Relationships between Quality of Life, Family Factors, Adherence, and Glycemic Control in Pediatric Patients with Type 1...

MISSING IMAGE

Material Information

Title:
Relationships between Quality of Life, Family Factors, Adherence, and Glycemic Control in Pediatric Patients with Type 1 Diabetes Mellitus
Physical Description:
1 online resource (71 p.)
Language:
english
Creator:
St Amant,Jay Benjamin
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology, Clinical and Health Psychology
Committee Chair:
Geffken, Gary R
Committee Members:
Waxenberg, Lori B
Janicke, David M
Barnett, Tracey E.

Subjects

Subjects / Keywords:
adherence -- control -- diabetes -- glycemic -- life -- pediatric -- psychosocial -- quality
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:
Type 1 diabetes is an endocrine condition typically diagnosed in childhood. In order to maintain satisfactory health status, patients with Type 1 diabetes are tasked with monitoring their caloric and carbohydrate intake, frequently monitoring their blood glucose level, and adjusting their insulin levels. The complexity of the recommended medical regimen can be particularly difficult to adhere to for youth and adolescents with Type 1 diabetes and their parents/caregivers. As a result, pediatric patients with this condition frequently experience health complications. Previous research has identified several behavioral, biological and psychosocial factors that appear to impact adherence to medical regimen as well as health status, referred to as glycemic control. This study examined predicted and mediated relationships among diabetes-specific family factors, adherence to medical regimen, quality of life, and glycemic control (health status) in youth and adolescents with Type 1 diabetes. Patient and parent/caregiver dyads (n = 70) completed a packet of measures assessing diabetes-specific family factors and quality of life. In addition, they were separately given a structured interview to assess medical regimen compliance. Health status measures were accomplished by trained medical staff obtaining the patient's glycemic control as part of a routine pediatric endocrinologist visit. Statistical analyses were conducted using correlations and hierarchical regressions. Combined measures accounted for 42% of the variance in glycemic control. Quality of life partially mediated adherence and glycemic control. Ethnic minority status was associated with poorer glycemic control than Caucasians. Limitations, implications, and future directions are discussed.
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 Jay Benjamin St Amant.
Thesis:
Thesis (Ph.D.)--University of Florida, 2011.
Local:
Adviser: Geffken, Gary R.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0043298:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 RELATIONSHIPS BETWEEN QUALITY OF LIFE, FAMILY FACTORS, ADHERENCE, AND GLYCEMIC CONTROL IN PEDIATRIC PATIENTS WITH TYPE 1 DIABETES MELLITUS By JAY B. ST. AMANT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Jay B. St. Amant

PAGE 3

3 To Amy You bring out the best in me. You motivate, inspire, and encourage. I am eternally grateful for your heart.

PAGE 4

4 ACKNOWLEDGMENTS I thank my chair, Dr. Gary Geffken, for giving me the opportunity to work in his research lab. I am grateful for your wisdom, guidance, and supervision. I also thank the members of my dissertation commi ttee, Drs. David Janicke, Lori Waxenberg, and Tracey Barnett, for their support and time. I owe a great deal of gratitude to Adam Reid, for heading up our recruitment team. I thank the undergraduate research assistants who helped with data collection and entry. Finally, I thank my family for their endless support throughout my graduate school journey.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 ABSTRACT ..................................................................................................................... 9 CHAPTER 1 INTRODUCTION .................................................................................................... 11 Overview ................................................................................................................. 11 Background and Significance ........................................................................... 11 Prevalence of Type 1 Diabetes Mellitus ........................................................... 12 Treatment of Type 1 Diabetes .......................................................................... 13 Adherence versus Glycemic Control ................................................................ 16 Diabetes Specific Family Factors ..................................................................... 18 Quality of Life ................................................................................................... 21 Adherence and Quality of Life .......................................................................... 24 Sp ecific Aims and Hypotheses ............................................................................... 27 Specific Aim 1 ................................................................................................... 27 Hypothesis 1 .............................................................................................. 27 Hypothesis 1a ............................................................................................ 27 Specific Aim 2 ................................................................................................... 28 Hypothesis 2 .............................................................................................. 28 Specific Aim 3 ................................................................................................... 28 Hypothesis 3 .............................................................................................. 28 Hypothesis 3a ............................................................................................ 29 Specific Aim 4 ................................................................................................... 29 Hypothesis 4 .............................................................................................. 29 2 METHOD ................................................................................................................ 30 Participants ............................................................................................................. 30 Procedure ............................................................................................................... 30 Measures ................................................................................................................ 31 Measures Completed by the Pediatric T1D Patient .......................................... 31 Measures Completed by the Parent/Primary Caregiver ................................... 32 Measures Completed by the Parent/Primary Caregiver and Pediatric T1D Patient ........................................................................................................... 32 Statistical Analyses ................................................................................................. 35 3 RESULTS ............................................................................................................... 39

PAGE 6

6 Demographics of Partic ipants ................................................................................. 39 Relationships between Demographic Variables and Study Variables ..................... 39 Glycemic Control (A1c) ..................................................................................... 39 Diabetes Specific Family Factors ..................................................................... 40 Quality of Life ................................................................................................... 41 Intercorrelations among Study Variables ................................................................ 41 Glycemic Control .............................................................................................. 41 Diabetes Specific Family Factors ..................................................................... 42 Quality of Life ................................................................................................... 43 Hypothesis Testing ................................................................................................. 43 Relationships between Quality of Life, Adherence, and Glycemic Control (Hypothesis 1) ............................................................................................... 43 Accounted Variance in Glycemic Control by the Model (Hypothesis 1a) .......... 45 Relationship between Quality of Life and Behavioral Variables (Hypothesis 2) ................................................................................................................... 46 Quality of Life and Adherence .................................................................... 46 Quality of Life and Diabetes Specific Family Factors ................................. 46 Relationship between Quality of Life and Biological Factors (Hypothesis 3) .... 47 Association between Glycemic Control and Ethnic Minority Status (Hypothesis 3a) ............................................................................................. 48 Child Self Report versus Parent Observed Report of Psychosocial Functioning (Hypothesis 4) ............................................................................ 48 4 DISCUSSION ......................................................................................................... 56 Limitations ............................................................................................................... 62 Implications and Fut ure Directions .......................................................................... 64 LIST OF REFERENCES ............................................................................................... 66 BIOGRAPHICAL SKETCH ............................................................................................ 71

PAGE 7

7 LIST OF TABLES Table page 3 1 Pa tient Demographic Data ................................................................................. 49 3 2 Child Sample Diabetes Specific Characteristics ................................................. 50 3 3 Child Reported Study Variables Correlation Matrix ............................................ 51 3 4 Parent Reported Study Variables Correlation Matrix .......................................... 52 3 5 Multiple Regression Analysis with Parent Report Measures Predicting A1ca ..... 53 3 6 Quality of Life Means and Standard Deviations by Child Report and Parent Report ................................................................................................................. 54

PAGE 8

8 LIST OF FIGURES Figure page 3 1 Mediation Model and Path Coefficients for Parent Reported Quality of Life, Adherence, and A1c ........................................................................................... 55

PAGE 9

9 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy RELATIONSHIPS BETWEEN QUALITY OF LIFE, FAMILY FACTORS, ADHERENCE, AND GLYCEMIC CONTROL IN PEDIATRIC PATIENTS WITH TYPE 1 DIABETES MELLITUS By Jay B. St. Amant August 2011 Chair: Gary Geffken Major: Psychology Type 1 diabetes is an endocrine condition t ypically diagnosed in childhood. In order to maintain satisfactory health status, patients with Type 1 diabetes are tasked with monitoring their caloric and carbohydrate intake, frequently monitoring their blood glucose level, and adjusting their insulin levels. The complexity of the recommended medical regimen can be particularly difficult to adhere to for youth and adolescents with Type 1 diabetes and their parents/caregivers. As a result, pediatric patients with this condition frequently experience he alth complications. Previous research has identified several behavioral, biological and psychosocial factors that appear to impact adherence to medical regimen as well as health status, referred to as glycemic control. This study examined predicted and m ediated relationships among diabetes specific family factors, adherence to medical regimen, quality of life, and glycemic control (health status) in youth and adolescents with Type 1 diabetes. Patient and parent/caregiver dyads ( n = 70) completed a packet of measures assessing diabetes specific family factors and quality of life. In addition, they were separately given a structured interview to assess medical regimen compliance. Health

PAGE 10

10 status measures were accomplished by trained medical staff obtaining the patients glycemic control as part of a routine pediatric endocrinologist visit. Statistical analyses were conducted using correlations and hierarchical regressions. Combined measures accounted for 42% of the variance in glycemic control. Quality of life partially mediated adherence and glycemic control. Ethnic minority status was associated with poorer glycemic control than Caucasians. Limitations, implications, and future directions are discussed.

PAGE 11

11 CHAPTER 1 INTRODUCTION Overview Background and Significance Type 1 diabetes mellitus (T1D) is a chronic endocrine condition that is typically diagnosed in childhood or adolescence. The primary characteristic of this condition is an inability of the pancreas to produce insulin, a hormone necessary for the process o f converting sugar, starches, and other foods into energy in the body. Youth and adolescents diagnosed with T1D, along with their families, are tasked by health care professionals with a challenging and restrictive treatment regimen, which they must follo w for the remainder of their lives. As is the case with many other chronic illnesses, the lifelong presence of T1D in patients has led health professionals and researchers to examine the many facets of proper condition management. Medication regimen mana gement in T1D can be particularly problematic, given the typically young age of onset combined with the regimens complexity. Previous research has examined many variables that potentially influence T1D management in young patients including characteristi cs of the disease (e.g., symptom severity, complications, duration of disease), of the patient (e.g., disease education, psychosocial factors), and of the supporting family (e.g., parent child conflict, family dynamics). Although there has been a great d eal of research in this area, as well as significant strides in subsequent treatment, providers and researchers continue to seek out the variables that are most effective in improving T1D management. This paper will review and synthesize the extant litera ture on T1D in pediatric patients. Specifically, it will examine the disease, its treatment, and evaluate previous research that has focused on

PAGE 12

12 variables related to its management. The goal of this project is to integrate and further the empirical knowledge base by examining a model hypothesized to predict health status in pediatric populations with T1D. Prevalence of Type 1 Diabetes Mellitus The American Diabetes Association (ADA) estimates that T1D affects 1 out of every 400600 children in the United S tates, making it one of the most common chronic diseases of school aged children (ADA, 2009). T1D has previously been referred to as juvenile diabetes due to its much younger typical onset in comparison to Type 2 Diabetes Mellitus (T2D; a.k.a. noninsulin dependent diabetes mellitus). However, with the advent of the obesity epidemic, younger and younger people have been diagnosed with T2D, thus necessitating a change in nomenclature to the current form, T1D, in order to avoid confusion. T1D is usually di agnosed in childhood, with a peak age of onset occurring in middle childhood. However, T1D can be diagnosed as late as middle adulthood. It results from autoimmune destruction of the pancreatic islet cells that cells). This destruction of cells leads to a cessation of insulin production by the pancreas. Without the production of insulin, the body is unable to regulate glucose metabolism and convert food into energy. Insulin also plays a significant role in physical growth, activity, brain function, and the bodys ability to heal wounds. Despite significant advances in technology, the precise cause of T1D remains unknown. However, current research suggests that this condition is the result of a combination of a genetic predisposition and environmental factors (ADA, 2010; Silverstein et al., 2005; Wysocki et al., 2003).

