implemented during th e week and adherence rates dropped from 87.5% at the end of baselin e monitoring to 71.4% and 75% during treatment weeks one and two. In session three, a reevaluation of the previously designed plan was 54

PAGE 55

conducted and both P1 and her guardian decided to devise a new plan. Through use of the fivestep problem solving approach in session three, it was discovered that P 1 often fell asleep before remembering to take her medication. Primary contributors to this problem were: 1) returning home late and tired from her pa rt-time job and, 2) the remote lo cation in which her medications were kept (on top of a tall en tertainment center in the family room). During the brainstorming portion of this session, numerous ideas were ge nerated to overcome this barrier, among which was the relocation of P1s medications to a place in which sh e was highly likely to encounter before falling asleep. Thus, P1s medications were moved from on top of the entertainment center to in front of the power button on the television in her bedroom. Both P1 and her guardian agreed that this would be a more effective strategy as P1 always turned her television on in order to go to sleep. After implementation of this plan, P1s adherence increased to 91.7% and 100% for treatment weeks three and four, respectively (F igure 2). From session th ree until the end of treatment, P1s average adhere nce rate was 94.3%. Her overall adherence for the treatment phase of the study was 88.3%. Improvements in adheren ce rates were maintained at three month follow-up (average adherence 95%). Physician rating of adherence was 86% at both posttreatment and at follow-up. Visual inspection The magnitude of change and rate of change across phases were vi sually inspected in Figure 2 to determine if change occurred due to the intervention. In evaluating magnitude of change, changes in the mean and level of pe rformance across phases were considered. Mean adherence increased from baselin e (64.7%) to treatment (88.3%) to follow-up (95%). Level of performance declined between th e end of baseline monitoring and the beginning of treatment, suggesting that the intervention did not produce a rapid shift in improving adherence. However, 55

PAGE 56

as this intervention was designed to be deliver ed across several sessions, a rapid change in behavior from baseline to the first treatm ent session was not expected. When problem solving was implemented and followed through by P1 a nd her guardian (session three), the level of performance experienced a shift from 75% to 91.7% adherence. Rate of change was examined by inspecting the trend and latency of change. Although adherence data from the baseline period had a generally positive trend (slope=0.23), the data were highly variable. Treatment phase data clearly illustrate a positive trend with increased stability in data points (slope =0.04). Follow-up data appear to continue the trend established during the treatment phase and a ceiling effect was seen in overall adherence by the end of the study (slope=-0.02). With regard to latency, chan ges in adherence data were visible after implementation of the adherence strategy agr eed upon during session three. Thus, visual inspection suggests that th e intervention was effective in prom oting behavior change over time. Statistical analyses Autocorrelations from one to 10 lags were calculated to identify any serial dependency in the baseline and treatment phase study data prior to all statistical analyses. As seen in the correlogram and its corresponding table (Figure 3), associations be tween data points for all lags were small, suggesting that the autocorrelations did not significantly deviate from zero. This suggests that the data points could be treated as independent observations. Because P1s data were not serially dependent, conventional statisti cal analyses were used to examine differences between baseline, treatment, and follow-up adherence data. Table 3 presents descriptiv e statistics of the CAS duri ng the baseline, treatment, and follow-up phases of the study. As seen in Tabl e 3, skewness and kurtosis values from the baseline and follow-up phases of the study were beyond acceptable limits, suggesting that the 56

PAGE 57

data were no t normally distributed. Because this basic assumption of th e general linear model was not met, non-parametric statistics were used to compare differences between phases. No statistically sign ificant differences in adherence were found between baseline and treatment (Mann-Whitney U=19, p>.05). Treatment phase adherence and follow-up phase adherence did not significantly differ (Mann-Whitney U=8, p>.05). Baseline and follow-up adherence data also did not statisti cally differ (Mann-Whitney U=12.5, p>.05). Virologic functioning At the beginning of baseline monitoring, P1 had a viral load of 2800. When treatment was initiated, laboratory results suggested that P1s viral load had increased to 3300. At the end of treatment, Participant 1s viral load had decrea sed to below baseline and pre-treatment levels (2700). This decrease in viral load continued into the three-month follow-up period (viral load at follow-up = 430). Though P1s viral load decrease from 3300 at pre-treatment to 430 at followup was 100 short of reaching clinical significance (1 log10 reduction), this change represented an 87% reduction in viral load. Knowledge of medication regimen On the Treatment Interview Protocol complete d at the end of the baseline period, P1 was able to correctly provide 100% of her prescr ibed antiretroviral medications, 100% of her prescribed dosing frequencies, and 100% of her prescribed dosing amounts. It must be noted, however, that this information was not recalled from memory but rather from a list of her medications that P1s caregiver had brought to clinic. Under guidelines of the Treatment Interview Protocol, use of a cheat sheet is allowed as a memory recall tool. At post-treatment, P1 did not have a cheat sheet. De spite this, she was able to r ecall the names of 100% of her prescribed medications, 100% of her prescribed dosing frequencies, and 75% of her prescribed dosing amounts. 57

PAGE 58

Barriers to adherence During the initial assessm ent, P1 reported a num ber of barriers to a dherence over the past month including: not getting medications refilled in time (frequent problem), forgetting (almost always a problem), not remembering if dose was already taken (frequent problem), falling asleep (frequent problem), and being away from home (har dly ever a problem). In total, P1s barrier severity score at baseline was 10. At post-treatmen t, P1 reported no barriers to her adherence. Parent-child conflict P1s score on the Conflict Behavior Questi onnaire (CBQ) at the be ginning of treatment suggested that her response pattern was similar to that of nondistressed families (CBQ=3). Her score on the CBQ at post-treatment was two. Caregiver report of conflict at the beginning of treatment was similar that of nondistressed families (CBQ=5). At the end of treatment, her response pattern indicated increased levels of conflict more similar to that of distressed families (CBQ=8). Participant 2 Background information Participant 2 was a 15-year-old biracial (African-American/Hispanic) female living with her biological mother. Caregiver reported annual family income fell between $40,000 and $49,999. Participant 2 (P2) was diagnosed with HIV at three years, 11 months and was informed of her diagnosis at this time. Her prescribed treatment regimen consisted of two NRTIs (Epivir 150 mg twice a day; Zerit 40 mg twice a day) a nd two tablets of one PI (Viracept 1,250 mg twice a day). According to P2, she had not taken any of her medications over the past two months. At the time of study enrollment, physician estimat ed adherence for the past month was 39%. 58

PAGE 59

Constructio n of composite adherence score P2s physician requested that Viracept be electronically monitored as this medication required the greatest number of pills to be swallo wed per day (2 pills at each dosing). Adherence data as measured by MEMS, pill count, and self-r eport for Viracept are presented in Figure 4. As seen in Figure 4, self-report estimates were ge nerally higher than both MEMS cap and pill count estimates. MEMS data were substantially correla ted with pill counts (58% ) and with self-report (57%). Pill counts and se lf-report estimates were strongly corr elated with one another (60%). Given the substantial-to-strong correlations between MEMS cap and other adherence assessment measures, MEMS data were used as the foundatio n for CAS for all but on e of the data points (session 3). Because pill count and self-report da ta for session three were more congruent with one another than with MEMS, pill count data were used for this time point only. Treatment and adherence Composite adherence data for P2 are presen ted in Figure 5. During the baseline monitoring period, P2s adherence ranged from 75% to 83.3%, with an average adherence rate of 78.5%. Physician estimate of adherence fo r the baseline period was 79%. During the baseline monitoring period, being away from home and falling asleep were the most commonly reported reasons for missing a dos e. These barriers were identified as targets for intervention in addition to other barriers that arose while working with P2 and her mother. Adherence dropped from 77.8% at the end of baseline monitoring to 60% in the week following the first treatment sess ion. In session two, P2 and her mother expressed interest in using the problem solving approach to help P2 improve her adherence to her morning medications. According to P2, she often forgot to take her morning dosings because she did not have a structured routine in place. During the brainstorming portion of the session, P2 and her mother were able to come up with numerous solutions to help overcome this barrier. The 59

PAGE 60

solution rated most positively by P2 and her mother was taking her medications along with breakfast in the morning. Individu al responsibilities for implementing this plan were identified including cooking breakfast in the morning and pl acing P2s dosings next to her meal (mother) and being willing to be woken up to take he r medicine if found sleep ing past 10 am (P2). Following implementation of this plan, P2s adherence increased from 60% to 86.7%. Unfortunately, increased responsib ilities coinciding with the start of the school year for both P2 and her mother led to greater day-to-d ay variability and decreased adherence for both morning and nighttime medicati ons. Adherence following session four dropped to 42.1% and problem-solving was re-implemented to overcome newly arisen barriers. Because P2s mother had begun leaving very early for work and could no longer cook breakfast for P2, the previously implemented solution no longer applied to the fa milys circumstances. At night time, P2 often returned home late and tired from cheerleadi ng practice and went stra ight to bed without remembering to take her medications. Because electronic monitoring indicated that night time doses were missed at a greater frequency than morning doses, session six focused on improving adherence to nighttime medications. Using the problem solving approach, P2 and her mother agreed to implement a plan in which P2 woul d take her medications upon arriving home after cheerleading practice. P2s mother agreed to be responsible for reminding P2 to take her medication. Unfortunately, this plan was not well implemented and adherence remained low for session weeks six and seven ( 44.4% & 50%, respectively). P2s overall adherence for the treatment phase of the study was 57.7%. Physician estimated adherence for the treatment phase of the study was 81.5%. For the follow-up period, P2s average adherence rate wa s 73.2%. Physician rating of adherence for the follow-up period was 91%. 60

PAGE 61

Visual insp ection Magnitude and rate of change across phases were visually in spected in Figure 5. Changes in the mean and level of performance across phase s were considered in evaluating magnitude of change. Mean adherence decreased from baseline (78.5%) to treatment (57.7%) but increased at follow-up (73.2%). As with P1, P2s level of pe rformance declined betw een the end of baseline monitoring and the beginning of treatment, sugge sting that the first intervention session (e.g., HIV education) did not produce a rapid shift in improving adherence. When problem solving was implemented in session two, adherence im proved to 86.7%, the highest adherence rate obtained thus far. However, increased barrier s coinciding with the start of school led to a decrease in adherence rates fo r sessions four through seven, br inging down the overall adherence rate for the treatment phase of th e study to below baseline levels. Overall, adherence data from the baseline peri od had a slight negative trend (slope =-0.01) with relatively good stability in data points. Treatment phase data were highly variable but also had a slight negative trend (sl ope=-0.05). This trend was conti nued at follow-up (slope=-0.05). With regard to latency, a change in adheren ce was seen following the session which introduced problem solving (session two). Thus, visual in spection suggests that the intervention was only temporarily effective in promoti ng behavior change but that this behavior was not sustained over time. Statistical analyses Autocorrelations from one to 10 lags were calculated to identify any serial dependency in the baseline and treatment phase study data prior to all statistical analyses. As seen in the correlogram and its corresponding table (Figure 6), associations betw een data points were greatest at one lag but did not significantly de viate from zero, suggesting that the data points 61

PAGE 62

could be treated as independent observations and conventional statis tical analyses could be used to exam ine differences between treatment phases. As seen in Table 3, kurtosis values from th e baseline and follow-up phases of the study were beyond acceptable limits, suggesting that th e data were not normally distributed and nonparametric statistics should be used to compare differences between phases. Significant differences were found between ba seline and treatment adherence rates (MannWhitney U=4, p<.05), suggesting that the average baseline adherence ra te was significantly greater than the average treatment adherence ra te. Treatment and followup adherence rates did not significantly differ (Mann-Whitney U=21.5, p>.05). Baseline and follow-up adherence data were significantly different (Mann-Whitney U=4, p>.05), with the average baseline rate being significantly greater than the follow-up period rate. Virologic functioning At the beginning of baseline monitoring, P2s viral load was 470. This number decreased to 77 by the end of the baseline period. At posttreatment, P2s viral load was 59. By the end of the follow-up period, P2 had an undetectable viral load, representing a clin ically significant (1 log10) decrease from baseline to follow-up. According to medical records, P2 had not been undetectable for over two years prior to enrolling in the study. Knowledge of medication regimen On the Treatment Interview Protocol complete d at the end of the baseline period, P2 was able to correctly name 100% of her prescribed antiretroviral medications. She was able to provide her prescribed dosing fr equencies but not her dosing amount s. At post-treatment, P2 was able to correctly name 100% of her medicati ons and 100% of her dosing frequencies. Though she was unable to report her dos ing amounts, P2 was able to distinguish medications which required two pills per dosing from those requiring one pill per dosing. 62

