|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 AN INTERNET BASED TAILORED FEEDBACK INTERVENTION FOR SMOKING CESSATION By ALANA M. ROJEWSKI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGR EE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Alana M. Rojewski
3 To my parents
4 ACKNOWLEDGMENTS I thank my mentor, Jesse Dallery, for his guidance t hrough this stage of my career and for challengi ng me to be a better researche r I would also like to acknowledge my past and current labmates who have provided much appreciated feedback on my work over the years. In particular, I would like to thank Allison Kurti and Brantley Jarvis for their editorial comments on portions of this paper. I would also like to thank the Behavior Analysis faculty for the education they affor d ed me. The intellectual and scientific culture shaped by their dedicated work is unparalleled to any other. In particular, I thank Dr. Timothy Hackenberg for posin g challenging philosophical questions and bringing a theoretical framework full circle for me. I thank my first scientific family at Denison University, Drs. Nestor Matthews, Cody Brooks, and Susan Kennedy, for providing me with many exciting opportunities at the beginning of my career. I thank my dissertation c ommittee members Drs. Timothy Vollmer, Sara Jo Nixon, and Drake Morgan for the helpful feedback regarding my work. I thank Spencer Gauert, who served as a research assistant on this project and Aa ron Dumas, who assisted with recruitment comfortable and well caffeinated environment where a majority of this dissertation was written. I am fortunate enough to have an ext ensive network of family members, including my parents, step parents, and sister that have provide d me with more love and support than I could ask for. Thank you for that and for the continuous enthusiasm for my pursuits from across the country Thank you especially to my parents, Maura Manger and Thomas Rojewski, for being the first to teach me how to answer the many questions I had about the world, for reinforcing my early interests in science and for always encouraging me to pursue what makes me happy. I would also like to thank the many friends I have made in Gainesville over the years for the thought provoking conversations and distractions I am certain that my graduate experience in
5 Gainesville would have been markedly less wonderful without you (es pecially you, Olympic Rings) Thank you to my furry family, Oliver, Dexter, Penny, and Jack, for all of the cuddles and playtime. Finally, I would like to thank Nick Ronnquist I am so fortunate to have a partner who has supported m y goals challenged me i ntellectually, and provided love and encouragement every step of the way
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 10 CHAPTER 1 G ENERAL INTRODUCTION ................................ ................................ .............................. 11 Cigarette Smoking in the United States ................................ ................................ .................. 11 A Behavioral Approach to Smoking ................................ ................................ ....................... 11 Tailored Feedback for Cigarette Smoking ................................ ................................ .............. 14 Behavior Assessment ................................ ................................ ................................ .............. 18 An Internet based Approach ................................ ................................ ................................ ... 20 Present Studies ................................ ................................ ................................ ........................ 21 2 S TUDY 1 ................................ ................................ ................................ ................................ 22 An Internet Based Tailored Feedback Intervention ................................ ............................... 22 Methods ................................ ................................ ................................ ................................ .. 22 Participants ................................ ................................ ................................ ...................... 22 Tailoring Assessments ................................ ................................ ................................ ..... 23 Tailored Feedback ................................ ................................ ................................ ........... 25 Experimental Procedures ................................ ................................ ................................ 27 Data Analysis ................................ ................................ ................................ ................... 30 Results ................................ ................................ ................................ ................................ ..... 31 Discussion ................................ ................................ ................................ ............................... 33 3 S TUDY 2 ................................ ................................ ................................ ................................ 4 4 Methods ................................ ................................ ................................ ................................ .. 44 Participants ................................ ................................ ................................ ...................... 44 Experimental Procedures ................................ ................................ ................................ 45 Dat a Analysis ................................ ................................ ................................ ................... 46 Results ................................ ................................ ................................ ................................ ..... 46 Discussion ................................ ................................ ................................ ............................... 49 4 G ENERAL DISCUSSION ................................ ................................ ................................ ..... 62 APPENDIX
7 A EMA ITEMS AND ABBREVIATIONS ................................ ................................ ................ 68 B FASTR ITEMS AND A BBREVIATIONS ................................ ................................ ............ 69 C SE SC ITEMS AND ABBREVIATIONS ................................ ................................ .............. 70 D FEEDBACK ITEMS ................................ ................................ ................................ .............. 71 L IST OF REFERENCES ................................ ................................ ................................ ............... 73 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ......... 79
8 LIST OF TABLES Table page 2 1 Participant demographics for Study 1 ................................ ................................ ................ 36 2 2 Feedback d elivery ................................ ................................ ................................ .............. 36 3 1 Participant demographics for Study 2 ................................ ................................ ................ 52 3 2 Feedback d elivery ................................ ................................ ................................ .............. 52
9 LIST OF FIGURES Figure page 2 1 Baselin e EMA data for each participant ................................ ................................ ............ 37 2 2 Basel ine FASTR data for each participant. ................................ ................................ ........ 38 2 3 Baseline SE SC data for each participant. ................................ ................................ ......... 39 2 4 Individual participant time series data.. ................................ ................................ ............. 40 2 5 Average participant CO by study phase. ................................ ................................ ............ 41 2 6 Participant acceptability ................................ ................................ ................................ ..... 42 2 7 The percentage of feedback items that each participant engaged in ................................ .. 43 3 1 Baseline EMA data f or each participant ................................ ................................ ............ 53 3 2 Baseline FASTR data for each participant. ................................ ................................ ........ 54 3 3 Baseline SE SC data for each participant ................................ ................................ .......... 55 3 4 Individual participant time series data ................................ ................................ .............. 56 3 5 Average participant CO by study phase. ................................ ................................ ............ 57 3 6 Difference scores on the EMA after red uctions in smoking from baseline ....................... 58 3 7 Difference scores on the FASTR after reduction s in smoking fro m baseline ................... 59 3 8 Difference scor es on the SE SC after reduction s in smoking fro m baseline ..................... 60 3 9 Participant acce ptability ................................ ................................ ................................ ..... 61
10 Abstract of Dissertation Presented to the Graduate School of th e University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A N I NTERNET BASED TAILORED FEEDBACK INTERVENTION FOR SMO KING CESSATION By Alana M. Rojewski August 2012 Chair: Jesse Dallery Major : Psychology Tailored feedback interventions for smoking cessation allow researchers to identify variables involved in smoking and to deliver personalized strategies for cessation. T he purpose of the present studies was to evaluate prel iminary efficacy f easibility and acceptability of an internet based tailored feedbac k intervention Individual variables associated with smoking were assessed and an internet based system was used to deliver tailored feedback contingent on smoking status The results of th ese studies suggest that the t ailored feedback intervention was feasible and acceptable but t he effect of the intervention on abstinence outcomes was mixed
11 CHAPTER 1 GENERAL INTRODUCTION Cigarette Smoking in the United States An estimated 19.3% of all adults (aged 18 years or older) in the United States smoke cigarettes despite the well documented negative health consequences (Centers for Disease Control and Prevention [CDC], 2011). T he CDC reports that smoking remains the single largest preventable ca use of death and disease in the United States. In addition to multiple types of cancer, smoking is associated with heart disease, pulmonary disease, the exacerbation of chronic health problems, and 443,000 deaths annually. Unfortunately, o nly 4 to 7 % of sm okers that attempt to quit smoking each year are successful (Fiore et al., 2008). Numerous behavioral and pharmacological treatments are available for smoking, but success is moderate and long term abstinence rates are low (Jorenby et al., 2006; Kassel & Y ates, 2002). Consequently, two questions arise: W hy do people continue to smoke and how can we improve smoking cessation interventions? A Behavioral Approach to Smoking One perspective from which to address these questions is behavior analysis One centra l tenant of a behavior analytic view is that behavior is a function of genetics, environmental h istory, and current environment. B ehavior whose future likelihood is affected by environmental consequences is known as operant behavior (Skinner, 1953). Like m ost human or animal activities, drug use can be characterized as operant behavior L aboratory research utilizing animal models of drug self administration has shown that drug seeking is controlled by reinforcement contingencies Rates and patterns of drug self administration in animals can be controlled by manipulating the schedule, rate, immediacy, and magnitude of reinforcement (Bigelow & Silverman, 1999). R esearchers can also manipulate laboratory based and clinical
12 drug self administration in humans usi ng similar procedures (Bigelow & Silverman, 1999 ; Griffiths, Bigelow, & Henningfield, 1980 ) For example, Stitzer & Bigelow (198 2 ) provided contingent monetary reinforcement in the amounts of $5 per day to regular smokers for reducing their afternoon breat h carbon monoxide (CO) levels to 50% or less of baseline values. CO levels were reduced during contingent reinforcement This effect was r eproduced in a subsequent study, which also demonstrated that increases in pay magnitude corresponded to increasing pe rcentages of occasions on which CO reductions were achieved (Stitzer & Bigelow, 1983) The behavior change arising from the contingencie s arranged in these st udies reflect findings from the animal literature, as well in which nondrug reinforcers decrease drug self administration and larger magnitude reinforcers are generally better at maintaining behavior than smaller magnitude reinforcers ( Griffiths et al. 1980 ; Stitzer & Bigelow, 1983 ). The well documented effects of reinforcement contingencies on drug self administration across animal and human studies provide a firm empirical basis for the conception of drug abuse as operant be havior Thus, human cigarette smoking is considered operant within this framework The future likelihood of behavior occurring will depend on the consequence for that behavior. An increase in the target behavior following the presentation of a stimulus is defined as positive reinforcement (Skinner, 1953). Positive reinforcement for smoking could be associated with the direct pharm acological effects of nicotine conditioned reinforcement or social reinforcement Nicotine produces subjective euphoric sensations often Further in a review on motivational influences of smoking, Baker at al. (2004) suggest that after repeated pairings stimuli associated with nicotine delivery can affec t self administration behaviors on their own. These stimuli are then referred to as conditione d reinforcers. S moking may be