PAGE 13

13 Treatment of Type 1 Diabetes Since there is no cure for T1D at present, patients are faced with a complex treatment regimen in order to best manage their condition. The core component of this regimen is the administration of subcutaneous insulin to compensate for the patients inability to produce insulin on their own. This insulin administration is executed either with the use of syringes or via an external insulin pump. The dosage of insulin required is a continuous variable that is dependent on many factors. In order to know how much insulin to administer, the patient must monitor their blood glucose level several times throughout the day. This is typically accomplished by pricking ones finger with a lancet in order to place a small amount of blood on a blood glucose meter strip. This strip is then inserted into a blood glucose meter which provides accurate information about blood glucose levels. Additionally, the patient must monitor his/her food intake and physical activity, as both of these can impact blood glucose levels. Regarding blood glucose levels, normal amounts range between 80 120mg/dl. The goal of tight control of the T1D treatment regimen is to maintain a blood glucose level within this range. The A1c is a blood test of glycosylated hemoglobin which measures the patients relative blood glucose functioning over a 2 3 month period. This is commonly referred to as glycemic control (a.k.a. met abolic control). In the context of diabetes research and outcomes, glycemic control represents a patients health status. Quantitatively, desirable A1c levels are those under 6.5%, which would be classified as good glycemic control (ADA, 2010). Given th e multiple contributors to proper maintenance of blood glucose levels (i.e., insulin administration, blood glucose testing, food intake monitoring, exercise monitoring), it is not surprising that maintaining good glycemic control in terms of A1c levels for patients with T1D may present obstacles.

PAGE 14

14 Repeated deviations from normal A1c can result in both short term and long term complications for the T1D patient. In the short term, patients can develop conditions such as hypoglycemia, seizures, and diabeti c ketoacidosis (DKA). DKA is a potentially life threatening condition that is the result of an absolute shortage of insulin. In terms of long term complications, conditions such as nephropathy, neuropathy, and retinopathy can result from repeated hypergl ycemia. Despite the presence of numerous, potentially lifethreatening complications, adherence to treatment regimens remains a significant problem for pediatric T1D patients. This is not surprising in light of epidemiological research indicating that as many as 50% of pediatric patients are nonadherent to their medical regimen (Rapoff, 1999; Matsui, 1997). Pediatric T1D patients appear to accurately reflect this lack of adherence. A 9year follow up study of T1D youth reported that 45% of the adolescents in th eir study were in some way nonadherent to their treatment regimen (Kovacs et al., 1992). Adherence issues vary across the different aspects of the T1D treatment regimen. Regarding insulin, research has suggested that 10% of patients report administering the wrong dose, 20% report administering an injection at the wrong time, and 19% report difficulty adhering to physician recommendations regarding adjustments to their insulin dose (Delamater et al., 1998). Despite the high prevalence of errors in insul in administration, most patients do report that they regularly take their insulin. Insulin administration does not appear to be the domain of T1D regimens with the highest rate of adherence problems. Blood glucose monitoring, for example, presents a high er rate of adherence problems such that 31% of pediatric patients do not adhere to recommendations about timing and frequency. Further, 48% of pediatric T1D patients

PAGE 15

15 do not adhere to recommended eating habits, as reported by their parents (Delamater et al ., 1998). As shown by research, many different areas of T1D treatment regimen present frequent barriers to adherence in youth and adolescents with this disease. In an effort to combat the problems related to adherence in T1D treatment regimens, the Diabetes Control and Complications Trial (DCCT; 1994; 1993) sought to identify the relationships between regimen adherence, glycemic control, and long term diabetes related complications. Findings from this groundbreaking 10year study continue to guide common approaches to T1D treatment regimens today. Participants in this study were randomized into two treatment groups: conventional therapy and intensive therapy. Conv entional therapy consisted of 1 2 daily injections combined with daily monitoring of blood glucose levels. Further, participants in the conventional therapy group did not make adjustments to their insulin dosages based upon food intake or blood glucose levels. In contrast, the intensive therapy group participants were required to administer 3 or more insulin injections per day or use an insulin pump. Additionally, they checked their blood glucose levels a minimum of 4 times per day. Participants in the intensive therapy group also made insulin dosage adjustments dependent on blood glucose lev els, physical activity, and food intake. Findings from this largescale study indicated that participants in the intensive therapy group not only were better able to achieve glycemic control, but they also experienced a delayed onset and slower progressi on of diabetes related medical complications. Specifically, risks of retinopathy, neuropathy, and nephropathy were reduced in the intensive therapy group by rates of 50% or higher. These findings further emphasized the importance of well maintained glycemic control in terms of impacting

PAGE 16

16 long term health. Implications from the DCCT study resulted in sweeping changes in the realm of standard care of patients with T1D. Since the study, intensive insulin therapy is frequently recommended for patients in order to better achieve and maintain good glycemic control, as well as to reduce the risk of subsequent diabetes related complications. The DCCT suggested that the majority of patients with diabetes, including youth, should utilize an intensive insulin reg imen to optimally treat their condition. Although the DCCT reported that patients on an intensive insulin regimen are at a two to threefold risk for hypoglycemic episodes, they reported that the risk benefit ratio still indicated that this regimen is the treatment of choice due to the improved glycemic control. Patients on an intensive regimen typically administer 4 6 daily insulin injections, using a combination of long acting and short acting insulin. Carbohydrates from each meal must be counted and taken into consideration regarding insulin dosage. It is recommended that patients using this treatment check their blood glucose levels 4 or more times daily, particularly before and after meals. There are many young children and adolescents who find t hat being prescribed an intensive regimen is neither desirable nor easy to follow. As such, parental involvement is vital to the success of treatment. Adherence versus Glycemic Control Thus far, much has been said about the relationship between adherenc e and glycemic control in T1D regimen management. The term adherence has previously been defined as the extent to which a patients behavior coincides with medical or health advice (Haynes, 1979). Given that the goal of treatment regimens is to improve health status (A1c), it would seem as though adherence to ones treatment regimen

PAGE 17

17 would be virtually perfectly correlated with glycemic control (health status). However, although the two are related, the relationship is not perfect. Most studies of ad herence in patients with T1D fail to document high correlations between reports of adherence and glycemic control (Harris et al., 2000; Johnson, 1994; Johnson, 1993; Johnson et al., 1992; La Greca, 1990). One possible explanation for this discrepancy may result from problems with measuring the construct of adherence. For example, if a patients prescribed medical regimen is not effective, even perfect adherence is not expected to result in good glycemic control. Nevertheless, researchers believe that adherence is a significant predictor of HbA1c. Glasgow and colleagues (1987) suggested that the relationship between adherence and glycemic control is not always straight forward, and that adherence must be examined in the context of other factors. Research has identified such factors including regimen characteristics (i.e., intensive vs. conventional), disease characteristics (e.g., duration, age of onset, related complications, etc.), and family variables (Rapoff, 1999; LaGreca & Schuman, 1995). Addition ally, it has been suggested that biological variables such as being a member of an ethnic minority can negatively influence glycemic control. A multi site study by Chalew and colleagues (2000) inferred that being AfricanAmerican was predictive of poorer glycemic control, although their analysis did not control for socioeconomic variables. Work by Petitti and associates (2009) produced similar findings with regard to glycemic control in ethnic minority youth with T1D, this time controlling for factors suc h as family income, family structure, and parental education. These findings certainly warrant further investigation in this domain.

PAGE 18

18 DiabetesSpecific Family Factors A great deal of research has indicated that diabetes specific family factors mediate the relationship between adherence and glycemic control in youth and adolescents with type 1 diabetes (Duke et al., 2008; Lewin et al., 2006; Anderson et al., 2002). Knowing the important role that family factors play in relationship to adherence, it is impor tant to identify the specific factors (positive and negative) in order to better develop effective interventions to improve adherence. As with most parent child relationships, parental involvement plays a key role. However, it is necessary for parents to be more than simply nonnegative. It is vital that they exert some level of positive, supportive involvement in their childs regimen. Research by Anderson and associates (1997) examined the relationship between blood glucose monitoring and adherence. Their study was conducted in the context of parental involvement in this task. Results from the study indicated that a higher frequency of blood glucose monitoring was associated with better adherence and glycemic control. Interestingly, factors contrib uting to these findings were not limited to the frequency of monitoring, but also included parent child agreement on monitoring, as well as parental support and involvement in monitoring. Positive and supportive parenting behaviors, such as those of this study, intuitively seem to be helpful in increasing adherence and subsequently glycemic control. Conversely, negative parenting behaviors seem counterproductive. As illustrated by Lewin and colleagues (2006), poor adherence was associated with several negative family processes including criticism and lack of warmth and support. This point is further supported by research from Duke and associates (2008) that identified parental nagging about diabetes specific behaviors as a strong predictor of poor adherence.

PAGE 19

19 Diabetes specific family factors have been continually shown to impact adherence to T1D regimens in youth. Hauser and colleagues (1990) conducted a four year longitudinal study examining medical regimen adherence in youth and adolescents with type 1 diabetes. Their study of over 300 participants indicated that the strongest predictor of long term adherence was family conflict as perceived by the child. Similar findings have been shown in research by Miller Johnson and associates (1994) examining parent child relationships as they pertain to regimen management in type 1 diabetes. Their findings indicated that a negative parent child relationship was predictive of both poor adherence and poor glycemic control. Duke and colleagues (2008) noted that both negative diabetes specific family behaviors (i.e., nagging about regimen) as well as child externalizing behaviors were both indicative of poorer adherence. Further illustrating the impact of family factors, Lewin and associates (2006) examined 100 y outh and adolescents with type 1 diabetes. Their results showed a significant impact from negative family processes (i.e., criticizing, lack of warmth & support) on subsequently lower adherence and glycemic control. In their model, family factors predic ted 50% of the variance in glycemic control. Due to the abundance of literature supporting the importance of diabetes specific family factors as it pertains to adherence and glycemic control in type 1 diabetes, many researchers have developed measures to assess family factors in an effort to identify and treat those that may preclude adherence. Most of the commonly used measures are self report measures given to both parent and child. Although designed with different areas of particular focus, most are assessing some level of family conflict. The Diabetes Family Conflict Scale (DFCS; Hood, et al., 2007) is one of the most commonly

PAGE 20

20 used diabetes specific family conflict measures. It was designed to evaluate the level of diabetes specific support that the adolescent receives from his/her family. Factor analyses showed that all of the items in this measure load on either direct (e.g., blood glucose monitoring) or indirect (e.g., discussing diabetes with friends) diabetes specific tasks. Analyses indicated that direct diabetes specific tasks are predictive of both adherence and glycemic control. Further, the indirect tasks were shown to be related to mood and quality of life. This measure could be beneficial in identifying the level of family conflict in order to assess what method of treatment would achieve optimal outcomes. Another family factor often assessed is social support. As demonstrated in research from Miller Johnson and associates (1994), positive parent child relationships are predictive of better adherence and glycemic control. It is feasible to propose that social support is an important part of this relationship. With this in mind, the Diabet es Social Support QuestionnaireFamily Version (DSSQ F; La Greca, et al., 2002) was developed as a self report measure to assess the level of diabetes specific support that the child receives from his/her family. Analyses of this measure indicate that the strongest predictor of adherence is parental support of the childs diabetes. Another example o f a commonly utilized measure of family factors in youth with type 1 diabetes is the Diabetes Family Behavior Checklist (DFBC; Schafer et al., 1986). Analysis of the DFBC shows that responses load onto either positive or negative family support factors. Both factors are predictive of adherence and glycemic control in respective directions. Although there are many other well established measures of family functioning, these were used to highlight the importance of assessing any and all family factors that may be contributing to problems with adherence. Further, the analyses of these measures

