|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
PROBLEM SOLVING AND THE MANAGE MENT OF OBESITY IN WOMEN FROM UNDERSERVED RURAL SETTINGS By MARY E. MURAWSKI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
Copyright 2007 by Mary E. Murawski
iii ACKNOWLEDGMENTS I would like to thank my mentor, Dr. Michae l Perri, as well as the entire TOURS Project staff, for their con tinual encouragement and assistance throughout the formation and completion of this dissertation.
iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................vi ii 1 INTRODUCTION........................................................................................................1 Overview....................................................................................................................... 1 Obesity in America................................................................................................1 Obesity in Rural Areas..........................................................................................2 The Treatment of Obesity......................................................................................4 The Maintenance Problem in Obesity Treatment..................................................7 Problem Solving as a Method of Coping............................................................10 The Problem-Solving Process.............................................................................11 Problem Solving and Mental Health...................................................................13 Problem Solving and Medical Conditions...........................................................14 Problem-Solving Therapy...................................................................................15 Current Study.......................................................................................................18 2 MATERIALS AND METHODS...............................................................................22 Participants.................................................................................................................22 Measures.....................................................................................................................23 Weight.................................................................................................................23 Social Problem Solvi ng Inventory-Revised........................................................24 Self-Monitoring Records.....................................................................................25 Procedure....................................................................................................................26 Recruitment.........................................................................................................26 Initial Assessment................................................................................................26 Schedule of Assessment......................................................................................27 Phase 1.................................................................................................................27 Phase 2.................................................................................................................29 Analyses......................................................................................................................3 2
v 3 RESULTS...................................................................................................................37 Participant Characteristics..........................................................................................37 Preliminary Analyses..................................................................................................37 Primary Analyses........................................................................................................39 Secondary Analyses....................................................................................................47 4 DISCUSSION.............................................................................................................62 LIST OF REFERENCES...................................................................................................76 BIOGRAPHICAL SKETCH.............................................................................................83
vi LIST OF TABLES Table page 1-1. Overview of the Problem-Solving Process................................................................21 3-1. Baseline Characteristics ( n = 298).............................................................................49 3-2. SPSI-R Scales and Abbreviations..............................................................................50 3-3. Characteristics of Phase 2 Completers by Treatment Condition ( n = 144)...............51 3-4. Weights (kg) of Phase 2 Completers by Treatment Condition ( n = 144)..................52 3-5. Percent Body Weight Change from 6 to 18 Months by Phase 2 Condition ( n = 144)........................................................................................................................... 53 3-6. Cohorts 1-3 SPSI-R Scores Across Phase 1 ( n = 273)..............................................54 3-7. SPSI-S Scores Across Phase 1 by Ca tegory of Percent Weight Loss ( n = 273)......55 3-8. SPSI-S Scores Across Trial by Phase 2 Treatment Condition ( n = 144)..................56 3-9. Phase 2 Completers SPSI-R Scores Across Trial (n = 144)......................................57 3-10. Phase 2 Completers Self-m onitoring by Treatment Condition ( n = 144)..............58