PAGE 63

Barriers to adherence During the initial assessm ent, P2 reported a num ber of barriers to a dherence over the past month. In total, P2s barrier severity/frequency score at baseline was 13. Barriers endorsed as being hardly ever a problem (1-2 times per month) included: not getting me dications refilled in time, too busy with school, and worried others would find out about HIV. Barriers endorsed as frequent problems (1-2 times per week) include d: forgetting, couldnt deal with it/needed a break, didnt think I needed me dications anymore, dont remember if dose was already taken, and falling asleep. At post-treatment, P2 repor ted fewer barriers to adherence. Her barrier severity/frequency score at post-treatment was five. Parent-child conflict P2s score on the Conflict Behavior Questi onnaire (CBQ) at the be ginning of treatment suggested that her response pattern was similar to that of distressed families (CBQ=6). Her score on the CBQ at post-treatment was four, midway be tween distressed and non-distressed families. Caregiver report of conflict at the beginning of treatment was similar that of nondistressed families (CBQ=5). At the end of treatment, her response pattern indicated increased levels of conflict more similar to that of distressed families (CBQ=8). Participant 3 Background information Participant 3 (P3) was a 13-year-old African -American female living with her biological father. Annual family income was between $5,000 and $9,999. P3 was diagnosed with HIV at birth and was informed of her diagnosis at age 11 Her prescribed treatment regimen consisted of two NRTIs (Viread 300 mg every morning; Epzi com 1 tablet every morning) and two PIs (Norvir 100mg every 12 hours; Lexiva 700 mg every 12 hours). Per the clinic social worker, who had been conducting pills counts prior to study enrollm ent, P3s average adherence was 40%. 63

PAGE 64

Though physician estimate of adherence was 73% at the time of study enrollm ent, her physician described P3s adherence as erratic. Construction of composite adherence score Per the suggestion of her physician, Lexiva was selected as the medication to be monitored electronically throughout the enti re study as this medication ha d one of the greatest dosing frequencies. Adherence data as measured by MEMS, pill count, and self-report for Lexiva are presented in Figure 7. As seen in Figure 7, se lf-reported adherence was usually higher than adherence as measured by pill counts or MEMS. ME MS data were very strongly associated with pill count data (77%). Self-repor t was moderately correlated w ith MEMS (27%) and pill counts (33%). Given the very strong correlation between ME MS cap and other pill counts data, MEMS was used as the foundation for CAS for all but one of the data points (session five). At this time, P3 was hospitalized for 1.5 days and had her me dication administered to her by hospital nurses. Because the hospital required that P3s medi cations come from their own supplies while hospitalized, MEMS and pill count (which are ba sed off of P3s personal medication supply) could not be used. Self-report was used for se ssion five data. This data, however, was only slightly greater than MEMS and pill count estimates. Treatment and adherence Composite adherence data for P3 are presen ted in Figure 8. During the baseline monitoring period, adherence was 100% for the first two weeks of monitoring. Adherence then dropped by 40% for week three. Suggesting that the effect s of being monitored were less powerful. After four weeks of monitoring, adherence became less erratic. This coincided with increased parent involvement and awareness of changes in adheren ce rates from week two to week three. Overall 64

PAGE 65

adherence for the baseline m onitoring period wa s 86%. Physician estimate for the baseline period was 90%. Because P3 often reported 100% adherence ra tes that were not supported by MEMS and pill count data, very few reasons for missing a dose were listed on baseline self-report adherence sheets. Discussions with P3 and her father about general reasons for missi ng a dose revealed that forgetting and falling asleep were the most commonly, though rare ly, encountered barriers to adherence. Following session one, which focused on HIV education and improved knowledge of the medical regimen, adherence improved to 100%. Session two focused on using problem solving to identify methods to ensure that P3 took her evening medications before falling asleep. Solutions agreed upon by both P3 and her father included having P3s father place all of her medications in a pill box (minus Lexiva which continued to be monitored by MEMS) and having her father verify with P3 that she had taken her medication by 11 pm by checking her pill box for any remaining doses. Adherence remained high following sessions two and three (92.9% and 93.8%, respectively). During session four, which focused on family communication, P3s father expressed interest in having P3 assume mo re responsibility for her treatment regimen. Specifically, he requested that P3 be in charge of pre-filling he r pill box every week and that she assume more responsibility for making sure her morning medications were taken. P3 expressed interest in assuming more respons ibility as she felt her father sometimes nagged her about taking her medicine. Because this often led to increased conflict, both P3 and her father agreed to shift responsibility for filling out pill boxe s to P3. As part of the agreement, P3s father also promised not to nag P3 for the next week about taking her medicine. Following implementation of this plan, P3s adherence experienced a slight decrease ( 87.5%). Further decreases in adherence were 65

PAGE 66

seen during week five (83.3%). During session six, this trend was brought to the attention of P3 and her father. P3s father vowed to resum e responsibility to P3 taking her medication by reimplementing the adherence plan agreed upon in session two. Despite this, adherence continued to decline. Disruptions in family rou tines during sessions six an d seven coincided with the Thanksgiving holiday and increased work hours for P3s father. P3s overall adherence for the treatment portion of the study was 86.2%. P hysician rating for the treatment phase of the study was 80%. During the follow-up period, P3 maintained an average adherence rate of 92.5%. Physician rating of adherence for this time period was 92%. Visual inspection The magnitude of change and rate of change across phases were vi sually inspected in Figure 8 to determine if change occurred due to the intervention. In evaluating magnitude of change, changes in the mean and level of pe rformance across phases were considered. Mean adherence remained constant from baseline to treatment (86% to 86.2%) but improved to 92.5% at follow-up. Level of performance increased between the end of baseline monitoring and the beginning of treatment, suggesting that sess ion one produced a rapid shift in improving adherence. Rate of change was examined by inspecting th e trend and latency of change. Adherence data from the baseline period had a neutral trend (slope = 0), with increased stabilization of data points following week four. Treatment phase data i llustrate a slight negative trend (slope = -0.05) over time. Follow-up data illustrate a very slight positive trend (slope = 0.03). With regard to latency, adherence improved following session on e but gradually declined over time. Thus, visual inspection suggests that the intervention ha d slight effectiveness in promoting sustainable behavior change. 66

PAGE 67

Statistical analyses Autocorrelations from one to 12 lags were calculated to identify any serial dependency in the baseline and treatment phase study data prior to all statistical analyses. As seen in the correlogram and its corresponding table (Figure 9), associations be tween data points for all lags were small, suggesting that the autocorrelations did not significantly deviate from zero and that the data points could be treated as independent observations. Table 3 presents descriptive statistics of the CAS during the baseline, treatment, and follow-up phases of the study. As seen in Table 3, kurtosis values from the baseline period and skewness and kurtosis values from the fo llow-up period were beyond acceptable limits, suggesting that the data were not normally dist ributed. As such, non-parametric statistics were used to compare differences between phases. No statistically sign ificant differences in adherence were found between baseline and treatment (Mann-Whitney U=35, p>.05). Treatment adherence and follow-up adherence did not significantly differ (Mann-Whitney U=8.5, p>.05). Ba seline and follow-up adherence data also did not statistically differ (Mann-Whitney U=12, p>.05). Virologic functioning P3 had a viral load of 48 at the beginning of baseline monitoring. When treatment was initiated, laboratory results suggested that P3s viral load had increased to 220. At the end of treatment, Participant 3s viral load had d ecreased to below pre-treatment levels (68), representing a 31% decrease in vi ral load. At follow-up, P3s viral load increas ed to 820. Though an increase was noted in viral load from post-treatment to follow-up, this increase did not represent a clinically significant change. 67

PAGE 68

Know ledge of medication regimen On the Treatment Interview Protocol complete d at the end of the baseline period, P3 was able to correctly provide 100% of her prescr ibed antiretroviral medications, 100% of her prescribed dosing frequencies, and 0% of her prescribed dosing amounts. Knowledge of medication name, dosing frequency, or dosing amount did not change from preto posttreatment. Barriers to adherence During the initial assessment, P3s barrier severity score was six, with forgetting, not having help remembering, being busy, not rememb ering if dose was already taken, and needing to take a break being the most commonly re ported barriers. Although P3 reported a fewer number of overall barriers at post-treatment, the barriers endorsed (i.e., forgetting, not remembering if dose was taken, and falling asleep ) were endorsed as being a problem with greater frequency (barrier severity score= 6). Parent-child conflict P3s score on the Conflict Behavior Questi onnaire (CBQ) at the be ginning of treatment suggested that her response pattern was simila r to that of distressed families (CBQ=9). Though her score on the CBQ at post-treatment decreased to six, this level of conflict remained more similar to distressed families than nondistressed families. Caregiver report of conflict at the beginning of treatment was similar that of nondistressed families (CBQ=4). However, at the end of treatmen t, parent report of conflict had increased to be more similar to that of distressed families (CBQ=11). This may have been due to increased opportunities for conflict as P3 and her father were spending more time together as a result of the intervention. 68

PAGE 69

Partic ipant 4 Background information Participant 4 (P4) was a 17-year-old Caucas ian male who had been living in the legal custody of his older sister for the past six y ears. Annual family income was reported to fall between $50,000 to $59,999. P4 was diagnosed with HIV at birth and was informed of his diagnosis at age seven. His pres cribed treatment regimen consisted of two NRTIs (Viread 300 mg once a day; Epzicom 1 tablet once a day) and two PIs (Norvir 100mg once a day; Reyataz 300 mg once a day), all of which were taken in the morning. P4 had a history of poor adherence to his medical regimen, with particular problem s adhering to dosings during the summer break and on weekends. Physician estimate of adherence for the month pr ior to enrolling in the study was 24%. Construction of composite adherence score Based on physician recommendation, Reyataz wa s selected as the medication to be electronically monitored. Adhere nce data as measured by MEMS and self-report for Reyataz are presented in Figure 10. Pill count estimates could not be reliably obtained for P4 as part of his medications were kept at school and part were kept at home. As an additional complication, unknown refill amounts were distributed to the school and P4 at unknown times by the clinic social worker, further complicating pill count ad herence computations. The barriers encountered in obtaining reliable pill count s did not affect MEMS measurements as two MEMS caps were used (one at school, one at home) to track P4s adherence. As seen in Figure 10, self-report estimates were generally higher than MEMS esti mates though they were strongly correlated with one another (70%). MEMS data were used as the foundation for CAS for all data points. 69

PAGE 70

Treatment and adherence Com posite adherence data for P4 are presented in Figure 11. During the baseline monitoring period, P4s adherence ranged from 57.1% to 87.5%, with an average adherence rate of 65.3%. Physician estimate of adherence for the baseline period was 80%. During the baseline monitoring period, b eing away from home and running out of medication were the most commonly reported reasons for missing a do se during the weekend. These barriers were identifie d as targets for intervention. Adherence dropped from 71.4% at the end of baseline monitoring to 33.3% in the week following the first treatment sess ion. In session 2, the five-ste p problem solving approach was used to improve P4s adherence to his weeke nd dosings. According to P4, primary reasons for missed doses included: 1) not being reminded to ta ke his medications, and 2) the remote location in which his medications were ke pt (in a kitchen cabinet that contained no other items). During the brainstorming portion of this session, numer ous ideas were generated to overcome this barrier, among which was the relocation of P4s me dications to a place in which he was highly likely to encounter them in the morning. In P4s case, this meant relocating his medications from the kitchen cabinet to a cabinet containing cereal boxes. Both P4 and his guardian agreed that this would be a more effective strategy as P4 always ate cereal in the morning. P4s guardian agreed to check in with P4 on weekends to ma ke sure he took his medication. To help P4s guardian remember to do this, she se t an alarm on her mobile phone. Immediately following implementation of this plan, P4s adherence rate improved to 100%. P4s adherence following session 3 was 85.7% (6 out of 7 doses taken according to MEMS). This estimate was lower than P4s self-r eport of 100% adherence due to P4s claim that he pocketed a dose over the weekend to avoid having to bri ng the MEMS cap to his friends house. However, because this c ould not be verified by pill co unt, MEMS data were used. 70

