School Psychology Quarterly 2008, Vol. 23, No. 3, 389 – 406 Copyright 2008 by the American Psychological Association 1045-3830/08/$12.00 DOI: 10.1037/1045-3830.23.3.389 Using Behavioral Interventions to Assist Children With Type 1 Diabetes Manage Blood Glucose Levels Kim Lasecki Daniel Olympia, Elaine Clark, William Jenson, and Lora Tuesday Heathfield Bellin Psychiatric Center This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. University of Utah Treatment and management of chronic disease processes on children occurs across multiple settings, placing demands for consultation and expertise on school personnel, including school psychologists. One such chronic condition in children is type I diabetes. Children with type I insulin dependent diabetes mellitus exhibit high rates of noncompliance to treatment, which can lead to a variety of medical problems. This study examined the effectiveness of a specific behavioral intervention using behavioral consultation (BC) and conjoint behavioral consultation (CBC) to reduce uncontrolled blood glucose levels in medically at-risk children. An intermittent reward procedure was utilized to reinforce individualized target behaviors associated with treatment noncompliance. Specific target behaviors were individually established for six patients ages 8-12 through behavioral consultation interviews. Each child was randomly assigned to a reward ⫹ BC or reward ⫹ CBC condition. Results of the study showed that all participants improved; with slightly greater gains shown in the CBC condition. Follow-up data for 3 of the 4 participants completing the study showed improved compliance and mental health status. Treatment acceptability date indicate the intervention was viewed positively by parents and school based nurses. Keywords: interventions, behavioral, Type 1 diabetes, children The contributions of school psychologists to the treatment and management of students whose medical conditions are frequently associated with educational and social difficulties has grown significantly in the last several years (Brown & DuPaul,1997). Indeed, schools are more often asked to address a wide range of behavioral issues or problem behaviors that About the authors: Kim Lasecki received his PhD in educational psychology at the University of Utah in 2000. Currently, he works as a licensed psychologist for Bellin Psychiatric Center where he previously served as Director of Outpatient Services. Dr. Lasecki has interests in outpatient psychological assessment, therapy, and behavioral interventions related to child and adolescent issues. Daniel Olympia is an Associate Professor in the Department of Educational Psychology at the University of Utah. He received his doctorate in 1992 from the University of Utah. Dr. Olympia’s research interests include academic and behavioral interventions, consultation strategies, practice, and ethical issues in applied settings and autism. Elaine Clark is Professor of School Psychology, and Chair of the Department of Educational Psychology at the University of Utah. Dr. Clark’s research interests include assessment and interventions with conditions that impact a child’s cognitive functioning, including neurodevelopmental and acquired conditions, academic performance, and psychosocial status. William Jenson is a Professor in the Department of Ed- ucational Psychology at the University of Utah. He received his doctorate from Utah State University in 1976, with a specialization in child clinical psychology. Dr. Jenson’s research interests include behavioral interventions for tough kids, parent training, generalization of treatment effects, and autism. Dr. Jenson is the author of The Tough Kid Book. Lora Tuesday Heathfield, PhD, is an Associate Professor in the Department of Educational Psychology at the University of Utah in Salt Lake City, UT. Her research interests focus on improving service delivery to children with or at risk for learning and/or behavioral difficulties and their families in the areas of prevention and intervention strategies, as well as assessment issues. Dr. Tuesday Heathfield received her PhD in School Psychology from the University of Oregon and completed post-doctoral training in Pediatric Psychology. Correspondence concerning this article should be addressed to Daniel Olympia, University of Utah, Department of Educational Psychology, 327 MBH, SLC, UT 84112. E-mail: dan.olympia@utah.edu 389 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 390 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD may be associated with specific health problems. Treatment or management of chronic disease processes in children also occurs across multiple settings, placing additional demands for consultation and expertise on school personnel (Nastasi, 2000). This is especially true of Type I diabetes, an insulin-dependent autoimmune disease that has serious implications for a child’s mental and physical health and for their ability to be successful in school settings (Daley, Woderich, & Hasan, 2004). Type I diabetes, hereafter referred to as TI diabetes, destroys insulin producing beta cells in the pancreas and makes it difficult for the body to use glucose and break down foods into simple sugars. Without insulin, cells are starved for fuel and dangerous levels of unabsorbed glucose build up in the body. As a result, individuals with TI diabetes have to inject insulin and effectively manage dietary intake (Foster, 1991). TI diabetes affects 1 in 500 children under the age of 18 (Sperling, 1990), making it the second most common chronic disease in school-age children, the first being asthma (Strawhacker, 2001). Characteristics of TI diabetes at the time of initial diagnosis include: high blood-sugar levels, excessive thirst and urination, weight loss, and vision problems. When blood-sugar (glucose) levels cannot be stabilized, other problems often occur. Although some problems are short-term (e.g., slowed information processing, poor concentration and memory, dizziness, moodiness, and fatigue), others are more long term (Chase, 1985) and can include a variety of mental health and psychosocial related disorders (Kovacs, Goldston, Obrosky, & Bonar, 1997; Kovacs, Obrosky, Goldston, & Drash, 1997; Smith & Baum, 1980). Adherence to Typical Treatment Regimens Treatment for TI diabetes consists of a combination of regular blood-glucose testing, insulin injections (when needed), daily exercise, and dietary management (Karam, 1996). Children have particular difficulty with this regimen for several reasons. Blood sugar levels need to be checked 3 to 5 times a day (before meals and bedtime snacks). In addition, insulin injections are often needed before breakfast and dinner. The targeted goal for most individuals is a blood sugar level between 70 and 150 mg/dl at least 50% of the time (Betteridge, 2000). This means some children will need to check their bloodsugar levels more often during the day, and inject insulin more frequently to achieve recommended levels. More often than not, blood testing and insulin injections are done at home; however, in some cases insulin has to be administered at school (e.g., before lunch). In addition to insulin injection, treatment also includes exercise and a proper diet. Children need to manage their food intake by eating regular meals and having appropriate snacks (e.g., midmorning and midafternoon). In many cases this requires monitoring by educational personnel. Lastly, treatment must include exercise, another activity that school staff may regularly control. It is difficult to estimate the extent of children’s adherence to the treatment regimen; however, Kovacs and colleagues report that children are at greater risk for noncompliance than other age groups (Kovacs, Goldston, Obrosky, & Iyengar, 1992; Kovacs, Mukerji, Iyengar, & Drash, 1996). There are a number of possible reasons for noncompliance, including poor communication between health providers and families, a lack of understanding of the complications associated with TI diabetes, and a lack of parental involvement and skill in managing the treatment (Delamater, 2000). Given these issues, it is critical that interventions involve multiple individuals (e.g., families and peers) and address multiple contexts (e.g., home, school, and community). Improving Compliance Using Behavioral Interventions and Consultation Most school psychologists are familiar with the use of behaviorally based interventions to address academic and behavioral issues. The increased prevalence of children with medically complex conditions or children who experience some aspect of a medical treatment in school has created both opportunities and expectations for school psychologists. For example, under both the Individuals with Disabilities Improvement Act (IDEIA) and Section 504 of the Rehabilitation Act and the Americans with Disabilities Act, chronic health issues must be addressed if they affect a child’s ability to participate in school activities or learning. Empirical support for interventions that rely on behaviorally based principles to effectively manage pediatric diabetes were recently sum- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES marized by Wysocki (2006). These include treatments targeting treatment compliance, social coping skills, and family based management techniques. However, there is no research that addresses behavioral management of the disorder across multiple settings (i.e., home, clinic, and school) or uses a consultation centered framework to develop individualized treatment protocols. Additionally, traditional school consultation research has often focused on interactions with only teachers and parents, not other related personnel (i.e., school nurses). Innovative behavior management strategies supported by behavioral consultation can be a means to ensure that interventions aimed at increasing treatment compliance involve critical parties in all settings (Sheridan, Welch, & Orme, 1996). Mystery motivator. The mystery motivator (Jenson, Rhode, & Reavis, 1995) uses the principles of positive reinforcement, performance feedback, and randomized contingencies to initiate and, more importantly, to maintain compliance and anticipatory motivation. A central element of the intervention consists of an envelope, labeled with a question mark and containing a card identifying a high value, high interest reward. The child gains access to several reward choices (including the mystery motivator) through the achievement of a daily individualized goal. Several studies have determined that the mystery motivator can effectively improve academic and behavioral goals (DeMartiniScully, Bray, & Kehle, 2000; Leblanc, 1999; Moore et al., 1994; Robinson & Sheridan, 2000; Valum, 1996). However, no published studies have applied this intervention to facilitate the management of a medical treatment across settings. Behavioral consultation. Behavioral consultation (BC) involves an indirect service delivery approach, typically with a parent or teacher as the consultee. Conceived by Bergan and Kratochwill (1990), BC involves indirect service delivery to a student through a consultee, such as a parent or teacher, by a consultant (e.g., psychologist). Therefore, BC involves three participants: consultant, consultee, and client. The consultant is generally a psychologist, counselor, or mental health worker. The consultee is typically a teacher or parent, but is not limited to these individuals. Consultees could also be nursing staff in a hospital or 391 school setting. Clients are often school-age children identified to be struggling with significant behavioral, emotional, social, or academic difficulties. BC’s approach to helping the consultee and client is based upon the identification of specific antecedents or consequent events that affect a target behavior. The BC model has its foundation in applied behavioral