vii LIST OF FIGURES Figure page 3-1. Proportion of Sample with Different Levels of Phase 1 Weight Change..................59 3-2. Percentage of Phase 1 Weight Lo ss Maintained During Months 6-18 as a Function of Extended Care Condition......................................................................60 3-3. Distribution of Base line Problem-Solving Scores.....................................................61
viii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROBLEM SOLVING AND THE MANAGE MENT OF OBESITY IN WOMEN FROM UNDERSERVED RURAL SETTINGS By Mary E. Murawski August 2007 Chair: Michael G. Perri Major Department: Clini cal and Health Psychology The prevalence of obesity in America has increased dramatically over the past few decades, with higher rates observed in rural than in urban areas. Lifestyle interventions focused on behavioral management of eati ng and physical activity patterns have been shown effective in producing cl inically significant weight reductions. However, the maintenance of lost weight continues to be a common problem after treatment has ended. The acquisition and implementation of problem -solving skills have been shown to be helpful in combating weight regain subsequent to participation in a lifestyle intervention. Yet, little information exists describing the relation of problem-solving skills to weight change during the course of lifestyle trea tment for obesity. This study assessed the impact of problem solving on weight loss with in the context of the Treatment of Obesity in Underserved Rural Settings (TOURS) Proj ect. TOURS is a randomized control trial, in which obese women, ages 50-75 years, fr om medically underse rved rural counties participated in 2 phases of an 18-month we ight management intervention. Phase 1
ix consisted of a 6-month intensive lifestyle intervention. Phase 2 entailed a 12-month extended care period with participants randomized to problem-solving based interventions delivered via in-person sessions, by telephone co ntacts, or by ma il (control). The Social Problem Solving Inventory-Revi sed was used to assess problem-solving abilities at baseline, 6 months, and 18 mont hs. The principal hypothesis of this study was that weight change during Phase 2 would be mediated by changes in problem-solving abilities. The results showed a significant bu t small effect for problem-solving skills on weight change, such that that Phase 1 we ight change was related to both baseline problem-solving skills ( R2 = -.02, p < .05) and improvements in problem-solving skills from baseline to 6 months ( R2 = -.05, p < .01). During Phase 2, the in-person problemsolving intervention significantly reduced weight gain compared to the mail-based (control) group ( p < .05), but the reduction in weight regain was not mediated by changes in problem-solving skills. Supplemental analys es revealed that the effect of Phase 2 treatment assignment on weight change was me diated by adherence to self-monitoring. Thus, the findings in this study suggest th at baseline problem-solving abilities and changes in problem-solving skills during initia l treatment are associated with magnitude of initial weight loss. The results also suggest that problem-solving based extended care programs facilitate adherence to self-monito ring, thereby enhanci ng the maintenance of lost weight.
1 CHAPTER 1 INTRODUCTION Overview Obesity in America Rates of overweight and obesity in the U. S. are rising to what have been called nationwide epidemic proportions (U.S. Depa rtment of Health and Human Services [USDHHS], 2001). Findings from the 2003-2004 National Health and Nutrition Examination Survey (NHANES) indicate that 66 % of U.S. adults are overweight or obese with a Body Mass Index (BMI) of 25 kg/m2 or above, with 32% being obese (BMI of 30 kg/m2 or over) and 5% falling into the Class III category (BMI of 40 kg/m2 or over) of very severe or morbid obesity (Ogden, Carroll, Curtin, McDowell, Tabak, & Flegal, 2006). These findings indicate the prevalence of obesity has doubled in the last 30 years. The high rate of obesity is particul arly alarming because of the numerous deleterious effects of obesity on health and longevity. Research has shown obesity is directly associated with cardiovascular disease (CVD), hypertension, diabetes, certain cancers (e.g., endometrial, colon, breast), gallst ones, and osteoarthritis (Stein & Colditz, 2004). Increased disability, health complicatio ns, and decreased qual ity of life are all significantly related to obesity, such that BMI is positively associated with negative health outcomes (National Heart, Lung, and Bl ood Institute, 1998). Furthermore, risk of death from all causes but particularly from CVD significantly increases in adults with BMIs 30 kg/m2 and over and continues to rise dram atically in those with a BMI of 40 kg/m2 and over (Field, Barnoya, & Colditz, 2002) Mortality rates from obesity-related