PAGE 71

Following session four, P4 ran out of m edications at home was unable to take his weekend doses, dropping his adherence down to 71.4%. Though P4s guardian was able to obtain more medicine by session five, this session coincided with the Thanksgiving holiday and P4 did not take his medications as regularly as he had been doing so when he had a routine in place. Session six focused on implementing the problem solving appro ach in order to help P4 overcome the barrier of running out of medication. Wh en providing her perception of the problem, P4s guardian reported that P4 often waited until he was out of medication in order to notify her. This usually happened on a weekend, a time in which P4s guard ian was unable to reach the clinic social worker to arrange more pills to be delivered, thereby leading to P4 to miss 29% of his weekly doses. Following the brainstorming and evaluation por tion of the session, P4 agreed to notify his guardian whenever he had less than three pills left of his medica tion. This number was chosen as it would provide P4 with enough medication to ma ke it through the weekend and would give his guardian five days in which to obtain additional supplies before the next weekend. To help remind P4 of this plan, a sticker containing the agreed upon plan and the name and number of the social worker was attached to his medication vial. Immediately following this session, adherence improved to 100%. Unfortunately, this was not maintained for session seven due to forgetting doses. Overall, P4s average adherence for the treatment portion of the session was 74.4%. Physician rating of adherence fo r the treatment period was 86%. P4s follow-up period adherence was characterize d by high levels of instability. Just prior to the beginning of the follow-up monitoring period, P4 and his family (six individuals in total) were displaced from their home by a fire that de stroyed all of their possessions. Throughout the follow-up monitoring period, P4 and his family liv ed in hotel rooms and with family friends. 71

PAGE 72

Lacking a stable daily routine, P4s adherence during the fo llow-up period dropped to 60.7%. Physician rating of adherence fo r the follow-up period was 73%. Visual insp ection Magnitude and rate of change across phases were visually in spected in Figure 11. Changes in the mean and level of performance across phase s were considered in evaluating magnitude of change. Mean adherence increased from baseline (65.3%) to treatment (74.4%) but decreased at follow-up (60.7%). As with P1and P2, P4s leve l of performance declined between the end of baseline monitoring and the beginning of treatme nt, suggesting that the first intervention session (e.g., HIV education) did not produce a rapid sh ift in improving adherence. When problem solving was implemented in session two, adhe rence improved to 100%, the highest adherence rate obtained up to that point. However, runnin g out of medication and th e disruptions in normal routine due to the Thanksgiving holi day led to a decrease in adhere nce rates for sessions four and five. These disruptions lowered the overall treatment adherence rate to 74.4%, though this was still above P4s baseline average. Overall, adherence data from the baseline peri od had a very slight negative trend (slope =0.01). Treatment phase data had a slight posi tive trend (slope=0.03). This trend was not continued at follow-up (slope=-0.01). With regard to latency, improvements in adherence were seen immediately following sessions two and si x, which had a major focus on problem solving. Overall, visual inspection sugge sts that the intervention was m odestly effective in promoting behavior change but that this beha vior was not sustained over time. Statistical analyses Autocorrelations from one to 11 lags were calculated to identify any serial dependency in the baseline and treatment phase study data prior to all statistical analyses. As seen in the correlogram and its corresponding table (Figure 12), associations between data points did not 72

PAGE 73

significantly deviate from zero, s uggesting that the data points c ould be treated as independent observations. As seen in Table 3, the kurtosis value from the baseline phase of the study was beyond acceptable limits. Thus, non-parametric statistics were used to compare differences between phases. No significant differences were found between baseline and treatment adherence rates (Mann-Whitney U=34, p>.05), suggesting that th e average baseline adhe rence rate was not significantly greater than the average treatm ent adherence rate. Treatment and follow-up adherence rates did not significantly differ (Mann-Whitney U=19, p>.05). Baseline and followup adherence data also did not significantly differ (Mann-Whitney U=19, p>.05). Virologic functioning At the beginning of baseline monitoring, P4s viral load was 56,000. By the end of the baseline monitoring period, his viral load had increased to 77,000. Unfo rtunately, this trend continued despite improved adherence. At post-treatment, P4s viral load was 190,000. Laboratory tests conducted at posttreatment indicated that P4 had developed high levels of resistance to all of his prescribed antiretrovi ral medications. By the end of the follow-up period, P4s viral load was 80,000. This change was not clinically significant. Knowledge of medication regimen On the Treatment Interview Protocol complete d at the end of the baseline period, P4 was able to correctly name 100% of his prescribed antiretroviral medications and 100% of his dosing frequencies. He was unable to correctly prov ide any of his dosing amounts. No changes in medication name, dosing frequencies, or dosing amounts were seen at post-treatment. 73

PAGE 74

Barriers to adherence P4 reported a num ber of barrier s to adherence during the initial assessment. Forgetting, not having family members help remind him, not remembering if dose was taken, falling asleep, and being away from home were identified as freque nt problems (1-2 times per week). Not getting medications refilled in time was identified as a less frequent problem (1-2 times per month). In total, P4s barrier severity scor e was 11 at pre-treatment. At pos t-treatment, P4 reported a fewer number of barriers to adherence (barrier severity score = 2). Parent-child conflict P4s score on the Conflict Behavior Questi onnaire (CBQ) at the be ginning of treatment suggested that his response pattern was similar to that of nondistressed families (CBQ=2). His score on the CBQ at post-treatment remained at two. Caregiver report of conflict at baseline and post-treatment were similar that of nondistressed families (CBQ=2 for both assessments). Treatment Satisfaction Out of a possible range of eight to thirty-two points, with eight bei ng very dissatisfied and thirty-two being very satisfied, adolescent average rating of the program was 30.25 (range=2632). Parent-reported program satisfaction was also very high (average ra ting=31; range=29-32). 74

PAGE 75

Table 3-1. D escriptive statistics of CAS data for all participants Participant 1 Phase N Mean SD Range Median Skewness Kurtosis Baseline 4 64.7 45 0-100 72.25 -1.60 2.66 Treatment 7 88.3 11.37 71.4-100 91.7 -0.61 -1.78 Follow-up 4 95 10 80-100 90 -2.00 4.00 Participant 2 Phase N Mean SD Range Median Skewness Kurtosis Baseline 4 78.5 3.47 75-83.3 77.9 1.03 2.09 Treatment 7 57.7 16.68 42.1-86.7 50 1.03 -0.15 Follow-up 4 73.2 9.14 66.7-86.7 69.6 1.86 3.56 Participant 3 Phase N Mean SD Range Median Skewness Kurtosis Baseline 10 86 11.4 60-100 85.2 -1.12 2.60 Treatment 7 86.2 10.4 71.4-100 87.5 -0.29 -1.15 Follow-up 4 92.5 15.0 70-100 100 -2.00 4.00 Participant 4 Phase N Mean SD Range Median Skewness Kurtosis Baseline 7 65.3 11.25 51.7-87.5 60 1.54 2.13 Treatment 8 74.4 23.49 33.3-100 71.4 -0.65 0.33 Follow-up 4 60.7 13.68 42.9-71.4 64.3 -0.85 -1.29 75

PAGE 76

76 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 1 00% B1B2B3B4S1S2S3SS5S6S7F1F2F3F4 Treatmet PhaseAdherence MEMS Pill Self-Report 4 n Figure 3-1. Com parison of individual adherence meas ures for Participant 1

PAGE 77

= Mean adh erence rate for treatment phase 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1B2B3B4S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence Figure 3-2. Com posite Adherence for Participant 1 77

PAGE 78

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 12345678910Lagr Time lag k ACF(k) p 1 -0.17 0.29 2 0.29 0.18 3 -0.03 0.46 4 0.05 0.43 5 -0.01 0.49 6 -0.07 0.41 7 -0.21 0.25 8 -0.06 0.42 9 -0.16 0.30 10 -0.14 0.33 Figure 3-3. Participan t 1s correlogram as a function of time lag indicating independence of adherence data points 78

PAGE 79

79 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1B2B3B4S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence MEMS Pill Self-Report Figure 3-4. Com parison of individual adherence meas ures for Participant 2

PAGE 80

80 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1B2B3B4S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence = Mean adh erence rate for treatment phase Figure 3-5. Com posite adhe rence for Participant 2

PAGE 81

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 12345678910Lag r Time lag k ACF(k) p 1 0.49 0.07 2 0.22 0.24 3 0.17 0.29 4 -0.20 0.27 5 -0.10 0.38 6 -0.18 0.29 7 -0.37 0.12 8 -0.26 0.21 9 -0.19 0.27 10 -0.10 0.36 Figure 3-6. Participan t 2s correlogram as a function of time lag indicating independence of adherence data points 81

PAGE 82

82 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1B2B3B4B5B6B7B8B9B10S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence MEMS Pill Self-Report Figure 3-7. Com parison of individual adherence meas ures for Participant 3

PAGE 83

83 0% 10% 2 0 % 0% 0% 0% 0% % 80% 90% 10% B1B2B3B4B5B6B7B8B9B10S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence0 3 4 5 6 70 = Mean adherence rate for treatment phase Figure 3-8. Com posite adhe rence for Participant 3

PAGE 84

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 123456789101112Lag r Time lag k ACF(k) p 1 0.11 0.33 2 -0.17 0.25 3 0.02 0.48 4 -0.15 0.28 5 -0.27 0.14 6 -0.01 0.49 7 -0.06 0.41 8 -0.22 0.18 9 -0.02 0.47 10 0.08 0.38 11 0.12 0.32 12 0.12 0.31 Figure 3-9. Participan t 3s correlogram as a function of time lag indicating independence of adherence data points 84

PAGE 85

85 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1 B2B3B4B5B6B7S1S2S3S4S5S6S7F1F2F3F4 Treatment PhaseAdherence MEMS Self-Report Figure 3-10. Com parison of individual adherence measures for Participant 4

PAGE 86

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B1B2B3B4B5B6B7S1S2S3S4S5S6S7F1F2F3F4Treatment PhaseAdherence Figure 3-11. Com posite adherence for Participant 4 86

PAGE 87

-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1234567891011Lag r Time lag k ACF(k) p 1 -0.28 0.16 2 0.00 0.50 3 0.04 0.44 4 0.12 0.33 5 -0.44 0.06 6 -0.04 0.45 7 0.10 0.36 8 -0.01 0.49 9 -0.05 0.43 10 0.01 0.49 11 0.14 0.30 Figure 3-12. Participant 4s co rrelogram as a function of time lag indicating independence of adherence data points 87

PAGE 88

CHAP TER 4 DISCUSSION The purpose of the current study was to examine the effect of an intervention to improve antiretroviral medication adherence among adoles cents with HIV. The current study expanded upon the limited extant literatu re through the use of a multi-method adherence approach, the inclusion of laboratory data at each assessmen t, and its exclusive focus on adolescents, a population known to have problems with adherence. All four adolescents who received the inte rvention had a history of poor compliance to their treatment regimen prior to enrolling in the study. When compared to pre-enrollment adherence information as reported by either the medical team, the patient, pre-study enrollment pill counts, or pharmacy records, all participants demonstrated im proved adherence by the end of treatment. However, because improvements in adherence varied across treatment phases and across participants, it is difficult to draw conclu sions regarding the overall effectiveness of the treatment program. Closer examination of i ndividual treatment components and changes in adolescent adherence may be helpful in iden tifying possible active treatment components. Possible Active Treatment Variables Several aspects of the treatment must be cons idered as potentially influencing adherence levels among participating adolescents. As part of this study, adherenc e was closely monitored through the use of electronic monitoring caps, pi ll counts, weekly self-report, and weekly home visits. This in-depth level of m onitoring is significantly greater than what is typically employed in standard clinical care and raises the possibili ty that being monitored may have contributed to increased rates of adherence. Within the greater intervention literature, res earchers have questioned the extent to which monitoring, specifically, electroni c monitoring, may affect patien t pill taking behavior. Findings 88