13 maintained because the stimuli associated with the act of smoking (e.g. the feel of smoke, handling the cigarettes or the lighters, etc.) have become reinforcing as well ( Rose & Levin, 1991). Indeed, smokers report satisfacti on, liking, positive effects, and reductions in withdrawal effects after smoking cigarettes even when they do not contain nicotine ( Butschky Bailey, Henningfield, & Pickworth, 1995) In addition to concomitant smoking stimuli, o ther environmental or soci al stimuli may also become conditioned reinforcers through repeated pairings with nicotine. Social consequences, such as attention and approval from others, may also exert control over behavior as reinforcers in their own right (Skinner, 1953). An increas e in the target behavior following the removal of a stimulus is defined as negative reinforcement (Skinner, 1953). Negative reinforcement for smoking could be related to escape from aversive activities or emotional or physiological states (Baker Piper, et al., 2004) Even brief c essation from smoking leads to craving for tobacco, irritability, anxiety, difficulty concentrating, restlessness, decreased heart rate, and increased eating (Hughes and Hatsukami, 1986). Once nicotine is administered, the aversive symptoms of tobacco withdrawal are ameliorated, and the continued use of nicotine is fostered. Similarly, smoking may be reinforced by removal of negative affective states arising from non pharmacological sources, such as stress (Baker et al., 2004). Ind eed, smokers frequently report that stress and negative emotional states often immediately precede a relapse episode ( Baer & Lichtenstein, 1988). In addition, responses may come under the control of antecedent stimuli. Antecedent stimuli will occasion smo king in that they signal the availability of reinforcement when the behavior is emitted (Skinner, 1953). For example, being in the presence of friends may signal that reinforcement is available for smoking Research on immediate antecedents of smoking have demonstrated that smoking is associated with a v a riety of setting and activity variables, including
14 d rinking coffee eating, drinking alcohol or being in the presence or other smokers ( Shiffman et al., 2002; Shiffman, Paty, Gw altney, & Dang 2004). Insofar as smoking is operant, the processes of reinforcement and stimulus control will mai ntain smoking for an individual. These contingencies are likely idiosyncratic in nature because every smoker has a different history of reinf orcement with respect to smoking. T he goal of smoking cessation treatment is the development of interventions that change individual human behavior by promoting new behaviors that are incompatible with smoking, modifying current behaviors that maintain smo king and fostering abstinence (DiClemente Marinilli, Singh, & Bellino, 2001) Tailored Feedback for Cigarette Smoking Numerous types of treatments for smoking exist, but the goal of enhancing treatment effects is universal. T ailor ing treatment to the in dividual is o ne strategy for enhancing treatment effects (Axelrod, 1991; Fiore et al., 2008; Kassel & Yates, 2002). Tailored treatment interventions include tailored feedback strategies based on individual assessment, where researchers gather information f rom an individual in order to create personalized treatment recommendations designed to meet their needs (Strecher, 1999). The main components of personalized feedback interventions for smoking cessation are: 1) a baseline assessment of characteristics rel smoking status, situations in which smoking frequently occurs etc.), 2) a message library of feedback (may describe personalized risks, normative comparisons, suggestions for how to change, et c.), 3) an algorithm to match tailored messages to each individual based on his/her assessment results and 4) a channel to deliver the feedback (de Vries and Brug, 1999; DiClemente et al., 2001). For example, messages may outline the amount of money that would be saved by quitting or describe the improved social implications of not smelling like smoke. Messages may offer suggestions on
15 avoiding high risk situations or how to seek out appropriate social support. This type of tailoring allows for the deliver y of helpful behavior change information to smokers attempting to quit. Tailored feedback interventions include a range of approaches derived from various frameworks outside of behavior analysis. One prevalent theoretical basis for tailoring is the transt heoretical model, also referred t o as the stages of change model. Five stages of change for people in the process of quitting smoking have been identified (e.g., contemplation phase), and a responses to questions a bout when they are planning on quitting (Prochaska & DiClemente, 1983). In the research stemming from this framework, the tailoring interventions are stage based and the goal is to move participants through the stages of change towards abstinence (Prochask a, DiClemente, Velicer, & Rossi 1993) An additional theoretical framework for tailoring is self efficacy. Self efficacy is defined as the belief that one can successfully execute the behavior required to produce an outcome (Bandura, 1977). According to the theory of self efficacy, perceived self efficacy in combination with motivation to obtain a particular outcome, directs behavior change (Condiotte & Lichtenstein, 1981) Thus, an effective way to increase abstinence is to deliver tailored feedback to increase the T ailored feedback interventions often use behavior change feedback messages that target behavioral variables (e.g., avoiding key antecedent stimuli) even if the approach is not explicitly behavior analytic in nature. However, n o interventions have structured feedback based on a behavior analytic approach. The results of recent research suggest that tailored feedback interventions can improve cessation rates ( Becoa & Vazquez, 2001 ; Dijkstra De Vries, & Roijackers 1998 1999 ; Prochaska et al., 1993 ; Strecher et al., 1994 ). For example, Becoa and Vazquez (2001) evaluated the effects of written feedback adapted to a self help mail intervention group compared
16 to a control group. Participants in a standard feedback group were mailed six weekly packets consisting of pamphlets with new strategies to stop smoking, a personalized letter of introduction with instructions for the tasks to be carried out during the week, and self monitoring and evalua tion forms to be mailed back at the end of the week. The treatment group received the same six weekly packets plus an additional two letters with computer generated personalized feedback determined by their responses to questions presented in the weekly fo rms. A control group was included that did not receive treatment until the end of the study period. The proportion of participants who self reported abstinence at the end of the study was 37% in the standard group, 51% in the treatment group, and 0% for th e control group. Further, the proportion of participants who were abstinent ( self reported not smoking in the last 7 days and CO < 9 ppm ) at the 3 month follow up was 22% in the standard group, 37% in the treatment group, and 1% for the control group. The results were interpreted as suggest ing that both the standard and treatment interventions more effectively promoted abstinence than a control group. Broad scale advances in tailored feedback delivery have been possible as technology continues to develop. Feedback can be delivered over the phone (R amelson, Friedman, & Ockene 1 999), via text message (Weitzel, Bernhardt, Usdan, Mays, & Glanz, 2007), or by email (Te Poel Bolman, Reubsaet, & de Vries 2009). For example, Te Poel et al. (2009) evaluated an e m ail delivered computer tailored smoking cessation intervention. Participants received either a computer tailored e mail (treatment group) or a generic, non tailored e mail (control group). The results 6 months after baseline showed that significantly more participants in the intervention group reported not having smoked in the last 24 hours (21.5%) and 7 days (20.4%) in contrast with participants in the control group (9.8 and 7.8%, respectively). In addition, in the alcohol treatment literature, Weitzel et al. (2007) evaluated a tailored intervention, in which text
17 messages were delivered to wireless handheld computers to reduce drinking among college students. Participants received daily tailored messages on their handheld computers about avoiding alcohol r elated consequences. The authors concluded that the ta ilored messages had small but positive effects on alcohol related behaviors P articipants in the treatment group reported significantly fewer drinks per drinking day than the control group (4 .86 drinks versus 6.41) These types of technologically advanced delivery systems allow for more immediate and continuous delivery of tailored feedback than postal mail. Despite treatment success using tailored feedback, there are also instances of no treatment effec ts (Curry McBride, Grothaus, Louid, & Wagner, 1995; Owen, Ewins, & Lee 1989, Strecher et al., 2005 ). Strecher ( 1999) conducted a review of ten randomized tr ials of tailored materials and demonstrated that n umerous possible combinations of generating and delivering tailored feedback have been evaluated Aspects of feedback that varied included theory (e.g., drive reduction, stages of change), length (e.g., short sentences, lengthy manuals), form (e.g., computer printout, letter, email, text), and frequency (e.g., delivering materials once, or at several time points; Strecher, 1999). Due to the variability in treatment design, and the mixed results yi elded by this work, the effectiveness of tailored feedback interventions is unclear. Further, across studies, the key treatment variables that result in success versus failure are unclear. Thus, the optimal method for implementing tailored feedback has not been established. In addition to the various ways of implementing tailored feedback, there is also inconsist ency in the literature in the definition of abstinence. Often, biochemical verification of abstinence was not collected or, if it was, a less stringent CO criterion was used (e.g., <10 ppm ; Becoa & Vazquez, 2001 ; Wetter et al., 2011 ). Research has demonst rated that when assessing breath carbon monoxide, ( Javors, Hatch, &
18 Lamb, 2005; Raiff, Faix, Turturici, & Dallery 2010 ) Several studies have also tailored feedback based on self reported smoking status, bu t self report may be inaccurate. Behavior Assessment The strategy of assessing characteristics in order to deliver individualized behavior change information is amenable to the goals of a behavior analytic intervention. I n order to develop strategies for a behavior change intervention understanding environmental variables involved in the maintenance of behavior on an individual level is critical (Hanle y, Iwata, & McCord, 2003; Iwata Kahng, Wa llace, & Lindberg, 2000). Upon inspecti ng the history of an indi vidual smoker, we would be able to develop hypotheses about the reinforcement contingencies which maintain to isolate such contingencies. Information provided by smokers about the events preceding (i.e., possible discriminative stimuli) and following (i.e., possible reinforcing consequences) smoking can be collected and used to make hypotheses about behavioral function (Iwata et al., 2000). Typically, assessmen t in tailored feedback interventions involves participants completing pre treatment questionnaires, and the subsequent tailored feedback is based on the results of tho se questionnaires. Questionnaires allow researchers to collect information regarding smok ing and other variables. While this method is generally effective and widely used, collecting data on smoking in real time is preferable (Shiffman et al., 1997) One real time behavior assessment technique is Ecological Momentary Assessment (EMA; Shiffman & Stone, 1998; Stone & Shiffman, 1994). Like other self monitoring methods, EMA uses monitoring strategies to assess variables at the moment they occur in natural settings, thus maximizing ecological validity while avoiding retrospective recall ( Stone & Sh iffman, 1994). Participants are prompted to complete questionnaires or report on the frequency of occurrence of specific events in their natural environment using electronic diaries, such as palmtop computers, cell phones, or other hand held
19 devices. Thus, with respect to smoking, stimuli that might be assessed include the presence of other smokers, having just eaten a meal, etc. This provides a snapshot of variables associated with smoking for each individual. EMA has been used as a method for tailoring fe edback for smokers seeking treatment Wetter et al. (2011) evaluated a relapse prevention intervention among women who had recently quit smoking using group therapy and nicotine replace ment therapy. Participants completed EMA procedures for one week follow ing their quit day. At Day 7, 72.2 % of participants in the control group were abstinent, versus 77.5% of the participants in the computer delivered treatment group (CDT). Relapse prevention treatment for the CDT group was tailored based on the EMA results and provided real time access to individualized, context specific coping strategies, motivational messages, and general quitting and relapse prevention information via personal palmtop computers (PPC). The CDT intervention did not improve abstinence rates relative to a standard treatment control group (64% versus 61% abstinent) However, interpretation of these results is difficult because abstinence rates were assessed using biochemically confirmed seven day point prevalence abstinence, defined as self rep ort ing abstinence during the previous seven days and a CO level of <10 parts per million ( ppm ) As previously mentioned, a more conservative definition of abstinence would be ( Javors et al. 2005; Raiff et al. 2010 ), so initial abstinence rates may have been inflated in this study. Despite this limitation the authors found that 96% of the treatment group used the PPC during the treatment period. Further, they concluded that heavier smokers were more likely to use PPC delivered treatment and that greater usage may be associated with more positive outcomes.