2 causes have been estimated at anywhere from 111,000 (Flegal, Graubard, Williamson, & Gail, 2005) to 400,000 (Mokdad, Marks, Stroup, & Gerberding, 2004). Thus, obesity has been shown to have a major impact on morbidity and mortality. Obesity has also been shown to have a si gnificant economic impact. It has been estimated that U.S. health expenditures related to obesity total 117 billion dollars each year, which includes both direct (e.g., docto r visits, medications, and hospitalizations) and indirect (e.g., lost wages due to illnes s) costs related to obesity (USDHHS, 2001). Furthermore, these costs continue to incr ease by degree of obesity. In the year 2000, more than 11 billion dollars was spent on medical costs for the 5% of the U.S. population who were morbidly obese (Arterbur n, Maciejewski, & Tsevat, 2005). Obesity also has a considerable impact on quality of life. Obese women are less likely to be hired, rated more negatively in job performance, and less likely to be promoted (Field et al., 2002; Puhl & Brow nell, 2001). Social stigma and body image issues, as well as lower quality of life, resu lting from decreases in physical functioning and vitality, often affect obese indivi duals (Field et al ., 2002; Wadden, Womble, Stunkard, & Anderson, 2002). Thus, obesity is a serious problem and has significant implications on physical health, mental wellness, and overall quality of life. Obesity in Rural Areas Obesity is of special concern in rural regi ons of the U.S. Rates of obesity vary by degree of urbanization and are f ound to be highest in rural area s (Eberhardt et al., 2001). Rural has been defined in various manners by different government agencies. The Centers for Disease Control and Prevention (C DC) utilizes the definition of rural as explicated in the system of county classification developed by the Office of Management and Budget (OMB), which classifies county areas as rural or urban based on their
3 population size and integration with large c ities (Ricketts, Johnson-Webb, & Taylor, 1998). The OMB showed that based on 2000 Cens us data, rural counties provide a home to over 17% of the U.S. population (U.S. Ce nsus Bureau, 2002). Those approximately 49 million people who inhabit rural counties find themselves living in some of the most medically-underserved areas in the country (E berhardt et al., 2001). Indeed, rural areas are known for not only higher rates of obesity bu t also higher rates of persons living with chronic diseases (Pearson, 1996), higher rates of persons living without health insurance, and more Health Professional Shortage Areas than in non-ru ral counties (U.S. Department of Health and Human Services, 2002). Furthermore, there is a special need to better understand and serve populations in these rural areas because of the disproportionately high hear t disease mortality rates. Rates of death from ischemic heart disease ar e higher in rural areas than in urban areas nationwide (264 vs. 248 per 100,000). Also, there is a notably striking disparity in heart disease mortality between Sout hern women from rural area s vs. urban areas, where the death rate is 211 vs. 187 per 100,000 (Eberhardt et al., 2001). Several factors may be contributing to these higher rates of mort ality, including the eff ects of higher poverty rates and lower educational attainment in re sidents of rural areas (Economic Research Services, 1993; Findeis, Henry, Hirschl, Lewis, Ortega-San chez, & Peine, 2001). These factors need to be considered because research has shown that coronary heart disease in the U.S. is inversely related to socioec onomic status (Cooper et al., 2000). Also, residents of rural areas often have to overc ome barriers to obtain access to preventive health and medical care services. As a result of the substantial travel distances between their homes and their health care providers, rural residents often only seek urgent care
4 and neglect health promotion activitie s (Economic Research Services, 1993 ; Schootman & Fuortes, 1999). Lastly, rural populations often partake in li festyle behaviors that can increase their risk for heart disease. Research has shown re sidents of rural areas, as compared to nonrural residents, have been slow er in making changes to lifest yle habits related to nutrition and physical activity, which are key behavioral contributors related to heart disease (Pearson & Lewis, 1998). The tr aditional high-fat, high-calor ie diets and higher rates of sedentary behavior in rural re sidents, which contribute to th e higher rates of obesity, also contribute to higher rates of coronary h eart disease mortality in rural populations (Eberhardt et al. 2001). T hus, finding ways to aid rura l populations, by developing and implementing interventions to help prevent a nd treat obesity that can provide subsequent health benefits, is of the utmost importance. The Treatment of Obesity Numerous important national agencies, incl uding the Institute of Medicine (IOM) in 1995, National Institute of Health/Na tional, Heart, Lung, and Blood Institute (NIH/NHLBI) in 1998, and American Heart Association (AHA) in 2004, have recognized obesity as a significan t risk factor for disease and disability and have put forth recommendations for the treatment of obes ity. The various guide lines indicate that lifestyle interventions, which target creating a moderately negative energy balance through behavioral modifications in diet and physical activit y, are appropriate for the treatment of overweight and obese individuals (Klein et al, 2004). Behavioral interventions generally produce good short-term results. Weight losses of approximately 8-10% in total body weight, with average losse s of about 7-9 kilograms over 4-6 months of treatment, are common in behavioral lifes tyle interventions (Wing, 2002). Losses of