PAGE 21

21 emphasizes the robust relationships between diabetes specific family factors and both adherence and glycemic control. Quality of Life The biopsychosocial model propos es that physical and psychological factors combine to produce health outcomes (Engel, 1977). With this in mind, physicians are increasingly encouraged to treat the whole patient, using a holistic approach to clinical treatment. The application of this multidimensional approach to health outcomes is frequently referred to as Quality of Life. Research on healthrelated quality of life (QOL) has increased over the past two decades. Going beyond mortality and morbidity, QOL measures typically focus on broad domains of functioning, such as social, physical, academic, and psychological well being. As with most chronic medical conditions, the presence of T1D is bound to have some negative effect on the patients QOL. Due to the developmental, social, and cognitive changes that are universally experienced in childhood through adolescence, we see a particularly dramatic fluctuation in QOL, regardless of health. With this in mind, Varni and colleagues (2007) sought to examine the impact of numerous chronic conditions on youth and adolescents quality of life and compare them to healthy controls. Their research involved over 2,000 youth and adolescents with health conditions including: diabetes (type 1 and 2), cancer, endstage renal disease, cardiac disease, asthma, juvenile rheumatoid arthritis, obesity, and cerebral palsy. They also included a cohort of youth with no chronic health conditions. Their findings shed a great deal of light on the impact of many diseases and subsequent quality of life. All of the disease groups in their study reported significantly lower overall quality of life (both self report and parent proxy) than the healthy control group. However, youth and

PAGE 22

22 adolescents in the diabetes group appear to be the least affected of the disease groups included in the study. Specifically, the diabetes group was the only disease group that did not significantly differ from the control group on subscales of social and school functioning. Breaking the group into specific types of diabetes, the authors showed that youth and adolescents with T2D showed significantly lower levels of overall quality of life, social functioning and academic functioning than those with T1D. Parents of the children with T2D also reported significantly lower levels for their children in the same domains in comparison to parents of children with T1D. This could partially be attributed to the fact that those children with T2D were more likely to be overweight or obese than those with T1D. To further illustrate this as a pos sible explanation, the children in the obesity group in this study performed significantly lower across quality of life measures than the diabetes group. Like many psychosocial factors, quality of life is a continuous variable. Elements that contribute to increases or decreases in quality of life change naturally as a result of developmental changes that occur from youth to adulthood. Patients with diabetes are no different in this context. The factors that affect quality of life in youth versus adolescents have both similarities and differences. In young children with diabetes, the primary source of responsibility for diseasespecific regimen typically involves the childs parent or guardian. As youth develop into adolescence, there is usually some lev el of transition of primary responsibility for disease specific regimen components from the parent(s) to the adolescent (Anderson & Laffel, 1997). Thus, throughout childhood and adolescence, there is typically a strong component of family factors involved. As shown in several research studies (e.g., Duke et al., 2008; Lewin et al., 2006), diabetes -

PAGE 23

23 specific family factors mediate the relationship between adherence and glycemic control. Since both glycemic control and adherence have been shown to be related to quality of life in numerous studies (e.g., Wagner et al., 2005; Hoey et al., 2001; Hesketh et al., 2004; Vanelli et al., 2003), it seems logical to posit that some key factors impacting quality of life in both youth and adolescents include family fact ors, adherence, and glycemic control. One factor influencing quality of life in patients with diabetes appears in adolescence and seems to persist through adulthood. Beginning in adolescence, we see that females with diabetes report lower levels of qual ity of life than their male counterparts. This point is further illustrated in research by the SEARCH for Diabetes in Youth Study Group (Naughton et al., 2008), which examined health related quality of life in youth and adolescents with type 1 and 2 diabetes. Their findings showed a general drop in quality of life across sex in adolescence when compared to youth, which appears to be a relatively universal phenomenon independent of disease (Varni et al., 2007). However, the SEARCH group also showed an age by sex interaction such that female adolescents reported significantly lower quality of life than their male counterparts. This apparent gender gap appears to remain into adulthood in terms of diabetes. When Wexler and colleagues (2006) identified cor relates of healthrelated quality of life in adults with T2D, being female was among the factors that predicted lower quality of life. As seen in previous research, having diabetes naturally impacts quality of life to an extent, although the effect appear s to be relatively low in comparison to some other chronic disease groups. This relationship is strengthened by many contributing factors,

PAGE 24

24 including diabetes specific family factors, adherence, glycemic control, sex, perceived burden, and diabetes specifi c complications. These factors may have a stronger or weaker impact on quality of life as the patient progresses from childhood through adolescence into adulthood. Beginning in adolescence, females with T1D appear to be at more risk for lower quality of life, which may make them more at risk for other associated factors such as poor adherence and glycemic control. Given the relationship between quality of life and many health outcome factors, it would behoove clinicians to utilize interventions that take quality of life into consideration when treating a person with diabetes, regardless of age. Adherence and Quality of Life There has been a growing amount of research in pediatric psychology focusing on medical adherence and potential contributing factor s to nonadherence or poor adherence. This increase in focus is due in part to epidemiologic research indicating that as much as 50% of pediatric patients are nonadherent to their medical regimen (Rapoff, 1999; Matsui, 1997). With this in mind, much of the research has focused on identifying predictors of adherence in this population. Quality of life has been thought to be associated with adherence. In a study examining children and adolescents with T1D, Wagner and associates (2005) sought to identify correlates of quality of life. Their findings indicated that lower age (preadolescence), better glycemic control, and an intensive insulin regimen were all associated with better quality of life. Of importance in these findings is the correlation between good glycemic control and quality of life, as adherence has also been shown to be associated with glycemic control. Research by Laffel and colleagues (2003) also focused on quality of life in youth with type 1 diabetes. Interestingly, their

PAGE 25

25 results sh owed a negative relationship between diabetes specific family conflict as perceived by the child and the childs self reported quality of life. A great deal of research has shown that diabetes specific family factors mediate the relationship between adher ence and glycemic control in youth and adolescents with type 1 diabetes (e.g., Duke et al., 2008; Lewin et al., 2006). Thus, evidence that quality of life is associated with diabetes specific conflict, a family factor, begins to develop a clearer picture of the relationship between quality of life and both adherence and glycemic control as it pertains to this disease population. In a larger scale study, researchers from the Hviodore Study Group on Childhood Diabetes (Hoey et al., 2001) sought to examine the relationship between glycemic control and quality of life in adolescents with type 1 diabetes. They expanded the scope of their study to include participants from 11 countries across Europe, North America, and Asia. The results of their study suggest ed that the relationship between glycemic control and quality of life appears to be somewhat universal and unaffected by cultural and ethnic differences. Participants in their study showed a gradual decrease in both quality of life and glycemic control when comparing younger adolescents to older ones. Although some of this reduction in quality of life is seen across adolescents regardless of disease (Varni et al., 2007), researchers also attribute part of this apparent drop to difficulties related to the transfer of diseasespecific regimen behaviors as the child transitions into young adulthood (Anderson & Laffel, 1997). Another interesting finding from the Hviodore Group showed that female adolescents report increased worry, lower life satisfaction, and worse physical health at an earlier age in adolescence than their

PAGE 26

26 male counterparts, although both show a gradual decrease when compared to younger patients. Naturally, whenever the discussion focuses on pediatric medical regimen adherence, it is important to discuss the involvement of their parents. The research is fairly clear that positive involvement by the parents is important, although over involvement has sometimes been shown to be counter productive. Graue and colleagues (2005) conducted a study examining the effect that parents may have on their childs glycemic control and their quality of life. As would be expected, their findings indicated that factors improving both glycemic control and quality of life include: parent child cohesion on disease specific tasks, parental support of diseasespecific behaviors, and a positive level of involvement. Their results also showed that both over involvement and negative support behaviors tend to have a negative effect on both glycemic control and qualit y of life. Research from Vanelli and colleagues (2003) produced similar results, suggesting that good glycemic control is associated with good quality of life. Further, they showed that parent child conflict predicted both poorer glycemic control and quality of life. There has been extensive research on pediatric T1D patients involving some combination of the variables discussed (i.e., QOL, adherence, family factors, glycemic control, and demographic information). However, the relationship between all o f these variables simultaneously has not been adequately investigated. It is with this in mind that the current study was proposed. Based upon the existing research, the current study aimed to examine all of the aforementioned variables and hypothesized a contribution to the literature by creating a model that accounts for a greater total amount

PAGE 27

27 of explained variance in glycemic control. With this information, more effective interventions can be developed in order to improve patient health status. Specif ic Aims and Hypotheses The current study proposed to investigate the relationships of child, parent, and family level variables on adherence, quality of life, and glycemic control in pediatric T1D patients. Long term goals of this area of research are to develop more effective interventions designed to improve overall QOL, adherence, and health status specifically for this patient population. Specific Aim 1 The first aim of the proposed study is to investigate the relationship between adherence, quality of life, and glycemic control in youth and adolescents with T1D. Hypothesis 1 Based on research indicating a positive relationship between quality of life and both adherence and glycemic control (Wagner et al., 2005; Laffel et al., 2003; Vanelli et al., 2003), it is expected that quality of life will mediate the relationship between adherence and glycemic control. For the current study, two regression models were conducted using the separate parent and child reports of adherence (DSMP) and quality of life ( P edsQL Core) scores. This was done in order to accurately portray these domains as assessed by the patients and their caregivers. Hypothesis 1a Based on previous research suggesting that adherence and family factors have accounted for as much as 40% of the variance in glycemic control (Duke et al., 2008), it is expected that the addition of quality of life to this model will explain at least 45% of the variance in glycemic control.

PAGE 28

28 Specific Aim 2 The second aim of the proposed study is to investigate the rel ationship between quality of life and other behavioral factors in youth and adolescents with T1D. Hypothesis 2 Given the extant research in this area (Duke et al., 2008; Graue et al., 2005), it is expected that the associations between behavioral factors ( adherence, diabetes specific family factors) and quality of life outcome measures ( PedsQL Core and Diabetes modules) will be statistically significant. Specific Aim 3 The third aim of this study is to investigate the relationship between quality of life and biological factors (i.e., age, sex, race, and diabetes related complications) in pediatric T1D patients. Hypothesis 3 Given the research suggesting that QOL (as measured in the current sample by the separate child and parent report PedsQL Core and Diabetes modules) in young patients with T1D is negatively impacted by both increasing adolescent age and being female (Naughton et al., 2008; Varni et al., 2007; Hoey et al., 2001), it is expected that there will be significant negative correlat ions between quality of life and both age and being female. Based on research examining QOL and diabetes specific complications (Huang et al., 2007; Wexler et al., 2007) it is expected that there will be a negative relationship between QOL and number of diabetes specific complications.