20 EMA could potentially be a useful tool for behavior assessment prior to tailoring. Numerous other studies h ave demonstrated the benefit of collecting EMA data for understanding smoking variables in cessation or relapse prevention interventions ( Husky, Mazure, Carroll, Barry, & Petry, 2008; Shiffman, Paty, Gnys, Kassel, & Hickcox, 1996; Shiffman 1982, 2005 ). Fur ther, we have developed a functional assessment method for assessing the variables that maintain smoking ( Rojewski & Dallery, unpublished data). The EMA included an assessment of not only antecedent events (completed before each cigarette), but also conseq uent events (completed after each cigarette) across time for a specified number of days Individual differences in the events associated with smoking were observed, suggesting that these methods can be used to effectively assess smoking behavior at an indi vidual level. Results from this type of assessment allow for h ypotheses to be made about reinforcement contingencies and specific maintaining variables which could inform individualized treatment plans. Particularly, these data could provide the basis for tailoring feedback of behavioral modification strategies such as avoidance of key antecedent stimuli (e.g., giving up coffee while trying to quit smoking). An Internet based Approach Easy and on going methods for abstinence verification have been developed in recent years (Dallery & Glenn, 2005; Stoops et al., 2009). For example, Dallery and Glenn (2005) describe methods for collecting CO samples via a computerized, Internet based system. Participants can record themselves and provide CO samples several times a day, for several weeks at a time. This type of data collection allows for low effort, continuous and objec tive evidence of smoking status which is a more exact measure than self re p ort. Furthermore, continuous evidence of smoking s tatus allows for continuous and contingent feedback to be delivered, which may be an important for promoting a successful quit attempt. A s smokers begin to reduce cigarette
21 consumption during the ir quit attempt, smoking could fluctuate day to day. This wou ld provide For example, t he first day a participant is abstinent, the feedback delivered could be relevant to maintaining abstinence as opposed to inducing abstinence. If participants continue to smoke, a bstain, or lapse, then behavior change messages could be supplied accordingly. Present Studies In the subsequent chapters, two studies are described that sought to investigate a novel internet based tailored feedback intervention for smoking cessation Th e intervention is derived from a behavior analytic approach, and tailoring assessments included an evaluation of potential antecedent stimuli and positive and negative reinforcers for smoking. The tailored feedback messages selected specifically addressed these antecedents and consequences of smoking. Further, objective evidence of smoking status was collected to determine smoking status. The internet based intervention also allowed for continuous assessment of smoking status such that tailored feedback cou ld be delivered as behavior changed over the course of treatment Study 1 was an investigation of the preliminary efficacy, feasibility, and acceptability of this intervention Study 2 further investigated the preliminary efficacy of the intervention by at tempting to make outcomes more robust and to improve the tailor ing procedures
22 CHAPTER 2 STUDY 1 An Internet Based Tailored Feedback Intervention The internet based component of the intervention was a novel method for collecting objective, continuous samp ling of smoking status, and permitted the delivery of feedback on a continuous basis. Further, feedback was delivered contingent on smoking status. That is, feedback was cessation related if the participant was still smoking and focused on maintaining abst inence if the participant was not smoking. In order to conduct a thorough behavior assessment and understand relevant smoking variables at the individual level, both questionnaires and EMA were used for tailoring. Although EMA has been used to tailor relap se prevention messages in a feedback intervention (Wetter et al., 2011), it has not been used to tailor cessation related information. Study 1 sought to: (1) evaluate the preliminary efficacy of implementing an internet based tailored feedback intervention for smoking, (2) evaluate individual differences in the variables associated with smoking by assessing the events most frequently reported by each individual, and (3) assess the feasibility and acceptability of the intervention. Methods Participants Parti cipants were six self reported smokers recruited via print and online advertisement. Participants qualified for participation if they were between 18 and 60 years old, had access to the internet in their place of residence, and reported a desire to quit sm 7 on a scale of 1 10 of how interested they were in quitting). Further, participants had to smoke at least five cigarettes per day, and smoking status was confirmed with an objective measure of smoki ng, breath carbon monoxide ( during the practice sample). Exclusion criteria included
23 self reported history of psychiatric illness within the past year (with the exception of depression or any psychiatric illness that had been well controlled for the last year), self reported illicit drug use (excluding marijuana), and self reported current use of nicotine containing products other than cigarettes. During the screening process, participants completed the tailoring assessments (described below) and a Psychosocial History, which contains question s on demographic information, health and smoking history and the Fagerstrom Test for Nicotine Dependence ( FTND; Fagerstrom & Schneider, 1989; Heatherton Kozlowski, Frecker, & Fagerstrom, 1991 ) The questionnaires were either completed in person in our lab or online using Qualtrics. Tailoring Assessments EMA. The EMA provided information about the frequency with which events or stimuli occur when the participant smokes in their natural environment The EMA collected data on p otential antecedent stimuli and positive reinforcers T he participants were asked to record each cigarette they smoked and answer brief questionnaires about the events preceding and following each cigarette for 2 days during baseline The EMA was completed using a Qualtrics q uestionnaire which they accessed on web enabled phones. If the participant did not own a web enabled phone, one was provided (prepaid Samsung Admire). The assessment items were derived from Axelrod (1991), and they were originally published in The Wellness Encyclopedia (Health Letter Associates, 1991). In addition, smoking cessation and applied behavior analysis experts, as well as long term smokers, were consulted during the construction of this questionnaire. The assessment items included in the present s tudy and their abbreviations are presented in Appendix A. The first 13 items constituted the Pre Cigarette Assessment, and the remaining six items constituted the Post Cigarette Assessment. Any item reported on the EMA was entered into the tailoring algori thm.
2 4 Functional Assessment of Smoking for Treatment Recommendations (FASTR) The FASTR was developed in our lab to determine the environmental variables associated with smoking and is a reliable and valid instrument (unpublished data) The assessment inclu ded a range of a ntecedent and consequent events associated with smoking including escape/avoidance, conditioned and social reinforcement, antecedent stim uli, and positive reinforcement. Participants reported whether they smoked in 28 situations and scores ranged from 0 (Never ) to 4 (Al ways). The items on the FASTR and their abbreviations are presented in Appendix B. S elf Efficacy and Confidence Questionnaire (SE SC) The SE SC provided information about high risk situations in which participants smoked or situations in which they were uncertain that they could resist smoking Self confidence was assessed with a visual analogue item an d was scored on a scale of 0 100 (Alessi, Badger, & Higgins 2004). Reports of Confiden ce in Abstinence Tomorrow (CAT) were not used in tailoring, but were assessed to determine if exposure to the tailored feedback intervention increased confidence in ability to abstain. T he Self Efficacy Questionnaire asked participants to report the probab ility that they would be able to resist the urge to smoke if they were in a specific situation. The questionnaire has been shown to have good reliability (Condiotte & Lichtenstein, 1981). The items on the SE SC and their abbreviations are presented in Appe ndix C. Probabilities were presented on a scale of 0 % (would smoke) to 100% (would not smoke) in 10% increments. The theory of s elf efficacy while not necessarily c onducive to a behavior analytic approach to behavior d oes provide that expectations of eff icacy are derived from past performance accomplishments (Condiotte & Lichtenstein, 1981). Thus, it is possible that reports of self efficacy vary with a history of consequences ( Ramonowich, Mintz, and Lamb, 2009). If smoking cessation attem pts have been successful in the past, the participant may report
25 a greater sense of self efficacy. On the other hand, if attempts at smoking cessation have been unsuccessful in that past, the participant may report l ower self efficacy. Thus, s elf efficacy measures may be useful in tailor ing behavior change information if they indicate history of success in abstaining in certain situations. Reinforcement Survey Schedule (RSS). The RSS assessed r einforcement derived from non drug activities. T he questionnaire was developed to identify activities and situations that are reinforcing for different individuals, and is a reliable and valid instrument (Cautela & Kastenbaum, 1967; Keehn, Bloomfield, & Hug, 1970; Thorndike & Kleinknecht, 1974). A short version of the RSS was used t o measure past month reinforcement from 47 different activities (e.g., going to class, work, exercise, dating, time with kids, etc. ). Scores for activity frequency ranged from 0 (Never) to 4 (Several times per day), and scores for activity enjoyment ranged from 0 (Not at all) to 4 (Very much). Information gathered from this questionnaire was used to structure some of the cessation and abstinence related feedback messages The goal was to redirect participants to naturally occu rring nondrug activities during their quit attempt. Reasons for Quitting (RFQ). In addition, the RFQ was administered to gather information about the participant s main motivations for quitting (Curry Wagner, & Grothaus 1990). The RFQ consisted of a lis t of 20 statements of various reasons for quitting smoking, such as being concerned about illness. Scores ranged from 0 (Not true at all) to 4 (Extremely true) for each statement. Information gathered from this questionnaire was used to structure some of t he cessation and abstinence related feedback messages. Participants could then be reminded of and redirected to the reinforcing outcomes of successfully quitting. Tailored Feedback The treatment consisted of tailored feedback messages. As described by de Vries and Brug (1999), the messages were generated based on diagnosis, a message library, an algorithm, and a
26 channel to deliver the message. The diagnosis of individual smoking related variables w as derived from the FASTR, the SE SC and the EMA. The RSS and the RFQ provided information about alternative reinforcing activities and reasons that the participant was quitting smoking. The message library was developed from materials from the National Cancer Institute, the American Cancer Society, and behavior modification strategies. Strategies included distraction, reminders of smoking consequences, reminders of their reasons for quitting, escaping the situation, soliciting social support, substituting behavior, and relaxation. The tailoring library is present ed in Appendix D. The algorithm consisted of a series of if then statements linking the diagnosis items to relevant items in the message library via Excel A composite of EMA, FASTR, and SE SC scores determined if each item in the library would be presente d in the list of possible outcomes for each participant. Any reported EMA variable, FASTR items scored as 2, 3, or 4 (meaning the item occurred Sometimes, Often, or Always) and SE SC scores below 40% (meaning that they were not certain they could abstain i n the situation) were considered frequently occurring (or high risk) variables and were included in the algorithm for the corresponding library items. For example, if a participant reported that he or she often smokes when drinking coffee on the FASTR, or that they would not be able to resist smoking when they drank coffee on the SE SC, then item Change up your routine. Drink tea, juice, or water instead of coffee fil e. In this way, the list of possible outcomes was specific to each participant. they left their sample. The feedback was not automatic upon sample submission, but samples were checked and verified frequently throughout the day. The feedback was a lso either texted or emailed to the participant (depending on their preference) in an attempt to expedite the contact
27 with the and varied based on smoking status. All tailor ed messages were structured by : 1) addressing the participant in a personal and supportive tone, 2) thanking them for their sample (if the sample was positive) or congratulating them on not smoking (if the sample was negative), 3) delivering either a cessa tion related message (if positive) or a n abstinence related message (if negative) from the tailored message library For example, if their sample was negative, the message would t on tobacco in a jar every day and then buy yourself a weekly treat. Keep up the good work rid of all cigarettes, ashtrays and lighters in yo ur home. Keeping them around you only gives you Experimental Procedures Following intake, participants began experimental procedures. The study started as soon as possible following a confirmation of their eligibility. The stu dy materials (CO mo nitor, webcam, manual) were either mailed to the participants or hand delivered if the participant was local to Gainesville. Research assistants scheduled a set up over the phone with participants once they received the materials to go over the study procedures and familiarize the participant with the equipment. The study began shortly after the phone meeting. All participants were provided with y the National Cancer Institute. Participants were provided with a CO monitor and access to a computerized, secure, Internet ts take a breath CO measurement using a Bedfont piCO + Smoker lyzer The participant blew into a small tube on the devic e to measure their CO level (indicating recent smoking) if they were > 4 ppm and negative (indicating recent abstinence) if they were 4 ppm.