5 8-10% may seem like small reductions in tota l body weight to patien ts in a behavioral intervention, who often enter a weight-loss program with e xpectations of losing over 30% of their body weight (Foster, Wadden, Vogt, & Brewer, 1997). However, an extensive review by the Institute of Me dicine (1995) has shown that reductions of approximately 5% or more of initial total body weight can improve or prevent many of the obesity related risk factors for corona ry heart disease, such as: in sulin resistance and type 2 diabetes, dyslipidemia, and hypertension. Pharmacotherapy has been suggested as an appropriate treatment for obese individuals who are unable to achieve adequate weight loss with conventi onal lifestyle modifications alone and who have no contrai ndications for drug therapy. Currently, only two medications, Sibutramine (Meridia) and Orlista t (Xenical), are approved for longterm use (Klein et al, 2004). Sibutramine is a combined adrenergic and serotonergic drug that was shown in a randomized controlled trial (RCT) by Smith a nd Goulder (2001) to facilitate 7% total body weight loss after 1 year. Sibutramine is however, associated with multiple side effects such as dry mouth, constipation, insomnia, and increased blood pressure. Sibutramine can also be lethal if combined with serotonergic medications and should not be prescribed to individuals on selective serotonin reuptake inhi bitors or monoamine oxidase inhibitors (Kle in et al., 2004). Orlistat selectively inhibits pancreatic lipase and reduces the absorption of fats by 30%. A review of various RCTs showed indi viduals on Orlistat experience total body weight losses at 1 year of approximately 8 to 10% (Hauptman, Lucas Boldrin, Collins, & Segal, 2000). Orlistat is also associated with side effects, including malabsorption of fat-
6 soluble vitamins, fecal urgency, and fecal in continence (Bray, 2002). Many patients find Orlistat to be an unacceptable option for wei ght management as a result of these latter side effects that can be aversive especially in social situations. Lastly, bariatric surgery is sometimes r ecommended for Class II obese individuals with obesity-related diseases and Class II I obese individuals, though it has been noted that surgery should only be considered for t hose patients who were unable to lose weight with more conventional methods and who have no contraindications for surgery (NHLBI, 1998). The Swedish Obese Subjects Intervention Study showed that at 2 years, patients who underwent gastric banding had a mean pe rcent body weight loss of 21%, patients who underwent vertical banded gastroplasty had 23% total body weight loss, and those who underwent gastric bypass had 33% total body weight loss (Sjostrom, Lissner, Wedel, & Sjostrom, 1999). After 10 years post-sur gery, those patients who received gastric banding had maintained a 13% mean total body weight loss, patients who received vertical banded gastroplasty had maintained a 17% total body weight loss, and those who received gastric bypass had maintained a 25% total body weight loss. Also, patients who had received surgery, as compared to cont rols, showed numerous health improvements including decreases in blood pressure, decrea ses in incidence of diabetes type 2, and decreases in trigylceride levels at both 2 and 10 year follow-up (Sjostrom, et al., 2004). Bariatric procedures produce substantial we ight losses but are also associated with: numerous necessary post-operative behavioral modifications to eati ng (e.g., limitations in quantity and types of food and required mu lti-vitamin and mineral supplements), postoperative problems (e.g., protein deficienci es, vitamin and mineral deficiencies, vomiting, and changes in mood) and a periop erative mortality rate within 30 days
7 following surgery of approximately 1% (P odnos, Jimenez, Wilson, Stevens, & Nguyen, 2003). Thus, guidelines suggest surgery should be considered as a last resort for the extremely obese (Klein et al., 2004; NHLBI, 1998). The Maintenance Problem in Obesity Treatment Perri and Corsica (2002) de scribed the maintenance pr oblem associated with obesity treatment. Specifically, when obesity treatment ends, most participants will gradually abandon their changes in eating and exercise habits and over time regain most of the weight initially lost in treatment. Typically, by 18 months following entry into a lifestyle intervention, par ticipants will regain 38% of the initial lost weight (Wing, 2002). Also, studies with longer followup periods have shown that pa rticipants usually continue to steadily regain their lost weight, often returning to th eir baseline weights (Kramer, Jeffery, Forster, & Snell, 1989). This is an im portant problem to try to redress because if lost weight is not maintained, the health bene fits of obesity treatment may be negligible (Perri & Corsica, 2002). Perri & Corsica (2002) identif ied an interaction of physiological, psychological, and environmental factors as the probable co ntributors to post-treatment weight gain. Biological factors, such as decreased energy needs and decreased metabolic rate, prime participants for regain. Psyc hological factors, such as unrea listic expectations, lack of reinforcement for maintenance, and failure to acquire maintenan ce skills, often lead participants to abandon their weight-control efforts. Environmental factors such as constant exposure to low-co st, highly palatable, energydense foods, and barriers to maintaining physical activity (e.g., time a nd convenience), increase the likelihood for participants to experience a sl ip or lapse in their new healthier lifestyle regimen. Thus, identifying effective strategies for long-t erm management of obesity, which can help