PAGE 29

29 Hypothesis 3a Based on previous research indicating poorer glycemic control in ethnic minorities (Petitti et al., 2009; Chalew et al., 2000), it is expected that ethnic minority status will be negatively associated with glyc emic control. Specific Aim 4 The fourth aim of the proposed study is to compare patient self reports and parent observed reports on the measures of QOL in pediatric T1D patients. Hypothesis 4 Given that parents of chronically ill children tend to rate their childs physical and psychosocial functioning as lower than the childs report (Guyatt et al., 1997), it is expected that parent observed reports on the PedsQL Core and Diabetes modules will rate childs psychosocial functioning more poorly than the chi lds self report in this domain.

PAGE 30

30 CHAPTER 2 METHOD Participants Participants were 70 youth with T1D and their parent(s)/caregiver recruited from the outpatient Pediatric Endocrinology Clinic at the University of Florida in Gainesville, Florida. Our orig inal goal was 100 youth/parent dyads. For reasons that are unknown, patient populations were sparser than had previously been seen at these clinics. Thus, recruitment was slow over the 12 months in which it occurred prior to the current analyses. Inclus ion criteria for study participation were 1) patient age 818, 2) diagnosed with T1D for at least six months, 3) living with and accompanied by their primary caregiver, 4) no other chronic medical conditions, 5) no diagnosis of mental retardation or psychotic disorder (child or caregiver), and 6) ability to read, write, and understand English (child and caregiver). Regarding demographics, it was expected that our sample population would be similar to distributions noted in recent census results for Alachu a County, Florida. These statistics suggested that our sample will likely be overwhelming Caucasian (~65 70%), with smaller popula tions of African Americans (~20 25%), Hispanic/Latino (~5 10%), and other ethnic minorities. Procedure Participants and their primary caregivers were recruited during their regular visits to the Pediatric Endocrinology Clinic. Clinic nurses identified patients who met inclusion criteria. Upon receiving this information from clinic nurses, trained members of the research team approached patients and their primary caregivers to explain the study. Qualifying participants received a $10 gift card at the completion of their involvement. The research staff obtained a signed inform consent from the legal guardi an of all

PAGE 31

31 participants in the study. Assent from participating children was obtained when possible. The pediatric patients and their primary caregivers were separately interviewed about T1D regimen adherence and independently completed subsequent questionnaires. It was expected that total completion time for participation in the study wo uld be approximately 30 45 minutes. Research team members were available throughout the testing process in order to provide clarification when needed. In order to obtai n measurement of glycemic control (A1c), pediatric endocrinology nurses obtained finger stick blood samples as part of the patients routine care at the clinic. Measures Measures Completed by the Pediatric T1D Patient Diabetesspecific family factors. The Diabetes Family Behavior Scale (DFBS; Waller et al., 1986) was used to assess T1D specific family functioning. This 60item self report measure contains three subscales (warmth/caring, guidance/control, and problem solving) that assess the childs perceptions of family support in the context of T1D self management tasks. From a psychometric standpoint, the DFBS has shown retest reliability for the 0.79 and 0.83, respectively) (Waller et al., 1986). Previous research has indicated that, while the warmth/caring and guidance/control subscales are associated with glycemic control (r = .36 and .50, respectively), there exists only a weak correlation between glycemic control and the problem solving subscale. As such, this subscale was not used in the current study. Glycemic control. Glycemic control is a biological marker of health status in patients with diabetes. Currently, the measure of choice f or glycemic control is the glycosylated hemoglobin (A1c) test. The A1c provides an accurate estimate of the

PAGE 32

32 patients glycemic control over the past two to three months. Higher A1c values are indicative of poorer glycemic control. For patients with T1D, the ADA suggests that optimal figures representing good glycemic control are those values of A1c less than 6.5% (ADA, 2010). In the current study, pediatric T1D patients had an A1c test done during their visit to the Pediatric Endocrinology Clinic as par t of routine care. Patients routinely have their blood drawn and A1c checked as part of their regularly scheduled appointments. Blood samples were analyzed by trained, qualified, and experienced technicians from the Pediatric Endocrinology Team using a Bayer DCA 2000+. For this study, blood samples were analyzed by trained members of the Pediatric Endocrinology Team. Measures Completed by the Parent/Primary Caregiver Demographics. The parent/primary caregiver was asked to complete a standard demographic i nformation form that solicited information including age (of child and caregiver), sex (of child and caregiver), education, income, duration of T1D, and family structure (e.g., one parent, twoparent, grandparent, etc.). Measures Completed by the Parent/ Pr imary Caregiver and Pediatric T1D Patient Diabetesspecific family factors. The Diabetes Family Behavior Checklist (DFBC; Schafer et al., 1986) (parent and child versions) was used to assess family functioning with regard to pediatric T1D treatment regime n. It is a 16 item self report measure containing both a positive/supportive and a negative/unsupportive scale. Higher scores indicate more negative diabetes specific family factors. Research on this measure has indicated that the 7item negative/unsupportive scale is most strongly associated with T1D adherence and glycemic control (Lewin et al., 2006; Shafer et al.,

PAGE 33

33 1986). Therefore, the current study only utilized the 7 negative/unsupportive items for analyses. The Diabetes Family Conflict Survey (DFCS ; Hood et al., 2007) was used as an additional measure of family functioning in the context of the patients T1D treatment regimen. It is a 19item self report measure (child and caregiver versions) that includes questions related to diet, insulin, blood glucose testing, and school. Rates of internal diabetes specific conflict) (Hood et al., 2007). Qualit y of life. The PedsQL Pediatric Quality of Life Inventory Core module (Child Self Report and Parent Report Versions) ( PedsQL ; Varni et al., 1999) were used to assess pediatric QOL. The PedsQL version used in this study includes the core module, measuring generic QOL in the last month with separate versions for children 8 12 years of age and teens 13 18 years of age. The parent version of the core module also measures generic QOL of the child, with separate versions for parents to answer for their 8 12 y ear old child or for their 13 18 year old teen. Both versions of this 23item module assesses four scales (Physical 8 items, Emotional 5 items, Social 5 items, and School 5 items) which generate three summary scores: (1) Total Scale Score23 items, (2) Physical Health Summary Score8 items, and (3) Psychosocial Health Summary Score 15 items. On the Child Self Report version, all items are statements to which the child is asked to rate their agreement on a 5item Likert scale ranging from Never to Alm ost Always, and include statements such as It is hard to pay attention in class, and I worry about what will happen to me. The statements on the Parent

PAGE 34

34 Report version are slightly modified from the Child Self Report version and ask the parent/guardian to rate how much of a problem their child has had in each of the on both versions for the Total Scale Score (Child Self = 0.90), Physical Health Summary Score (Child Self 0.88), and Psychosocial Health Summary Scale (Child Self validity, the P edsQL differentiates not only between healthy children and acutely or chronically ill children but also between levels of disease severity within a given medical condition. Higher scores on the PedsQL indicate better quality of life. The PedsQL Pediatric Quality of Life Inventory Diabetes module (Varni et al., 2003) was given to both child and parent to assess diabetes specific QOL. It is a 28 item, self report measure that assesses functioning on five scales (Diabetes Symptoms, Treatment Barriers, Treat ment Adherence, Worry, Communication). The Diabetes module was found to have satisfactory internal consistency reliability (Child Self Report patients with diabetes. Adheren ce. The Diabetes Self Management Profile (DSMP; Harris et al., 2000) was used to assess T1D regimen adherence. This structured interview consists of 23 questions in the domains of insulin administration/dose adjustment, blood glucose monitoring, exercise, diet, and management of hypoglycemia. This measure assesses adherence in each of these domains over the past three months. The DSMP interview was administered separately to pediatric patients with T1D and their respective

PAGE 35

35 parent/caregiver by trained me mbers of the research team. Completion of the interview is typically accomplished in approximately fifteen minutes. Responses are coded on scales ranging from either 0 to 1 or 0 to 4, with higher ratings indicating better regimen adherence. The combined ratings are summed to create a total adherence score, which can range from 0 79. Previous psychometric research on the DSMP has shown observer agreement (94%) (Harris et al., 2000). Higher scores on the DSMP indicate better adherence. In order to be properly trained to administer the DSMP, research team members were required to complete a series of trainings. They were first given background materials on diabetes in order to educate them o n the disease. The DSMP was reviewed with them and scoring was demonstrated. Team members were then required to observe and score mock interviews, and to practice until able to meet specific criteria. Once in the Pediatric Endocrinology Clinic, they wer e required to independently score observed clinic interviews until such time that there was a minimum of 90% inter rater reliability. When this level of agreement was attained, the team member was allowed to conduct live interviews, with the first two bei ng supervised to ensure continuity. Throughout the data collection phase, all team members were checked 3 times per year for reliability. For a subset of the current sample, inter rater reliability was 93%. Statistical Analyses All data processing and st atistical analyses were conducted using the statistical software package, PASW Statistics 18.0.0 by SPSS. Parallel analyses were conducted on the childreport measures and the parent report measures. Analyses were conducted on both child and parent report versions in order to accurately assess the domains measured. Although our desire was to trust in

PAGE 36

36 the accuracy of child self report on psychosocial measures, there is research to suggest that chronically ill children are more susceptible to social desirability response bias than their parents (Logan et al., 2008). Thus, parent report versions were separately analyzed as an additional source of information. Previous research suggests that parent reported psychosocial measures of chronically ill children are a reliable and accurate method of assessment (Le Coq et al., 2000). Preliminary analyses were conducted to test for relationships between independent and dependent variables for purposes of control in subsequent analyses. To identify significant relationships between respective continuous factors, a correlation matrix was computed. Dichotomous variables were compared using independent samples t tests. This also eliminated demographic items that were not significantly related to study measures in sub sequent analyses. Descriptive statistics were obtained, including mean age, sex ratio, and mean duration of disease. In addition, these statistics identified variables that may have influenced results through skewness or kurtosis. As correlational analy ses were conducted, the criterion value for significance was adjusted according to the number of comparisons in order to avoid any family wise errors. In order to examine the relationships between variables for purposes of control, regression analyses were used with glycemic control (A1c) as the dependent variable. Separate regressions were conducted using child and parent report versions of study measures. Independent variables were comprised of demographic information, diabetes specific family factors, adherence, and QOL. The possible presence of mediation effects was tested using hierarchical linear regression in accordance with Baron and Kennys (1986) guidelines: 1) the predictor should be significantly associated

PAGE 37

37 with the outcome variable, 2) the pr edictor should be significantly associated with the mediator, 3) the mediator should be associated with the outcome variable, and 4) the addition of the mediator to the full model should significantly reduce the relationship between the predictor variable and the outcome variable. Standardized coefficients were calculated to examine changes in path coefficients with the addition of the mediator to the model. Sobels Z Scores were calculated to assess significance of changes in path coefficients (Sobel, 1988). Additionally, Pearsons correlations were conducted on the proposed factors to examine any possible relationships between them. For the regressions, only those variables from each block that were shown to be statistically significantly related to the dependent variable (glycemic control) were included. There were not a sufficient amount of significant correlations between study variables when examining the childreport versions to justify a regression analysis. However, the parent report versions o f study variables were correlated to a necessary degree to conduct regression analyses. The parent report hierarchical regression was conducted as follows: Block 1Significant Demographics (Race); Block 2Significant Family Factors (DFBC); Block 3 Adherence (DSMP); Block 4 Q uality of Life ( PedsQL Core module). It was expected that this full model would account for at least 45% of the variance in glycemic control (Hypothesis 1a) To examine the relationship between patient self report and their respective p arent observed reports, independent samples t tests were conducted comparing the PedsQL Child and Parent Proxy versions. Since multiple statistical analyses were conducted on the same data, a modified Bonferroni correction factor was implemented

PAGE 38

38 to contro l for any family wise error and the resulting criterion level of significance was reduced accordingly.