PAGE 89

in the HIV literature regarding a m onitoring in tervention effect are mixed and limited only to adult populations. While one research study docum ented a monitoring intervention effect that waned over a 40-day period (Deschamps, Van Wijngaerden, Denhaerynck, Vandamme, & De Geest, 2006), others have found that electronic monitoring without consequences has not been found to alter adherence in patients w ith HIV (Wagner & Ghosh-Dastidar, 2002). During the baseline monitoring period of th is study, adherence data were discretely obtained and no consequences were implemented so that the effect of being monitored on patient adherence could be minimized. De spite this, it is still possible that being monitored may have impacted adherence. Taking patient history and baseline monitoring data for P1 and P2 into account, some evidence exists for monitoring alone as an active treatmen t component. Prior to enrolling in the study, P1 had not had her medica tions refilled in over three months while P2 openly admitted to being completely non-adherent for a two-month period. However, during the baseline monitoring period, both P1 and P2 had adherence rates greater than what would be expected given their history. Changes in physician estimate of adherence fr om preto post-bas eline monitoring may also provide support for monitoring-alone as an adherence improving st rategy. Assuming that monitoring had no effect on adhere nce, it would be expected that physician rating of adherence, if accurate, would remain the same at both the beginning and end of this period. However, across all participants, physician rating of adherence increased by an av erage of 40 percentage points, suggesting that physicians perc eived adherence to improve during the baseline monitoring period. This change occurred despit e that fact that physicians we re aware that the intervention program had not yet been delivered. 89

PAGE 90

Though m onitoring was intended as an assessment t ool in the current st udy, it is likely that monitoring contributed to some extent to improved medication adherence by increasing adolescent awareness. In order to obtain a better understandi ng of how monitoring-alone may impact adolescent treatment adherence, the incl usion of a monitoring-only comparison group is needed in future treatment designs. Caregiver involvement is another potential mech anism of change. Prior to enrolling in the study, caregivers of participating adolescents ap peared to be less involved in their childs treatment regimen. As part of the program, careg ivers and teens met on a weekly basis with the intervention team to discuss the teens adherence. It is possible that these weekly meetings led to greater caregiver awareness and involvement in the teens treatment re gimen. Among two of the families enrolled in this study (P3 and P4), car egivers spontaneously began keeping their own pill counts to obtain a better understanding of thei r childs adherence. Increased awareness of problems with adherence may have led to addi tional involvement. This may have provided additional support for the adolescent to take their medication. Because greater medication knowledge has been associated with higher rate s of adherence among children and adolescents with HIV (Martin et al., 2007), the first session of the trea tment program was designed to improve patient knowledge of th eir illness, their medication regimen, and the importance of adherence. It was expected that such an impr oved understanding would pr ovide adolescents with additional motivation to adhere to their regi men. Unfortunately, data suggest that among the three oldest adolescents in the study, adherence rates decreased rather than increased following this session. Though this drop in adherence may be due to extraneous factor s, it is also possible that the educational component may have played a role. Adolescents often hear about the importance and consequences of being non-adherent at every medical appointment. They also 90

PAGE 91

likely hear this inform ation on a regular basis from their family when they miss doses. Though this information is given with good intentions, ad olescents may consider this to be a form of lecturing and may react negatively. Drops in ad herence following the educational session raises the possibility that the educational component in this treatment may have had unintended consequences. Obtaining adolescent feedback on educational approaches to improve adherence may provide valuable information on the efficacy of this approach in future interventions. Comparison of an education-only approach to a more practical medication management approach (such as problem solving-only or educa tion + problem solving) ma y help to clarify the effect of education on adolescent adherence. A major component of the intervention program was the use of problem solving to help adolescents and their caregivers overcome commonly experienced ba rriers to adherence. Using a structured approach, it was expected that caregivers and teens would be able to move beyond failed attempts to improve adherence and arrive at innovative and effective solutions. Adherence improved for all adolescents following the implem entation of a plan agreed upon during problem solving. Successful solutions ge nerally involved finding a way in which the medication regimen could be incorporated into the adolescents dail y routine rather than getting the adolescents routine to bend to the demands of the treatment regimen. Such practical medication management approaches have been associated with improve ments in adherence among adults (Rueda et al., 2006). In the case of P1 and P4, this meant phys ically moving their medications to a place in which the adolescent would likely come across th eir medication while engaging in their regular routine. Unfortunately, the long-te rm effectiveness of the plans implemented were limited by the extent to which adolescent rout ines maintained stability. 91

PAGE 92

As seen in the data for P 2, P3, and P4, ch anges in the adolescen ts routine, upon which previously implemented plans were based, led to decrea ses in adherence. For P2, this change occurred with the commencement of school, de manding after school activ ities, and increased parent work responsibilities. Fo r P3 and P4, school breaks/holiday s and a destructive home fire (P4) interfered with their normal routines. When such changes occurred, participants tended to deviate from their previously established plans and adherence decreased. These patterns are consistent with prior research s howing that changes in daily routin e are associated with decreases in adherence among adolescents (Murphy et al ., 2003) and highlight a critical area for intervention. Within the immediat e context of the intervention problem solving should be tailored to help adolescents plan ahead for how to handle expected changes in routine so that the extent to which such changes negatively imp act adherence are minimized. Additionally, to maximize the likelihood that behavior changes ar e maintained over time, interventions should focus on helping caregivers and adolescents intern alize the problem solving framework so that they can implement this approach on their ow n after the conclusion of the intervention. Unlike prior research documenting an associ ation between lower so cioeconomic status (SES) and poorer adherence rates in individuals with HIV (for a review, see Mehta, Moore, & Graham, 1997), no clear relationship between these two variables were seen in the current study. In the current sample, caregiv er report of annual family income varied widely. Of all participating adolescents, P1 (who had an annual family income ranging from $20,000 to $29,999) demonstrated the greatest improvement in adherence throughout the course of the study. Conversely P4, whose annual family inco me was the highest and fell within the $50,000 to $59,999 range, did not demonstrate such impr ovements in adherence. Prior research documenting the link between lower SES and poorer adherence has cited a lack of resources 92

PAGE 93

(e.g., reliable transportation to a ttend m edical appointments, adequate medical insurance to afford medications) as a primary contributor to the SES/adherence relationship. In the current study, the impact of such barriers may have been minimized through the delivery of sessions within the home and the provision of no-cost medication coverage through state-sponsored insurance programs offered to all children and adolescents with HI V. Given the limited research on adherence in adolescents with HIV, further res earch is needed to better elucidate any existing relationship between SES and adherence in this population. Changes in Viral Load, Barriers to Adh erence, Medication Knowledge, and Family Conflict In addition to examining the impact of the intervention on adolescen t adherence, the study also aimed to reduce patient viral load. Overall, reductions in viral load were seen for three out of the four participants in the study. P1 and P2s viral load decreas ed from preto post-treatment, with further decreases seen at follow-up. P3 experi enced a decline in viral load from preto posttreatment but experienced an in crease in viral load at followup despite a 92.5% adherence rate. Though adherence is a predictor of viral load, research has dem onstrated that fully adherent patients do not always achieve or maintain viral load suppression (Williams et al., 2006). This may be due to individual variability in medi cation resistance and patient immune functioning (such as illness). Conversely, patients with less than optim al adherence (below 95%) may achieve an undetectable viral load if their pa rticular viral strain remains sensitive to their treatment regimen. Prior development of resistan ce and current immune functioning may impact the extent to which improved adherence leads to changes in vira l load. Medication resist ance may have played a role in P4s viral load throughout treatment. De spite improved adherence rates, P4s viral load continued to increase throughout the intervention. This may have been due to high levels of 93

PAGE 94

m edication resistance and suppressed immune fu nctioning. Though recent testing indicated that P4s virus was highly resistant to all of his prescribed regimens it must be noted P4s improved adherence still fell well below the optimal adherence level of 95%. Thus, it is possible that a combination of less than optimal adheren ce, immune suppression, and high medication resistance may have played a role in P4 s increasing viral load throughout the study. A third aim of the current study was to reduce the severity of reporte d adolescent barriers to adhering to their medication regimen. Overall, every participant report ed a reduction in their barriers to adherence. Prior to beginning treatment, adolescents reported experiencing a number of different barriers, with forgetting, not re membering if dose was taken, not getting refills in time, falling asleep, and being away from home being the most commonly reported. These results are consistent with research doc umenting that non-adhere nt adolescents tend to experience barrier clusters rather than one single barrier to adhe rence (Rudy, Murphy, Harris, Muenz, & Ellen, 2009). As such, past attempts to improve adherence using a single strategy may not have adequately addressed the numerous ba rriers adolescents experi ence. Through the use of problem solving, these barrier clusters were di sentangled and addresse d individually. At posttreatment, the overall number of barriers reported and the severity with which these barriers interfered with adhe rence were reduced. Despite reductions in barriers to adherence, adolescent knowledge of their medication regimen did not appear to improve from preto post-treatment. This may have been due to the high level of knowledge already possessed at the beginning of the study. All adolescents were able to accurately name their medications and their dosing schedule. Though participants were unable to provide specific dosing amounts, they were able to identify how many pills were 94

PAGE 95

required for each dose. T hus, adolescents appeared to possess sufficient knowledge of their medication regimen in order to adhere appropriately. Future research may benefit from a broa der assessment of adolescent HIV knowledge including knowledge of the imm unologic and virologic benefits of adherence, the consequences of poor adherence, different routes of transmi ssion, the types (and limited number of) treatment options, etc. In addition to allowing for a mo re thorough assessment of adolescent knowledge from pre-to post-treatment, such an assessment would provide valuable information that can be used to tailor the delivery of any HIV educational component. A final aim of the current study was to examin e changes in family c onflict. Three out of the four adolescents reported a decrease in confli ct from preto post-treatment. One participant (P4) reported very low levels of conflict at pretreatment that were unchanged at post-treatment. Though teens reported either no chan ge or a decline in conflict, three caregivers reported an increase in conflict from preto post-tre atment. One caregiver reported no change (P4s guardian). Caregiver report of in creased conflict may have been re lated to increases in caregiver involvement with regard to th e adolescents treatment regimen. As part of the intervention, adults met with their teen on a weekly basis to discuss their adherence. This may have provided increased opportunities for conflict. However, given that only car egiver, and not teen, report of conflict increased over time, this hypothesis is unlikely. Another possibility is that th e intervention had a differential impact on caregiver and teen perception of conflict. As part of the intervention, maladaptiv e patterns of communication were immediately interrupted by the therapist, who assisted in helping families communicate their thoughts, opinions, and feelings in a more effec tive manner. Overall, participating caregivers were more likely to engage in negative communi cation strategies than teenagers. This often 95

PAGE 96

occurred in the for m of a lecture or interrup ting/talking over the adolescent. Intervening when caregivers engaged in these negative forms of communication may have been viewed positively by the adolescent, as they were spared being lect ured or nagged and may have gained additional opportunities to express themselves. This may have led to perceptions of decreased conflict over time. Caregivers, on the other hand, may have pe rceived this aspect of the intervention differently. Although they were guided in expre ssing thoughts previously communicated in the form of nagging or lecturing in a manner in which the adolescent woul d be more responsive, caregivers may have felt that this aspect of the in tervention limited the extent to which they were able to express themselves. This may have led to caregiver perceptions of increased conflict. Another possibility is that ch anges in family conflict levels as measured by the Conflict Behavior Questionnaire, were unrelated to the study. Instead, th ese changes might have been reflective of the natural ebb and flow of conflict levels that occur within a caregiver-adolescent relationship over time (Robin & Foster, 1989). Because the current study employed a global measure of family conflict, it is difficult to de termine to what extent conflict levels were associated with the adolescents adherence or with other factors. Use of a specific adherence conflict measure may help to clarify this issue in future research. Although a family conflict measure specific to HIV-related care does not curre ntly exist to our knowledge, disease-specific family conflict measures developed among othe r populations, such as the Diabetes Family Conflict Scale (Rubin, Young-Hyman, & Peyrot, 1989) may help guide the development of a similar measure for use with individuals with HIV. Utility of Multi-Method Adherence Assessment The use of a multi-method adherence approach was helpful in obtaining a comprehensive evaluation of the adolescents medication taking behavior. MEMS provided valuable information regarding adherence patterns that otherwise would not have been detected with pill count or by 96