8 participants overcome these numerous hinde ring factors leading to regain, becomes essential if the health benefits of weight loss are to be maintained after treatment (Perri & Corsica, 2002). In a review, Byrne (2002) examined the possible link between psychological factors and their impact on longterm weight management strate gies. First, she noted that only few studies conducted have assessed psyc hological factors prospectively in weightloss program participants. Thus little is known about factor s at pre-treatment that may impact weight loss outcomes at post-treat ment. Furthermore, Byrne noted those prospective studies that do exis t, often utilize samples in uni versity-based settings that may not be representative of the greate r national population w ho are attempting to manage their weight. Next, Byrne provided so me insight gleaned from previous studies examining the characteristics or features of pa rticipants who were able to maintain their weight loss after treatment versus those who regained their weight. Byrne found a main difference between maintainers versus reg ainers in various studies was related to problem-solving abilities, such that regain ers were less likely to use problem-solving skills to cope with stressful life events but rather practice more escape or avoidance of problems and challenges. Maintainers, how ever, were shown to often utilize direct coping through problem solving to overcome barriers and maintain their weight loss. Indeed, Kayman, Bruvold, and Stern (1990) found that rega iners had an increased likelihood of over-eating as a response to ne gative emotions than did maintainers who were better able to cope w ith stressors. Byrne (2002) also noted that as compared to weight regainers, maintainers were more confid ent in their ability to control their weight and caloric intake and had an increased recognition and response to the need for
9 continued vigilance in weight management efforts. Byrne suggested the need for additional research into the nature of the va rious factors and mechanisms that may impact weight management, emphasizing the need to th en utilize that insight learned from the research to develop more effective interventions for obesity treatment. As a result of the numerous factors influencing regain and the significant difficulties encountered by so many of those at tempting to maintain lost weight, several theories exist as to what may be the most e ffective type of long-term weight management intervention. Wilson (1994) posited that pharmacologica l treatment rather than participation in standard lifestyle interv entions is necessary for effective weight management. Similarly, Williamson (1999) expre ssed a belief that lifestyle interventions have been ineffective in reducing the preval ence of cardiovascular disease and obesity and thus alternate forms of in tervention are needed. Next, some researchers believe that lifestyle interventions have failed because they simply have not addressed issues, including cognitive and behavioral factors potentially responsible for weight regain, necessary for effective long-term weight main tenance. Thus, they suggest that those cognitive-behavioral factors would have to be addressed within the treatment to create more efficacious weight management interv entions (Cooper & Fairburn, 2001). Lastly, various researchers (Latner, Stunkard, Wils on, Zelitch, & Labouvie, 2000; Perri, Nezu, & Viegener, 1992) have suggested that lifestyle interventions have failed because they have not provided care suitable for the chronic proble m that is obesity. They have argued that the maintenance of weight lost in behavior al treatments of obesity can be improved by developing a continuous-care problem-solv ing approach to obesity management.
10 Problem Solving as a Method of Coping Problem solving has been defined as the self-directed cognitiv e-behavioral process by which a person attempts to identify or disc over effective or adaptive solutions for specific problems encountered in everyday living (DZurilla & Nezu, 1999). This concept of problem solving, and its applicati on within a variety of research areas, has generated increasing in terest within the past decades. DZurilla and Goldfried (1971), who first de veloped a prescriptive model of social problem solving, proposed within their conceptual framework a set of principles detailing how people could be trained to better cope with problems. Their goal was to help enhance peoples overall social competen ce and to maximize their problem solving success in a variety of social s ituations. The descriptor of social was used to emphasize that the focus was on coping with problems w ithin natural social environments. Thus, their model of social problem solving can be applicable across numerous problems associated with daily living. The DZurilla and Goldfried (1971) m odel of social problem solving views effective problem solving as a process involvi ng a combination of speci fic skills that are learned, rather than a trait or solitary ability a person possesses. Thus, two main reasons exist that would explain why people may not ha ve adequate social problem-solving skills. Either the person may not have learned the n ecessary skills to be an effective problem solver or the person learned the appropriate sk ills but fails to perform them when coping with certain problems as a result of poor psychological adjustment (i.e., presence of anxiety or depression) that may lead to mala daptive behavior such as the inhibition of proper skill utiliza tion in particular situations (Nezu, Nezu, & Perri, 1989).