analysis; as such, data are collected to determine if the intervention is effective. Typically, the model involves four distinct stages (problem identification, problem analysis, plan implementation, and problem evaluation). Large scale metaanalytic studies of behavioral consultation have reported effect sizes as high as .95 (Bergan & Kratochwill, 1990; Kratochwill, Elliott, & Busse, 1995; Medway & Updyke, 1985; Sibley, 1986). Conjoint behavioral consultation (CBC). The CBC model is a variant of BC. Although it uses an indirect form of service-delivery (with distinct stages for problem identification, analysis, etc.), it makes use of a broader range of consultees. For example, parents, teachers, and other support staff from schools work with a consultant (often a school psychologist) to address problem behaviors. These individuals form an alliance to provide assistance in the environments where children are having difficulty. Working with a consultant, multiple consultees share responsibility for developing and implementing the intervention, and evaluating its effectiveness (Sheridan & Kratochwill, 1992; Sheridan, Kratochwill, & Bergan, 1996). These partnerships are considered to be critical in that each consultee is viewed as bearing partial responsibility for meeting the needs of a student because of different perspectives and connections with different environments (e.g., home and school). Given an emphasis on a broad-based multisource, multisetting, multimethod assessment approach (Sheridan, Kratochwill, & Bergan, 1996), the CBC model seems to be particularly well suited to address the problem of treatment compliance in TI diabetes. Parents and medical professionals, together with school personnel (e.g., teachers and school psychologists), can help identify specific issues associated with treatment noncompliance problems in various settings, including the school, home, and community. In other models, including the BC model and mental health consultation, the atten- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 392 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD tion of the consultant is often focused on an identified client with a problem in a particular milieu (and with a single consultee associated with that milieu). Critical information, and individuals, may be omitted in treatment planning and implementation, thereby limiting the effectiveness of the treatment plan. The CBC approach addresses this problem by increasing sources for information and the number of potential interventionists (e.g., parents, teachers, and other health professionals) needed to ensure that an effective treatment is designed, implemented, and maintained (Sheridan et al., 1996). The CBC model has also proven to be effective in improving a wide array of academic, behavioral, and social problems. Sheridan and her colleagues have found effect sizes to range from ⫺0.62 to 6.31 (e.g., Galloway & Sheridan, 1994; Sheridan, Colton, Fenstermacher, Lasecki, & Wilson, 1996; Sheridan, Eagle, Cowan, & Mickelson, 2001; Weiner, Sheridan, & Jenson, 1998). This study examined the usefulness of a specific behavioral intervention (mystery motivator) under two variations of behavioral consultation to improve treatment compliance for TI diabetes. To date there has been no systematic research that examines the effectiveness of these methods, singly or in combination, to help children with TI diabetes comply with the treatment regimen needed to stabilize their bloodglucose level. A single reinforcement strategy (mystery motivator) based on the principles of reinforcement unpredictability (when positive reinforcement becomes available) and reinforcement value (what is earned) was employed across two consultation conditions. Two variations of behavioral consultation, traditional Behavioral Consultation (BC) and Conjoint Behavioral Consultation (CBC) were implemented across four participants with well established difficulty in maintaining appropriate blood glucose levels. Method Participants Children who participated in the study were selected from a patient roster in a hospital affiliated with an outpatient clinic for children with diabetes in a large, urban Midwestern city. Each participant met the following criteria: chronological age between 8 years-0 months and 12 years-0 months at the time of the study, a diagnosis of TI diabetes made 4 to 6 years earlier, no other chronic illnesses, and most importantly, evidence of an uncontrolled bloodglucose level (i.e., recurrent hyperglycemia using hemoglobin A1c or glycolated hemoglobin ratings above 250 mg/dl). In cases where there was no recent hemoglobin A1c, participants were included if their blood-glucose levels averaged above 250 mg/dl (documented by daily glucose testing). Six children were randomly selected from a list of children who met these criteria. Half of the children were randomly assigned to the mystery motivator ⫹ BC condition and the other half assigned to the mystery motivator ⫹ CBC condition. Two children dropped out of the study during the Baseline phase, due in the first case to a psychiatric hospitalization (and later placement in a residential treatment program) and, in the second case, to parental inability to participate in the study due to a job schedule change. Two participants remained in the BC condition and 2 participants in the CBC condition. An abbreviated summary of participant demographic information is provided in Table 1. Participant 1 (CBC). Participant 1 was a 9 year-old, third grade male who had been TI diagnosed with diabetes for 4 years, 1 month at the time of the study. Since diagnosed, the child had not been consistently compliant with treatment, and therefore, was having frequent episodes of hyperglycemia. No specific intervention had been used to improve the child’s compliance with the treatment regimen for his diabetes. The child was living with his biological mother and younger sibling at the time of the study (his mother had divorced his stepfather 2 years earlier). Participant 1 had also been diagnosed with an Attention Deficit Hyperactivity Disorder (ADHD), combined type, and enrolled in a special education program for students with Emotional Disturbance (SED). The child’s mother described him as having significant behavior problems, including oppositional behaviors that were difficult to control. The school nurse further reported that Participant 1 was having academic problems as well as difficulty with lying, being manipulative, failing to follow rules, refusing (at times) to eat, and not bringing his diabetic monitor to school on a BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES 393 Table 1 Age, Gender, Grade Level, Family Status, Length of Diagnosis and Treatment Target Behavior for Each Participant Age/Gender Grade Family Status This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Length of Diagnosis Target Behavior Participant 1 Participant 2 Participant 3 Participant 4 9/Male 3rd Low income/single parent 4 years-1 month Percent of insulin injections monitored by an adult 11/Male 6th Low income/single parent 5 years-5 months Incidents per day of snacking without parent permission 11/Female 5th Middle income/ two parent 4 years-4 months Percent compliance with monitoring blood glucose levels and taking insulin 10/Male 4th Middle income/ two parent 4 years-4 months Percent compliance with monitoring blood glucose levels and taking insulin consistent basis. No formal, or structured, interventions were being used at school to assist with these behaviors (except the general special education program). Participant 2 (BC). Participant 2 was an 11 year-old, sixth grade male who had been diagnosed with TI diabetes for 5 years 5 months. Prior to the study, the child had been eating inappropriate snack food and had an elevated blood-glucose level. Variable supervision at home had reportedly contributed to Participant 2 having access to inappropriate foods (e.g., unsupervised snacking after school). Although the child (an only child) and his biological mother had received counseling for behavior problems before the study, his mother reported using mostly negative feedback to manage negative behaviors. Participant 2 had been diagnosed with ADHD and was prescribed Adderall. In addition, he was classified as “other health impaired” (OHI) and was receiving special education services in the schools. Primary concerns identified for Participant 2 included: poor academic performance, a poor relationship with his mother, and noncompliance with house rules and management of his diabetes. Participant 3 (BC). Participant 3 was an 11 year-old, fifth grade girl. She was diagnosed 4 years 4 months earlier with diabetes. Prior to the study, the child (the oldest of 3 children) had been noncompliant with treatment. Her noncompliance was associated with her parents’ work schedules, a lack of supervision in terms of her snacking (reported to be excessive), and poor monitoring of her blood sugar level. Participant 3’s mother reported that she and her husband differed in their parenting styles, that is, she was lenient and her husband was strict and at times negative. Although the child’s parents reported using time-out to help manage inappropriate behaviors at home, the use was inconsistent. Academically, Participant 3 was described by her teachers as being “average” and reported to be outgoing and having an engaging personality. She was also reported to be motivated by rewards. Participant 4 (CBC). Participant 4 was a 10 year-old, fourth grade male. He had been diagnosed with TI diabetes 4 years 4 months before the study. Since the initial diagnosis, the child was frequently noncompliant with monitoring his diabetes, leading to numerous episodes of hyperglycemia. Prior to the study, Participant 4’s mother had met with a nurse and dietitian at a local diabetes clinic. The child, the middle of three, was living with his siblings and biological parents at the time of the study. Participant 4’s relationship with his parents was observed to be positive. He did not have any psychiatric diagnosis and had not received any special education service. The school nurse who served as a consultee, however, reported some concern about children teasing him and about his “lying” about his blood-sugar levels. Consultant and Consultees The consultant for this study was a male doctoral graduate student in the School Psychology Program at the University of Utah. He received extensive supervised training to mastery through specific coursework and supervised practice through a two year placement in a federally funded training grant in evidence This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 394 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD based intervention strategies and BC and CBC models. To conform to the definition of each consultation strategy, consultees in the study included a “parent only” in the BC condition, and both a parent and school representative in the CBC condition. In both conditions, however, the child’s biological mother served as the parent consultee. The average age of the “parent consultees” in the study was 35 years of age (range ⫽ 26 to 41). None of the consultees reported familiarity with CBC or BC (and no one had any prior experience with these approaches). School nurses served as consultees during the CBC condition (i.e., for Participants 1 and 4). The nurse working with Participant 1 was 44 years-old and had 17 years of experience in the school district. The nurse working with Participant 4 was 32 years-old and had 6 years of experience. Neither nurses participating in the study was familiar (or had experience) with CBC or BC. Setting Consultation cases were primarily conducted at a hospital-based diabetes clinic and/or in the child participant’s school. For Participant 2, it was necessary on one occasion to meet for consultation at the