PAGE 39

39 CHAPTER 3 RESULTS Demographics of Participants A total of 70 child/parent dyads were enrolled in this study. The participants in this study were predom inantly Caucasian (75.0%), with 18.8% identifying themselves as African American, 1.6% as Hispanic, and 4.7% as other ethnicities. This sample was similar in ethnic diversity to what was expected when compared to the demographic makeup of Alachua County, Florida. Children in the current study ranged in age from 9 to 18 ( M = 13.72, SD = 2.56), with parents/caregivers ages 24 to 68 ( M = 43.84, SD = 9.46). The pediatric patients in this study were 59.4% male and 40.6% female. Caregivers who participated in this study self identified as 56.7% mothers, 25.4% fathers, 9% grandparents, 3% stepfathers, and 6% other. The sex distribution of the caregivers was 67.8% female and 32.2% male. Detailed descriptive statistics are shown in Table 3 1. Relationships b et ween Demographic Variables and Study Variables Glycemic Control (A1c) Correlational analyses and independent samples t tests were conducted to identify significant relationships between demographic variables (e.g., age, race, sex, number of adults in the home) and the independent variable, glycemic control (A1c). A Pearsons r correlation matrix was conducted comparing all of the continuous demographic variables (i.e., number of children in the home, years since diagnosis, number of missed school days in the past year, and number of phone calls to the endocrinologist in the past month). A total of five demographic compari sons were run in this manner, necessitating a modified Bonferroni adjustment to avoid family wise errors. Since five

PAGE 40

40 comparisons were run, the criterion for statistical significance was reduced to p = .01 (.05/5). None of the aforementioned continuous de mographic variables were significantly associated with glycemic control T he number of episodes of diabetic ketoacidosis was significantly associated with glycemic control ( r (42) = .509, p = .001). It is intuitive that fewer episodes of DKA, a potential consequence of poor glycemic control, would be associated with better glycemic control, with which the findings in the current sample were consistent. The medical nature of DKA as well as the high, yet intuitive, correlation with glycemic control precludes this variable from being included as a variable in regression analyses. Independent samples t tests were conducted comparing the dichotomous demographic variables to glycemic control. Initially, sex was the only dichotomous demographic variable. However, several other demographic variables (i.e., marital status, race, relationship to child, with whom the child resides, level of parental education, household income, age of the child, and number of endocrinologist appointments in the past year) were collapsed into dichotomous variables due to the low frequency of some responses as well as for clarification. A total of nine t tests were conducted, resulting in a criterion for significance reduced to p = .006. None of the demographic variables were significantly associated with glycemic control at this criterion. However, a t test comparing Caucasians to all other races suggested a trend toward significant differences on A1c levels ( t (62) = 2.56, p = .013). DiabetesSpecific Family Factors Multiple correlations were run to examine any significant relationships between the demographic variables and our measures of diabetes specific family factors and adherence. Separate correlations wer e run using the child and parent reported

PAGE 41

41 versions of these measures when applicable. However, following necessary adjustments to the criterion level for significance, no significant relationships were identified between demographic variables and these st udy measures. A trend toward significance was seen in the t test comparing childreported adherence between households with or without both parents ( t (66) = 2.39, p = .02). A trend was also seen comparing parent reported adherence to whether or not thei r child is a teenager ( t (54) = 2.45, p = .018). Quality of Life Relationships between quality of life and demographic variables were assessed in the hypothesis testing section of this paper, as this is our third hypothesis. All demographic variables demons trating significant relationships with the dependent variable were controlled for in subsequent analyses. Intercorrelations among Study Variables Glycemic Control Correlational analyses were conducted to identify study variables that were significantly ass ociated with the dependent variable (A1c) for subsequent analyses. Separate analyses were again conducted using the child and parent reported versions of study variables. Analyses of the childreported versions of study variables involved six comparisons, therefore the criterion for significance was reduced to p = .008. Analyses demonstrated significant relationships between our dependent variable of glycemic control (A1c) and the childreported Diabetes Family Behavior Checklist (DFBC; r (68) = .348, p = .004), a measure of diabetes specific family conflict (family factors). Glycemic control was significantly associated with our childreported measure of adherence to the

PAGE 42

42 medical regimen, the Diabetes Self Management Profile (DSMP; r (68) = .403, p = .001) Neither child reported version of the PedsQL (Core and Diabetes modules) was significantly associated with glycemic control. Correlational analyses comparing the parent reported versions of study measures to the dependent variable of glycemic control involved four comparisons. Thus, the criterion for significance was reduced to p = .013. Results showed a significant relationship between glycemic control and parent reported diabetes specific family conflict on the DFBC ( r (68) = .297, p = .012). A sig nificant negative relationship was exhibited between glycemic control and parent reported adherence on the DSMP ( r (65) = .386, p = .001). An additional negative correlation was shown between glycemic control and parent reported quality of life as measured on the PedsQL Core module ( r (56) = .457, p = .001). The parent reported version of the PedsQL Diabetes module score was not significantly associated with glycemic control in the current sample at our adjusted criteria ( r (56) = .270, p = .044), although a trend towards significance was demonstrated. DiabetesSpecific Family Factors In addition to the aforementioned correlations with the dependent variable of glycemic control, several significant associations were found between the parent reported dia betes specific family factor measure (DFBC) and other independent variables after adjusting for the number of comparisons on the same data. Scores from the parent reported version of the DFBC were negatively associated with parent reported adherence to medical regimen (DSMP; r (67) = .354, p = .003), overall quality of life ( PedsQL Core; r (55) = .401, p = .002), and diabetes specific quality of life ( PedsQL Diabetes; r(55) = .488, p < .001). No significant relationships were shown

PAGE 43

43 between childreported measures of diabetes specific family factors and other test variables. Quality of Life Significant relationships between the various quality of life outcome measures and other independent variables are addressed below as this was our second hypothesis. A comprehensive description of all correlations between childreported study variables is presented on Table 3 3. A correlation matrix of parent reported study variables is shown on Table 3 4. Hypothesis Testing Relationships b etween Q uality of L ife, A dher ence, and G lycemic C ontrol ( Hypothesis 1) To examine the hypothesis that quality of life will mediate the relationship between adherence and glycemic control, Baron and Kennys (1986) steps for mediation were conducted. Parallel regressions were conducted using the childreported and parent reported versions of the measures. Examining the childreport versions of the measures, a regression was conducted using adherence (DSMP) as a predictor of glycemic control (A1c). This analysis showed that adherence was significantly correlated with glycemic control ( r (67) = .403, p = .001), thus satisfying the first requirement of mediation. The path coefficient (c) was computed to be .081, SE = .022 The second step of Baron and Kennys test of mediation requi res a significant relationship between the initial variable (adherence) and the mediator variable (quality of life). Separate regressions were subsequently conducted using the PedsQL Core module and the PedsQL Diabetes module as the dependent variable and adherence (DSMP) as the independent variable.

PAGE 44

44 No significant correlation was demonstrated between childreported adherence and the PedsQL Core module ( r (55) = .214, p = .113). Therefore, the second requirement of mediation was not met with the PedsQL Co re module as the measure of quality of life. Utilizing the PedsQL Diabetes module as a measure of quality of life also produced a nonsignificant correlation with the DSMP ( r (54) = .164, p = .231). Since the second requirement for mediation was not met using the childreported versions, no additional mediation analyses were run using these versions. Mediation analyses were conducted on the parent reported versions of the current measures. The first step of the mediation process, demonstrating a relations hip between the initial variable (adherence) and the outcome variable (A1c), showed a significant relationship ( r (67) = .386, p = .001). The path coefficient (c) was .069, SE = .02. Since the parent report PedsQL Core module was the only quality of lif e measure significantly associated with the dependent variable (A1c) at the adjusted level of significance, it was used in the second step of the mediation process. Results indicated that there was a significant relationship between the parent reported versions of the PedsQL Core module and DSMP ( r (56) = .389, p = .003), satisfying the second step of mediation. The path coefficient (a) between adherence and quality of life was .534, SE = .177. The third step in this mediation process requires showing that the mediator (quality of life) has an effect on the outcome variable (A1c) while controlling for the initial variable (adherence). A regression was run using A1c as the dependent variable and both parent reported adherence (DSMP) and quality of life ( Pe dsQL Core total score) as predictors. Analyses showed that, after controlling for adherence, quality of life had a significant negative relationship with glycemic control ( r (56) = .457, p = .011).

PAGE 45

45 Additionally, the path coefficient (b) between quality of life and glycemic control was .053, SE = .016. These findings met the requirements of the third step. The final step of Baron and Kennys mediation testing examines whether partial or complete mediation is occurring. The path coefficient (c) between adherence and glycemic control was .041, SE = .022. Since the subtraction of c from c ( .069 ( .041) = .028) does not equal zero, partial mediation was demonstrated. To test the significance of the indirect effect of adherence on glycem ic control via quality of life, a Sobels Z statistic test was run. For this sample, Sobels Z = 2.23, SE = .013, p = .026. Thus, in this sample, quality of life significantly mediated the effect of adherence on glycemic control, which supports our first hypothesis. The mediation model is shown in Table 3 5. Accounted Variance in Glycemic Control by the Model (Hypothesis 1a) To examine the hypothesis that the current model (demographic variables, family factors, adherence, and quality of life) would account for at least 45% of the variance in glycemic control, a hierarchical multiple linear regression analysis was conducted using parent report versions of the study measures. Since race was the only demographic variable previously identified as trending toward a significant association with the dependent variable (A1c), it was the only variable entered into the first block of the equation. The DFBC score was entered in the second block to represent diabetes specific family factors, as it was significantly correlated with glycemic control in the current sample. The parent report DSMP score, our study measure of adherence to medical regimen, was entered into the third block. It was also significantly associated with glycemic control. The parent report PedsQL Core module score, which was significantly correlated with glycemic control, was entered into the fourth block as our measure of quality of life. The overall model accounted for approximately 41.6% of the

PAGE 46

46 variance in glycemic control in the current sample ( r = .416; F (4,44) = 7.12, p < .001). In order to account for any multicollinearity between study variable, t he variance inflation factor (VIF) was calculated during the regression. VIF values in the current regression were much less than 10 (maximum VIF = 1.397), which signified no potential problems with multicollinearity in the current model. Although this model accounted for a significant portion of the variance, it did not meet our expectations and thus Hypothesis 1a was not supported. The regression model is shown in Table 3 6 Relationship b etween Q uality of L ife and Behavioral V ariables (Hypothesis 2) In order to test our second hypothesis that quality of life outcome measures would be significantly related to other behavioral factors relevant to diabetes (i.e., adherence to medical regimen and diabetes specific family factors), Pearsons r correlations w ere conducted on these continuous variables. Separate comparisons were conducted using child and parent report versions of measures when applicable. Quality of Life and Adherence The child report PedsQL Core and Diabetes modules were not significantly associated with any of the other childreport study measures. However, the parent report PedsQL Core was significantly associated with our parent report measure of regimen adherence (DSMP; r (55) = .389, p = .003). The parent report PedsQL Diabetes module was also significantly associated with parent reported adherence in this sample ( r (55) = .494, p < .001). Quality of Life and DiabetesSpecific Family Factors Correlational analyses identified multiple sig nificant relationships between parent reported quality of life measures and the parent report measure of diabetes specific family conflict in the current study. The parent report PedsQL Core was significantly

PAGE 47

47 negatively associated with our parent report m easure of diabetes specific family conflict (DFBC; r (55) = .401, p = .002). The parent report PedsQL Diabetes module was also significantly associated with the parent reported DFBC ( r (55) = .488, p < .001). Overall, significant correlations were identi fied in the current sample between parent observed quality of life and both adherence and diabetes specific family factors, lending support to our second hypothesis. Relationship b etween Quality of Life and Biological Factors (Hypothesis 3) Independent sam ples t tests were run to compare the dichotomous biological variables to the quality of life measures used in this study with a Bonferroni corrected criterion for significance of p = .005. Once again, separate analyses were conducted using the child and parent report versions of the quality of life measures. No significant relationships were demonstrated between either version of the PedsQL modules and whether or not the child was a teenager. The number of episodes of diabetic ketoacidosis (DKA) in the past year was negatively associated with the childreport PedsQL Core score ( r (42) = .436, p = .004). There were no significant relationships demonstrated between the parent re port quality of life measures and number of DKA episodes. Analyses in the current study did not identify any differences between sexes on either child or parent reported measures of quality of life. Overall, significant negative correlations were identi fied between childreported quality of life and the number of diabetes specific complications in the current sample. No quality of life differences by sex or age were identified in the current study. Thus, our third hypothesis is only partially supported.