11 The Problem-Solving Process DZurilla and Nezu (1982) posited that there are five major components of an effective problem-solving process, which pe ople can be trained to implement, these include: (1) problem orientat ion, (2) problem definition and formulation, (3) generation of alternatives, (4) decision making, and (5) solution implemen tation and verification (see Table 1-1, located at the end of the section, for a summary of the five major components of problem solving). Problem orientation ha s been described as a motivational process that involves the immediate cognitive-affectiv e reactions of an individual when first confronted with a problem (Nezu & Nezu, 2001). Essen tially, this process allows individuals to recognize problem s as encountered within da ily living and consists of individuals generally stable beliefs and assumptions regardin g their ability to effectively solve problems (Nezu, 2004). Th eir model posits that how an individual perceives his or her problem-solving ability, either positively or negatively, can impact his or her outlook relevant to five major problem orientati on variables, includi ng: problem perception, problem attribution, problem appraisal, percei ved control, and time/effort commitment. Thus, if an individual holds a positive problem orientation, it can facilitate more effective problem solving because the individual will be more accurate in recognizing the presence of a problem, will be more correct in attr ibuting the cause of the problem to the appropriate source, will be mo re likely to appraise the problem as a challenge rather than a threat, will be more likely to perceive the problem as solvable, will be more likely to accurately estimate the time and effort necessary to solve a problem, and will be more likely to devote the necessary time to effectively problem solve. The four remaining component s contributing to the problem-solving style have been described as those core cognitive-behavioral activities that people engage in when
12 attempting to cope with problems in living (Nezu, 2004). Unlike the problemorientation component that targets a pers ons motivation related to problem solving, problem-solving style (also referred to as p roblem-solving proper) involves the actual application of four problem-s olving skills that will enhance effective problem resolution through goal-directed tasks. Fi rst, in the task of problem definition and formulation, the goal is to gather all relevant information to clarify details of the situation so as to facilitate the most well defined conceptuali zation of the problem. Furthermore, it is important to specify realistic problem-solvi ng goals and set forth reasonable objectives. Next, in the task of generati ng alternative solutions, the goa l is to generate a mass of possible solutions by brainstorm ing alternatives. Seeking th e best solution possible, the purpose is to produce an abundance of alte rnatives, as the quantity of solutions is important in providing the most opportunity for generating a successful solution. Also, judgment of the alternatives is deferred until the decision-making process, so as not to stunt the generation of ideas. It is important in this step that imagination and innovation are utilized so as to create a wide variety of potential problem-solvi ng solutions, to again increase the likelihood of successful problem resolution. Next, in the decision making step, the goal is to evaluate all of the possi ble alternatives, system atically noting potential costs and benefits, to ultimately select the most effective solution. Lastly, in solution implementation and verification, the goal is to implement the chosen solution within the problematic situation, observ e the consequences, evaluate the effectiveness of the solution employed, and return to a previous st age of problem solving if necessary to find a more efficacious solution to the problem. Thus, individuals are taught to only withdraw
13 from the problem-solving process when a successful solution has been achieved (DZurilla & Nezu, 1999). Problem Solving and Mental Health Acknowledging the proposed importance of effective implementation of problemsolving skills on successful problem solving and social competence, it is important to assess and evaluate how appropriate skill ve rsus skill deficit impact individuals across various situations and circumstances. Early research, conducted in the 1970s by Spivack and colleagues, examined the association between psychopathology and defi cits in problem-solving skills. They assessed social problem-solvi ng skills by specifically eval uating a persons means-ends thinking, which they describe d as the ability to explain th e individual steps necessary to effectively problem solve in sp ecific situations (Spivack, Pl att, & Shure, 1976). Using the Means-Ends Problem-Solving (MEPS) Procedure (Platt & Spivack, 1975), testtakers skills in means-ends thinking are evaluated by administeri ng a set of hypothetical problems