mother’s home due to transportation difficulties. During the treatment phase for both conditions, the specific behavioral intervention (mystery motivator) was provided only within the home environment for all participants in the study. Design A single subject multiple-baseline design across participants was used to evaluate the effectiveness of the mystery motivator intervention under two consultation methods, (i.e., mystery motivator ⫹ Behavioral Consultation and mystery motivator ⫹ Conjoint Behavioral Consultation). The behaviorally based reinforcement technique (mystery motivator) was used to reinforce the reduction of hyperglycemia in the participants. The baseline was initiated for all participants simultaneously. Given that 2 participants failed to complete the study, baseline data for Participants 1 and 4 (in the CBC condition) were not paired, therefore, treatment ended up beginning at the same time. Partici- pants 2 and 3 (in the BC condition) did not begin treatment at the same time. A brief follow up period for each participant occurred 1 month after treatment to obtain data for target behaviors and blood glucose levels. Intervention Procedures In this study, the mystery motivator was used in combination with a “spinner” (a small cardboard circle divided into unequal, numbered segments with a small arrow or pointer, which is placed at the center and allowed to rotate). For each participant, the following steps were used to implement the mystery motivator program. Rewards to be used with each participant were selected during both the problem identification and problem analysis phases of BC and CBC. Consultees and child participants were given opportunities to list rewards they believed would be most effective and reasonable. After rewards were selected, a menu listing the rewards was completed. The menu identified numbers on the spinner corresponding to specific rewards. The menu also identified a space on the spinner that was designated as the “mystery” section. Parent consultees were asked to determine a “mystery” reward for each participant in the study prior to the intervention. Once the mystery reinforcer was determined, consultees were instructed to write the reward on a piece of paper, fold the paper, and place it in an envelope to maintain the secrecy of the reward from the participants. Table 2 displays the rewards including the mystery motivator selected and used during the intervention phase for each participant. Both the CBC and BC models involved the four typical stages: problem identification, problem analysis, treatment implementation, and treatment evaluation. Within CBC model, the consultant and the two consultees participated in three interviews: Conjoint Problem Identification Interview (CPII), Conjoint Problem Analysis Interview (CPAI), and Conjoint Treatment Evaluation Interview (CTEI). In the BC condition, only the consultant and parent consultee participated in three interviews: Problem Identification Interview (PII), Problem Analysis Interview (PAI), and Treatment Evaluation Interview (TEI). Each interview was completed in a single session and followed a similar standard protocol (Sheridan et al., BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES 395 Table 2 Reinforcement Used by Participants during Mystery Motivator This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Reinforcer Menu Space Participant 1 Participant 2 1 Pick t.v. show for the evening 2 Spend night at grandparents house Use cell phone to make 10 minute call. Have friend stay over night w/ permission 3 Stay up 1⁄2 hour later 4-Mystery Trip to local amusement park Pack of Pokemon cards 5 Staying up 1⁄2 hour later Baseball Batting Glove Picking appropriate dessert for dinner 1996). Each interview was scheduled by the consultant. For CBC cases, coordination of the nurse and parent schedules was necessary to allow both to be involved in each phase of the consultation process. All consultation sessions were conducted using a standardized set of questions adapted from Sheridan et al. (1996) for CBC and from Kratochwill and Bergan (1990) for BC. Children in the study participated in the problem analysis and treatment evaluation interviews (for both CBC and BC conditions). However, children in this study were excluded from the problem identification phase (in either CBC or BC condition) to prevent potential reactivity during collection of baseline data. Target behaviors. Parent consultees for all children in the study were given behavioral charts to write down the target behavior identified during the problem identification interview that was most critical to facilitate adherence to the proscribed treatment. One target behavior was identified for each child in the study. Parent consultees were instructed to keep the chart and menu in a visible location in the home such as posting it on the refrigerator. For CBC, school consultees did not take an active role in the mystery motivator intervention other than to supply data with the use of a homenote adapted from the Tough Kid Tool Box (Jenson, Rhode, & Reavis, 1995) to parents regarding the target behavior observed within the school environment. For Participant 1, the school nurse reported daily to the parents whether or not insulin injections were necessary and if so whether they were observed. For Participant 4, the Participant 3 Participant 4 Special shopping trip to mall with mom No chores for week Staying up 1 hour later at night Staying up 1 hour later for the night Friend sleep over Getting to control what is on tv. for night $25 Gift Certificate for clothes Permission to use phone for 1⁄2 hour. Soccer Ball 1 hour of computer time school nurse was active in giving daily feedback to parents about whether or not blood glucose levels were checked at school and if insulin was taken if necessary. On the first weekly chart, stars were randomly placed in boxes corresponding to the days of the week. Stickers were also used to cover each box to conceal where stars were located. Hidden stars were placed in 4 of the 7 possible locations corresponding to the days of the week. If the child participant attained the daily goal, they were then allowed to peel off the sticker to determine if they had earned a spin of the spinner. If the child uncovered a star, they were allowed to spin one time for a reward. If the child participant landed on the mystery section of the spinner, they were immediately allowed to open the mystery motivator envelope with the enclosed reward. All parent consultees were instructed to begin fading the program after the child had successfully completed their goal of performing the target behavior across an entire week without error. The fading method employed by parent consultees involved reducing the number of hidden stars by one per week after each successful week of target behavior performance. All parent consultees were contacted by telephone to collect follow-up data one month after consultation was completed. Data for the most recent four days of blood-glucose levels and target behavior were collected. Data collection. Baseline data collection procedures were discussed with consultees during the problem identification phase. After agreement and definition of a target behavior for This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 396 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD each case, consultees were asked to unobtrusively collect baseline data. For CBC cases, baseline data were collected in both the home and school environment. For BC cases, baseline data were collected in the home environment only. In all cases, the parent consultees were responsible for managing the collection of baseline data related to blood-glucose levels. Each parent was then responsible for bringing their children to the diabetes clinic to have bloodglucose data downloaded from the child’s blood-glucose monitor by the first author. During problem identification interviews within CBC and BC formats, all consultees consistently expressed concern about noncompliance toward treatment of diabetes. Within each case, however, the specific target behavior differed due to factors specific to each individual child’s case. The problem identification phase was designed to evaluate baseline data and environmental factors contributing to diabetes treatment noncompliance. For each case, antecedent conditions and setting events, consequent conditions, and environmental/sequential conditions associated with target behaviors were analyzed during the problem identification interview. Table 1 summarizes the target behaviors that were monitored for each child as well as demographic variables such as marital status of parents, income level, age of child, and length of diagnosis. The target behaviors listed in Table 1 were reviewed by the consultant and consultees and determined to be the most significant factor affecting each participant child at the time of the study’s problem identification interviews. To increase the likelihood of consultees understanding and properly implementing the procedures of this study, each consultee was contacted by telephone or in person weekly during the intervention phase. Consultees were given opportunities to discuss questions about data collection or procedures during those contacts. Additionally, parent consultees completed a treatment plan worksheet that highlighted the key components in carrying out the intervention procedures. Dependent Measures Blood glucose levels taken daily by each child participant under the supervision of consultees, were collected as baseline data and subsequently during intervention to establish glu- cose levels prior to, during, and after treatment. One-month from the point where consultation was terminated, follow-up measurements of blood glucose levels were also taken. Each participant’s target behaviors were also recorded across baseline, treatment, and follow-up conditions and plotted in relationship to changes in blood glucose levels. Additional data was collected to assess the social validity of treatment outcome, treatment acceptability, and treatment integrity. Goal Attainment Scaling (GAS) was utilized to provide an index of social validation (Kiresuk, Smith, & Cardillo, 1994). The GAS ratings range from ⫹ 2 to ⫺2 (⫹2 ⫽ behavior goal fully met, ⫹1 ⫽ goal partially met, 0 ⫽ no progress, ⫺1 ⫽ behavior somewhat worse, ⫺2 ⫽ behavior significantly worse). All consultees including parents and school nurses (during CBC) completed the GAS to assess the level of attainment of stated treatment goals for each consultation case. The acceptability of the behavioral intervention and consultation services used across cases was assessed via the Behavioral Intervention Rating Scale – Revised (BIRS-R; Elliott & Von Brock Treuting, 1991). The BIRS-R is comprised of 24 items scored on a 6 point Likert scale (1 ⫽ not at all acceptable; 6 ⫽ highly acceptable). Factor analysis of the BIRS-R has yielded three factors: Acceptability (15 items); Effectiveness (7 items); and Time to Effectiveness (2 items) (Elliott & Von Brock Treuting, 1991). Parent and school consultees completed the BIRS-R-R after the successful completion of the treatment evaluation phase of both types of consultation. Likewise, child participants completed the Children’s Intervention Rating Profile (CIRP) (Witt & Elliott, 1985) to assess acceptability of the intervention. Treatment integrity for this study was assessed in two ways. For each session of consultation, the consultant quantitatively evaluated treatment integrity using behavioral checklists adopted from Sheridan and Colton (1994). Checklists were used as a self-monitoring technique to determine the percentage of compliance in following the intended structure for each type of consultation (Sheridan et al., 1996; Kratochwill & Bergan, 1990) for CBC and BC. The consultant completed the appropriate