28 Participants were not told what the abstinence criterion was : They were o nly told that the CO result would fluctuate depending on their recent smoking. Th e purpose of this was to avoid criterion) and to allow participant behavior t o contact the feedback. Participants were asked to make video recordings of themselves blowing into the CO monitor participant exhaling into the CO monitor, presenting the final reading to the camera, and self reporting how much they had smoked since their last sample. Leaving a sample too k approximately one to three minutes. The research staff accessed the videos over the secure server to verify smoking status Participants were provided with a webcam if they did not already have one. The treatment was assessed using a non concurrent multi ple baseline design ( Baer, Wolf and Risley, 1968; Watson & Workman, 1981) This design was s elected in order to determine the effects of the independent variable in a small sample. In a multiple baseline design, t he experimental contingency is applied to o ne participant at a time, such that the baseline phases are longer for each subsequent participant enrolled. In the case of a non concurrent baseline, the b aselines do not necessarily occur at the same point in time. Effectiveness due to the treatment alon e would be demonstrated if the behavior changed only when the intervention was introduced (Watson & Workman, 1981) the treatment was implemented, one would more confidently conclude that the decrease was a result of the treatment and not extra experimental variables. Each phase of the study required participants to submit CO samples twice a day. The first phase, b aseline, was conducted for a minimum of 2 days and participants received no f eed back
29 for their samples. A p articipant s baseli ne had to meet two requirements before the treatment phase began: The baseline had to be separated from other participant baselines by at least two samples and it had to be stable Stability was determined by v isual analysis (i.e., n o decreasing trends, and no extreme variability) over the last four samples All participants also complete d the EMA for 2 days during the b aseline phase. The participants had a counseling session over the phone with the PI to discus s the results of the EMA and ensure that the participant felt that the EMA and the assessments captured the instances in which they most frequently smoke The Feedback phase followed b aseline and the feed back treatment was implemented until stability was achieved. Stability was determined by visual analysis (i.e., n o decreasing trends, and no extreme variability) over the last 6 samples If the Feedback phase was extended because of trends or extreme variability and over time the variability itself wa s st able, this too satisfied visual stability Once stability was achieved in the Feedback phase, a quit day was set for two days later and the participant was notified. For the Feedback + Goal s phase the participant continued to receive tailored feedback and was given a CO goal of This phase was included to evaluate the effect of adding a goal to the tailored feedback in inducing abstinence. The participant was encouraged to reduce their smoking (if they had not already) and to use the feedback they had received thus far to achieve the goal of quitting. Stability criteria for this phase were the same as for the Feedback phase. To be consistent across participants, all participants experienced the Feedback + Goal s phase, even if they had already achiev ed abstinence. At th e end of t reatment all participants completed the lab assessments In addition, the participants completed the Feedback Follow up questionnaire on a weekly basis to assess whether and how frequently they followed each item of feedback or whether they followed a different strategy that they found helpful. This questionnaire was administered on a
30 weekly basis and included feedback items from the previous 7 days. A Treatment Acceptability Questionnaire, administered at the end of the t re atment phase, assessed the acceptability of the treatment. Most questions were answered according to a Likert scale. The questionnaire assessed how helpful the treatment was in changing smoking behavior, thoughts and feelings related to the treatment, and items related to the EMA. Finally, several open ended questions were asked, such as how the treatment could be improved, particular challenges encountered, etc. All participants had the chance to earn a total of $45 for completing assessments and meeting s with the research team. Data Analysis EMA, FASTR, SE SC, data were compiled and entered into the algorithm for tailoring. For the EMA, frequency counts were obtained for each variable to assess the number of times each item was associated with an instanc e of smoking. Because each participant smoked a different number of cigarettes each day, proportions were calculated based on the frequency counts and the total number of cigarettes smoked. Assessments in which the participant completed both the Pre and P ost Cigarette Assessment were considered acceptable assessments. Assessments that did not include a Post Cigarette Assessment following initiation of the Pre Cigarette Assessment were considered unacceptable and dropped from the analysis. Similarly, assess ments that did not include a Pre Cigarette Assessment prior to a Post Cigarette Assessment were dropped from the analysis. In order to evaluate treatment effectiveness, within subject changes in CO were evaluated across time following the implementation o f tailored feedback. The average CO for each phase of the study was also calculated for each participant. The results of the TAQ were evaluated to assess the feasibility and acceptability of implementing tailored feedback.
31 Results A total of six particip ants completed the study. Participant demographics are presented in Table 2 1. The length of each phase varied for each participant. The average length of time of the F eedback pha ses (Feedback and Feedback+Goal ) was 42 .6 7 days ( SD = 8.73 ). Feedback delivery information for individual participants is presented in Table 2 2. The percentage of samples completed refers to the number of scheduled CO samples that were completed by each participant. Average completed CO samples ranged from 60 to 98 % for each partici pant. The number of times feedback was delivered refers to the feedback scheduled to be delivered, not including feedback missed due to a missed sample by the participant. Of the scheduled samples submitted, average samples delivered ranged from 94 100% fo r each participant The instances in which 100% feedback delivery were not achieved was due to the timing of the sample submission by the participant. There were several instances in which the participant left a sample just before the midnight deadline, an d left a sample exactly 8 hours later. Occasionally samples were not checked after midnight and before 8:00 a.m., thus it was not possible to deliver feedback for the evening CO sample before the morning sample was left. In these instances, only one feedb ack message was delivered for the morning sample. The latency to feedback delivery refers to the average number of minutes between the CO sample submission and feedback delivery for each participant. Average latency to feedback delivery ranged from 93 to 2 66 minutes per participant. Individual subject tailoring was based on the results of the EMA, FASTR, and SE SC baseline assessments. These d ata are presented in Figure 2 1, 2 2, and 2 3 (respectively). For the baseline EMA data in Figure 2 1, higher scores reflect variables m ore frequently reported when a participant was smoking. For the baseline FASTR data in Figure 2 2, scores ranged from 0 (Never) to 4 (Always), thus higher scores reflect more frequent smoking in those situations. For
32 the baseline EMA da ta in Figure 2 3, scores ranged from 0 % (would smoke) to 100% (would not smoke), thus lower scores reflect a lower probability that they could abstain in those situations. Individual differences were apparent for all assessment outcomes. Individual s ubjec t CO levels across ti me are presented in Figure 2 4 Reductions in smoking were observed for three of the six participants (CH03, MG02, and NO07). One of those three participants, MG02, demonstrated s ustained abstinence during the F eedback and Feedback + G oals phases. For the rema ining three participants, no changes in CO were observed following the onset of treatment. Only one participant, NO07, demonstrated a response to the added goal in the Feedback + Goal s phase. Abstinence was briefly achieved by goal setting but this effect did not maintain. Average CO levels by phase for each participant are presented in Figure 2 5 Participant point is excluded for ease o f viewing the other samples. These data confirm the reductions i n smoking observed in Figure 2 4 T he data in Figure 2 5 also average CO levels decreased from b aseline. However, the role of the feedback intervention in the reductions in CO levels for PC01 is unclear because similar decreases were observed during b aseline. Participant acceptability data are presented in Figure 2 6 The tailored feedback components received slightly higher average scores than the counseling and Motiv8 com ponents although all components received average scores above 50%. There was a great deal of variability necessarily correspond to treatment outcomes. In other words, not all participants who demonstrated reduct ions in CO rated the treatment highly On the other hand, some participants who did not reduce CO levels gave very high scores.
33 Feedback follow up data (percentages of cessation related feedback items that each participant engaged in) are presented in Fig ure 2 7 If a participant reported engaging in the feedback at all (even once), that item was scored as having been completed by the participant. The numbers above the bars represent the total number of cessation related f eedback items issued to each parti cipant over the course of treatment. Most participants engaged in a majority of the feedback items at least once. The feedback engagement seems to be unrelated to treatment outcome. Participants who reported trying a majority of items did not necessarily s how reductions in CO level (e.g., LC04). CH03 was the only participant who engaged in fewer than half of the feedback items issued to her. When asked if they liked receiving the 4 ppm goal, all participants except LC04 and TN05 said yes. Those participants that liked receiving a goal reported that the goal helped to increase their motivation Those participants who received Discussion The participants were mostly complian t with the CO sample submission schedule as the average percentage of samples completed ranged from 60 to 98% per participant. Further, almost all of the scheduled feedback was delivered to participants when samples were submitted. The average percentage of scheduled feedback samples delivered ranged from 94 100% per participant, suggesting that the imp lementation of the fe edback intervention was feasible. Average latency to feedback delivery ranged from 93 to 266 minutes per participant. Individual patterns in variables associated with smoking on the EMA, FASTR, and SE SC were apparent. There was some o verlap in the types of variables reported, but the degree to which the variables were reported as being relevant to smoking varied by participant. On the TAQ, a majority of participants reported that they liked the tailored feedback. Further, they reported that the tailored feedback was helpful to their quit attempt, that it helped them in
34 understanding the reasons they smoke, and that it reminded them what they needed to do to quit. The assessments seemed to have captured potential maintaining variables of smoking for each individual. The results of this study are mixed with respect to treatment efficacy. CO trends during the interventi on varied across participants. MG02 achieved sustained abstinence Gradual decreases in smoking, but not complete cessation were observed at the onset of the treatment intervention for CH03 and NO07 The three remaining participants maintained rates of smoking similar to their baseline levels The added CO goal in the final phase of the intervention did little to promote addi tional abstinence, as only one participant achieved CO reductions during that phase. The intervention was generally well accepted by participants, as all components were given average scores above 50%. One limitation of the intervention is that its effects on smoking were not robust. Some participants were unresponsive to the treatment. F or those who were responsive to treatment, decreases in smoking following the onset of the intervention were gradual. There are several possibilities for why this may have occurred One gradual reduction in smoking is that the variables maintaining their smoking may have been eliminated gradually over time after the participant contacts successive feedback messages. If the participant is compliant with the feedback and stops smoking in some situations additional situations in which they are still smoking may still persist if the feedback has not addressed those variables yet Participants may require additional feedback messages relevant to the remaining variables before complete cessation is observed. Unfortunately, the Feedback Follow up only provided information regarding which suggestions participants tried, and did not reveal how their behavior changed in response to receiving feedba ck. An other possible
35 reductions in smoking is that participant s were smoking less in general That is, they may continue to occasionally smoke in all of their high risk situations. T he data collected in this study do no t allow for hypotheses to be made about how smoking variables changed over time Further, w ithout information about changes in smoking variables over time the feedback messages may not have been relevant once participants reduced smoking The feedback was structured such that the feedback items were selected for delivery based on the baseline assessments Data collection did not include a re evaluation of smoking situations after the beginning of the t reatment phase Thus, s ome of the feedback delivered la ter in the treatment phase may have no longer applied if the participant ha d reduced smoking. In sum, Study 1 sought to evaluate preliminary efficacy of tailored feedback intervention for smoking cessation. The results demonstrated feasibility and accepta bility of the intervention, although the treatment effects were weak and the role that feedback played in behavior change was unclear Based on these findings, three areas for improvement emerged: 1) Understanding how feedback affects smoking behavior, 2) improving accuracy of feedback, and 3) promoting a more robust treatment effect.
36 Table 2 1. Participant demographics for Study 1 ID Sex Race Education Age Duration CO FTND CAT PC01 Male Caucasian Some college 34 14 38 5 49.5 MG02 Female Caucasian High school 25 10 9 5 91 CH03 Female Caucasian Some college 21 2 23 4 21.6 LC04 Female Caucasian Some college 29 13 7 5 20 TN05 Female African American Some college 40 16 28 5 14 NO07 Male Turkish Graduate school 24 6 38 7 100 Note. Duration refers to the number of years the participant has been smoking. CO (carbon monoxide) refers to a measure of exhaled carbo n monoxide in parts per million (ppm). FTND refers to the score on the Fagerstrom Test for Nicotine Dependence. in Abstinen SC. Table 2 2. Feedback Delivery PC01 MG02 CH03 LC04 TN05 NO07 Samples Completed 81 9 8 89 60 8 3 83 Feedback Delivered 100 97 94 97 100 100 Latency to Delivery 93 202 198 209 205 266 Note. Sam ples Completed refer to the twice daily scheduled CO samples. Feedback Delivered refers to the scheduled feedback delivered, not including feedback missed due to a missed sample. The Latency to Delivery refers to the average number of minutes between the CO sample submission and feedback delivery.