PAGE 97

self -report, which is subject to social desirability. MEMS data helped guide the intervention to target problematic trends in medication taki ng. In the case of P4, this meant focusing on developing strategies to improve adherence to weekend doses a nd developing a plan to order refills in a timely manner. For P2, this meant better regulating her su mmer break routine to improve morning medication adherence. Though overall MEMS were considered a valuab le addition to the study, the use of MEMS had its drawbacks. By nature of its programming MEMS caps did not record any vial opening of less than five seconds duration or any openi ng within a 15-minute time-frame of another opening. These unchangeable program features may have led to an underestimation of adherence. This was likely the cas e for P1 who reported that she of ten forgot to keep her bottle open for more than five seconds. MEMS may have also underestimated adherence by interfering with previously established adherence strategies such as pill boxes and the pocketing of doses. Due to the cumbersome size of the MEMS cap, pa tients may have opted to remove their doses from the bottle ahead of time to avoid having to carry their bottles with them. This occurred several times with P4, who preferred to transpor t his medications discretely when staying with friends on the weekend, possibly leading to an underestimation of adherence. An additional barrier to the use of MEMS relate s to the reliability of the data. MEMS data can be compromised by caps malfunctioning or being lost. Though all participants were instructed in proper MEMS care, MEMS were sometimes dropped onto hard surfaces (as with P1). In the current study, approximately 23.8% of MEMS observations were discarded due to suspected validity issues. However, fewer substitu tions were made in this study compared to Liu et al. (2001), who discarded appr oximately 40% of all MEMS obser vations in construction of his widely used composite adherence score. 97

PAGE 98

Because of the aforem entioned limitations with MEMS, pill counts were used as a back-up adherence assessment strategy. Though pill count estimated adherence was inexpensive and generally easy to compute, barriers encountered with this method included working with patients who had more than one supply of the same medication and unexp ected increases, rather than decreases, in pill count. While these issues were generally resolv ed by helping patients consolidate their different sources of pills and referring to pharmacy refill data, numerous factors affected the ability to calculate pill counts with P4. Fortunately, all of P4s MEMS data were considered to be valid thereby pr eventing the need to rely on self -report in the absence of MEMS and pill count data. Self-report data collected also had strengths and weaknesses. A primary reason for the inclusion of self-report was that these data were generally very easy and inexpensive to obtain. Additionally, because participants were asked to provide a reason for each missed dose, selfreport data provided valuable information on barrier s to adherence that would not otherwise have been obtained if only MEMS and pill counts were us ed. This information played a crucial role in informing the intervention. However, self-report data were considered a last-resort strategy in calculating a composite adherence score due to it s tendency to consistently provide higher rates of adherence than MEMS or pill count estimates. As discussed above, the use of a multi-method adherence approach provided many benefits to the design of treatment. An additional advantage to the use of this approach was the elimination of missing data through the constr uction of a composite adherence score. High correlations between MEMS, pill counts, and self -report data, along with prior research on the use of a multi-method adherence approach (Liu et al., 2001), supported the use of this substitution approach in constructin g a composite adherence score. 98

PAGE 99

Limita tions Although the current study is a promising and innovative attempt to improve adherence among adolescents with a longstanding history of less than optimal a dherence to their HIV treatment regimen, several limitations prohibit the extent to which definitive conclusions can be drawn regarding the effectivene ss of the intervention. A few limitations have been discussed in context of the discussion above. The following are additional limitations and suggestions for further investigation. First, by nature of the single-case experi mental design, which lacks a control group and randomization to treatment, it is not possible to confid ently conclude that the changes observed in adolescent adherence were due to the intervention or were a re sult of extraneous variables. Given the small sample size, it is not possible to determine if the treatment or response pattern obtained can be generalized to the greater population of adolescen ts with HIV, including those adolescents who were behavi orally-infected. Unfortunatel y, given the small number of adolescents with HIV who also have problems with adherence and are living in any one area, the feasibility of conducting a randomized clinical trial with sufficient statistical power and generalizability is somewhat limited unless multi-site collaborations are established. An additional limitation with regard to the study design was a deviation from the nonconcurrent multiple baseline design. Because of a scheduling conflict, one participant began the intervention out of sequence. Thus, instead of having participants with four, six, eight, and 10week baseline periods, two participants had a fou r-week baseline and no participant had an eightweek baseline. This deviation from the multiple baseline design limited the extent to which changes in adherence could be compared across participants. A third limitation relates to the collection of data through th e delivery of the intervention via alternating home and te lephone sessions. Though this desi gn was chosen in order to 99

PAGE 100

m inimize treatment burden experienced by families the extent to which adherence could be accurately assessed during telephone session weeks was somewhat limited. Because the intervention was guided by adhere nce data obtained, and the on ly adherence data collected during telephone sessions was a dolescent self-report, data coll ected during telephone sessions was often an over-report of adhere nce. Thus, actual slips in adhe rence were often not detected until MEMS data were downloaded at the following home session. By this time, up to two weeks may have passed and opportunities for early intervention were lost In the current study, MEMS data were collected via a singl e personal laptop computer bro ught to each pa rticipants home during home treatment sessions. In a study with greater levels of financia l support, it may be possible to equip each participant with a lapt op, MEMS transponder, and the proper computer software to enable them to upl oad their MEMS data each week. This would allow researchers to more quickly detect slips in adherence a nd may allow for more immediate intervention. An additional limitation of the study design re lates to the timing of the initiation of treatment. According to study cr iteria, treatment was initiated after two data points in which adherence was not improving were obtained. Once th e decision was made to initiate treatment, the patient was scheduled for an in-clinic appoi ntment. Due to the busy nature of the medical clinic, same-day appointments were not possible. All patients had to be scheduled at least one week in advance. As a result, the baseline mon itoring period continued for one additional week after treatment initiation criteria had been met. For three out of the four adolescents, adherence improved during this week. These improvements may have been reflective of the general trend for adherence to improve just prior to havi ng a medical appointment (Farmer, 1999). Thus, temporal social demands may have played a role in improvements in adherence. 100

PAGE 101

Because of the high expense associated with ME MS, the current study was limited to only having one medication per participant electr onically monitored. Though many studies have employed a one MEMS cap per person approach in the HIV literature as a money-saving strategy, this limits the extent to which diffe rential adherence across medications can be monitored. In the current study, pill counts were used to exam ine potential differences in adherence across medications. Overall, similar adherence rates were noted across medications with dosing schedules the same as the medi cation that was monitored with a MEMS cap. Because the medication selected for electronic monitoring was often the one with the more demanding dosing schedule, it wa s expected that medications with less demanding schedules would have greater levels of adherence. Pill c ount data for other medications appear to support this hypothesis. Strengths Despite the aforementioned limitations, the study possessed several strengths that may help guide the development of future researc h. Through its exclusive focus on adolescents, the current study addressed a critical gap in the research l iterature. Because adolescence is a time in which lifelong positive and risky health behaviors are established (Holmbeck, 2002), interventions targeting adolescents who have poor adherenc e to their HIV regimen may have direct implications for improvements in patient quality of life and longevity. As previously discussed, the use of a mu lti-method adherence approach significantly enhanced data collection. This approach allowed for the collection of a wealth of information that would not have been otherwise obtained had the single-method approaches used in prior studies been employed. In addition to minimizing missing data, MEMS, pill count, and selfreport provided valuable information regard ing adherence patterns and reasons for nonadherence. These data allowed for a more individualized treatment approach. 101

PAGE 102

An additional strength of the current study is the use of problem solving to tailor the intervention to address the unique barriers to adherence experienced by the adolescent and their family. Through use of this individualized appro ach, parents and teens were able to generate innovative strategies to help them overcome l ongstanding barriers. This individualized approach may have contributed to the high level of program satisfaction reported by all participants and the absence of participan t attrition. Given that participants we re enrolled in the study for at least six months, the lack of attrition is a significant strength of the current study. The studys innovative design, which allowed for the delivery of treatment across home, clinic, and telephone settings, is also a strength. Compared to costly prior multisystemic therapy (MST) approaches which have employed two to three home visits per week across several months (Cunningham et al., 2006; Ellis et al., 2006), the current study was less intrusive to study participants and required fewer resources on be half of the intervention team. Given that improvements in adherence were observed among all study participants to varying degrees, it is possible that interventions util izing alternating weekly home and telephone sessions may be a more cost-effective alternative to MST approach es. Further research is needed, however, to determine how the current treatment approach w ould compare to MST or a no-treatment control group. The studys high level of integration within the clinical setting is also a significant strength. Compared to prior inte rvention research which has inte rfered with the delivery of clinical services (Rogers et al., 2001), the current interv ention was designed to be minimally obtrusive yet effective in data collection. The strong collaborative relations hip existing between psychology and the medical team provided a strong foundation upon which the study was designed and allowed for the colle ction of laboratory data and phys ician rating of adherence at 102

PAGE 103

103 each major assessment. Given that no known intervention study for children and adolescents has been able to collect and obse rve changes in viral load and physician rating of adherence over time, the inclusion of these data in the current study is a significant strength. Conclusions and Future Directions The current study is a promising beginning to what is hoped to be a growing area of research. Limitations of the current study, as well as with prior researc h, are reflective of the general challenges researchers experience when wo rking with adolescents with HIV. Helping the adolescent find ways to integrate their medical regimen within their regula r routine appears to be an effective adherence improving strategy. Howeve r, this is only eff ective as long as the adolescents routine remains stab le. Using problem solving to pr oactively plan for changes in routine may help to minimize declines in adherence. Despite the many challenges associated w ith intervention research among adolescents with HIV, it is important that invention research with this population continue. Multi-site research collaborations are greatly needed to allow for the design of randomi zed clinical trials to more clearly investigate treatment outcome. Multi-method adherence assessment approaches, along with data on patient physiological functioning, should also be included in future research as these data provide valuable information th at can guide treatment design and inform the interpretation of treatment outcome data.

PAGE 104

APPENDIX A SAMPLE GRAPH OF PATIENT ADHERENCE FROM POWERVIEW SOFTWARE 104

PAGE 105

APPENDIX B HIV MYTHS HANDOUT HIV, the human immuno deficiency virus, is the virus that causes AIDS. For many reasons, AIDS is a disease that is commonly misunderstood and, as a result, unduly feared.1 In order to combat this fear, knowledge concerning the virus is the BEST weapon! Following are ten of the most common myths about HIV/AIDS and the reality of these misconceptions Myth #1 : Women with HIV cant have children without infect ing them with the virus. Reality : This used to be true, but not anymore. Women living with HIV can and do have families. While certain steps and precautions have to be taken, women can now have the families they always dreamed about! Myth #2 : People with HIV can get rid of the virus if they take all of their medication like they are supposed to. Reality: Current treatments for HIV are better than ever, but the bottom line is there these treatments only help prolong life, not cure the disease itself. When the treatments work, there is so little of the virus in the blood that blood tests ca nt detect it. Even though blood tests will say the virus is undetectable, the virus is still there, hiding in a sleep-like state in your body. People whose HIV is in this state must continue taking their medicine to stay well. Myth #3 : People with HIV do not need to take their medications when they do not feel sick. Reality : Even when people with HIV are feeling great, HIV is making billions of copies of itself everyday and attacking their immune system. When they finally start to feel sick, HIV has already damaged their immune system and nothing can fully bring it back to normal. 105

PAGE 106

Myth #4 : You can tell if someone has HIV just by looking at them. Reality : Have you heard the old saying, looks can be deceiving? A person with HIV may not show any symptoms for up to 10 years. Since HIV affects each person differently, many people with HIV can look and feel healthy for years and not even know they are infected! Without knowing, that person may be practic ing unsafe sex and may be spreading the virus to others. Myth #5 : HIV can be spread by kissing or hugging someone who is HIV positive. Reality : Though there are other means of transmission, there are four main ways HIV spreads: 1. Through unprotected sexual contact 2. Through blood transfusion 3. By sharing needles through injection drug use 4. From mother to child You cant contract HIV by kissing or hugging someone. For more information, visit: http://www.cdc.gov/hiv/resources/brochures /liv ingwithhiv.htm 106

PAGE 107

APPENDIX C PICTORIAL REPR ESENTATION OF MEDICATION REGIMEN ZE RI T EPIVIR ZE RI T EPIVIR VIRACEPT VIRACEPT VIRACEPT VIRACEPT 107