selfmonitoring checklist immediately after the consultation session had ended and percentage of compliance was calculated. Treatment plan This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES worksheets completed by the parent consultees were also used as a self-monitoring checklist, and consisted of a review/verification of each step of the mystery motivator intervention. Instrumentation. The One Touch FastTake compact blood glucose monitoring system was used by all child participants. The One Touch FastTake system was developed by Lifescan (retrieved from http://www.lifescan.com/ company/, May 15, 2006) and used across all participants to minimize variability in measurement of glucose levels due to instrumentation. All participants were trained how to properly use the One Touch FastTake compact blood glucose monitoring system. FasTtake Test Strips were used by participants to measure blood-glucose levels. Test strips from each participant were placed in the meter after a small amount of blood was applied to the white target area of the FastTake Test Strip. Blood glucose test results appear after the meter counts down from 45 to 1 second. The FastTake Meter displays results between 20 and 600 mg/dL. The FastTake Meter also stores up to 250 blood glucose test results with date and time for later recording and review by the first author. Data Analysis Procedures Behavioral data were evaluated using effect sizes and visual analysis to detect level change. Immediacy of change and level stability also were calculated. Percentage of nonoverlapping data points (PND) was not used, as baseline data were too variable. Effect size (ES) was calculated using the Busk and Serlin (1992) “no assumptions” method. Baseline and treatment levels were determined by summing average daily blood glucose levels (mg/dL) and dividing by the number of data points in each phase. Comparisons of mean baseline and treatment levels were made visually to determine treatment effects. Immediacy of change was used to calculate the strength of intervention by determining the difference between the first treatment data point and last baseline data point for each participant (Tawney & Gast, 1984). As the difference between these two data points becomes greater, the immediacy of change (and the effectiveness of the intervention) can be ascertained. Level stability also was determined for baseline and treatment phases. Data were considered stable in each phase if at least 80% of data 397 points fell within a 15% margin surrounding the mean level for each baseline and treatment condition. Throughout the study, specific insulin doses or medication as prescribed each participant’s physician and the treatment team at the diabetes clinic were not changed. The most significant side effect of intensive treatment anticipated was an increase in the risk for low blood sugar episodes severe enough to require assistance from another person. To guard against any increased risk of hypoglycemia, strict control over blood sugar levels was not exercised through increased use of insulin injections beyond the protocol normally in place at the diabetes clinic. Results Blood Glucose Levels Figure 1 displays baseline, treatment, and follow-up data for each participant’s daily blood glucose levels. Each participant in this study was required through their normal regimen to check blood glucose levels four times per day. Data points represent daily averages for blood glucose levels calculated for each participant during baseline, treatment, and follow-up phases. Participant 1. For Participant 1, baseline blood-glucose levels were calculated to be an average of 273.0 mg/dL (range ⫽ 95– 460), with a treatment average of 141.3 mg/dL (range 35– 425). An effect size for change in average bloodglucose levels for Participant 1 was calculated to be 1.23. At follow-up, Participant 1 was calculated to have an average daily blood glucose level across four days of 161.3 mg/dL. Participant 2. For Participant 2, baseline blood-glucose levels were calculated to be an average of 302.0 mg/dL (range ⫽ 70 –599), with a treatment average of 185.8 mg/dL (range ⫽ 44 – 499). An effect size for Participant 2 was calculated to be 0.89. At follow-up, Participant 2 was found to have an average daily blood-glucose level across four days of 214.0 mg/dL. Participant 3. Baseline blood-glucose levels were calculated to be an average of 255.1 mg/dL (range ⫽ 74 –598). A treatment average was found to be 159.8 mg/dL (range ⫽ 69 – 245), generating an effect size of 1.48. However, at follow-up, Participant 3 was found to have an average daily blood-glucose level across four days of 267.69 mg/dL, representing a return to baseline levels. 398 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD Treatment Follow-up 400 Subject 1 300 CBC 200 100 23 21 19 17 15 13 11 9 7 5 3 1 0 500 Subject 2 400 Behavioral 300 Consultation 200 100 23 21 19 17 15 13 11 9 7 5 3 1 0 500 Subject 3 400 Behavioral 300 Consultation 200 100 23 21 19 17 15 13 11 9 7 5 3 1 0 400 Subject 4 300 CBC 200 100 23 21 19 17 15 13 11 9 7 5 3 0 1 Blood Glucose Level (mg/dL) Blood Glucose Level (mg/dL) Blood Glucose Level (mg/dL) This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Blood Glucose Level (mg/dL) Baseline Day of Study Figure 1. Changes in blood glucose level by subject across experimental phases. Participant 4. For Participant 4, baseline blood-glucose levels were calculated to be an average of 270.6 mg/dL (range ⫽ 79 – 644), with a treatment average of 154.7 mg/dL (range 57–259). An effect size for change in average blood-glucose levels for Participant 4 was calculated to be 2.27. At follow-up, Participant 4 was calculated to have an average daily bloodglucose level across four days of 218.5 mg/dL. Target Behavior Response Figure 2 Displays Behavioral Data for Each Participant Across Baseline, Treatment, and follow-up Phases. Participant 1. For Participant 1, the target behavior identified during the PII of CBC was to have the Participant take insulin while observed by his mother, the school nurse, or a BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES while allowing his mother, school nurse, or grandparent to observe the event occurred an average of 33.33% (0 –50%, std ⫽ 25.82). Treatment data was calculated to be an average of 94.44% (range ⫽ 50 –100%; std ⫽ 16.17) with a corresponding effect size of 2.37. At follow-up, Participant 1 was calculated to have Treatment Follow-up 100 80 Subject 1 60 CBC 40 20 23 21 19 17 15 13 11 9 7 5 3 1 0 12 10 Subject 2 8 6 4 2 Behavioral Consultation 23 21 19 17 15 13 11 9 7 5 3 1 0 100 80 Subject 3 60 Behavioral Consultation 40 20 23 21 19 17 15 13 11 9 7 5 3 1 0 100 80 Subject 4 60 CBC 40 20 23 21 19 17 15 13 11 9 7 5 3 0 1 Percent Compliance with Checking BS Level and Tak ing Insulin Incidents/Day of Snacking without Permission Percent of Insulin Injections Monitored by Adult Baseline Percent Compliance with Checking BS Level and Tak ing Insulin This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. grandparent. Prior to the study, this was not occurring and was of concern because it was feared that this boy was not taking his insulin correctly. Participant 1’s mother felt that there were times when he did not take insulin or took only a portion of the prescribed dosage. For Participant 1, baseline data for taking insulin 399 Day of Study Figure 2. Target behavior response by subject across experimental phases. LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD a 75% average rate of compliance (std ⫽ 28.9). At follow-up, the school year was completed and data from the school nurse was not possible to obtain because she did not have daily contact with Participant 1 during the day to observe insulin injections. Participant 2. For Participant 2, the identified target behavior consisted of asking permission from an adult before snacking on foods, particularly in the afternoon after school. Prior to the study, Participant 2 was eating food from the refrigerator or cupboards without permission. This behavior was contributing to his hyperglycemia based upon information gathered during the Behavioral Consultation PII. For Participant 2, baseline data was calculated to be 7.37 (range ⫽ 4 –11; std ⫽ 2.56) episodes per day of taking food from the refrigerator or cupboards without asking permission from his mother. Treatment data was calculated to be an average of 1.33 episodes per day (range ⫽ 0 –3; std ⫽ 1.0) with an effect size of 2.36. At follow-up, Participant 2 was calculated to be taking food without permission 3.25 (std ⫽ 1.71) times per day. Participant 3. For Participant 3, the target behavior consisted of checking blood glucose levels four times per day and taking insulin appropriately. Prior to the study, Participant 3’s mother was concerned that her daughter was inconsistently checking her blood glucose levels and inappropriately administering her insulin. For Participant 3, baseline data was calculated to be 64.29% compliance with checking blood sugar levels and taking insulin (range ⫽ 0 –100; std ⫽ 28.95). Treatment data was calculated to be an average of 93.75% compliance with treatment (range ⫽ 75–100; std ⫽ 11.57) with an effect size of 1.02. At follow-up, Participant 3 was calculated to be complying with treatment at a reduced rate of 50% (std ⫽ 45.6). Participant 4. For Participant 4, the identified target behavior consisted of checking blood glucose levels four times per day and taking insulin appropriately. Prior to the study, Participant 4 was inconsistently checking his blood glucose levels. He was also suspected of lying about his blood glucose levels so he would not have to take insulin. For Participant 4, baseline data was calculated to be 48.21% compliance with checking blood sugar levels and taking insulin (range ⫽ 0 –75; std ⫽ 18.25). Treatment data was calculated to be an average of 86.11% compliance with treatment (range ⫽ 50 –100%; std ⫽ 18.16) with an effect size of 2.08. At follow-up, Participant 4 was calculated to be complying with treatment at a reduced rate of 56.25% (std ⫽ 23.94). Comparing BC to CBC To determine the level of effectiveness for the two consultation models used, effect size (ES) was calculated for each using the Busk and Serlin (1992) “no assumptions” approach. ES was calculated for both CBC and BC cases to determine the strength of change related to blood glucose levels and noncompliance behaviors. Figure 3 shows the ES calculated for each 2.5 2 Effect Size This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 400 1.5 Target Behaviors 1 Blood Glucose Levels 0.5 0 CBC Behavioral Consultation Figure 3. Differences in effect size calculations for CBC and BC related to changes in blood glucose level and target behavior response. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES type of consultation related to changes in blood glucose levels and target behaviors. ES calculations were determined based upon changes from baseline to the end of treatment and exclude follow-up data. Reviewing the data, all participants made positive gains during intervention. The average ES for target behaviors was higher than the average ES for blood glucose levels. In Figure 3, CBC was calculated to have average effect sizes of 1.75 and 2.22 for changes in blood glucose level and target behaviors from baseline to treatment, respectively. Behavioral Consultation was found to have average effect sizes of 0.88 and 1.69 for changes in blood glucose level and target behaviors, respectively. The average ES for CBC and BC were found to be 1.99 and 1.29, respectively. This data, while interesting, is limited to only four subjects and does not by itself support the utility of one approach over another. Consumer Satisfaction Consumer satisfaction measures were collected in a variety of ways as part of this research. Consultees were asked to complete the Behavioral Intervention Rating Scale—Revised (BIRS-R) to assess the acceptability of using mystery motivator as the primary intervention and to assess the acceptability of the consultation model used. Children in the study were asked to complete the Children’s Intervention Rating Profile (CIRP) to assess how agreeable they were to the use of mystery motivator as the intervention in the study. Acceptability of mystery motivator. Consultees rated their acceptability of mystery motivator by completing the Behavioral Intervention Rating Scale (BIRS-R). Items on the BIRS-R are rated on a 6-point Likert scale, with total possible scores ranging from 24 to 144. High scores on the BIRS-R reflect a high degree of acceptability. Average acceptability responses for parent and school consultees were calculated to be 124 and 133, respectively. Individual item means for the parent and school consultee responses to the BIRS-R were calculated to be 5.17 and 5.54, respectively. These results suggest that both parent and school nurse consultees believed that mystery motivator was, overall, highly acceptable as an intervention. Children in the study rated their acceptability 401 of mystery motivator with the CIRP. Items on the CIRP are rated on a 5-point Likert scale, with low scores indicating high acceptability. The individual item mean for all participants was 2.72. This suggests a rather neutral opinion as to their acceptability of mystery motivator in treating identified behaviors related to hyperglycemia. Consultee acceptability of BC and CBC. Consultees also rated their acceptability of the type of consultation used (either BC or CBC) by completing a revised format of the Behavioral Intervention Rating Scale (BIRS-R). The BIRS-R is rated on a 6-point Likert scale, with total possible scores ranging from 24 to 144. High scores on the BIRS-R reflect a high degree of acceptability. Parent acceptability of the consultation procedures resulted in an overall mean of 119.25 with an individual item mean of 4.97. School consultee acceptability of the consultation procedures resulted in an overall mean of 129.0 with an individual item mean of 5.38. Acceptability ratings by all consultees for CBC were found to be 126.5 (individual item mean of 5.27). Acceptability ratings by parent consultees for Behavioral Consultation were found to be 114.5 (individual item mean of 4.77). Treatment Integrity Intervention integrity. Treatment plan worksheets were given to each of the parent consultees and completed on a daily basis to determine the reported percent compliance in carrying out the mystery motivator intervention in the proper manner. All parents completed the forms for each day of intervention. Across all four parent consultees, a 96% compliance with the outlined steps of mystery motivator was calculated. Thus, parents reported a high level of compliance in following the steps of mystery motivator. Consultation integrity. All consultation sessions were conducted using a standardized set of questions adapted from Sheridan et al. (1996) for CBC and from Kratochwill and Bergan (1990) for BC. The standard formats were used to help ensure that both BC and CBC procedures used were consistent with each model’s original intent and standards. After each session of consultation, the consultant quantitatively evaluated treatment integrity using behavioral checklists. Checklists were This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 402 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD used as a self-monitoring technique to determine the percentage of compliance in following the intended structure for each type of consultation. Results for the problem identification, problem analysis, and treatment evaluation phases of BC and CBC treatment were determined to be 96, 98.5, and 100%, respectively. Based upon this self-monitoring technique, treatment integrity for each of the three phases of either BC or CBC were high and suggest consistent adherence to each consultation approaches’ intended formats. No significant differences were found between treatment integrity data for BC and CBC (98.3% vs. 98%). Social Validity Goal attainment scaling procedures provided a means of assessing parent and school consultee perceptions of case outcomes. The scale requires parents and school nurses to report the degree of goal completion, with a possible response range of ⫺2 (goal not at all met) to ⫹ 2 (goal fully met). In all cases for this research, parents and school nurses reported that either the goal was fully met (⫹2) or partially met (⫹1) for all cases at the time that consultation was discontinued. Parent consultees reported for 75% of the cases (N ⫽ 3) that the consultation goal was fully met and for 25% (N ⫽ 1) that the goal was partially met. Of the two school nurse consultees, one rated the goal as fully met and the other rated the goal as partially met. The BIRS-R effectiveness factor (7 items) was also utilized as a way to summarize social validity data. The BIRS-R effectiveness factor as rated by both parent and school consultees yielded a mean of 5.21 across all participants. Thus, consultee ratings suggest that for the mystery motivator intervention and consultation models used in the study, both were rated as highly effective. Discussion This study assessed a specific behavioral intervention (mystery motivator) in conjunction with the use of CBC and BC strategies to alleviate difficulties children face due to Type I diabetes. There is no previous research that utilizes specific behavioral modification strategies based on intermittent reinforcement to lead to better control of blood glucose levels for children with T1 diabetes. Likewise, no study previously used either CBC or BC to assist in implementing a behavioral strategy focusing on children with diabetes. Results show that the behavioral intervention implemented within the context of either BC or CBC significantly reduced hyperglycemia in children with T I diabetes. All four participants in this study improved from baseline levels reflecting uncontrolled blood sugar levels to more normalized levels during treatment. The average overall reduction in blood glucose measured daily by participants was 115.3 mg/dL. In other words, on an average day during treatment, participants were experiencing markedly less hyperglycemia than they did at baseline. It is important to note that all participants made positive gains during intervention. While the average ES for target behaviors was higher than the average ES for blood glucose levels, this may be attributed to the fact that even when target behaviors improved, not all variables affecting blood glucose are impacted and therefore do not result in equally large improvements or gains. Follow-up results obtained 1 month after treatment suggest mild to moderate maintenance of improvement across time. Participants 1, 2, and 4 all continued to achieve better blood glucose measures compared to their baseline levels by an average of 83.8 mg/dL. Only Participant 3 declined significantly from treatment gains made when follow-up data was gathered. Participant 3 went from baseline blood glucose levels of 255.07 mg/dL to 267.69 mg/dL at follow-up. This suggests that Participant 3 reverted to a very similar pattern of treatment noncompliance after consultation was discontinued. Participants 1, 2, and 4 continued to do much better in controlling blood glucose levels compared to baseline rates. The reduction of hyperglycemia demonstrated in this research has a potential for directly reducing the risk of both short and long term medical complications for children with diabetes. Secondary medical complications such as tissue damage leading to retinopathy, neuropathy, and nephropathy as well as diabetic ketoacidosis may have a reduced occurrence due to better blood glucose control. Past research indicates that keeping blood sugar levels This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES as close to normal also slows the onset and progression of eye, kidney, and nerve diseases caused by diabetes (Wysocki, 1999). Results also show that mystery motivator, combined with either variation of consultation, significantly decreased noncompliance related target behaviors associated with each participant’s hyperglycemia. For instance, at baseline Participant 1 was rated as allowing his mother to supervise injections of insulin only 33% of the time whereas during the treatment phase, this Participant became much more motivated to do so. In the treatment phase, Participant 1 allowed an adult to supervise injections of insulin 94% of the time. This represents a particularly important gain in light of health risks and the fact that this participant was not taking insulin properly prior to treatment or indicating that he was taking insulin when in fact he was not. Participant 1’s ability to dramatically improve compliance with adult supervision of injections correlates significantly with gains in improved blood glucose level across the same time. Participant 1 may have demonstrated even better gains if the school year continued and allowed follow-up data to be collected there. Since all monitoring responsibilities were shifted to the home environment, slightly lower levels of compliance may have been more likely. In addition, gains were even more impressive in light of the student’s history of having behavioral problems warranting special education classification. Participant 2 also made positive gains related to the behavior targeted during consultation. Participant 2’s gains were acquired by complying with the rule targeted during consultation that permission must be gained by an adult prior to snacking. During baseline, Participant 2 had been snacking without permission on foods that played a role in elevating his blood sugar levels. Prior to consultation, it was felt that after school hours were most problematic for this participant. With increased supervision and use of mystery motivator, snacking levels from baseline to treatment were reduced six fold, and corresponded directly to improved blood glucose levels. Interestingly, Participant 2’s mother also restricted food availability by not buying some items that she had in the past. Thus, during treatment, Participant 2’s snacking not only was reduced by the mystery motivator intervention, but also by food being removed from home. This may 403 have reduced opportunities for Participant 2 to snack at home without parental permission. Participant 2’s mother also seemed to restrict typical activities outside of the family home after baseline. This may have allowed for better control over inappropriate snacking behaviors. Participant 3, the only female child in the study also improved in both target behaviors of checking blood glucose levels and taking insulin. This again was associated with better blood glucose levels during the treatment phase. At follow-up, Participant 3’s compliance with monitoring blood glucose and taking insulin injections had faltered slightly when compared to gains made during treatment. Participant 3’s mother indicated that her use of the mystery motivator had declined, likely leading to Participant 3’s decline in performance at follow-up. Unfortunately, this also impacted this Participant’s blood glucose levels, which reverted to baseline levels when follow-up data was gathered. Participant 4 who participated in CBC also improved in monitoring blood glucose levels and taking insulin compared to pretreatment levels for these targeted behaviors. Participant 4 responded very well to the procedures and may have benefited most from the positive nature of the intervention. Prior to the study, Participant 4 was described as more anxious and possibly depressed about managing his diabetes. Mystery motivator, by reinforcing positive treatment behaviors, appeared to boost the confidence of this participant, in turn improving