37 Figure 2 1. Baseline EMA data for each participant. Pre cigarette assessment items are on the left of the dashed vertical line, and Post cigarette assessment items are on the right of the line.
38 Figure 2 2. Bas eline FASTR data for each participant.
39 Figure 2 3. Baseline SE SC data for each participant.
40 Figure 2 4 I ndividual participant time series data. Data points represent individual CO samples. Horizontal dashed lines represent the 4 ppm abstinence crit erion. Vertical black dashed lines represent phase changes. Vertical red dashed lines represent the point at which the participant was informed of their upcoming quit day.
41 Figure 2 5 Average participant CO by study phase. Data points represent individ ual CO samples. Bars represent the average CO for each study phase for each participant.
42 Figure 2 6 Participant acceptability on items relating to the counseling, tailored feedback, and internet based components of th e intervention. Black data points represent individual participant scores. Gray bars represent the average of the participant data points for each item.
43 Figure 2 7 The percentage of feedback items that each participa nt engaged in. The numbers above the bars represent the total number of cessation related feedback items issued to each participant over the course of treatment.
44 CHAPTER 3 STUDY 2 Two main areas for improvement w ere highlighted in Study 1 and methodologi cal changes were implemented for Study 2 accordingly. One goal was to u nderstand whether feedback has a selective effect on smoking behavior and to improve feedback for those who are responsive to treatment Thus, as participants began to reduce their smok ing, more frequent assessments were implemented. The goal was to capture changes in smoking related variables (as assessed on the EMA and FASTR) and ability to abstain in high risk situations (as assessed on the SE SC) as participants began to smoke less f requently and to adjust the tailored feedback message library accordingly This may reveal a gradual elimination of maintaining variables over time feedback. Further, this may also allow feedback to target only those variables that are relevant as smoking behavior changes If participants are no longer smoking in certain situations, some smoking related variables assessed at baseline will no longer apply. Second, in an attempt to make the treatment more robust, additional congratulatory messages for reductions in smoking were implemented The goal was to deliver praise for decreases in smoking in an attempt to reinforce small steps towards abstinence. Methods Participa nts Participants were six self reported smokers recruited via print and online advertisement. Inclusion and exclusi on criteria participant screening procedures, and the lab assessments were the same as in Study 1.
45 Experimental Procedures F ollowing intake participants began the experimental procedures. The Motiv8 system was employed for sample collection and participants completed the EMA for the first 2 days of baseline. The treatment consisted of tailored feedback messages as described in Study 1. In ad average baseline CO was calculated and 25%, 50% and 75% reduction targets were established. Once the treatment phase was initiated a moving ave rage of the last four CO samples was calculated. Once the participant achieved each r eduction target, an additional battery of assessments (EMA, FASTR, and SE SC) was completed and a congratulatory message was delivered for reducing their smoking by the specified amount. Because the EMA required the participant to have actually smoked, th is assessment was only completed if the participant smoked the day that they achieved the reduction criterion. Study 2 was assessed us ing a combined multiple baseline and changing criterion design A changing criterion design includes stepwise change s in c riterion rate for the target behavior (Hartmann & Hall, 1976) The CO reduction targets are stepw ise changes in smoking behavior that must be met before the additional congratulatory messages is delivered, and, thus, constitute a changing criterion. This s tudy includ ed Baseline and Feedback phases, and t he Feedback + Goals phase was removed to target the feedback alone. Each phase of the study required participants to submit CO samples twice a day. The first phase, b aseline, was conducted for a minimum of 2 days and participants receive d no feedback for their samples. The stability criteria were the same as in Study 1. All participants also completed the EMA for 2 days during the baseline phase. The participants had a counseling session over the phone with t he PI to discuss the results of the EM A and ensure that the participa n t felt that the EMA and the assessments captured the instances in which t hey most frequently smoke d The Feedback phase followed b aseline and the fee dback treatment was implemented until stability was reached. The stability
46 criteria for the treatment phase were the same as in Study 1. At the end of treatment all participants completed the TAQ All participants had the chance to earn a total of $45 for completing assessments and meetings with the research team. Data Analysis FASTR, SE SC, RFQ, and RSS data were compiled and entered into the algorithm for tailoring. For the EMA, frequency counts were obtained for each variable to assess the number of times each item was associated with an instance of smoking. Because each participant smoked a different number of cigarettes each day, proportions were calculated based on the frequency counts and the total number of cigarettes smoked. These proportions were also entered in to the tailoring alg orithm. Assessments in which the participant completed both the Pre and Post Cigarette Assessment were considered acceptable assessments. Assessments that did not include a Post Cigarette Assessment following initiation of the Pre Cigarette Assessment wer e considered unacceptable and dropped from the analysis. Similarly, assessments that did not include a Pre Cigarette Assessment prior to a Post Cigarette Assessment were dropped from the analysis. In order to evaluate treatment effectiveness, within subje ct changes in CO were evaluated across time following the implementation of tailored feedback. The average CO for each phase of the study was also calculated for each participant. The results of the TAQ were evaluated to assess the feasibility and acceptab ility of implementing tailored feedback. Results A total of six participants completed the study. Participant demographics are presented in Table 3 1. The length of each phase varied for each participant. The average length of time of the Feedback phase was 24 days ( SD = 5.18 ). Feedback delivery information for individual participants is presented in Table 3 2. The average percentage of completed CO samples ranged
47 from 8 0 97%. Of th e scheduled samples submitted, the average number of feedback messages deliv ered was 83 96%. As in Study 1, t he instances in which 100% feedback delivery was not achieved was due to the timing of the sample submission by the participant. The average latency to feedback delivery ranged from 220 326 minutes. Participant acceptabilit y data are presented in Figure 3 9 All treatment components received average scores above 70%. Individual subject tailoring was based on the results of the EMA, FASTR, and SE SC baseline assessments. Thes e data are presented in Figure 3 1 through 3 3 (re spectively). For the baseline EMA data in Figure 2 1, higher scores reflect variables more frequently reported when a participant was smoking. For the baseline FASTR data in Figure 2 2, scores ranged from 0 (Never) to 4 (Always), thus higher scores reflect more frequent smoking in those situations. For the baseline EMA data in Figure 2 3, scores ranged from 0 % (would smoke) to 100 % (would not smoke), thus lower scores reflect a lower probability that they could abstain in those situations I ndividual differ ences were apparent for all assessment outcomes. Individual subject CO levels across time are presented in Figure 3 4 For two of the six participants, CG08 and MH10, no changes in baseline CO were observed following the onset of treatment Thus, their ind ividual subject data are not presented with the participants who were responsive to treatment. One additional participant AB14, demonstrated reductions in CO prior while she was still in the baseline phase. Her reductions in smoking during the Feedback phase are likely not due to the feedback per se, but perhaps to being in a smoking cessation study, monitoring her own behavior, and/or setting a quit date. Therefor e, her treatment data are similar to her baseline data and this was considered a lack of treatment effect. Her data are presented in Figure 3 4 because she did reach one reduction target. Reductions in smoking were observed for
48 three participants, CI13, JW 15 and AS16 Of the three responsive participants, CI13 demonstrated sustained absti nence during the Feedback phase Average CO levels by phase for each parti cipant are presented in Figure 3 5 These data confirm the reductions i n smoking observed in Figu re 3 4. Participants completed additional assessments following 25%, 50%, or 75% CO reductions from baseline. Changes in scores for each participant on the EMA, FASTR, and SE SC from basel ine are presented in Figure 3 6 3 7 and 3 8 (respectively). In F igure 3 6 de creases in EMA proportions indicat e less frequent smoking in those situations. JW15 was the only participant who completed an EMA assessment on the day s she achieved the reduction criteria. For the 25% and 50% reductions, she reported less fre quent smoking in several situations. The 75% reduction assessment was completed after several days of abstinence, and a lapse happened to occur on that day. The increases in smoking for the 75% criterion reflect the variables present during that lapse. In Figure 3 7 decreases in FAS TR scores in dicate the participant smoked less frequently in those situations. T here are several instances in which participants report increases in smoking frequency for several variables, despite the fact that they had achieve d reductions in CO. In Figure 3 8 inc reases in SE SC scores indicate that the participant is more confident that they could abstain from smoking in those situations. For example, AB14 is more confident that she could resist smoking in almost all of the si tuations assessed, which may not be surprising as she was abstinent for several days at a time. Interestingly, JW15 reported a decrease in self reported probability of abstinence when she achieved 25% reduction in CO. This may be because she was still smo king at the time. However, once she had reached 50% and 75% reduction criteria, she reported feeling increasingly more able to abstain i n several situations. Missing data for participants is due to either the participant
49 not achieving the CO reduction crit erion (e.g., AB14 never achieved 50% reduction) or the participant not complying with requests to complete the assessments (e.g., CI13 did not complete 25% or 50% reduction assessments). Discussion The participants were mostly complian t with the CO sample submission schedule as the average percentage of samples completed ranged from 80 97% per participant. Further, almost all of the scheduled feedback was delivered to participants when samples were submitted. The average percentage of scheduled feedback s amples delivered ranged from 83 96% per participant, suggesting that the implementation of the feedback intervention was feasible Average l atencies to feedback delivery ranged from 220 326 minutes per participant. As in Study 1, individual patterns in var iables associated with smoking on the EMA, FASTR, and SE SC were apparent. The results of this study are mixed with respect to treatment efficacy. CO trends during the intervention varied across participants. CI13 achieved sustained abstinence levels Gra dual decreases in smoking, but not complete cessation, were observed following the onset of the treatment intervention for JW15 and AS16 CG08, MH10, and AB14 maintained rates of smoking similar to their baseline levels (which meant abstinence levels for A B14 ). All treatment components re ceived average scores above 70% suggesting that the intervention (particularly the tailored feedback and Motiv8 components) was well accepted by participants. All participants rated the treatment highly on the TAQ and the y particularly liked the tailored Additional EMA, FASTR, and SE SC assessments were implemented after participants achieved each of the three CO reduction criteria (25%, 50%, and 75 %) The purpose was to capture changes in smoking as they began to smoke less frequently. The most robust outcomes
50 from these assessments were the increases in confidence to abstain following reductions in smoking on the SE SC for each participant. The results from the FASTR are less clear. Each participant reported s ome decreases, but also many increases in smoking frequency from baseline for several variables, despite the fact that they had achieved reductions in CO. This may be due to the wording of the questions on the FASTR. For example, the measures ask participa nts to report smoking frequency in the presence of the variables (Never, Rarely, Sometimes, Often, or Always). Smoking in the certain situations following a 25% reduction may be relatively higher due to decreases in frequencies for ot her variables. Smoking situations For example, at baseline the way i t makes me feel around other people. Unfortunately, we do not have EMA data to confirm the presence of other people during her smoking bouts. JW15 is the only participant who rep orted EMA data. She reported decreases in smoking in several situations relative to baseline. In addition, her data provide a clear picture of her lapse episode, which are the variables that increased from baseline for the 75% reduction criterion. The as sessment data suggest that participants are smoking less in certain situations (but still smoking in others), and that they are more confident in their ability to abstain in some situations (but not all). Further, the reduction assessment data suggest that the situations in which participants are most likely to continue to smoke and are less confident in their ability to abstain are the high risk smoking situations that they reported at baseline. This lends support to the
51 interpretation that participants ar e gradual ly smoking less in specific situations, as opposed to globally smoking less, as they contact successive feedback and begin to cut b ack.