PAGE 108

APPENDIX D PROB LEM SOLVING HANDOUT Problem Solving Techniques Problem solving consists of five steps: (1) defining the problem, (2) generating di fferent solutions, (3) making a decision, (4) planning to use the solution, and (5 ) making changes over time. These five steps can be used whenever disagreements over managing medical care occur. Step One Defining the Problem There are three goals when defining a problem: 1) Each family member should express clearly to the others his/her view about the issue. 2) Each family member should understand and be able to explain the others view(s). 3) The topic being discussed should be simple rather than complex. Remember : Differences of opinion are no rmal and healthy, not necessa rily a sign of rebellion! All family members should take turns defining the problem as they see it. Each person then checks how much the others understand what was ju st said by having family members repeat it back in their own words. If the family memb ers explanation does not match up with what the person was trying to say, the person should explai n the definition again more clearly. Here are some tips: Be specific. Try to be brief. Avoid angry language, accusa tions, or blaming others. Describe behaviors, feelings, and situa tions, not personality characteristics of individuals. Only discuss the problem at hand. Do not bring up the past or other issues. Start with I rather than you (e.g., I feel angry). 108

PAGE 109

Example: CHRIS: My problem is that I want to be able to look after my own medicines. I feel like a baby because I havent been allowed to keep my medicine in my bathroom or remember to take them on my own. MRS. SMITH: So, what you are trying to say is that it bothers you when you have little control over your medicine. You feel like keeping your meds in your bathroom and remembering on your own to take them would make you feel more responsible. CHRIS: Yes, thats it. Step Two Set A Goal Setting a goal involves stating wh at you would like to happen, or the end result that you want to work towards. Family members may need to discuss the goal and write it down in clear, achievable terms. Example: CHRIS: I want my mom to trust that I can take my meds without constantly reminding me. MRS. SMITH: I agree. I want to trust that Chris to take his meds without me nagging him. Step Three Listing Different Solutions The goal here is to think of as many ideas as possi ble for ways to fix the problem defined in Step One. When coming up with different solutions dont worry about how good the ideas are until later in the discussion. Be sure to suggest creative and extreme ideas too; anything goes! Family members should take turns suggesting ideas and one family member should write down all of the ideas. Try to list at least 4-5 solutions for each problem. 109

PAGE 110

Example: MRS. SMITH: Maybe you could keep your morning meds in your bathroom, but I will still remind you when to take them before you go to school. CHRIS: You could put my meds in a place where I can get to them on my own without them being in my bathroom. Maybe the kitchen would be a good place. MRS. SMITH: Maybe during the week when youre busy with school, I could keep the meds and remind you when to use them, but you could try doing it on your own on the weekends. CHRIS: You could stop telling me when to take my medicine, but I could mark on a calendar when Ive taken them, so you will know. Step Four Making a Decision When making a decision, there are three goals : 1) Discuss the good and bad points of each idea together as a group. 2) Have each person rate each idea as good or bad overall. 3) Agree on the best solution and put it in place. Each family member should take turns pointin g out the good and bad points of the solution for himself/herself and for the rest of the family. Consider both potential short-term and long-term effects of the solution, and then rate the solution (give it either a plus or minus) based on how good the idea is, not who came up with it. When all of the solutions have been discussed and rated by each family member, see if everyone rate d any of the ideas a plus. If there is agreement by all family members on one or more ideas, select one or co mbine several ideas to use. If there is no agreement, begin to discuss a compromise. Example: CHRIS: I rate the first solution as minus becau se I think that I would still feel like a baby if you have to keep reminding me to take my medicine. The second solution is a plus because I dont really need to keep my meds in my bathro om, just in a place where I can get to them myself without you having to get them for me. The third solution is a minus because I still will have to keep the old rout ine five out of seven days a week, and that 110

PAGE 111

doesnt give m e much of a chance to show that I can be responsible. The fourth solution is a plus because I have more control over when I take the medicine, and I dont mind marking it on a calendar to let you know that I di d. Maybe then you will start to see that I am dependable. MRS. SMITH: I dont like the first soluti on very much because it makes me uneasy knowing that you have the medication in your bath room. I will rate the first solution as a minus. I will rate the second solu tion as a plus because I dont mind keeping the inhalers in a location where you can get to them. I know how much you want to show me that you are responsible and I think it will help if you can get to the medicines on your own. I will rate the third solution as a plus because I think two days a week is a good place to take care of your own medicines. I will rate the fourth solution as a plus because at least I will know whether you have taken your medicines or not and it will help me build confidence in you. CHRIS: It looks like weve both rated the second and fourth solution as plus. Lets try using those two solutions. Step Five Planning to Use the Solution When planning to use a solution, your goals are to: (1) Consider the details that are necessa ry to put the solution into place (2) Try to predict difficulties that may come up when the solution is used. Identify the behaviors that each family member needs to do in order to show that he/she is following through with the solution. Assign specific tasks to particular family members. Try to come up with ways to keep track of whether each family member is following through with their part of the solution. Plan for how the family will deal with any problems that come up. Example: Chris and Mrs. Smith agree that keeping the me dication in the cabinet in the kitchen is a good place where both can get to them. Mrs. Smith will place them in the cabinet and check once a week to make sure that they are still there. They agree that Ch ris will record the time that he takes his medicines on a chart on the refrigerator. When Mrs. Smith checks on the medications and notices that they are still in the cabinet, she will record this on the chart too. 111

PAGE 112

Step Six Making Changes Over Time If it seems that the first solution didnt work well, go back and examine step-by-step where the problems may have occurred. Avoid blaming any particular family member for the solutions failure. Start from Step One or Step Two of the problem solving process and go through the steps again. Continue this process until all family members are happy with another possible solution to try again. 112

PAGE 113

Problem-Solving Worksheet Problem to be solved: __________ ____________________ __________________ ________________________ __________________________________________ ________________________ __________________________________________ ________________________ __________________________________________ Solutions Family member 1 Family member 2 1. 2. 3. 4. 5. 113

PAGE 114

Agreed-upon solution:_____________ __________________ _________________ ________________________ __________________________________________ ________________________ __________________________________________ ________________________ __________________________________________ ________________________ __________________________________________ Responsibilities for each family member (list name & responsibilities): Measurements/checks (note who, when, & how): Date to re-evaluate how well the solu tion is working: _______________________ Signatures: 114

PAGE 115

APPENDIX E COMMUNI CATION HANDOUT Parents can do a great deal to help and support their youth with their medical care! There are th ree main ways to do so: instructional help, informational help, an d emotional support Instructional help is a hands-on way of helping your youth to manage his or her medical care. Some examples of instructional help include: Helping your child to find a way to simp lify his or her medication management. Keeping track of the doctor appointments and being sure that your child has enough medication refills. Making sure that your child takes his or her medication when he or she goes away for the evening or a weekend. Making sure your child takes his or her medication at the right times. 115

PAGE 116

Informational help is a way to educate your yo uth about his or her medical condition. This should always be done in a positive or neutral manner to avoid parent-child conflic t. Some examples of informational help include: Providing your child with reading material on their medical condition. Keeping a list of questions that you or yo ur child may have to ask your childs physician. Encouraging family discussions on you r childs health condition and its treatment. Emotional support is a way of helping your child to feel good both about himself or herself and his or her medical care. Some ways to provide your child with emotional support include: Praising your child often for adhering to their treatment plan. Avoiding treating your child differen tly because of his or her illness. Taking an active interest in your childs life, learning about his or her likes and dislikes. 116

PAGE 117

Why is communication between parents and youth important? Positive Communication Between Parents and Youth Adolescence is a time of rapid change, with youth seeking more independence from their parents. Parents sometimes face new issues that need to be discussed with their children rather than decided by the parents alone. The way parents and youth communicate with each other can affect the quality of their relationship. Positive Communication Positive communication helps parents and youth communicate effectively and avoid unnecessary conflicts. Telling one a nother how you feel is a very useful way to communicate. It is best to use st atements that do not blame others when discussing how you feel about a problem or difficult situation. These statements are called I statements because they fo cus on how the speaker is feeling rather than on what the listener is doing wrong. I statements often take the form of I feel ____ when _____ happens. For example, I feel worried when you forget to take your medicine. List an example for you and your family: 117

PAGE 118

Roadblocks to communication One of the first steps of positive communication is to remove all of the obstacles, or roadblocks. These roadblocks are statements that hurt people and make them feel defensive, instead of making them want to change their behavior. Here is a list of some common roadbloc ks to communication, as well as some positive ways to stay away from these roadblocks: Roadblocks Positive Talk Commands & orders Discuss possible solutions Threats Compromise Interrupting Listen and summarize Accusing, blaming, or shaming Use I statements Sarcasm Use a neutral tone of voice 118

PAGE 119

COMMUNICATION WORKSHEET Family Name ____________________________ Week of ________________ to _______________ (Dates) Communication Behavior to Target: Day Roadblock Positive Communication Change 119

PAGE 120

APPENDIX F FAMILY R OLES HANDOUT Family Roles Parent : Provide support and encouragement to children and spouse/partner Create an environment where children can develop skills (physical, emotional, educational, and social) Provide basic needs for children (food, housing, clothing, safety) Discipline children for not following household rules Help with daily chores Follow household rules What are some that are specific to your home: ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ Child/Teens : Follow household rules set by parent(s) Help with daily chores Not worry about getting basic needs met (food, housing, clothing, safety) Provide support and en couragement to other family members What are some that are specific to your home: ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ 120

PAGE 121

Medication Parent : _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ Child/Teen: _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ _____________________________________________________________________________ 121

PAGE 122

APPENDIX G DEVEL OPMENTAL CHANGES AND ADHERENCE HANDOUT The Need for Parental Involvement W ith school-aged children, parents often wonder how much involvement they need to have in their childs treatment regimen. It is likely that pare nts will need to maintain at least some degree of involvement in their childs treatment, and finding the appropriate level of involvement that works best for you and your child will be key. Children Can Forget Adults have better memories than children do. A dults also are used to dealing with everyday tasks and duties. One of the main reasons why parents need to maintain a certain degree of involvement in their childs treatme nt is that children simply forg et about taking their medicine. Youth May Not Understand When explaining a treatment plan to families, do ctors often u se words or phrases that youth may not understand. The steps involved in properly ta king medicine also can be long and complex. Children often rely upon parents to learn the pr ocedure and then teach it to them later. Additionally, it is not uncommon for youth to require repe ated instruction by parents. Parents May Be More Assertive Doctors, nurses, or others in positions of power can easily intim idate youth. Children may be too frightened to ask questions or too embarrassed to admit that they do not understand something. Often, adolescents expect their parents to speak up in these situations. 122

PAGE 123

Parental In volvement May Be Required There are some aspects of childrens treatment plans that youth cannot handle on their own. Parental involvement often is required in these cases. For example, issues related to health insurance paperwork, purchasi ng prescriptions, and transpor tation to and from doctors appointments all call for some de gree of parental involvement. Parents Normally Set the Rules Parents almost always hold positions as heads of the household. As part of that position, they normally create the household rules and routines that children are expected to follow. Youth are accustomed to having parents provide structure in many areas of their lives and will expect parents mistake send the message that treatment adherence is not important. Youth May Have Many Commitments As youth becom e more independent, they naturally begin to take on extra ac tivities outside of the home (for example, sports, af ter-school jobs, dating). As th e number of these activities increases, many young adolescents find that they have less time to give to their medical care and sometimes simply forget to take their medicine at all. Adolescents May Be Too Bold Many adolescents have an attitude that they are not at risk a nd that nothing bad will ever happen to them. They have difficulty seeing the possible negative consequences of their actions, particularly if these negative ou tcomes are far off in the future. Sometimes youth underestimate their need for medicine. Something bad may not happen right after they a ccidentally forget to take a dose, so some youth will fall into the habit of skipping doses often. Parents sometimes respond to this problem by nagging their children to take their medicine. Thus, medical care, particularly remembering medications, can become an area of great conflict. 123

PAGE 124

Parents Can Act as Coaches Deciding how m uch parental involvement is needed for your indi vidual child will likely be based upon a number of factors. Parents need to consider the childs ag e, the severity of his or her medical condition, the difficulty of the treatment plan, and any prior successes or failures with independence. Additionally, parent s need to maintain proper exp ectations concerning what their child can be expected to handle. Parents are best viewed as the coaches of a medical care team, with level of involvement changing as needed. Greater Freedom May Lead to Misunderstandings Some parents have difficulty accepting their child s new sense of independence and worry more about their childs health. Sometimes youth overestimate the amount of responsibility that they can handle. Misunderstandings a bout parental involvement in treatment can result. Some parents demand too much involvement, and so me youth demand too much freedom. Striking a compromise would be the ideal situation. Setting Realistic Expectations for Children Pre-Adolescents and Young Adolescents (Ages 9 through 12) This age is the perfect tim e for teaching children about how to cope with medical regimens. As children develop, they learn to imag ine situations they have not tr uly experienced. At this point, children can begin to consider what if s ituations. Because youth at this stage of development place a lot of value on their personal abilities, their ability to learn new things and develop the skills required to take care of themselves is important. So, parents can begin to teach young adolescents how to manage their own medical care. Yet, youth at this stage still require a great deal of parental supervision, so parents should begin slowly. 124