his confidence level around peers. Improvements were also anecdotally noted for participants related to psychological functioning. Participant 1’s self-reports of depression were noticeably reduced and confirmed by parental report. Participant 2 self-reported significant improvements in anxiety reduction, attention, and social relationships. Participants 3 and 4 also self-reported improvement in selfesteem. Some of the patterns of psychological functioning found in the literature for children with TI diabetes were also observed among participants and their families in this study. For instance, Participant 1 experienced significant levels of depression often found in children with diabetes (Kovacs et al., 1997). All participants acknowledged risk factors related to feelings of poor social competence, interpersonal problems, or poor relations with parents. These findings are consistent with typically reported dif- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 404 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD ficulties children with diabetes experience socially, emotionally, or in family relationships. However, as Kovacs et al., (1992) points out, variables such as social competence, selfesteem, and aspects of family functioning do not predict noncompliance with treatment. For example in this study, Participant 3, with the fewest identified psychological risk factors, was later found to be the least compliant with treatment at follow-up. Adverse psychosocial factors such as lowincome level, single parent family situations, and so forth may have also contributed to poor psychological adjustment for Participants 1 and 2. Participants 1 and 2 both came from single parent, low income families compared with Participants 3 and 4, each coming from two-parent families with middle income levels. When comparing BC and CBC, it was found that CBC was slightly more effective in reducing levels of hyperglycemia and improving treatment compliance. This limited data on only four subjects, while interesting, does not by itself support the utility of one approach over another, although results are consistent with previous research (Sheridan, et al., 1990). Although CBC was found to be slightly more effective, either method of consultation appears effective because of the positive gains made within each model. BC may be particularly useful when it is determined that the identified problem related to uncontrolled blood glucose levels is isolated to only one environment. CBC on the other hand may be most useful when it is felt that a child and his or her family require assistance from a team of individuals across a variety of settings because the identified problem is not specific to one environment. CBC in particular demands an interplay between team members in treating problems related to diabetes. As such CBC appears effective in going beyond traditional child-centered strategies when implementing effective interventions for helping children by incorporating strategies that consider the organizational and systems variables contributing to the identified problem of diabetes. School psychologists and other school based mental health professionals may find that a careful analysis of setting factors in the Problem Identification Interview will identify the most appropriate strategy. Based upon the feedback from treatment acceptability measures, parent and school consultees reported the mystery motivator intervention and either type of consultation as highly accept- able. This was not the case when child participants were asked to give feedback on the use of the mystery motivator. Their level of endorsement was more modest or reflected a neutral appraisal. Goal Attainment Scaling measures also provide a high degree of social validity for changes related to decreases in blood glucose levels and improvement in treatment compliance. This suggests that approaches utilized in this study are likely to find a receptive audience among parents and caregivers who work with children who struggle with the medical management of diabetes. Additionally, the procedures employed in this study are very compatible with actual practice guidelines in both school and medical clinic settings. Despite their improvements, more substantive feedback is needed from child participants to understand how they experience the motivational aspects of the intervention. Limitations of the Study Results of this research suggest that an intervention such as the mystery motivator, applied in the context of either BC or CBC is an effective management strategy for children with diabetes who have difficulty maintaining adherence to a treatment protocol. However, only four participants were studied, making it difficult to generalize results to a larger population. It is not clear how the addition of an additional consultee (school nurse) in the CBC condition may be more effective than BC due to the fact that only two participants were involved in each consultation condition during the study. The study involved school nurses rather than teachers as consultees. A single school psychologist served as a consultant in both treatment conditions and for all consultees and clients, thus was not blind to the model used. The study was also managed from a medical/clinic setting by the first author, although some meetings took place in a school setting. While it is likely that the consultation strategies and specific interventions could have been implemented completely in a school setting, there may be certain expectations or other aspects of the clinical treatment setting that could not be replicated in a school setting. Further studies originating out of a school setting and replicating the procedures used in this study with similarly trained consul- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. BEHAVIORAL MANAGEMENT OF TYPE 1 DIABETES tant school psychologists would address these issues. Another limitation of study was the relatively short intervention period coupled with a brief fading technique used with all participants. Based upon follow-up results, it is possible to assume that better generalization across time for improved treatment compliance and blood glucose levels could have been achieved by giving participants more opportunities for reinforcement. Continued consultee contact with the consultant across additional sessions may be important to assure that intervention procedures are carried out consistently and without fading too quickly or extinguishing the program completely. This may have been a factor in the case of Participant 3, who at follow-up reverted back to baseline behavior (failure to check blood glucose and/or take insulin as prescribed by her treatment team). Interestingly, school consultees rated consultation as slightly more acceptable than parents. This may be due to the fact that school consultees were required to implement less of the actual intervention than parents and this led to higher reported levels of acceptability for both consultation strategies. Implications for Future Research This study provides support for a specific intervention (mystery motivator) used in concert with behaviorally based consultation strategies to treat behaviors associated with poor adherence to a treatment protocol for children with TI diabetes. Future research should determine the efficacy of these strategies across a greater length of time to determine the impact beyond a 2-month period. Chronic conditions such as TI diabetes may require a more continuous stream of intervention support or periodic “tune-ups” to insure that interventions are maintained or periodically adjusted. Future research is needed to determine how variables such as age, gender, and length of diagnosis impact the effectiveness of these strategies. Additionally, further research may provide some understanding as to how psychological and family variables influence treatment compliance in the context of these strategies. 405 References Bergan, J. R., & Kratochwill, T. R. (1990). Behavioral consultation and therapy. New York, Plenum Press. Betteridge, J. (2000). Diabetes: Current perspectives. London: Martin Dunitz. Brown, R. T., & DuPaul, G. W. (1999). Introduction to the Mini-series: Promoting school success in children with chronic medical conditions. School Psychology Review, 28, 175–181. Busk, P. L., & Serlin, R. C. (1992). Meta-analysis for single-case research. In T. R. Kratochwill & J. R. Levin (Eds.), Single case research design and analysis: Applications in psychology and education (pp. 187–212). Hillsdale, NJ: Erlbaum. Chase, H. P. (1985). Avoiding the short and long term complications of juvenile diabetes. Pediatrics in Review, 7, 140 –149. Daley, K. B., Woderich, D. L., & Hasan, K. (2004). Classroom based and computerized measures of attention among children with Type 1 Diabetes mellitus: The influence of glucose instability. Paper presented at annual convention of the American Psychological Association, Honolulu. Delamater, A. M. (2000). Quality of life in youths with diabetes. Diabetes Spectrum, 13, 42. DeMartini-Scully, D. D., Bray, M. A., & Kehle, T. J. (2000). A packaged intervention to reduce disruptive behaviors in general education students. Psychology in the Schools, 37, 149 –156. Elliott, S. N., & Von Brock Treuting, M. (1991). The Behavior Intervention Rating Scale: Development and validation of a pretreatment acceptability and effectiveness measure. Journal of School Psychology, 29, 43–51. Foster, D. W. (1991). Diabetes Mellitus. In J. D. Wilson, E. Braunwald, K. J. Isselbacher, R. G. Petersdorf, J. B. Martin, A. S. Fauci, & R. K. Root (Eds.), Harrison’s Principles of Internal Medicine (12th ed.) (pp. 1739 –1759). New York: McGrawHill. Galloway, J., & Sheridan, S. M. (1994). Implementing scientific practices through case studies: Examples of using home-school interventions and consultation. Journal of School Psychology, 32, 385– 413. Jenson, W. R., Rhode, G., & Reavis, H. K. (1995). The Tough Kid Tool Box. Longmont, CO: Sopris West. Karam, J. H. (1996). Diabetes mellitus and hypoglycemia. In L. Tierney, Jr., S. McPhee, & M. apadakis, (Eds.), Current Medical Diagnosis and Treatment (35th ed.), (pp. 1030 –1068). Stamford: Appleton & Lange. Kiresuk, T., Smith, A., & Cardillo, J. (1994). Goal attainment scaling: Applications, theory, and easurement. Hillsdale, NJ: Erlbaum. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 406 LASECKI, OLYMPIA, CLARK, JENSON, AND TUESDAY HEATHFIELD Kovacs, M., Goldston, D., Obrosky, D. S., & Bonar, L. K. (1997). Psychiatric disorders in youths with IDDM: Rates and risk factors. Diabetes Care, 20, 36 – 44. Kovacs, M., Goldston, D., Obrosky, D. S., & Iyengar, S. (1992). Prevalence and predictors of ervasive noncompliance with medical treatment among youths with insulin-dependent diabetes mellitus. Journal of the American Academy of Child and Adolescent Psychiatry, 31, 1112–1119. Kovacs, M., Mukerji, P., Iyengar, S., & Drash, A. (1996). Psychiatric disorder and metabolic control among youths with IDDM. A longitudinal study. Diabetes Care, 19, 318 –323. Kovacs, M., Obrosky, D. S., Goldston, D., & Drash, A. (1997). Major depressive disorder in youths with IDDM. A controlled prospective study of course and outcome. Diabetes Care, 20, 45–51. Kratochwill, T. R., & Bergan, J. R. (1990). Behavioral consultation in applied settings: An individual guide. New York: Plenum Press. Kratochwill, T. R., Elliott, S. N., & Busse, R. T. (1995). Behavior consultation: A five year evaluation of consultant and client outcomes. School Psychology Quarterly, 10, 87–117. LeBlanc, D. M. (1999). Mystery motivator versus reward menu: An investigation of the effects of home-based reinforcement delivery systems used with home-school notes on disruptive/disengaged classroom behavior, Unpublished doctoral dissertation, U Southern Mississippi, Hattiesburgh. Medway, F. J., & Updyke, J. F. (1985). Metaanalysis of consultation outcome studies. merican Journal of Community Psychology, 13, 489 –505. Moore, L. A., Waguespack, A., Wickstrom, Witt, J., & Gaydos, G. A. (1994). Mystery motivator: An effective and time efficient intervention, School Psychology Review, 23, 106 –118. Natasi, B. (2000). School psychologists as health-care providers in the 21st century: Conceptual framework, professional identity, and professional practice., School Psychology Review, 29, 540 –554. One Touch FastTake. Lifescan Systems, Inc (2006, May 15). Retrieved May 15, 2006, from http:// lifescan.com/productmeters/select/ Robinson, K. E., & Sheridan, S. M. (2000). Using the mystery motivator to improve child bedtime compliance. Child & Family Behavior Therapy, 22, 29 – 49. Sheridan, S. M., & Colton, D. L. (1994). Conjoint behavioral consultation: A review and case study. Journal of Educational and Psychological Consultation, 5, 211–228. Sheridan, S. M., Colton, D. L., Fenstermacher, K., Lasecki, K., & Wilson, K. (1996, August). Efficacy of conjoint behavioral consultation as a vehicle for inclusion. Paper presented at the annual convention of the American Psychological Association, Toronto. Sheridan, S. M., & Kratochwill, T. R. (1992). Behavioral parent-teacher consultation: Conceptual and research considerations. Journal of School Psychology, 30, 117–139. Sheridan, S. M., Kratochwill, T. R., & Bergan, J. R. (1996). Conjoint behavioral consultation: A procedural guide. New York: Plenum Press. Sheridan, S. M., Kratochwill, T. R., & Elliott, S. N. (1990). Behavioral consultation with parents and teachers: Delivering treatment for socially withdrawn children at home and school. School Psychology Review, 19, 33–52. Sheridan, S. M., Welch, M., & Orme, S. (1996). Is consultation effective? Remedial and Special Education, 17, 341–354. Sheridan, S. M.: Eagle, J. W., Cowan, R. J., & Mickelson, W. (2001). The Effects of Conjoint Behavioral Consultation Results of a 4-Year Investigation. Journal of School Psychology, 39, 361–385. Sibley, S. (1986). A meta-analysis of school consultation research. Unpublished doctoral dissertation, TX Woman’s University, Denton. Smith, G. A., & Baum, J. D. (1980). Emotional, behavioral, and educational disorders in diabetic children. Archives of Disease in Childhood, 55, 371–375. Sperling, M. A. (1990). Diabetes Mellitus. In S. A. Kaplan (Ed.), Clinical pediatric endocrinology (pp. 127–164). Philadelphia: W. B. Saunders. Strawhacker, M. T. (2001). Multidisciplinary teaming to promote effective management of Type 1 diabetes for adolescents. Journal of School Health, 71, 213–217. Tawney, J. W., & Gast, D. L. (1984). Single Participant research in special education. Columbus, OH: Merrill. Valum, J. L. (1996). Student managed study skills teams: Academic survival for adolescents at risk of school, Unpublished doctoral dissertation, University of Utah, Salt Lake City. Weiner, R. K., Sheridan, S. M., & Jenson, W. R. (1998). The effects of conjoint behavioral consultation and a structured homework program on math completion and accuracy in junior high students. School Psychology Quarterly, 13, 281–309. Witt, J. C., & Elliott, S. N. (1985). Acceptability of classroom intervention strategies. In T. R. Kratochwill (Ed.), Advances in school psychology (Vol. 4; pp. 251–288). Hillsdale, NJ: Erlbaum. Wysocki, T. (1999, June). Early information on an ongoing study of intensive therapy and Children: Advances in behavioral medicine. Paper presented at the meeting of the American Diabetes Association, San Diego, CA. Wysocki, T. (2006). Behavioral assessment and intervention in pediatric diabetes. Behavior Modification, 30, 72–92. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Psychological Bulletin 2009, Vol. 135, No. 1, 121–141 © 2009 American Psychological Association 0033-2909/09/$12.00 DOI: 10.1037/a0014210 What Works in Coping With HIV? A Meta-Analysis With Implications for Coping With Serious Illness Judith Tedlie Moskowitz and Jen R. Hult Cori Bussolari University of California, San Francisco University of San Francisco Michael Acree University of California, San Francisco Knowledge of effective ways of coping with HIV is critical to help individuals with HIV maintain the best possible psychological and physical well-being. The purpose of the present article is to determine, through meta-analysis, the strength of the evidence regarding 2 questions: (a) Which types of coping are related to psychological and physical well-being among people with HIV? and (b) Do contextual (pre–post introduction of highly active antiretroviral therapies [HAART]; time since diagnosis), measurement (HIV-related event vs. generic prompts for coping measurement), or individual (gender) variables affect the extent to which coping is related to physical and psychological well-being? The authors’ analysis demonstrates that Direct Action and Positive Reappraisal were consistently associated with better outcomes in people coping with HIV across affective, health behavior, and physical health categories. In contrast, disengagement forms of coping, such as Behavioral Disengagement and Use of Alcohol or Drugs to Cope, were consistently associated with poorer outcomes. The findings also indicate that in some cases, coping effectiveness was dependent on contextual factors, including time since diagnosis and the advent of HAART. Keywords: coping, positive affect, negative affect, health behaviors, physical health the question of what works in coping with HIV. The larger literature on coping with serious illness, which suffers from the same heterogeneity, provides little guidance. As in the general coping literature, the majority of research on coping with HIV has been based on a transactional theory of stress and coping (Lazarus & Folkman, 1984), in which coping is viewed as part of a process that unfolds in response to the demands of the situation and includes cognitive appraisals of the significance of the event and associated emotional responses. Coping is defined as efforts to deal with demands taxing or exceeding the resources of the person (Lazarus & Folkman, 1984). In other words, coping is a cognitive or behavioral response to something appraised as stressful. It is the subjective appraisal, more than the objective characteristics of the situation, that determines how stressful the event is for an individual and subsequently drives the coping response. Many aspects of HIV are potentially stressful (Moskowitz & Wrubel, 2005). For example, fear of death, the need to adhere to complex medication regimens, side-effects of the treatments, interactions with a complex medical system, symptoms associated with disease progression, financial difficulties, stigma, and the need to incorporate a new identity as someone with a serious illness are all potential stressors associated with being HIVpositive. It is important to note that these stressors are not necessarily unique to HIV. People living with other illnesses—such as cancer, diabetes, and arthritis—are subject to many of the same concerns. Thus, although our focus is on HIV, the findings have potential application to coping with other illnesses, and our methods could serve as a model for meta-analyses of coping with other serious illnesses. Twenty-five years after the identification of the first AIDS cases, it is estimated that more than one million people in the United States are living with HIV, the virus that causes AIDS, and 40,000 people become infected each year (Centers for Disease Control and Prevention, 2007). Although great strides have been made toward extending length of life and improving quality of life for people living with the virus, for many people HIV remains a stressful and demanding illness (Remien et al., 2006). Stress has deleterious effects on mental health and can hasten the progression of HIV disease (Bottonari, Roberts, Ciesla, & Hewitt, 2005; Remor, Penedo, Shen, & Schneiderman, 2007; Sledjeski, Delahanty, & Bogart, 2005; Weaver et al., 2005). Therefore, knowledge about effective ways of coping with the stress of HIV is key to helping individuals with HIV maintain the best possible psychological and physical well-being. The literature on coping with HIV is large and heterogeneous with respect to measures of coping, outcomes, and conclusions. There is little consistency across studies regarding types of coping and types of outcomes that are assessed and reported. Even a careful reading of the literature does not provide a clear answer to Judith Tedlie Moskowitz, Jen R. Hult, and Michael Acree, Osher Center for Integrative Medicine, Department of Medicine, University of California, San Francisco; Cori Bussolari, Department of Counseling Psychology, University of San Francisco. Correspondence concerning this article should be addressed to Judith Tedlie Moskowitz, Osher Center for Integrative Medicine, University of California, San Francisco, 1701 Divisadero, Suite 150, San Francisco, CA 94115. E-mail: moskj@ocim.ucsf.edu 121 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. MOSKOWITZ, HULT, BUSSOLARI, AND ACREE 122 The purpose of the present article is to critically review the literature on coping with HIV and to determine, through metaanalysis, the strength of the evidence regarding two questions: (a) Which types of coping are related to psychological and physical well-being among people with HIV? and (b) Do contextual (pre– post introduction of highly active antiretroviral therapies [HAART]; time since diagnosis), measurement (HIV-related event vs. generic prompts for coping measurement), or individual (gender) variables affect the extent to which coping is related to physical and psychological well-being? Coping One of the biggest challenges to a comprehensive summary of the coping literature is the diversity of measures and labels for the various ways of responding to stress (Schwarzer & Schwarzer, 1996). Conceivably, there are a countless number of ways of coping, from making a plan of action or fantasizing about an ideal outcome, to reminding oneself of the good that will come out of the situation or pretending the stressful event did not happen. In their comprehensive review of the literature on the structure of coping, Skinner, Edge, Altman, and Sherwood (2003) listed 400 coping responses. Most coping measures are based on Folkman and Lazarus’s (1980, 1988) Ways of Coping Questionnaire. The Ways of Coping Questionnaire consists of 50 items that generally factor into eight subscales: Confrontive, Distancing, Self-Controlling, Seeking Social Support, Self-Blame, Escape/Avoidance, Planful ProblemSolving, and Positive Reappraisal. Another frequently used measures is the COPE (Carver, Scheier, & Weintraub, 1989), which is a 52-item questionnaire designed to measure 14 types of coping. The COPE subscales are Active Coping, Planning, Suppression of Competing Activities, Restraint, Seeking Social Support for Instrumental Reasons, Seeking Social Support for Emotional Reasons, Positive Reinterpretation and Growth, Acceptance, Turning to Religion, Focus on and Venting of Emotions, Denial, Behavioral Disengagement, and Mental Disengagement. The major differences among these and other commonly used coping scales lie primarily in the number and specificity of types of coping assessed. For example, the Ways of Coping assesses more general seeking social support, whereas the COPE assesses two more specific types of seeking support: Seeking Social Support for Emotional Reasons and Seeking Social Support for Instrumental Reasons. The COPE also includes subscales that are not tapped by the Ways of Coping, such as Suppression of Competing Activities, and Focus on and Venting of Emotions. In the HIV literature, although many studies use the Ways of Coping or the COPE (e.g., Ashton et al., 2005; Blaney et al., 2004; Carels, Baucom, Leone, & Rigney, 1998; Deschamps et