52 Table 3 1. Participant demographics for Study 2 ID Sex Race Education Age Duration CO FTND CAT CG08 Male Afr ican American Some college 50 35 49 6 2 MH10 Male Caucasian Graduate school 31 16 27 8 15.7 CI13 Male Hispanic Some college 37 24 16 6 75 AB14 Female Caucasian College graduate 22 4 7 0 50 JW15 Female African American Graduate school 55 25 14 1 50 AS1 6 Male Caucasian College graduate 23 2 42 4 10.8 Note. Duration refers to the number of years the participant has been smoking. CO (carbon monoxide) refers to a measure of exhaled carbon monoxide in parts per mi llion (ppm). FTND refers to the score on the Fagerstrom Test for Nicotine Dependence. SC. Table 3 2. Feedback Delivery CG08 MH10 CI13 AB14 JW15 AS16 Samples Completed 82 89 95 95 97 80 Feedback Delivered 83 96 95 84 94 88 L atency to Delivery 222 298 229 326 295 220 Note. Samples Completed refer to the twice daily scheduled CO samples. Feedback Delivered refers to the scheduled feedback delivered, not including feedback missed due to a missed sample. The Latency to Delivery refers to the average number of minutes between the CO sample submission and feedback delivery.
53 Figure 3 1. Baseline EMA data for each participant. Pre cigarette assessment items are on the left of the dashed vertical line, and Post cigarette assessmen t items are on the right of the line.
54 Figure 3 2. Baseline FASTR data for each participant.
55 Figure 3 3. Baseline SE SC data for each participant.
56 Figure 3 4. Individual participant time series data. Data points represent individual CO samples. Horizontal dashed lines represent the 4 ppm abstinence criterion. Vertical black dashed lines represent phase changes. Horizontal red solid lines represent th e 4 day average where the participant reached the 75%, 50% and 25% reduction criterion.
57 Figure 3 5 Average participant CO by study phase. Data points represent individual CO samples. Bars represent the average CO for each study phase for each participant.
58 Figure 3 6 Difference scores on the EMA after a 25% 50%, or 75% reduction in smoking from baseline for JW15
59 Figure 3 7 Difference scores on the FASTR after a 25%, 50%, or 75% reduction in smoking from baseline for each participant
60 Figure 3 8 Difference scores on the SE SC after a 25% 50%, or 7 5% redu ction in smoking from baseline for each participant
61 Figure 3 9 Participant acceptability on items relating to the counseling, tailored feedback, and internet based components of the intervention. Black data points represent individual participant scores. Gray bars represent the average of the participant data points for each item.
62 CHAPTER 4 GENERAL DISCUSSION Study 1 demonstrated the initial feasibility and acceptability of the tailored feedback interventi on, but the effects on smoking abstinence were not robust. One participant achieved sustained abstinence, two demonstrated gradual reductions in CO across time (some achieving abst inence but eventually lapsing) and three showed no difference from baseline. This gradual reduction in smoking, as opposed to complete cessation, was an interesting phenomenon which was pursued f urther in Study 2. In Study 2, additional assessment points were implemented to further elucidate individual patterns of behavior that oc curred following ongoing contact with the feedback. In an attempt to make the treatment effects more robust additional congratulatory messages for fixed reductions in CO levels were included Much like the pattern of results obtained in Study 1, one parti cipant achieved sustained abstinence, two others demonstrated gradual reductions in CO across time, and three sho wed no change from baseline The results of these studies suggest that the tailored feedback intervention is feasible and acceptable, but the e fficacy of the tailored feedback on abstinence outcomes is unclear. In line with some previous research ( Curry et al., 1995; Owen et al. 1989, Strecher et al., 2005), t he present study also suggests no clear effect of tailored feedback. However, drawing parallels between the results of this study and the extant literature is difficult. There are many methods for generating and delivering tailored feedback, and the methods used in this study are unique. This study is the first to use an explicitly behavior analytic approach to tailoring. Previous studies that provide d a description of the theoretical approach used often referenced the transtheoretical model of change or the self efficacy model. In addition, t he assessments selected for gathering data on pos sible maintaining variables of smoking are also novel compared to other studies The possibility that different assessments result in different types of tailoring cannot be
63 overlooked. For example, tailoring based on EMA data has been successful ( Wetter et al., 2011 ), but relevant smoking variables may not have been accurately assessed on our version of the EMA or the FASTR Even if they were, it is also possible that the variables evaluated in the present study via the EMA or FASTR may not have been sensit ive to treatment using feedback. Further, the present tailoring algorithm is likely different from previous studies and may be responsible for the differences compared to those studies that reported improved abstinence outcomes. Further, tailored feedback studies are typically evaluated using a between groups design G iven the small sample used in the present studies, the efficacy of the tailored feedback strategy evaluated is unclear. In addition, given the single subject design employed in the present stu dies (and no control group comparison), the role of feedback in isolation is unknown. Thus, the impact of the treatment variables other than feedback in the outcomes cannot be discounted. For example, participants reported liking the CO monitor and the qua ntitative progress graphs and these components may have contributed to success among those participants who achieved abstinence. In order to have a clearer understanding of the effects of feedback or the effects of our tailoring system, a study utilizing a 2 (tailored vs. untailored feedback) X 2 (tailoring alone vs. tailoring + Motiv8) between subject design should be conducted. One advantage of the current study was the use of objective biochemical verification of smoking. Many previous studies collected only self report, or used less stringent CO criteria These methods are much less precise and may have resulted in inflated outcomes. The CO reports and false negative CO samples. Given the availability of internet based methods for simplified collection of
64 biochemical verification, using CO to confirm smoking status should no longer be considered impractical. The only baseline assessment that seemed to show a relationship with treatment succ ess was the SE SC. For the most part, those participants who were able to quit or reduce smoking reported Confidence in Abstinence Tomorrow (CAT) scores above 50 (MG02, NO07, CI13, and JW15). This was not true for all participants, as CH03 and AS16 demonst rated reductions in CO levels but not CAT scores above 50. F urther those participants who were able to quit or reduce smoking reported higher self effic acy in more situations (MG02, NO07, JW15, CI13, and AS16). However, this was not true across all partic ipants, as CH03 did not report many situations in which she thought they could definitely abstain These results are consistent with research that has demonstrated that smokers who feel more confident in their ability to abstain and who report greater self efficacy are more likely to demonstrate post treatment abstinence ( Baer, Holt, & Lichtenstein, 1986; Coelho, 1984 ; Condiotte & Lichtenstein, 1981; Mclntyre, Lichtenstein, & Mermelstein, 1983; Gwaltney, Metrik, Kahler & Shiffman, 2009 ). Gwaltney et al. (2 009) report in a meta analysis on self efficacy and smoking cessation that those participants who are abstinent post treatment report significantly higher self efficacy scores at baseline than those individuals who resume or continue smoking. Further, the present study suggests that self efficacy changes as participants reduce smoking as indicated by the Study 2 reduction assessment data Other studies have demonstrated that self efficacy increases during a quit attempt (Coelho, 1984; Lamb, Morral, Galbick a, Kirby, & Iguchi, 2005 ; Schnoll et al., 2011 ), possibly because participants gain a better understanding of thei r high risk situations and their ability to abstain in those situations. Ramonowich et al. (2009) suggest ed that reductions in smoking predict ed later increases in s moking cessation self
65 efficacy, but not vice versa. They further suggest that self own behavior, and not a predictive factor of behavior change. Regardless of the predictive ability, this measure a ppear ed to reflect changes in smoking behavior over time and allowe d for the further refinement of tailored feedback in this intervention The role of this assessment for prediction of treatment success and behavior change should be investigated further in additional studies. There are several possible explanations for the observed variability in treatment outcomes and the weak overall effect. The assessment data suggested that feedback played a role in the reduction of smoking, but the strength of tailored feedback may lie in the additive effects. Those participants who were responsive to treatment exhibited CO reductions across time as they contacted successive tailored feedback items. Thus, a lengthier treatment period may be required before participants are able to achieve complete abstinence. Al ternatively, delivering feedback relevant to only one high risk variable per day or per week may be preferable. Participants would then have an opportunity to implement behavior change strategies for that variable exclusively before addressing the next one For example, an effective intervention by Becoa and Vazquez ( 2001 ) consisted of six weekly packets containing pamphlets with new strategies to stop smoking, a personalized letter of introduction with instructio ns for the tasks to be carried out during the week, self monitoring and evaluation forms and two letters with computer generated personalized feedback Having strategies to work on over the course of the week may have been the essential component for the successful reductions in smoking reported in the study. Another possible explanation for the variability in treatment success may be the differences able to con trol whether they remain in the presence of other smokers, but controlling the amount
66 of stress experienced at work may not be possible. The data suggested that, in response to feedback, participants smo ked less in specific situations as opposed to globall y smoking less. The degree to which the remaining high risk variables were present in the environment likely varied. In other words, some high risk variables may have been present at a low rate across any given day, making abstinence more probable. On the other hand, it may have been more feasible for some participants to avoid their high risk situations. Once the intervention methods are solidified, moving to an automated system would be ideal. F or each participant, a research assistant had to check the p smoking status, and then deliver the appropriate feedback. This occasionally result ed in a substantial lag between sample submission and feedback which may be another possible limitation of this intervention No data are availa ble on the optimal frequency or latency of feedback delivery, thus it is unclear if a shorter latency to feedback delivery would have increased the efficacy of the intervention. Nevertheless, an automated system would make implementation more feasible. Thi s intervention employed an internet based method for collecting objective evidence of smoking and delivering contingent tailored feedback. The literature suggests that some, but not all, tailored feedback interventions promote abstinence from smoking (Stre cher, 1999). Definitive conclusions cannot be made about the efficacy of this intervention, but t he data from the present studies suggest that this tailored feedback intervention shows promise F urther refinement and investigation may be required. An evalu ation of several variables, including frequency of feedback delivery, tailored versus untailored feedback, and Motiv8 components versus tailoring alone, is warranted. Understanding the roles of these variables may allow us to
67 make the effects o f this inter vention more robust such that greater rates of abstinence are achieved by more participants.