PAGE 48

4 8 Association between Glycemic Control and Ethnic Minority Status (Hypothesis 3a) As previously mentioned, an independent samples t test comparing Caucasians to all other races suggested a trend toward significant differences on A1c levels after adjusting the criterion for significance ( t (62) = 2.56, p = .013). Post hoc analyses in the current study showed that Caucasian children had significantly lower A1c levels ( M = 8.90, SD = 2.05) than African American children ( M = 10.33, SD = 2.17; t (58) = 2.14, p = .036). Together, these findings add support to our hypothesis that there is an association between being an ethnic minority and having poorer glycemic control. Child Self Report v ersus Parent Observed Report of Psychosocial Functioning (Hypothesis 4) Independent samples t tests were conducted to examine the relationship between patient self report and their respective parent observed reports on the PedsQL Core and Diabetes modules. Multiple comparisons were conducted on the same data; necessitating a modified Bonferroni correction to avoid family wise error. This was accomplished by dividing the standard level for significance, .05, by the number of comparisons, which was eight. Therefore, for these analyses values for significance were set at p = 006. Statistical analyses indicated that parents rated their childs functioning significantly lower than the patient self report on the PedsQL Core total score ( t (111) = 3.76, p < .001) and the PedsQL Diabetes module score ( t (110) = 4.07, p < .001). Sev eral subscales were also significantly different. These results support our fourth hypothesis. Means and standard deviations for PedsQL total scores and subscale scores by parents and patients are shown in Table 3 7.

PAGE 49

49 Table 3 1 Patient Demographic Dat a N Valid% M SD Range Child Age 12 or Younger 13 or Older Sex ( n = 69) Male Female Race/Ethnicity ( n = 64) Caucasian African American Hispanic Other 57 21 33 41 28 48 12 1 3 38.9 61.1 59.4 40.6 75.0 18.8 1.6 4.7 13.72 2.56 9 18 Parent/Guardian Relationship to child ( n = 67) Mother Father Grandparent Step Father Other Marital Status ( n = 55 ) Single/Never Married Married/Remarried Separated/Divorced Widowed Annual family income ( n = 54 ) Below $9,999 $10,000 $19,999 $20,000 $39,999 $40,000 $59,999 $60,000 $79,999 $80,000 and above Number of Children in Home ( n = 54) One Two Three Four Five Eight 38 17 6 2 4 13 32 9 1 11 11 13 7 7 5 18 17 16 1 1 1 56.7 25.4 9.0 3.0 6.0 23.6 58.2 16.4 1.8 20.4 20.4 24.1 13.0 13.0 9.3 33.3 31.5 29.6 1.9 1.9 1.9

PAGE 50

50 Table 32. Child Sample Diabetes Specific Characteristics M Range SD Years Since Diagnosis ( n = 30) A1c Level Today ( n = 69) Number of Hospitalizations ( n = 54) Number of Episodes of Diabetic Ketoacidosis (DKA) ( n = 42) Number of Endocrinologist Appointments in Past Year ( n = 50) Number of Missed Endocrinologist Appointments in Past Year ( n = 51) Average Number of Times Called Endocrinologist in One Month ( n = 50) Number of Missed Days of School in Past Year ( n = 52) 5.44 9.27 0.87 0.86 4.00 0.43 0.74 7.83 4.31 2.15 0.99 1.37 2.31 0.76 0.88 10.08 .88 15.00 5.40 14.00 0 6 0 5 1 12 0 3 0 4 0 55

PAGE 51

51 Table 33. Child Reported Study Variables Correlation Matrix p < .05 ** p < .01 A1c DFBC DFCS DFBS DSMP PedsQL Core PedsQL Diabetes A1c DFBC .348** DFCS .047 .004 DFBS .116 .101 .009 DSMP .403** .189 .159 .021 PedsQL Core .224 .154 .124 .037 .214 PedsQL Diabetes .196 .057 .016 .019 .164 .676**

PAGE 52

52 Table 3 4. Parent Reported Study Variables Correlation Matrix p < .05 ** p < .01 A1c DFBC DSMP PedsQL Core PedsQL Diabetes A1c DFBC .297* DSMP .386** .354** PedsQL Core .457** .401** .389** PedsQL Diabetes .270* .488** .494** .777**

PAGE 53

53 Table 35 Multiple Regression Analysis with Parent Report Measures Predicting A1ca Step Variables(s) R 2 R 2 F 1 Demographics Race .117 .117 5.73* .354** 2 Family Factors DFBC .156 .038 3.87* .018 3 Adherence DSMP .276 .120 5.20** .217 4 Quality of Life PedsQL Core Score .416 .140 7.12*** .425** a p < .05 ** p < .01 *** p < .001

PAGE 54

54 Table 3 6 Quality of Life Means and Standard Deviations by Child Report and Parent Report PedsQL Scale Child Report ( n = 57) M SD Parent Report ( n = 56) M SD Total Scale Score 82.61** 11.23 72.00*** 18.05 Psychosocial Health Summary Score 79.74** 12.75 69.85** 18.77 Physical Health Summary Score 88.02** 11.20 77.26** 22.09 Diabetes Module Total Summary Score 75.52*** 10.20 66.44*** 13.24 Diabetes Symptoms Summary Score Diabetes Treatment Summary Score Diabetes Worry Summary Score Diabetes Communication Summary Score 65.40* 14.29 85.28*** 10.45 71.58 21.37 80.80* 20.84 58.08* 16.43 74.10*** 14.28 66.07 25.07 70.98* 28.52 Values are significantly different at: p < .05 ** p < .01 *** p < .001

PAGE 55

55 Figure 3 1 Mediation Model and Path Coefficients for Parent Reported Quality of Life, Adherence, and A1c Quality of Life ( PedsQL Core) Adherence (DSMP) Glycemic Control (A1c) b = .0 53 SE = .01 6 r ( 56) = 457** a = 534 SE = 177 r (56) = .389** c = .069, SE = .02 r (68) = .386** Sobels Z = 2. 23* SE = .0 13 p < .05 ** p < .01

PAGE 56

56 CHAPTER 4 DISCUSSION The goal of the current study was to examine the relationships between diabetes specific family factors, adherence to medical regimen, quality of life, and glycemic control. This study is unique in that it is one of the first studies to incorporate all of these independent variables in an examination of the psychosocial factors that influence glycemic control in youth and adolescents with Type 1 diabetes. Given the complexity of a typical T1D medical regimen, combined with the serious potential complicati ons associated with poor glycemic control, it is extremely important to examine factors that may predict better disease management in this population. This study found that, in the current sample population, quality of life partially mediated the relationship between adherence to medical regimen and glycemic control. A portion of adherence accounted for a change in quality of life, which in turn accounted for a change in glycemic control above and beyond the change in glycemic control explained by adhere nce. This is an important finding because although intuitively we would expect there to be a perfect relationship between adherence and glycemic control, this is not the case in reality. Therefore, there must exist other elements beyond adherence that in fluence glycemic control. Previous research identified some diabetes specific family factors as mediators of glycemic control (Duke et al., 2008; Lewin et al., 2006; Anderson et al., 2002). Quality of life has been shown to be significantly associated wi th glycemic control in multiple studies (Grau et al., 2005; Vanelli et al., 2005; Wagner et al., 2005; Hoey et al., 2001). Prior to the current study, there has been no published research attempting to show that quality of life mediates the relationship b etween adherence and glycemic control. By identifying quality of life

PAGE 57

57 as a mediator of the relationship between adherence and glycemic control, we further clarified the need for quality of life to be taken into consideration with regard to interventions i n this area. However, despite the findings of our mediation model, the explanation of this relationship is not entirely clear. It appears from our model that quality of life mediates the relationship between adherence and glycemic control; however, thi s is not a certainty. What can be said with certainty is that the three constructs are heavily interrelated. When discussing the involvement of adherence and quality of life in this model, the inevitable chicken or egg quandary presents itself. Is it plausible that poor quality of life unrelated to diabetes results in youth with T1D being less motivated and concerned with their complex medical regimen, thereby resulting in poor adherence and poor glycemic control? Is it possible that poor adherence du e to unexplained factors leads to poor glycemic control which then result in poor quality of life? Is it plausible that the possibility of poor glycemic control spurns poorer adherence and quality of life? At this point, a final understanding of these possibilities remains unknown. The question could likely be askedwhy is quality of life important in patients with Type 1 diabetes? Given the complexity, inconvenience, and discomfort of an intensive diabetes regimen, psychosocial variables such as quality of life are the factors that can likely influence motivation and adherence. As Richard Rubin (2000) stated in his editorial, poor quality of life may to contribute to a to hell with it! attitude which could lead to poor self care, poor adherence, a nd ultimately poor diabetes health status. Intangible factors such as quality of life appear to make a difference with regard to

PAGE 58

58 adherence. This is particularly important during periods in which adherence could be more difficult such as when feeling ill, tired, or angry. Similar to results from previous research, the current study showed that negative diabetes specific parenting as reported by both the child and parent (e.g., nagging about diet, criticizing for poor exercise) had a significant negative as sociation with glycemic control such that increased diabetes specific family conflict was associated with increased A1c values or decreased/poorer glycemic control. Both child and parent reported adherence also demonstrated a significant negative relationship with glycemic control (better adherence associated with lower A1c or better glycemic control). Parent reported quality of life had a similar negative correlation with glycemic control, although child report versions did not. This negative relationship suggested that increased (better) quality of life was associated with decreased (better) glycemic control. Childreported quality of life was significantly negatively associated with the number of episodes of diabetic ketoacidosis. This relationship i s rather intuitive, as it would be expected that children with a high number of DKA episodes would likely report lower quality of life. The results in our sample with respect to quality of life, adherence, diabetes specific family factors, and glycemic c ontrol essentially mirror the relationships between these variables as shown in previous research. In the current sample, the parent reported versions of the study variables produced more meaningful relationships than the childreport versions. One possi ble explanation for this occurrence is that since the study measures were predominantly facevalid, it is possible that the children in our study were more susceptible to social desirability response bias than their parents,