al., 2004; Folkman, Chesney, Pollack, & Coates, 1993), others have developed scales specific to coping with HIV (e.g., Moneyham et al., 1998; Murphy, Rotheram-Borus, & Marelich, 2003). Even though these scales were developed specifically to measure coping with HIV, they have subscales that are similar to the Ways of Coping and the COPE. For example, the scale developed by Murphy et al. (2003) has seven factors: Positive Action, Passive Problem Solving, Self-Destructive Escape, Social Support, Spiritual Hope, Depression/Withdrawal, and Nondisclosure/Problem Avoidance. Moneyham et al.’s (1998) scale includes Seeking Peer Support, Living Positively, Managing HIV Disease, Seeking Support of Family/Friends, Isolation/Withdrawal, Spiritual Activities, Denial/ Avoidance, and Seeking Information. Higher Order Coping Classifications As evidenced by the above review, working with even a half dozen coping subscales can become unwieldy for reporting and analysis, so researchers often group the subscales, either empirically or theoretically, into factors. These second or higher order coping distinctions include Problem-Focused/Emotion-Focused (Folkman & Lazarus, 1980), Approach/Avoidance (Roth & Cohen, 1986), Active/Passive (Billings & Moos, 1981; Mercado, Carroll, Cassidy, & Cote, 2000; Moneyham et al., 1998), Engagement/ Disengagement (Tobin, Holroyd, Reynolds, & Wigal, 1989), and Cognitive/Behavioral (Billings & Moos, 1981), among others (see Skinner et al., 2003, for a review and critique of approaches for classifying coping). In the HIV literature, a few studies rely on the EmotionFocused/Problem-Focused dichotomy (Gore-Felton et al., 2002; Pakenham & Rinaldis, 2001); a larger number use an Approach/ Avoidance distinction (Avants, Warburton, & Margolin, 2001; Halkitis, Kutnick, & Slater, 2005; Heckman et al., 2004; Schmitz & Crystal, 2000; Semple, Patterson, & Grant, 2000) or a variation on this distinction: Avoidant/Proactive (Nicholson & Long, 1990), Avoidant/Adaptive (Simoni & Ng, 2000), and Maladaptive/ Adaptive (Safren, Radomsky, Otto, & Salomon, 2002). For example, Heckman et al. (2004) conducted a principal components analysis of the items from the Ways of Coping Checklist (Folkman & Lazarus, 1980) in their sample of 329 men and women living with HIV. They found two factors and labeled them Active (“I made a plan of action and followed it”) and Avoidant (“I refused to believe that it had happened”). Leslie, Stein, and RotheramBorus (2002) selected two factors from the seven measured by the Coping with Illness Questionnaire (Murphy et al., 2003) to assess Active and Passive coping in their sample of parents living with HIV or AIDS. In summary, a number of general and disease-specific coping measures are available, most of which are based on the Ways of Coping Questionnaire. The extensive variability in the specific types of coping that are measured and reported across studies presents a challenge to the synthesis of the literature on coping effectiveness. Meta-Analyses of Coping Previous meta-analytic studies of coping have addressed the challenge of wide variation in types of coping either by focusing on a few select measures (e.g., Jordan & Revenson, 1999; Penley, Tomaka, & Wiebe, 2002) or grouping the subscales into established categories based on the descriptions of types of coping provided in the articles (Aldridge & Roesch, 2007; Ano & Vasconcelles, 2005; Connor-Smith & Flachsbart, 2007; Duangdao & Roesch, 2008; Littleton, Horsley, John, & Nelson, 2007; Mullen & Suls, 1982; Nes & Segerstrom, 2006; Roesch & Weiner, 2001; Suls & Fletcher, 1985; Tamres, Janicki, & Helgeson, 2002). For example, in a meta-analysis of coping and physical and psychological health, Penley et al. (2002) restricted their analysis to 11 ways of coping measured either with a version of the Ways of This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. COPING WITH HIV Coping (Folkman & Lazarus, 1985; Folkman, Lazarus, DunkelSchetter, DeLongis, & Gruen, 1986) or the Ways of Coping Checklist scales developed by Vitaliano, Russo, Carr, Maiuro, and Becker (1985) based on the Ways of Coping (Folkman & Lazarus, 1980). This approach of restricting the sample to a few select coping measures simplifies the task of combining findings across disparate studies. However, this efficiency may come at the expense of comprehensiveness; some potentially important types of coping are not included on the selected coping measures. In a meta-analysis of coping and personality, Connor-Smith and Flachsbart (2007) classified coping hierarchically with three levels (from broadest to most specific): (a) Engagement versus Broad Disengagement; (b) Primary Control Engagement, Secondary Control Engagement, and Narrow Disengagement; and (c) Specific Coping Strategies (problem solving, social support, emotion regulation, distraction, cognitive restructuring, acceptance, religious coping, avoidance, withdrawal, and wishful thinking). In addition, at the broadest level, they also included Negative Emotion Focused (emotion regulation and uncontrolled expression) and Mixed Emotion Focused coping (a mix of controlled and uncontrolled emotion regulation and expression strategies). The results supported the value of distinguishing among specific lower order coping strategies because they were differentially related to facets of personality; these findings would have been obscured had the authors restricted their analysis to the higher levels of the coping hierarchy. Meta-Analyses of Coping With Serious Illness Meta-analyses of studies of coping with serious illness and other medical procedures can inform the present work on coping with HIV and provide a metric against which the effects sizes can be compared. Roesch and colleagues conducted meta-analyses of coping in people with diabetes (Duangdao & Roesch, 2008), men with prostate cancer (Roesch et al., 2005), and people who had a chronic illness or were undergoing a medical procedure (Roesch & Weiner, 2001). Roesch and Weiner (2001) classified coping three ways: (a) Approach/Avoidance; (b) Cognitive Approach, Cognitive Avoidance, Behavioral Approach, and Behavioral Avoidance; and (c) Problem-Focused/Emotion-Focused. The authors noted that although they chose this classification to deal with the diversity of coping measured in the primary studies, the drawback is that the unique effectiveness of lower order forms of coping cannot be determined. Results indicated that Approach, Cognitive Approach, Behavioral Approach, Problem-Focused and EmotionFocused forms of coping were associated with better psychological adjustment, and Avoidance and Cognitive Avoidance were associated with poorer psychological adjustment. According to Cohen’s (1988) criteria, effect sizes were generally small to medium (ranging from r ⫽ .13 for Behavioral Approach to r ⫽ ⫺.30 for Cognitive Avoidance). Roesch et al. (2005) and Duangdao and Roesch (2008) used an Approach/Avoidance classification and a Problem-Focused/ Emotion-Focused classification. In the sample of men with prostate cancer (Roesch et al., 2005), Approach coping was significantly associated with better overall adjustment (r ⫽ .23), and Avoidance coping was significantly associated with poorer overall adjustment (r ⫽ ⫺.21), both small to medium effects. Among people with diabetes (Duangdao & Roesch, 2008), Approach coping was associated with better overall adjustment (r ⫽ .20), less 123 anxiety (r ⫽ ⫺.16), less depression (r ⫽ ⫺.19), and better glycemic control (r ⫽ .17). Avoidance coping was not related to any of the four outcomes. On the basis of these previous meta-analyses of coping with serious illness, it appears that effects sizes tend to be small and that Approach coping is associated with better outcomes. Avoidance coping is less consistently related to outcomes; however, when there is a significant effect, Avoidance coping tends to be associated with poorer outcomes. These studies did not, however, examine effects of any lower order forms of coping, possibly obscuring significant unique effects of these more specific types of coping. Classification of Coping In the present meta-analysis, we use two approaches to address the problem of coping classification. The first approach uses a lower order classification based on specific subscales from the Ways of Coping and the COPE, and the second uses a higher order, more macroscopic classification that relies on the Approach/ Avoidance distinction. We selected Approach/Avoidance because in the literature on coping with HIV, most of the higher order classifications rely on some form of this dichotomy, and this classification allows us to compare with previous meta-analyses of coping in seriously ill samples that relied on an Approach/ Avoidance classification (Duangdao & Roesch, 2008; Roesch et al., 2005; Roesch & Weiner, 2001). The use of these two approaches allows us to determine the potentially unique effects of lower order coping subscales, but it also addresses the question of whether and under what conditions the more manageable Approach/Avoidance distinction might be preferable. Selection of Outcomes: What Is Effective Coping? Effective coping can be defined a number of ways. Zeidner and Saklofske (1996) listed eight criteria that could be used to define effective coping: resolution of the stressful situation, reduction of physiological reactions, reduction of psychological distress, normative social functioning, return to prestress activities, well-being of self and others affected by the situation, maintaining positive self-esteem, and perceived effectiveness. However, these generic outcomes are unlikely to apply in all situations. According to stress and coping theory, one approach to determining coping effectiveness is to examine the effect of coping on positive and negative outcomes that are relevant and appropriate in the situation under study (Folkman, 1992). In the present review, we focus on the situation of having HIV and examine four sets of outcomes that are most commonly used in the HIV stress and coping literature: positive affect, negative affect, health behaviors, and physical health. As in the more general coping literature, studies of coping with HIV most frequently rely on psychological outcomes, such as depressive mood (e.g., Folkman et al., 1993; Patterson et al., 1995; Penedo et al., 2001; Schmitz & Crystal, 2000), anxiety (Catz, Gore-Felton, & McClure, 2002; Commerford, Orr, Gular, Reznikoff, & O’Dowd, 1994; Kelly et al., 2000), or other negative affective states, such as anger (Weaver et al., 2005) or perceived stress (Koopman et al., 2000), as the primary outcome. Physical health outcomes, such as CD4 (an indicator of strength of the This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 124 MOSKOWITZ, HULT, BUSSOLARI, AND ACREE immune system; Ashton et al., 2005; Patterson et al., 1995), viral load (Avants et al., 2001), and symptoms (Pakenham & Rinaldis, 2001; Phillips, Sowell, Rush, & Murdaugh, 2001), are also often included. Some studies examine health behavior outcomes, such as adherence to antiretroviral medications (Avants et al., 2001; Deschamps et al., 2004; Ironson et al., 2005), safer sexual behaviors (Gore-Felton et al., 2002; Semple et al., 2000), and decreased drug use and smoking (Collins et al., 2001). Finally, as in other areas of psychology, researchers in the area of coping with HIV are beginning to include positive affective outcomes, such as life satisfaction and positive mood, as variables separate from depression and other negative affects (Fleishman et al., 2000; Siegel & Schrimshaw, 2005). Contextual, Individual, and Methodological Influences on Coping Effectiveness Dispositional and Situational Coping Measurement As noted above, a central tenet of stress and coping theory is that coping is not inherently adaptive or maladaptive (Lazarus & Fol...
Purchase answer to see full attachment