68 APPENDIX A EMA ITEMS AND ABBREVIATIONS Items Abbreviation 1. I am with other people who are smoking 2. I am with other people who are NOT smoking. 3. I am wait ing for someone or something. 4. I have nothing to do. 5. I just ate a meal or a snack. 6. I am taking a break from work. 7. I am in my car. 8. I am drinking a beer or other alcoholic beverage. 9. I am drinking coffee. 10. I am talking on the phone. 11 I am at home. 12. I am at work. 13. I have an urge to smoke 14. I liked handling the cigarettes and cigarette container. 15. I liked handling the lighter or matches. 16. I liked the way smoking made me feel around other people. 17. I liked the way peop le looked at me when I smoked. 18. I liked the way people interacted with me with I smoked. 19. I socialized with other when I smoked. With smokers With non smokers Waiting Nothing to do Meal Break from work Car Alcohol Coffee Phone At h ome At w ork Urge Handling Cigarettes Handling lighter Feel around others Looked at me Interacted with me Socialized with others
69 APPENDIX B FASTR ITEMS AND ABBREVIATI ONS Items Abbreviation 1. I smoke to give myself a lift. 2. I smoke because it's relaxing. 3. I smoke be cause it's satisfying. 4. I smoke because it makes me happy. 5. I smoke because it makes me feel content. 6. I smoke to help deal with anger. 7. I smok e to help deal with frustration. 8. I smoke to cope with feeling sad or depressed. 9. I smoke when I feel anxious. 10. I smoke because I get the urge to do so. 11. I smoke when I am with other people who are smoking. 12. I smoke when I am with other people who are NOT smoking. 13. I smoke because I like handling the cigarettes and cigarette container. 14. I s moke because I like handling a lighter or matches. 15. I smoke because I like the way it makes me feel around other people. 16. I smoke because I like the way people look at me when I am smoking. 17. I smoke because I like the way people interact with me w hen I'm smoking. 18. I smoke to socialize with others. 19. I smoke when I wait for someone or something 20. I smoke when I have nothing better to do. 21. I smoke after my meal or snacks. 22. I smoke to take a break from work. 23. I smoke in my car. 24. I smoke when I drink alcoholic beverages. 25. I smoke when I drink coffee. 26. I smoke when I am on the phone. 27. I smoke when I am at home 28. I smoke when I am at work Lift Relaxing Satisfying Happy Content Anger Frustration Depressed Anxious Urge Wit h smokers With non smokers Handling cigarettes Handling lighter Feel around others Look at me Interact with me Socialize with others Waiting Nothing to do Meal Break Car Alcohol Coffee Phone At home At work
70 APPENDIX C SE SC ITEMS AND ABBREVIATIONS Items Abbreviation 1. When you feel impatient 2. When you feel frustrated. 3. When you are waiting for someone or something. 4. When you want something in your mouth. 5. When you are worried. 6. When you feel restless. 7. When you feel upset. 8. When you want to keep yourself busy. 9. When you feel tense. 10. When you want to take a break f rom work or some other activity. 11. When someone offers you a cigarette. 12. When you want something to do with your hands. 13. When you feel nervous. 14. When you are drin king an alcoholic beverage. 15. When you feel angry. 16. When you want to keep slim. 17. When you are trying to pass time. 18. When you want to avoid sweets. 19. When you feel depressed. 20. When you feel annoyed. 21. When you feel bored. 22. When you see others smoking. 23. When you become aware of the fact that you aren't smoking. 24. When you feel angry with yourself. 25. When you have finished a meal or snack. 26. When you are drinking coffee. 27. When you feel anxious. Impatient Frustrated Waiting for someone Want something in mouth Worried Restless Upset Want to keep your self busy Tense Take a Break Offers you a cigarette Something t o do with hands Nervous Alcoholic beverage Angry Keep slim Pass time Avoid sweets Depressed Annoyed Bored See others smok ing Aware you aren't smoking Angry with self Meal or snack Drinking coffee Anxious
71 APPENDIX D FEEDBACK ITEMS Library of f eedback i tems f or a p ositive CO s ample 1. Practice saying, "No thank you, I don't smoke." 3 Set up a support system for your quit attempt. This could be a group program or a friend or family member who has successfully quit and is willing to help you. Ask family and friends who still smoke not to smoke around you or leave cigarettes out where you can see them. Tell your friends and f amily you need their help and support! 3. Each morning, wait an extra 15 minutes before you have your first cigarette. 4. Get rid of all lighters and cigarettes at work. Do not to take cigarettes to work with you. 5. Do not smoke with other people or soci alize with others when you smoke. 6. A void smoking in your car. Clean out your car of all smoking paraphernalia (e.g., lighters) and clean out your ashtrays. 7 If you need to smoke while driving, pull over and get out of your car to smoke. 8. Get rid of all cigarettes, ashtrays and lighters in your home. Keeping them around you only gives you more opportunities to smoke. 9. U sing oral substitutes such as sugarless gum, carrot sticks, hard candy, cinnamon sticks, coffee stirrers, straws, and/or toothpicks. 10. K eep active during your quit attempt. Go for a walk or [insert RSS hobby here]. 11. Reduce or avoid alcohol. Drinking may lower your chance of success 12. Avoid people who are smoking. 13. Change up your routine. Drink tea, juice, or water instead of coffee. 14. Do [insert RSS item here] to keep your hands busy, which can help distract you from the urge to smoke. 15. When the urge to smoke strikes, breathe deeply and picture your lungs filling with fresh, clean air. 16. If you feel that you are about to light up, hold off. Tell yourself you must wait at least 10 minutes. Often this simple trick will allow you to move beyond the strong urge to smoke. 17. Instead of smoking after a meal, brush your teeth im mediately after eating and take a walk. 18. Was h your hands or the dishes when you want a cigarette very badly. 19. Have your coffee at a different time of day or in a different place. 20. Take public transportation or carpool to avoid smoking in your car. 21. Have a list of things you can do at a mome nt's notice. For example, you could text a friend or delete old emails. Something to k eep you busy during down time! 22. Learn to relax quickly and deeply. Make yourself go limp. Think about a soothing, pleasing situation, and imagine yourself there. Get a way from it all for a moment. Focus on that peaceful image and nothing else. 23. R efrain from doing other activities while smoking, such as talking on the phone. 24. Do more [insert RSS item here]. It is something you enjoy and it may help you refrain from smoking. 25. You mentioned th at you wanted to quit because [insert RFQ item and how importan t it is to focus on this during a quit attempt]
72 Library of feedback items for a negative CO sample 1. Put the money you would have spent on tobacco in a jar e very day and then buy yourself a weekly treat. 2. Since you haven't smoked in a few days, you may notice your senses of taste and smell are better. 3. Since you haven't smoked in a few days, you may notice you can breathe easier. 4. Since you haven't smoke d in a few days, you may notice your "smoker's cough has started to go away. You may keep coughing for a while, tho ugh, as your body heals itself. 5. You have added healthy, full days to each year of your life. You've greatly lowered your risk of death fr om lung cancer and other diseases including stroke and emphysema. 6. Since you haven't smoked in a few days, you may notice that [insert RSS item] is easier for you to do. 7. After [insert time] of not smoking, you have saved [insert savings calculated fro m smokefree.gov]. After 1 year, you will have saved [insert savings ] by not smoking! 8. You mentioned that you wanted to quit because...[insert item and how they can contact naturally occurring contingencies stemming from cessation] 9. If you are still get ting the urge to smoke, remember to breathe deeply and picture your lungs filling with fresh, clean air. 10. Continue to practice relaxation and deep breathing so you are prepared if the urge to smoke strikes. 11. You don't need to smoke to do activities y ou love, such as [enter RSS item]. Keep doing these things and enjoying them! 12. Continue to practice saying "No thank you, I don't smoke." You don't smoke anymore, but people may still offer. Be prepared to be strong! Library of feedback Items for a po sitive CO sample after a period of abstinence 1. The difference between a slip and a relapse is within your control. A slip is a one time mistake that is quickly corrected. A relapse is going back to smoking. You can use the slip as an excuse to go back t o smoking, or you can look at what went wrong and renew your commitment to staying away from smoking for good. Don't be discouraged! 2. Everyone slips up, but you can get back on track! The most important thing is that you are still motivated to quit (refe rence previously reported level of motivation).
73 LIST OF REFERENCES Alessi, S. M., Badger, G. J., Higgins, S. T. (2004). An experimental examination of the initial weeks of abstinence in cigarette smokers. Experimental and Clinical Psychopharmacology, 2( 4), 276 287. Axelrod, S. (1991). Smoking cessation through functional analysis. Journal of Applied Behavior Analysis, 24, 717 718. Bandura, A. (1977). Self efficacy: Toward a unifying theory of behavior change. Psychological Review, 84, 191 215. Baer, D. M., Wolf, M. M., & Risley, T. R. (1968). Some current dimensions of applied behavior analysis. Journa l of Applied Behavior Analysis, 1, 91 97. Baer, J. S., Holt, C. S., & Lichtenstein, E. ( 1986 ). Self e fficacy and smoking r eexamined: Construct v alidi ty and c linical u tility Journal of Consulting and Clinical Psychology, 54, 846 852 Baer, J. S., & Lichtenstein, E. (1988). Classification and p rediction of smoking relapse e pisodes: An e xploration of i ndividual d ifferences Journal of Consulting and Cli nical Psychology 56, 104 110 Baker, T. B., Brandon, T. H., & Chassin L. (2004). Motivational influences on cigarette s moking Annual Review of Psychology 55 463 491 Baker, T B. Piper, M E. McCarthy, D E. Majeskie, M. R. & Fiore, M C. (2004 ). Addiction m otivation r eformulated: An a ffective p rocessing m odel of n egative r einforcement. Psychological Review, 111(1), 33 51. Becoa, E., & Vazquez, F. L. (2001). Effectiveness of personalized written feedback through a mail intervention for smokin g cessation: A randomized controlled trial in Spanish smokers. Journal of Consulting and Clinical Psychology, 69, 33 40. Bigelow G. E., & Silverman, K. ( 1999 ). Theoretical and empirical foundations of contingency management treatments for drug abuse. In S. T. Higgins, & K. Silverman (Ed s.), Motivating behavior change among illicit drug abusers: Research on contingency management interventions (pp. 15 31). Washington, DC : American Psychological Association Butschky M F Bailey D Henningfield J E ., & Pickworth W B. ( 1995 ) Smoking without nicotine delivery decreases withdrawal in 12 hour abstinent smokers. Pharmacology Biochemistry and Behavior, 50, 91 96. Cautela, J. R., & Kastenbaum, R. (1967). A Reinforcement Survey Schedule for use in thera py, training, and research. Psychological Reports, 20(3), 1115 1130.
74 Centers for Disease Control and Prevention. Vital Signs: Current Cigarette Smoking United States, 2005 2010 Morbidity and Mortality Weekly Report 2011;60(33):1207 12. Coelho, R. J. (1984). Self efficacy and cessation of smoking. Psychological Reports, 54, 309 310. Condiotte, M. M., & Lichtenstein, E. (198 1). Self efficacy and relapse in smoking cessation programs. Journal of Consulting and Clinical Psychology, 49(5), 648 658. Curry S. J., McBride, C., Grothaus, L. C., Louid, D., Wagner, E. H. (1995). A r andomized t rial of s elf h elp m aterials, p ersonali zed f eedback, and t elephone c ounseling w ith n onvolunteer s mokers Journal of Consulting and Clinical Psychology, 63(6), 1005 1014 Curry, S., Wagner, E. H., & Grothaus L. C. ( 1990 ). Intrinsic and e xtrinsic m otivation for s moking c essation Journal of Co nsulting and Clinical Psychology, 58(3 ) 310 316 Dallery J. & Glenn I. (2005 ). E ffects of an i nternet based voucher reinforcement program for smoking abstinence: A feasibility study. Journal of Applied Behavior Analysis, 38, 349 357 de Vries H., & Brug, J. ( 1999 ). Computer tailored interventions motivating people to adopt health promoting behaviours: Introduction to a new approach Patient Education and Counseling, 36, 99 105. DiClemente C. C., Marinilli, A. S., Singh, M., & Bellino, L. E. (200 1). The r ole of f eedback in the process of health behavior change. American Journal of Health Behav ior, 25(3), 217 227. Dijkstra, A., De Vries, H., & Roijackers, J. (1998). Long term effectiveness of computer generated tailored feedback in smoking cessat ion Health Education Research, 13(2), 207 214 Dijkstra, A., De Vries, H., & Roijackers, J. (199 9 ). Targeting s mokers with l ow r eadiness to c hange with t ailored and nontailored s elf help m aterials Preventive Medicine 28, 203 211. Fagerstrom, K. O., & Schneider, N. G. Measuring nicotine dependence: A review of the Fagerstrom Tolerance Questionnaire Journal of Behavioral Medicine, 12, 159 182. Fiore, M. C., Jaen, C. R., Baker, T. B., Bailey, W. C., Benowitz, N. L., Curry, S. J., et al. (2008). Treat ing tobacco use and dependence: 2008 update. Clinical practice guideline. Rockville, MD: US Department of Health and Human Services, Public Health Service.