PAGE 125

Here are so me examples of reasonable expectatio ns for children in this age range: Parents should expect that childre n require a great deal of PARENTAL SUPERVISION. They often will forget to take medicine on their own. Parents should expect that they may h ave to SET LIMITS with their children. Parents should expect to need A LOT OF PATIENCE when teachi ng their children how to manage their medicine. Children will have problems following complex medication routines the w ay that adults can. Although giving youth chances to be more ac tive in their health care is helpful in their development, it is most realistic that decisions be shared between parents and youth. Later Adolescents (Ages 13 through 18): Most adolescents within this age range are beginn ing to think abstractly and consider different ideas and possibilities. As th ey get older, adolescents beco me increasingly independent of parents and are held responsible for parts of their own lives. Adolescence provides many chances for children to prepare for what they have to do as adults. For example, children in this ag e range begin to develop the ability to manage their own medical care with less supervision by their parents. Ini tially however, they may n eed to rely upon parents to help them learn how to plan in advance for occasional troubles in routine. Younger adolescents may welcome pare ntal advice. As children grow into late adolescence, they de velop a stronger sense of their independence and identity. At this stage, many adolescents welc ome being put in charge of their own medical treatment. They tend to appreciate when parents ask their opinions and allow them a large role in the decision-making process. However, ch ildren may react negatively to parents who repeatedly remind them or offer advice without being asked for it. What a parent views as helpful, an adolescent may see as either naggi ng or a sign that the pa rent doesnt trust the adolescent. 125

PAGE 126

Here are so me examples of reasonable expectations for older adolescents: Parents should expect that adolesce nts may want to MANAGE THEIR HEALTH ON THEIR OWN. Some adolescents view parental involvement as a clue that the adolescent is not trusted. Be prepared to discuss independence issues and give responsibility to youth when appropriate. Parents should expect that youth WILL ARGUE. Issues of fairness often are at the center of many parent-adolescent argume nts. Be prepared to problem solve the fairest/most reasonable solution. Parents should expect that PEER PRES SURE will be a majo r factor in their childrens lives. For many adolescents, th e opinions and values of friends begin to be more important than those of family. Be ready to come across issues regarding what is or is not cool when addressi ng your youths medical care management. The adolescent years are a time of increased peer pr essure and feelings of self-consciousness. As such, youth tend to avo id anything that they believ e will make their friends and classmates view them negatively. Sometimes, children do not like to take their medication in front of others because they are embarrassed or do not want to answer questions about why they need to take medicine. Other times, peer pressure can lead to adolescents picking up bad habits, like smoking or drinking, which can interfer e with their medications. 126

PAGE 127

LIST OF REFERE NCES Adam, B. D., Maticke-Tyndale, E., & Cohen, J. J. (2003). Adherence practices among people living with HIV. AIDS Care, 15, 263-274. AIDSinfo (2005). Glossary of HIV/AIDS-related terms, 5th edition. Department of Health and Human Services. 1-174. Retrieve d on November 6, 2007, from http://aidsinfo.nih.gov/ ContentFiles/GlossaryHIV-relatedTerm s_FifthEdition_en.pdf Altice, F. L., & Friedland, G. H. (1998) The era of adherence to HIV therapy. Annals of Internal Medicine, 129, 503-505. APREX (1998). Medication event monitoring system: users guide (version 2.61) Menlo Park, USA: APREX. Armstrong, F. D., Willen, E. J., & Sorgen, K. (2003) HIV and AIDS in children and adolescents. In M. Edwards (Ed.), Handbook of Pediatric Psychology, 3rd edition (pp. 358-374). New York: Guilford Press. Barlow, D. H. & Hersen, M. (1984). Single case experimental desi gns: Strategies for studying behavior change, 2nd edition. New York, Pergamon Press. Becker, M. H., Drachman, R. H., & Kirscht, J. P. (1972). Predicting mothers compliance with pediatric medical regimens. Journal of Pediatrics, 81, 843-854. Becker, M. H., Maiman, L. A., Kirscht, J. P ., Haefner, D. P., & Drachman, R. H. (1977). The health belief model and prediction of dietary compliance: A field experiment. Journal of Health and Social Behavior, 3, 125-135. Belzer, M. E., Fuchs, D. N., Luftman, G. A., & Tucker, D. J. (1999). An tiretroviral adherence issues among HIV-positive adolescents and young adults. Journal of Adolescent Health, 25, 316-319. Berrien, V. M., Salazar, J. C., Reynolds, E., & McKay, K. (2004). Adherence to antiretroviral therapy in HIV-infected pediatric patients improves with home-based intensive nursing intervention. AIDS Patient Care and STDs, 18(6), 355-363. Bjelland, I., Dahl, A. A., Haug, T. T., & Dag, N. (2002). The va lidity of the Hospital Anxiety and Depression Scale; an updated review. Journal of Psychosomatic Research, 52(2), 6977. Bova, C. A., Fennie, K. P., Knafl, G. J., Di eckhaus, K. D., Watrous, E., & Williams, A. B. (2005). Use of electronic monito ring devices to measure antire troviral adherence: Practical considerations. AIDS and Behavior, 9(1), 103-110.bova Brackis-Cott, E., Mellins, C. A., Abrams, E., Reval, T., & Dolezal, C. (2003). Pediatric HIV medication adherence: The views of medical providers from two primary care programs. Journal of Pediatric Health Care, 17, 252-260. 127

PAGE 128

Brown, L. K., Lourie, K. J., Pao, M. (2000). Children and adolescents living with HIV and AIDS: A review. Journal of Child Psychology and Psychiatry, 41, 81-96. Brown, R. (2002). Society of Pedi atric Psychology presidential ad dress: Toward a social ecology of pediatric psychology. Journal of P ediatric Psychology, 27(2), 191-201. Centers for Disease Contro l and Prevention (2005). HIV/AIDS Surveillance Report. Retrieved April 27, 2007, from http://www.cdc.gov/hiv/topics/surve illance/resources/ reports/ 2005report/default.htm Chadwick, E. G., & Yogev, R. (1995). Pediatric AIDS. Pediatric Clinics of North A merica, 42, 969-992. Clingempeel, W. G. & Henggeler, S. W. (2002) Randomized clinical trials, developmental theory, and antisocial youth: Guidelines for research. Development and Psychopathology, 14, 695-711. Davies, G., Koenig, L. J., Stratford, D., Palm ore, M., Bush, T., Golde, M., et al. (2006). Overview and implementation of an interven tion to prevent adherence failure among HIVinfected adults initiating antiretroviral ther apy: Lessons learns from Project HEART. AIDS Care, 18(8), 895-903. De Civita, M., & Dobkin, P. L. (2004). Pediat ric adherence as a multidimensional and dynamic construct, involving a triadic relationship. Journal of Pediatric Psychology, 29(3), 157169. Deschamps, A. E., Van Wijngaerden, E., Denh aerynck, K., Vandamme, A. & De Geest, S. (2006). Use of electronic monitori ng induces a 40-day interventi ons effect in HIV patients. Journal of Acquired Immune Deficiency Syndromes, 43(2), 247-249. Dolezal, C., Mellins, C., Brackis-Cott, E., & Ab rams, E. J. (2003). Reliability of reports of medical adherence from children with HIV and their caregivers. Journal of Pediatric Psychology, 28(5), 355-361. Dong, B. J. (2007). The HIV epidemic and treatment strategies: Where are we now? U.S. Pharmacist, 32, 95-102. Farley, J., Hines, S., Musk, A., Ferrus, S., & Tepper, V. (2003). Assessment of adherence to antiviral therapy in HIV-infected children using the Medication Event Monitoring System, pharmacy refill, provider assessment, care-gi ver self-report, and appointment keeping. Journal of Acquired Immune Deficiency Syndromes, 33, 211-218. Farmer, K. C. (1999). Methods for measuring and monitoring me dication regimen adherence in clinical trials and clinical practice. Clinical Therapeutics, 21, 1074-1090. Feingold, A. R., Rutstein, R. M., Meislich, D ., Brown, T., & Rudy, B. J. (2000). Protease inhibitor therapy in HIV-infected children. AIDS Patient Care and STDs, 14, 589-602. 128

PAGE 129

Fiese, B. H., & Everhart, R. S. (2006). Medica l adherence and childhood ch ronic illness: Fam ily daily management skills and emotional climate as emerging contributions. Current Opinion in Pediatrics, 18, 551-557. Frey M. A., & Naar-King, S. (2000). The challe nge of measuring adhe rence in children and adolescence. Journal of Child and Family Nursing, 4, 296-300. Gavin, P. J., & Yogev, R. (2002). The role of pr otease inhibitor therapy in children with HIV infection. Paediatric Drugs, 4(9), 581-607. Goode, M., McMaugh, A., Crisp, J., Wales, S., & Ziegler, J. B. (2003). Adherence issues in children and adolescents receiving highly active antiretroviral therapy. AIDS Care, 15(3), 403-408. Harris, M. A., & Mertlich, D. (2003). Piloting home -based behavioral family systems therapy for adolescents with poorly controlled diabetes. Childrens Health Care, 32(1), 65-79. Haynes, R. B. (1979). Introduction. In R. B. Ha ynes, D.W. Taylor, & D. L. Sackett (Eds.), Compliance in health care (pp. 1-7). Baltimore, MD: Johns Hopkins University Press. Henggeler, S. W., Schoenwald, S. K., Borduin, C. M., Rowland, M. D., & Cunningham, P. B. (1998). Multisystemic treatment of antisocial behavior in children and adolescents. New York: Guilford Press. Herrmann C. (1997). Internationa l experience with th e Hospital Anxiety and Depression Scale: A review of validation data and clinical results. Journal of Psychosomatic Research, 42, 17-41. Holmbeck, G. N. (2002). A developmental perspective on adolescent hea lth and illness: An introduction to the special issues. Journal of Pediatric Psychology, 27(5), 409-415. Johnson, M. O., Gamarel, K. E., & Rose, C. D. (2006). Changing HIV treatment expectancies: A pilot study. AIDS Care, 18(6), 550-553. Joint United Nations Prog ramme on HIV/AIDS (2004). AIDS Epidemic Update 2006, 1-94. Retrieved on August 21, 2007, from http://www.unaids.org/en/HIV_data/epi2006/ default.asp Katko, E., Johnson, G. M., Fowler, S. L., Turner, R. B. (2001). Assessment of adherence with m edications in human immunodefici ency virus-infected children. Pediatric Infectious Disease Journal, 20, 1174-1176. La Greca, A. M. & Bearman, K. (2003). Adheren ce to pediatric treatment regimens. In M. Edwards (Ed.), Handbook of Pediatric Psychology, 3rd edition (pp. 119-140). New York, NY: Guilford Press. LaFleur, J., & Oderda, G. M. (2004). Methods to measure patient compliance with medication regimens. Journal of Pharmaceutical Care in Pain & Symptom Control, 18(3), 81-87. 129