PAGE 59

59 subsequently portraying themsel ves in a more positive light. This possibility is supported by research showing that chronically ill youth are more vulnerable to this response bias than their parents, particularly when self reporting psychosocial problems (Logan et al., 2008). The fact that this study revealed a mediation effect of quality of life on glycemic control through adherence could be used to tailor future interventions. There has been some research (Hirsch, 2000; Rubin, et al., 1993) suggesting that brief interventions aimed at improving diabetes specific coping skills and self management have resulted in improved some aspects of quality of life (i.e., empowerment, emotional well being, diabetes self efficacy). Knowing that quality of life partially mediates the relationship b etween adherence and glycemic control further emphasizes the need for interventions that target quality of life in addition to other factors. The findings of the current study were inconsistent with a majority of current research in two respects. There w as not a difference by sex or age on child or parent report measures of quality of life in the current sample. The interpretation of this finding that seemed most plausible to us was that this was likely to be a reflection of the relatively small sample s ize. This study included measures assessing diabetes specific family factors, as these have previously been shown to account for variance in glycemic control (Duke et al., 2008). In the current sample, the only family factors measure that was significantl y predictive of glycemic control was both the child and parent reported versions Diabetes Family Behavior Checklist (DFBC). The items on this measure assess negative and unsupportive diabetes specific behaviors. The child version asks the child to rate h ow

PAGE 60

60 often their primary caregiver engages in these behaviors, while the parent version asks the parent to self rate their own behavior. The fact that both versions of this measure were so strongly predictive of glycemic control is very telling of the impac t of negative diabetes specific behavior. Given that this measure has consistently been shown to be predictive of glycemic control, it stresses the importance of interventions for the parents/guardians of children with diabetes to teach the importance of positive diabetes specific behaviors as well as the health consequences of negative behaviors. Findings from the current study with regard to the effect of race on glycemic control appear to suggest a possible racial disparity with regard to health status. Previous research (Petitti et al., 2009; Chalew et al., 2000) has shown that racial/ethnic minorities, particularly AfricanAmericans, have poorer glycemic control than Caucasians even after controlling for other variables such as parent income, family structure, parent education, form of insurance, and duration of diabetes. However, recent research has suggested that there is a disparity in the accuracy of A1c test results in African American children (Kamps et al., 2010). Specifically, this research suggested that African American children with higher A1c levels than Caucasian children often have similar mean blood glucose levels. The authors opined that A1c levels alone are not an accurate reflection of health status in African American patients wi th diabetes. Had this information been validated and replicated prior to this study, we might have included an additional measure of health status, such as mean blood glucose level, so that we could control for any inaccuracies on our assessment of health status. Although it is possible that the African American children in our population may have had higher A1c levels due to psychosocial or socioeconomic reasons, the findings

PAGE 61

61 of Kamps and colleagues suggest the possibility that the interpretation of a di fference in health status is merely an inaccurate measuring mechanism. If this is true in the case of youth with Type 1 Diabetes, African American ethnic status may not be associated with poorer glycemic control. Kamps and colleagues commented on the misi nterpretations of A1c levels in African American patients with diabetes as potentially harmful consequences. However, findings from a single study would need significant replication prior to accepting them. In light of the Kamps study, further research on racial differences in diabetes is warranted. Previous research in pediatric literature suggests that parent reports of their chronically ill childs functioning are variable in their accuracy (Guyatt et al., 1997). Consistent with our hypothesis, parents in the current sample reported significantly lower general and diabetes specific quality of life than their children self reported. This reporting gap across the various quality of life domains is noteworthy for clinicians when inquiring about a patie nts functioning. This difference in observed functioning of the pediatric Type 1 diabetes population extends knowledge of similar parent/patient gaps in other chronically ill child populations. Although this data does not imply that parent report or chi ld report of the childs quality of life is more accurate per se, it once again highlights differences in perception of well being, among other factors, between a child and their parent. Given other findings of different answers to the same questions from children and their parents (e.g., adherence, negative diabetes specific behaviors) once they have been differentiated, it may have implications for possible changes in health care provider behavior. Although further research is needed to clarify this iss ue, it is possible that more accurate health care information with regard to day to day

PAGE 62

62 responsibilities, such as adherence, might be obtained by discussing these topics separately with both parties. This may particularly relevant for questions to which socially desirable responses may be more likely. Overall, the findings from this study supported the majority of our hypotheses. In conducting this study, we sought to better understand the relationships among some of the factors that have been useful in u nderstanding differences in glycemic control. It is our hope that our study will contribute to improved psychosocial interventions for improving glycemic control in the pediatric T1D patient. Limitations We need to highlight some of the limitations of the current study. While the use of self report questionnaires is common in psychological research, the data they provide are subjective in nature. However, it is also important to state that many of the domains assessed in the current study would be impossi ble or impractical to objectively measure, making subjective self report measures necessary. The impact of social desirability needs to be taken into consideration. By examining both the patient self report and parent observed versions of measures in the c urrent study, some of these limitations have been addressed. Over the twelve months of recruitment, enrollment in the current study was slower than previous studies of similar design conducted at the same pediatric endocrinology clinics. This contributed to the relatively small sample size in the current study. Regardless of the reason(s) for this slow rate of enrollment (e.g., fewer clinics due to economic constraints, reduced patient attendance at clinics due to financial problems), the small sample size places limitations on the scope of our investigation as well as any generalizability of conclusions. Participants in the current sample were recruited from

PAGE 63

63 the catchment areas for our clinic. A larger and more diverse sample would provide for a larger variety of patient attributes that might further clarify predictors of glycemic control in pediatric T1D patients. Sample size limits the studys power, as do Bonferroni corrections to avoid family wise errors on multiple correlations of the same data. It is possible that further group differences were not identified due to these limitations. Despite the sample size, quality of life was shown to partially mediate the relationship between adherence and glycemic control. Given the pattern of results seen in the current sample, it seems plausible that stronger conclusion could be drawn from a larger sample. Missing data can be a particular problem confronting clinical research. This was true for demographic data in the current study. The study would have benefitted from improved procedures to ensure that all questions had been answered in the study packets prior to terminating the interview process. Such procedures could have increase the amount of useable data for the number of participants enrolled in the study. This would have further improved power in analyses already affected by sample size. The crosssectional design of the study in nature did not allow for more helpful information that could have been obtained through a longitudinal design. For example, many of the variables examined in this study likely fluctuate over time. A longitudinal study with repeated measure within participant could have demonstrated how quality of life, adherence, A1c, and other variables change over time. Additionally, findings of a mediation model that remain valid in a longitudinal study can allow causal relationships to be identified. A longitudinal study would enhance confidence in conclusions about

PAGE 64

64 quality of life mediating the relationship between adherence and gl ycemic control, as well as numerous other significant correlations that were identified. The current study may have been strengthened by utilizing other methods of assessing diabetes regimen adherence in addition to the DSMP. The interview format of the DSMP may contribute to socially desirable responses for participants. Separating parents and their children during their respective adherence interviews and emphasizing the importance of being honest in their answers rather than concerned with giving the right answer were procedures used in the current study to address some aspects of social desirability. Information about the complexity of diabetes treatment regimens did not allow the use of objective measures of adherence used in some other pediatric disease populations (e.g., pill counts, prescription records). The DSMP has been well validated and used in numerous studies. The high degree of validation of the DSMP through repeated use in clinical research provides greater confidence in the measure des pite concerns related to the self report nature of this adherence measure. The use of only the A1c for health status in Type 1 diabetes may be considered another limitation, although it is not a subjective, self report measure. However, the A1c is conside red the goldstandard in diabetes research and is universally used as it was in the current study. Implications and Future Directions The current study identified an original finding in the partial mediation by quality of life in the relationship between adherence and glycemic control in youth and adolescents with Type 1 diabetes. Additionally, we replicated the findings of previous stu dies indicating that adolescent patients with T1D are at higher risk for lower quality of life than preadolescent patients with T1D. A larger, longitudinal study with multiple

PAGE 65

65 assessment points would likely clarify the magnitude of quality of lifes involvement in the relationship between adherence and glycemic control. As previously discussed, the chicken or the egg paradox remains with regard to the constructs of adherence, glycemic control, and quality of life. Further research on the findings of t he current study with a larger and more diverse population is warranted. Extensions of this research as well as research on other psychosocial factors may contribute to the development of more effective and accurate interventions aimed at stabilizing glyc emic control. Given the large overlap between some of these study variables, an intervention aimed at improving one domain would likely result in improvement in others. The findings from the current study suggest the value in identifying factors that imp act quality of life when developing interventions to improve adherence and, ultimately, glycemic control in youth and adolescents with T1D. The current study accounted for roughly 42% of the variance in glycemic control. Further exploratory studies shoul d seek to detect additional factors, whether psychosocial, biological, or otherwise, that influence glycemic control in youth with T1D. Additionally, future research clarifying the relationships between any predictive variables would be vital with regard to the development of interventions. The better understanding of the factors impacting glycemic control the more effective interventions that can be developed to improve the physical and emotional well being of youth with T1D as they cope with this chronic condition over the course of their lives.

PAGE 66

66 LIST OF REFERENCES American Diabetes Association. (2010). Diagnosis and classification of diabetes mellitus. Diabetes Care, 33(Suppl. 1), S62S69. American Diabetes Association. (2009). Total prevalence of diabetes & prediabetes. Retrieved July 25, 2009, from http://www.diabetes.org/diabetes statistics/prevalence.jsp Anderson, B. J., Ho, J., Brackett, J., Finkelstein, D., & Laffel, L. (1997). Parental involvement in diabetes management tasks: Relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulindependent diabetes mellitus. The Journal of Pediatrics, 130(2), 257 265. Anderson, B. J., & Laffel, L. M. B. (1997). Behavioral and psychosocial research w ith school aged children with type 1 diabetes. Diabetes Spectrum, 10, 281285. Anderson, B. J., Vangsness, L., Connell, A., Butler, D., Goebel Fabbri, S., & Laffel, L. M. B. (2002). Family conflict, adherence, and glycaemic control in youth with short d uration Type 1 diabetes. Diabetic Medicine, 19, 635642. Baron, R. M., & Kenny, D. A. (1986). The moderator mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 52, 11731182. Chalew, S. A., Gomez, R., Butler, A., Hempe, J., Compton, D., Mercante, T., et al. (2000) Predictors of glycemic control in children with Type 1 diabetes: The importance of race. Journal of Diabetes and Its Complicati ons, 14, 71 77. Danielson, K. K., Drum, M. L., Estrada, C. L., & Lipton, R. B. (2010). Racial and ethnic differences in an estimated measure of insulin resistance among individuals with Type 1 diabetes. Diabetes Care, 33(3), 614 619. Delamater, A. M., Applegate, B., Eidson, M., & Nemery, R. (1998). Increased risks for poor metabolic control in minority youths with Type I diabetes. Diabetes, 47 (Supp. 1), A326. Diabetes Control and Complications Trial Research Group. (1993). The effect of inte nsive treatment of diabetes on the development and progression of long term complications in insulin dependent diabetes mellitus. New England Journal of Medicine, 329(14), 977986. Diabetes Control and Complications Trial Research Group. (1994). Effect of intensive diabetes treatment on the development and progression of long term complications in adolescents with insulindependent diabetes mellitus: Diabetes control and complications trial. The Journal of Pediatrics, 125 177 188.