75 Griffiths, R. R., Bigelow, G. E., & Henningfield, J. E. (1980). Similarities in animal and human drug taking behavior. In N. K. Mello (Ed.), Advances in substance abuse (pp. 1 90). Greenwich, CT: JAI Press. Gwaltney, C J ., Metrik, J., Kahler C. W., & Shiffman, S. (2009). Self e fficacy and smoking c essation: A meta an alysis Psychology of Addictiv e Behaviors, 23(1), 56 66 Hanley, G. P., Iwata, B. A., & McCord, B. E. ( 2003 ). Functional analysis of problem behavior: A review. Journal of Applied Behavior Analysis, 36, 147 185. Hartmann D. P., & Hall, R. V. ( 1976 ). The changing criterion design. J ournal of Applied Behavior Analysis, 9(4), 527 532. Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerstrom, K. O. (1991). The Fagerstrom t est for n icotine d ependence: A revision of the Fagerstrom t olerance q uestionnaire. British Journal of Ad diction 86, 1119 1127. Husky M. M., Mazure, C. M., Carroll, K. M., Barry, D., Petry, N. M. ( 2008 ). Using the experience sampling method in the context of contingency management for substance abuse treatment. Journal of Applied Behavior Analysis, 41(4), 635 644. Hughes, J. R., & Hatsukami, D. (1986). Signs and symptoms of tobacco withdrawal. Archives of General Psychiatry, 43, 289 294. Iwata, B. A., Kahng, S. W., Wallace, M. D., & Lindberg, J. S. (2000). The functional analysis model of behavioral as sessment. In J. Austin, & J.E. Carr (Eds.), Handbook of applied behavioral analysis (pp.61 89). Reno, NV: Context Press. Javors, M. A., Hatch, J. P., & Lamb, R. J. (2005). Cut off levels for breath carbon monoxide as a marker for cigarette smoking. Addi ction, 100, 159 167. Jorenby, D. E., Hays, J. T., Rigotti, N. A., Azoulay, S., Watsky, E. J., Williams, K. E., et al. vs placebo or sustained release buproprion fo r smoking cessation: A randomized controlled trial. Journal of the American Medical Association, 296, 56 63. Kassel, J.D., & Yates, M. (2002). Is there a role for assessment in smoking cessation treatment? Behavior Research and Therapy, 40, 1457 1470. K eehn, J. D., Bloomfield, F. F., & Hug, M. A. (1970). Use of the Reinforcement Survey Schedule with alcoholics. Quarterly Journal of Studies on Alcohol, 31, 602 615. Lamb, R. J., Morral, A. R., Galbicka, G., Kirby, K. C., & Iguchi, M. Y. ( 2005 ). Shaping r educed smoking in sm okers w ithout c essation p lans. Experimental and C linical Psychopharmacology, 13(2), 83 92.
76 Mclntyre, K. O., Lichtenstein, E., & Mermelstein, R. J. ( 1983 ). Self efficacy and r elapse in smoking ce ssation: A replication and ex tension J ournal of Consulting and Clinical Psychology 51, 632 633. Owen N., Ewins, A., Lee, C. ( 1989 ). S moking cessation by mail: A comparison of standard and personalized correspondence course formats. Addictive Behaviors, 14, 355 363. Prochaska, J. O., & DiC lemente, C. C. (1983). Stages and processes of self change of smoking: Toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51(3), 390 395 Prochaska J O DiClemente C C Velicer W F., Rossi J S. (1993). Standa rdized, individualized, interactive and personalized self help programs for smoking cessation. Health Psychology, 12, 399 405. Raiff, B. R., Faix, C., Turturici, M., & Dallery, J. (2010). Breath carbon monoxide output is affected by speed of emptying the lungs: Implications for laboratory and smoking cessation research Nicotine & Tobacco Research, 12(8), 834 838 Ramelson H. Z., Friedman, R. H., Ockene, J. K. ( 1999 ). An automated telephone based smoking cessation education and counseling system Patie nt Education and Counseling, (36), 131 144. Ramonowich, P., Mintz, J., & Lamb, R. J. (2009). The r elationship between se lf efficacy and r eductions in s moking in a c ontingency m anagement p rocedure Experimental and Clinical Psychopharmacology, 17(3), 139 145 Rose, J. E., & Levin, E. D. (1991). Inter relationships between conditioned and primary reinforcement in the maintenance of cigarette smoking. British Journal of Addiction, 86 605 609. Schnoll, R A. Martinez, E., Tatum K. L., Glass M., Bernath, A., F erris D., & Reynolds P. ( 2011 ). Increased self efficacy to quit and perceived control over withdrawal symptoms predict smoking cessation following nicotine dependence treatment Addictive Behaviors 36, 144 147 Skinner B F. (1953). Science and human behavior. New York: Macmillan. Shiffman S. (1982). Relapse following smoking cessation: A situational analysis. Journal of Consulting and Clinical Psychology, 50, 71 86. Shiffman, S. (2005). Dynamic influences on smoking r elapse p rocess Journal o f Personality, 73, 1715 1748
77 Shiffman, S., Gwaltney, C. J., Balabanis, M. H., Liu, K. S., Paty, J. A., Kassel, J. D., Hickcox, M., & Gnys M. (2002). Immediate a ntecedents of c igarette s moking: An a nalysis f rom e cological m omentary a ssessment Journa l of Abnormal Psychology, 111, 531 545 Shiffman, S., Hufford, M., Hickcox, M., Paty, J. A., Gnys, M., Kassel, J. D. (1997). Remember that? A c omparison of r eal t ime v ersus retrospective r ecall of smoking l apses Journal of Consulting and Clinical Psycho logy, 65 292 300 Shiffman S., Paty, J. A., Gnys, M., Kassel, J. A., Hickcox, M. ( 1996 ). First l apses to s moking: Within s ubjects a nalysis of real time r eports Journal of Consulting and Clinical Psychology, 64, 366 379 Shiffman, S., Paty, J. A., Gwa ltney, C. J., & Dang Q. (2004). Immediat e antecedents of cigarette s moking: An a nalysis of u nrestricted smoking p atterns Journal of Abnormal Psychology, 113, 166 171 Shiffman, S., & Stone, A. A. (1998). Ecological momentary assessment: A new tool for behavioral medicine research. In D. Krantz & A. Baum (Eds.), Technology and methods in behavioral medicine (pp. 117 131). Mahwah, NJ: Erlbaum. Stitzer, M. L., & Bigelow, G. E. (1982). Contingent reinforcement for reduced carbon monoxide levels in ciga rette smok ers. Addictive Behaviors, 7, 403 412. Stitzer M. L., & Bigelow, G. E. (1983) Contingent payment for carbon monoxide reduction: Effects of p ay a mount Behavior Therapy, 14, 647 656 Stone, A. A., & Shiffman, S. (1994). Ecological momentary ass essment (EMA) in behavioral medicine. Annals of Behavioral Medicine, 16, 199 202. Stoops, W. W., Dallery, J., Fields, N. M., Nuzzo, P. A., Schoenberg, N. E., Martin, C. A., et al. (2009). An internet based abstinence reinforcement smoking cessation inte rvention in rural smokers. Drug and Alcohol Dependence, 105, 56 62. Strecher, V. J. (1999). Computer tailored smoking cessation materials: A review and discussion. Patient Education and Counseling, 36, 107 117 Strecher, V. J., Kreuter, M., Den Boer, D., Kobrin, S., Hospers, H. J., Skinner, C. S. (1994). The effects of computer tailored smoking cessation messages in family practice settings. Journal of Family Practice, 39, 262 270. Strecher, V. J., Marcus, A., Bishop, K., Fleisher, L., Stengle, W., Levi nson, A. Fairclough, D. L., Wolfe, P., Morra, M., Davis, S., Warnecke, R., Heimendinger, J., & Nowak, M. (2005). A r andomized c ontrolled t rial of m ultiple t ailored m essages for s moking c essation among c allers to the Cancer Information Service Journal of Health Communication, 10, 105 118.
78 Te Poel, F., Bolman, C., Reubsaet, A., & de Vries, H. (2009). Efficacy of a single computer tailored e mail for smoking cessation: results after 6 months. Health Education Research, 24, 930 940. Thorndike, R. M., & K leinknecht, R. A. (1974). Reliability of homogeneous scales of reinforcers: A cluster analysis of the reinforcement survey schedule. Behavior Therapy, 5, 58 63. U.S. Department of Health and Human Services. Reducing Tobacco Use: A Report of the Surgeon General. Atlanta, Georgia: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2000. Watson, P. J., & Workman, E. A. (1981). The non concurrent multiple baseline across individuals design: An extension of the traditional multiple baseline design. Journal of Behavior Therapy and Experimental Psychiatry, 12 257 259. Weitzel, J. A., Bernhardt, J. M., Usdan, S., Mays, D., Glanz, K. (2007). Using wireless handheld computers and t ailored t ext m essaging to reduce ne gative c onsequences of d rinking a lcohol Journal of Studies on Alcohol and Drugs, 68(4), 534 537. Wetter D. W., McClure, J. B., Cofta Woerpel, L., Costello, T. J., Reitzel, L. R., Businelle, M. S., & Cinciripini, P. M. (2011) A randomized c linical t rial of a p almtop computer d elivered t reatment for smoking relapse p revention a mong w omen Psychology of Addictive Behaviors, 25(2), 365 371
79 BIOGRAPHICAL SKETCH Alana M. Rojewski was born in Zanesville, Ohio. She graduated from Bishop Rosecrans High School in 2002, where she took an early interest in psychology. She was able to pursue this major at Denison University in Granville, Ohio, graduating magna cum laude with her B.A. in psychology in 2006. From her early research experiences at Denison, she also developed an interest in drug abuse research. Alana matriculated in the behavior analysis program of the Psychology Department at the University of Florida under Dr. Jesse Dallery in the fall of 2006. She received her M.S. from the University of Florida in the fall of 2009 and her Ph.D in August 2012. Her research interests include smoking cessation and contingency management. Recently, she have taken an interest in smoking among populations where the choice to smoking is directly detrimental to immediate health (e.g., with cancer patients who smoke) and evaluating cessation methods for such populations. She will have the opportunity to pursue these research intere sts at Yale University, where she has accepted a Post Doctoral Fellowship.