along with successful resolutions th at are missing the descriptions of how the problems were solved and asking the test take rs to provide the act ual steps or means that led to the effective problem-solving e nds. Platt and Spivack (1973) found that adult psychiatric patients showed significan t deficits on MEPS performance as compared to non-clinical adult controls. Patients w ith current substance abuse problems (Platt, Scura, & Hannon, 1973) as well as adolescent ps ychiatric patients were also shown to have deficits in MEPS test performance scores (Platt, Spivack, Altman, Altman, & Peizer, 1974), while higher levels of soci al competence were associated with higher MEPS scores (Platt & Spivack, 1972). This suggests that effective problem-solving
14 abilities are indeed related to social competence and inversel y associated to maladaptive behavior and psychopathology (Nezu, Nezu, & Perri, 1989). More recent research also demonstrated the significant impact of problem-solving abilities on various psychologi cal conditions (Douglass, 2000; Nezu & DZurilla, 1989). Specifically, numerous studies have shown an association between deficits in problem solving and conditions of anxiety and depr ession (Priester & Cl um, 1993; Zebb & Beck, 1998). Studies have shown a relationship be tween negative probl em orientation and anxiety (Dugas, Gagnon, Ladouceur, & Freeston, 1998), poor problem-solving confidence and anxiety (Davey & Levy, 1999) and ineffective problem solving and anxiety (Bond, Lyle, Tappe, Seehafer, & DZurill a, 2002). Associations have also been explicated regarding negative problem or ientation and depression (Frye & Goodman, 2000; Marx & Schulze, 1991) and ineffective problem solving and depression (Miner & Dowd, 1996; Nezu, 1985). Nezu (2004) e xpounded upon the relationship of problem solving and negative affectiv e conditions, explaining psycho logical distress is often a result of impaired problem-solving abilities, including: inappropriate appraisal of the problem, negative self-efficacy, negative pr oblem orientation, and ineffective or unsuccessful attempts at problem resolution. Furthermore, those deficits in problem solving when taken to the extreme may have implications for suicidal behavior in individuals experiencing severe psyc hological distress (Sheehy & OConnor, 2002). Thus, problem-solving abilities have an impor tant role in unders tanding psychological deficits associated with indi viduals experiencing distress. Problem Solving and Medical Conditions Patients problem-solving skills also signifi cantly impact their ability to cope with various medical conditions, including: chro nic pain (Kerns, Rosenberg, & Otis, 2002),
15 inflammatory bowel disease (Dudley-Brow n, 2002), and multiple sclerosis (Pakenham, 2001). Attention has been especially afforded to areas of chronic illness, such as diabetes, where daily problem solving is of ten relevant to self-management of the condition (Hill-Briggs, 2003). Elliott, Shew chuk, Miller, and Richards (2001) discussed the immense importance of behavioral adju stment, including implementing daily selfcare regimens, in the effective treatment of the chronic illne ss diabetes. They hypothesized that problem-solving abilities in individuals coping w ith diabetes would significantly impact the management of the ch ronic illness. Ind eed, after assessing twohundred fifty-nine persons with diabetes, util izing the Social Probl em Solving InventoryRevised (SPSI-R) questionnaire, they found that persons who were most distressed and unskilled were those with a profile characterized by lower positive problem orientation, higher negative problem or ientation, lower rational problem solving, higher impulsiveness/carelessness, and higher avoida nce of problem solving. They also found that those individuals with a higher positiv e problem orientation, lower negative problem orientation, higher rational pr oblem solving, lower impulsive ness/carelessness, and lower avoidance of problem solving were ideal problem solvers, who were most optimally adjusted to managing their diabetes. Thes e results suggested that problem-solving abilities have a profound impact on self-manag ement of diabetes. Elliott et al. (2001) used the insight provided by their study to ge nerate interventions to teach persons with diabetes effective problem-solving skills, which could facilitate better coping and adjustment to daily diabetes management. Problem-Solving Therapy Problem-solving therapy (PST), which focu ses specifically on teaching efficacious problem solving through instruction about the necessary tasks and skills within the five