PAGE 67

67 Duke, D. C., Geffken, G. R., Lewin, A. B., Williams, L. B., Storch, E. A., & Silverstein, J. H. (2008). Glycemic control in youth with Type 1 diabetes: Family predictors and mediators. Journal of Pediatric Psychology, 33(7), 719 727. Engel, G. L. (1977). The need for a new m edical model: a challenge for biomedicine. Science, 196, 129136. Glasgow, R. E., McCaul, K. D., & Schafer, L. C. (1987). Self care behaviors and glycemic control in type 1 diabetes. Journal of Chronic Diseases, 40(5), 399 412. Graue, M., Wentzel Larse n, T., Bru, E., Hanestad, B. R., & Sovik, O. (2004). The coping styles of adolescents with Type 1 Diabetes are associated with degree of metabolic control. Diabetes Control, 27(6), 13131317. Graue, M., Wentzel Larsen, T., Hanestad, B. R., Batsvik, B., & Sovik, O. (2003). Measuring self reported, healthrelated quality of life in adolescents with type 1 diabetes using both generic and diseasespecific instruments. Acta Paediar, 92 11901196. Graue, M., Wentzel Larsen, T., Hanestad, B. R., & Sovik, O. (2005). Healthrelated quality of life and metabolic control in adolescents with diabetes: The role of parental care, control, and involvement. Journal of Pediatric Nursing, 20(5), 373 381. Guyatt, G. H., Juniper, E. F., Griffith, L. E., Feeny, D. H., & Ferrie, P. J. (1997). Children and adult perceptions of childhood asthma. Pediatrics, 99 165168. Harris, M. A., Wysocki, T., Sadler, M., Wilkinson, K., Harvey, L. M., Buckloh, L. M., et al. (2000). Validation of a structured interview for the asses sment of diabetes self management. Diabetes Care, 23, 1301 1304. Hauser, S. T., Jacobson, A. M., Lavori, P., Wolfsdorf, J. I., Herskowitz, R. D., Milley, J. E., Bliss, R., Wertlieb, D., & Stein, J. (1990). Adherence among children and adolescents with insulin dependent diabetes mellitus over a four year longitudinal follow up: II. Immediate and long term linkages with the family milieu. Journal of Pediatric Psychology, 15, 527542. Haynes, R. B. (1979). Introduction. In R. B. Hayes, D. W. Taylor, & D. L. Sackett (Eds.), Compliance in Health Care (pp. 110). Baltimore, MD: Johns Hopkins University. Hesketh, K. D., Wake, M. A., & Cameron, F. J. (2004). Healthrelated quality of life and metabolic control in children with Type 1 diabetes: A prospective cohort study. Diabetes Care, 27(2), 415420. Hirsch, A. (2000). Diabetes care in Germany: New roles for professionals and patients. Diabetes Spectrum, 13 3.

PAGE 68

68 Hoey, H., Martul, P., Aanstoot, H. J., Matsuura, N., Chiarelli, F., McGee, H., M., et al., for the Hviodore Study Group on Childhood Diabetes. (2001). Good metabolic control is associated with better quality of life in 2,101 adolescents with Type 1 diabe tes. Diabetes Care, 24(11), 19231928. Hood, K. K., Anderson, B. J., Butler, D. A., & Laffel, L. M. B. (2007). Updated and revised diabetes family conflict scale. Diabetes Care, 30(7), 17641769. Huang, E.S., Brown, S.E.S., Ewigman, B.G., Foley, E.C., & Meltzer, D.O. (2007). Patient perceptions of quality of life with diabetes related complications and treatments. Diabetes Care, 30(10): 24782483. Johnson, S. B. (1993). Methodological issues in diabetes research: Measuring adherence. Diabetes Care, 15 16581667. Johnson, S. B., (1994). Health behavior and health status: Concepts, methods, and applications. Journal of Pediatric Psychology, 19, 129 141. Johnson, S. B., Kelly, M., Henretta, J. C., Cunningham, W. R., Tomer, A., & Silverstein, J. H. (1992). A longitudinal analysis of adherence and health status in childhood diabetes. Journal of Pediatric Psychology, 17, 537 553. Kamps, J. L., Hemphe, J. M., & Chalew, S. A. (2010). Racial disparity in A1c independent of mean blood glucose in childr en with Type 1 diabetes. Diabetes Care, 33(5), 10251027. Kovacs, M. (1992). Childrens Depression Inventory North Tonawanda, NY: Multi health Systems, Inc. Kovacs, M., Goldston, D., Obrosky, S., & Iyengar, S. (1992). Prevalence and predictors of pervasive noncompliance with medical treatment among youths with insulin dependent diabetes mellitus. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 11121119. La Greca, A. M., Bear man, K. J. (2002). The Diabetes Social Support QuestionnaireFamily Version: Evaluating adolescents diabetes specific support from family members. Journal of Pediatric Psychology, 27, 665676. La Greca, A. M., Follansbee, D. M., & Skyler, J. S. (1990). Developmental and behavioral aspects of diabetes management in youngsters. Childrens Health Care, 19 132 139. La Greca, A. M., & Schuman, W. B. (1995). Adherence to prescribed medical regimens. In M.C. Roberts (Ed.), Handbook of Pediatric Psychology (pp.5583). New York: Guilford Press.

PAGE 69

69 Laffel, L. M. B., Goebel Fabbri, A., Connell, A., Mansfield, A., Vangsess, L., & Anderson, B. J. (2003). General quality of life in youth with Type 1 diabetes. Diabetes Care, 26(11), 30673073. Le Coq, E. M., Boek e, A. J. P., Bezemer, P. D., Bruil, J., & Van Eijk, J. T. M. (2000). Clinimetric properties of a parent report on their offsprings quality of life. Journal of Clinical Epidemiology, 53(2), 139 146. Lewin, A. B., Heidgerken, A. D., Geffken, G. R., Willi ams, L. B., Storch, E. A., Gelfland, K. M., & Silverstein, J. H. (2006). The relation between family factors and metabolic control: The role of diabetes adherence. Journal of Pediatric Psychology, 31 (2), 174 183. Logan, D. E., Claar, R. L., & Scharff, L (2008). Social desirability response bias and self report of psychological distress in pediatric chronic pain patients. Pain, 136(3), 366372. Matsui, D. M. (1997). Drug compliance in pediatrics. Clinical and research issues. Pediatric Clinics of Nor th America, 44,1 14. Miller Johnson, S., Emory, R. E., Marvin, R. S., Clarke, W., Lovinger, R. & Martin, M. (1994). Parent child relationships and the management of insulin dependent diabetes mellitus. Journal of Consulting and Clinical Psychology, 62, 6 03610. Naughton, M. J., Ruggiero, A. M., Lawrence, J. M., Imperatore, G., Klingensmith, G. J., Waitzfelder, B. W., et al., for the SEARCH for Diabetes in Youth Study Group. (2008). Healthrelated quality of life of children and adolescents with Type 1 or Type 2 diabetes mellitus. Archives of Pediatric and Adolescent Medicine, 162(7), 649657. Petitti, D. B., Klingensmith, G. J., Andrews, J. S., Bell, R. A., Dabelea, D., Imperatore, G., et al., for the SEARCH for Diabetes in Youth Study. (2009) Glycemi c control in youth with diabetes: The SEARCH for diabetes in youth study. Journal of Pediatrics, 155 668 672. Rapoff, M. A. (1999). Adherence to pediatric medical regimens Dordrecht, Netherlands: Kluwer Academic Publishers. Rubin, R. R. (2000). Diabetes and quality of life. Diabetes Spectrum, 13, 21. Rubin R. R., Peyrot, M., Saudek, C. S. (1993). The effect of a diabetes education program incorporating coping skills training on emotional well being and diabetes self efficacy. Diabetes Education, 19 210 214. Schafer, L. C., McCaul, K. D., & Glasgow, R. E. (1986). Supportive and nonsupportive family behaviors: Relationships to adherence and metabolic control in persons with type I diabetes. Diabetes Care, 9, 179185.

PAGE 70

70 Silverstein, J., Klingens mith, G., Copeland, K., Plotnick, L., Kaufman, F., Laffel, L., et al.; American Diabetes Association. (2005) Care of children and adolescents with type 1 diabetes: A statement of the American Diabetes Association. Diabetes Care, 28(1), 186212. Sobel, M E. (1988). Direct and indirect effects in linear structural equation models. In J. S. Long (Ed.), Common Problems/Proper Solutions: Avoiding Error in Quantitative Research. 46 64. Beverly Hills CA: Sage. SPSS (2009). PASW Statistics (vols. 18.0.0). Chicago, IL: SPSS Inc. Vanelli, M., Chiarelli, F., Chiari, G., & Tumini, S. (2003). Relationship between metabolic control and quality of life in adolescents with Type 1 diabetes. Acta Bio Medica, 74(1), 13 17. Varni, J. W., Burwinkle, T. M., Jacobs J.R., Gottschalk, M., Kaufman, F., & Jones, K. L. (2003). The PedsQL in Type 1 and Type 2 diabetes: Reliability and validity of the Pediatric Quality of Life Inventory generic core scales and Type 1 diabetes module. Diabetes Care, 26, 631 637. Varni, J. W., Limbers, C. A., & Burwinkle, T. M. (2007). Impaired healthrelated quality of life in children and adolescents with chronic conditions: A comparative analysis of 10 disease clusters and 33 disease categories/severities utilizing the PedsQL 4.0 Generic Core Scales. Health and Quality of Life Outcomes, 5(43). Varni, J. W., Seid, M., & Rode, C. A. (1999). The PedsQL : Measurement model for the pediatric quality of life inventory. Medical Care, 37(2), 126139. Wagner, V. M., Muller Godeffro y, E., von Sengbusch, S., Hager, S., & Thyen, U. (2005). Age, metabolic control and type of insulin regime influences healthrelated quality of life in children and adolescents with Type 1 diabetes mellitus. European Journal of Pediatrics, 164, 491496. Waller, D. A., Chipman, J. J., Hardy, B. W., Hightower, M. S., North, A. J., Williams, S. B., et al. (1986). Measuring diabetes specific family support and its relation to metabolic control: A preliminary report. Journal of the American Academy of Chil d and Adolescent Psychiatry, 25 415 418. Wexler, D.J., Grant, R.W., Wittenberg, E., Bosch, J.L., Cagliero, E., Delahanty, L., et al. (2006). Correlates of healthrelated quality of life in Type 2 diabetes. Diabetologia, 49, 1489 1497. Wysocki, T., Greco, P., & Buckloh, L. M. (2003). Childhood diabetes in psychological context. In M. C. Roberts (Ed.), Handbook of Pediatric Psychology, Third Edition. New York: Guilford Press.

PAGE 71

71 BIOGRAPHICAL SKETCH Jay St. A mant received his bachelors degree from Mount Olive College in Mount Olive, North Carolina in 2006. He majored in psychology and graduated summa cum laude. He received his masters degree in psychology from the University of Florida in 2008. He received his Ph.D. in psychology from the University of Florida in August 2011.