PAGE 130

Larsen, D. L., Attkisson, C. C., Hargreaves, W A., and Nguyen, T. D. (1979). Assessment of client/patient satisfaction: Development of a general scale. Evaluation and Program Planning, 2, 197-207. Lindegren, M. L., Byers, R. H., Thomas, P., Davis, S. F., Caldwell, B., Rogers, M., et al. (1999). Trends in perinatal transmission of HIV/AIDS in the United States. JAMA: Journal of the American Medical Association, 282(6), 531-538. Liu, H., Golin, C. E., Miller ,L. G., Hays, R. D., Beck, K., Sanandaji, S., et al. (2001). A comparison study of multiple measures of adherence to HIV protease inhibitors. Annals of Internal Medicine, 134(10), 968-977. Marhefka, S. L., Farley, J. J., Rodrigue, J. R., Sandrik, L. L., Sleasman, J. W., & Tepper, V. J. (2004). Clinical assessment of medication adherence among HIV-infected children: Examination of the Treatment Interview Protocol (TIP). AIDS Care, 16(3), 323-337. Marhefka, S. L., Tepper, V. J., Farley, J. J., Sleasman, J. W., & Mellins, C. A. (2006). Brief report: Assessing adherence to pediatric antiretroviral regi mens using the 24-hour recall interview. Journal of Pediatric Psychology, 31(9), 989-994. Martin, S., Elliot-DeSorbo D. K., Wolters, P. L ., Toledo-Tamula, M. A., Roby, G., Zeichner, S., et al. (2007). Patient, caregiver and regimen characteristics associated with adherence to highly active antiretroviral therapy among HIV-infected children and adolescents. The Pediatric Infectious Disease Journal, 26(1), 61-67. Martinez, J., Bell, D., Camacho, R., Henry-Reid L. M., Bell, M., Watson, C. et al. (2000). Adherence to antiviral drug regimens in HIV-infected adolescent patients engaged in care in a comprehensive adolescent and young adult clinic. Journal of the National Medical Association, 92(2), 55-61. Mehta, S., Moore, R. D., & Graham, N. M. ( 1997). Potential factors affecting adherence with HIV therapy. AIDS, 11(14), 1665-1670. Melbourne, K. M., Geletko, S. M., Brown, S. L ., Willey-Lessne, C., Chase, S., & Fisher, A. (1999). Medication adherence in patients w ith HIV infection: A comparison of two measurement methods. AIDS Reader, 9(5), 329-338. McConnell, M. S., Byers, R. H., Frederick, T., Pete rs, V. B., Dominguez, K. L., Sukalac, T. et al. (2005). Trends in antiretroviral therapy use and survival ra tes for a large cohort of HIVinfected children and adolescen ts in the United States, 1989-2001. Journal of Acquired Immune Deficiency Syndrome, 38(4), 488-494. Mellins, C. A., Brackis-Cott, E., Dolezal, C., & Abrams, E. J. (2004). The role of psychosocial and family factors in adherence to antire troviral treatment in human immunodeficiency virus-infected children. Pediatric Infectious Disease Journal, 23, 1035-1041. Miller, I. W., Bishop, D. S., Herman, D. S., & Stein, M. D. (2007). Relationship quality among HIV patients and their caregivers. AIDS Care, 19(2), 203-211. 130

PAGE 131

Murphy, D. A., Sarr, M., Durako, S. J., Moscicki A., W ilson, C. M., & Muenz, L. R. (2003). Barriers to HAART adhere nce among human immunodeficiency virus-infected adolescents. Archives of Pediatrics a nd Adolescent Medicine, 157, 249-255. Murphy, D. A., Wilson, C. M., Durako, S. J., Nu enz, L. R., Belzer, M., & the Adolescent Medicine HIV/AIDS Research Network (2001). Antiretroviral medication adherence among the REACH HIV-infected adolescent cohort in the USA. AIDS Care, 13, 27-40. Naar-King, S., Frey, M., Harris, M., & Arfken, C. (2005). Measuring adherence to treatment of paediatric HIV/AIDS. AIDS Care, 17(3), 345-349. National Institute of Allergy and Infectious Diseases (2005). HIV infection and AIDS: An overview. Retrieved on August 27, 2007, from http://www.niaid.nih.gov/ factsheets/hivinf.htm Panel on Antiretroviral Guidelines for Adult and Adolescents (2006) Guidelines for the use of antiretroviral agents in H IV-infected adults and adolescents. Depart ment of Health and Human Services. 1-113. Retrie ved on August 21, 2007, from http://www.aidsinfo.nih.gov/Conten tFiles/AdultandAdol escentsGL.pdf. Paterson, D. L., Swindells, S., Mohr, J., Brester, M., Vergis, E. N., S quier, C. et al. (2000). Adherence to protease inhibito r therapy and outcom es in patients with HIV infection. Annals of Internal Medicine, 133(1), 21-30. Prochaska, J. O. & DiClemente, C. C. (1983) St ages and processes of self change of smoking: Toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51, 390-395. Puga, A. M. (2006). Adherence in the pediatri c HIV population. In J. Beal, J. Orrick, & K. Alfonso (Eds.), HIV/AIDS: Primary care guide. Norwalk, CT: Crown House Publishing Limited. Quittner, A. L., Drotar, D., Iveres-Landis, C., Slocum, N., Seidner, D., & Jacobsen, J. (2000). Adherence to medical treatments in adolescent s with cystic fibrosis: The development and evaluation of family-based interventions. In D. Drotar (Ed.) Promoting Adherence to Medical Treatment in Chroni c Childhood Illness: Concepts, Methods, and Interventions (pp. 367-381). Nahwah, NJ: Lawrence Erlbaum Associates Publishers. Rao, D., Kekwaletswe, T. C., Hosek, S., Martinez J., Rodriguez, F. (2007). Stigma and social barriers to medication adherence in urban youth living with HIV. AIDS Care, 19(1), 28-33. Rapoff, M. (1999). Adherence to pediatric treatment regimens. New York: Kluwer Academic/Plenum Publishers. Reddington, C., Cohen, J. M., Baldillo, A., Toye, M., Smith, D., Kneut, C., et al. (2000). Adherence to medication regimens among ch ildren with human immunodeficiency virus infection. Pediatric Infectious Di seases, 19(12), 1148-1153. 131

PAGE 132

Re mien, R., Hirky, A. E., Johnson, M. O., Weinhardt, L. S., Whittier, S., & Le, G. M. (2003). Adherence to medication treatment: A qualitativ e study of facilitators and barrier among a diverse sample of HIV+ men and women in four US cities. AIDS and Behavior, 7(1), 6172. Remien, R. H., Stirratt., M. J., Dolezal, C., Dogni n, J. S., Wagner, G. J., Carballo-Dieguez, A., et al. (2005). Couple-focused support to improve HIV medication adherence: A randomized controlled trial. AIDS, 19(8), 807-814. Robin, A. L. (1998). ADHD in adolescents: Diagnosis and treatment. New York: Guilford Press. Robin, A. L. (2003). Behavioral Family Systems Therapy for adoles cents with anorexia nervosa. In A. Kazdin (Ed.) Evidence-based Psychotherapies for Children and Adolescents (pp. 358-373). New York: Guildford Press. Robin, A. L. & Foster, S. L. (1989). Negotiating parent-adolescent conflict: A behavioral-family system approach. New York: Guilford. Rogers, A. S., Miller, S., Murphy, D. A., Tanney, M. & Fortune, T. (2001). The TREAT (Therapeutic Regimens Enhancing Adherence in Teens) Program: Theory and preliminary results. Journal of Adolescent Health, 29S, 30-38. Rubin, R., Young-Hyman, D., & Peyrot, M. (1989). Parentchild responsib ility and conflict in diabetes care. Diabetes, 38(S2) 28A (abstract). Rudy, B. J., Murphy, D. A., Harris, D. R., Muenz, L. & Ellen, J. (2009). Patient-related risks for nonadherence to antiretroviral therapy among HI V-infected youth in the United States: A study of prevalence and interactions. AIDS Patient Care and STDs, 23(3), 185-194. Rueda, S., Park-Wyllie, L. Y., Bayoumi, A. M., Tynan, A. M., Antoniou, T. A. Rourke, S. B., et al. (2006). Patient support and education for promotion adherence to highly active antiretroviral therapy for HIV/AIDS. Cochran Database of Systematic Reviews, 19(3), CD001442. Schwed, A., Fallab, C. L., Burnier, M., Waeber, B., Kappenberger, L., Burnand, B. et al. (1999). Electronic monitoring of compliance to lipi d-lowering therapy in clinical practice. Journal of Clinical Pharmacology, 39, 402-409. Secord, E., & Cotronei-Cascardo, C. (2007). Pediatric and adoles cent HIV disease. Annals of Allergy, Asthma, and Immunology, 98, 405-413. Sethi, A. K., Celentano, D. D., Gange, S. J., M oore, R. D., & Gallant, J. E. (2003). Association between adherence to antiretroviral therapy and human immunodeficiency virus drug resistance. Clinical Infectious Diseases, 37(8), 1112-1118. 132

PAGE 133

Sim oni, J. M., Montgomery, A., Martin, E., New, M., Demas, P. A., & Rana, S. (2007). Adherence to antiretroviral therapy for pediatric HIV infection: A qualitative systemic review with recommendations for re search and clinical management. Pediatrics, 119(6), 1371-1383. Speilberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Steele, R. G., Anderson, B., Rindel, B., Dreyer, M. L., Perrin, K., Christensen, R., et al.(2001). Adherence to antiretroviral therapy among HIVpositive children: Examination of the role of caregiver health beliefs. AIDS Care, 13(5), 617-630. Steele, R., & Grauer, D. (2003). Adherence to antir etroviral therapy for pe diatric HIV infection: Review of the literature a nd recommendations for research. Clinical Child and Family Psychology Review, 6(1), 17-29. Steele, R. G. & Nelson, T. D. (2007). Psychosocia l functioning of children with AIDS and HIV infection: Review of the literatur e from a socioecological framework. Journal of Developmental and Behavioral Pediatrics, 28(1), 58-69. Tervo, R. C., Estrem, T. L., Bryson-Brockma nn, W., Symons, F. J. (2003). Single-case experimental designs: Applications in developmental-behavioral pediatrics. Developmental and Behavioral Pediatrics, 24(6), 438-448. Wagner, G. J., & Ghosh-Dastidar, B. (2002). El ectronic monitoring: Adherence assessment or intervention? HIV Clini cal Trials, 3(1), 45-51. Watson, D. C., & Farley, J. J. (1999). Efficacy of and adherence to highly active antiretroviral therapy in children infected with hu man immunodeficiency virus type 1. Pediatric Infectious Disease Journal, 18, 682-689. Weiss, D. & Marmar, C. (1997). The Impact of Event Scale -Revised. In J. Wilson & T. Keane (Eds), Assessing psychological trauma and PTSD. New York: Guildford. White., D., Leach, C., Sims, R., Atkinson, M., & Co ttrell, D. (1999). Validation of the HADS in adolescents. British Journal of Psychiatry, 175, 452-454. Williams, P. L, Storm, D., Montepiedra, G., Nic hols, S., Kammerer, B., Sirois, P. A., et al. (2006). Predictors of adheren ce to antiretroviral medications in children an d adolescents with HIV infection. Pediatrics, 118(6), 1745-1757. Wiener, L., Battles, H. B., & Heilman, N. (2000). Public disclosure of childs HIV infection: Impact on children and families. AIDS Patient Care & STDs, 14, 485-497. Wiener, L., Mellins, C. A., Marhefka, S. & Battle s, H. B. (2007). Disclosure of an HIV diagnosis to children: History, current re search, and future directions. Journal of Developmental and Behavioral Pediatrics, 28(2), 155-166. 133

PAGE 134

134 Wiener, L., Riekert, K., Ryder, C., & Wood, L.V. (2004). Assessing medication adherence in adolescents with HIV when electron ic monitoring is not feasible. AIDS Patient Care and STDs, 18(9), 572-538. World Health Organization. HIV/AIDS and Ad olescents.Child and Adolescent Health and Development,2000. www.who.int/ch ild-adolescenth ealth/HIV/ HIV_adolescents.htm#_ftn (L ast accessed April 27, 2007). World Health Organization (2006). Taking stock: HIV in children. http://www.who.int/ hiv/toronto2006/takingstockchildren.pdf Wysocki, T., Greco, P., & Harris, M. A. ( 2000). Behavioral Fa mily Systems Therapy for adolescents with diabetes In D. Drotar (Ed.) Promoting Adherence to Medical Treatment in Chronic Childhood Illness: Conc epts, Methods, and Interventions (pp. 367-381). Nahwah, NJ: Lawrence Erlbau m Associates Publishers. Wysocki, T., Harris M. A., & Buckloh, L. M. (2006). Effects of Behavioral Family Systems therapy for diabetes on adolescents family relationships, treatment adherence, and metabolic control. Journal of Pediatric Psychology, 31(9), 928-938.

PAGE 135

BIOGR APHICAL SKETCH Wendy Gray received a Bachelor of Science degree from Duke University in 2004, where she majored in developmental psychology and biological anthropology and anatomy. She earned a Master of Science degree in psychology from the University of Florida in 2006. She is currently a Psychology Resident at Cincinnati Childrens Hospital and received her doctoral degree in Clinical Psychology from the University of Florida in 2010. 135