16 components of the problem-solving process, ha s been shown to be beneficial in the treatment of various conditions (DZurilla & Nezu, 1999). PST has been effectively used in smoking cessation interventions, in the tr eatment of marital problems, and most notably, has been utilized in the successful treatment of mo od disorders (Arean et al., 1993; DZurilla & Nezu, 1999). Nezu and Pe rri (1989) showed that problem-solving therapy, where participants were trained in sk ills related to effec tive problem orientation as well as the four main components of pr oblem-solving style, significantly decreased depression scores in participan ts with unipolar depression as compared to those in a waitlist control condition. Nezu and colleagues recent research ( 2003) examined the impact of PST on adult cancer patients in Project Genesis. They assessed the efficacy of a PST individual treatment condition and also a PST individual plus significant other condition (vs. a waitlist control condition) to decrease psychologi cal stress in adult cancer patients. The sample included 132 adult cancer patients, who were randomized to receive either 10 sessions, 1.5 hour/weekly of PST in an indivi dual setting, or 10 sessions, 1.5 hr/weekly of PST with a significant other who was to act like a problem-solvi ng coach by providing support for the patient, or no treatment for 10-12 weeks (with treatment being offered to these patients at the end of the trial). They found that partic ipants in either of the PST conditions showed significantly better im provements in psychological well-being (as measured by a variety of psychological que stionnaires including the POMS, BSI, and CARES) as compared to the wait-list contro l condition. Thus, Nezus research showed that PST was effective in aiding medical pa tients with their psychological distress.
17 Little research has examin ed the impact of problem -solving abilities and PST on initial weight loss. Some research has evaluated problem solving and its impact on weight maintenance. Perri and colleagues ( 2001) assessed the effec tiveness of extended treatment interventions, in adults 21 60 years of age, subsequent to a 20-week standard behavioral weight-loss program Investigating the impact between a relapse prevention training (RPT) intervention, a problem-solvi ng therapy (PST) intervention, and a nofurther contact condition (BT only), the authors believed that the extended treatment interventions would prove superior in fac ilitating successful weight maintenance as compared to the BT only condition. In fact, they found no differences between RPT and BT only or between RPT and PST, but they di d find significant diffe rences between BT only and PST. Those participants who had b een randomized into the PST condition were more likely after one year posttreatment to maintain their in itial weight loss as compared to those participants in the BT only group. Furthermore, of thos e in the PST condition, 35% achieved clinically signifi cant weight losses as compared to 6% in the BT group. The authors also showed that participants in the PST condition had better adherence (as measured with a 7-point Likert scale self-report measure of ni ne key behavioral strategies including self-monitoring and stimulus contro l) to the healthy be haviors/weight-loss strategies taught in the initial weight-loss treatment as co mpared to those participants in the BT condition. Lastly, the study showed that adherence to behavior al strategies was a partial mediator of the treatment condition e ffect such that the long-term success of participants in the PST cond ition was partially accounted fo r by their better adherence to behavioral weight-loss strategi es (Perri et al., 2001). C onsequently, this study suggests
18 that problem-solving skills, as well as adhere nce to behavioral wei ght-loss strategies, are significantly associated with weight maintenance. Current Study Thus, recognizing the significant adverse impact of obesity in the U.S. and particularly in rural areas, the problem w ith weight maintenance, the influence of problem-solving skills on preventing weight regain, and the sparse research examining the effects of pre-intervention problem-solving abilities during the initial weight-loss intervention, this study endeavors to redres s gaps in the literature by providing insight into the impact of problem solving on th e management of obesity in women from medically-underserved Southern rural se ttings. The women, ages 50-75 years, participated in 2 phases of an 18-month we ight management intervention. Phase 1 consisted of a 6-month intensive lifestyle intervention. Phase 2 entailed a 12-month extended care period with participants randomized to problem-solving based interventions delivered via in -person sessions, by telephone cont acts, or by ma il (control). Specifically, the aim of the current study was to examine the impact of problem solving on weight loss at various points thr oughout the weight-loss pr ocess, investigating the following questions: (1) Do participants pre-intervention probl em-solving abilities impact weight change at 6 months? (2) Do participants problem-solving abilities improve following the 6-month intervention? (3) Do improvements in participants problem-solving abilities following the 6-mont h intervention impact weight change at 6 months? (4) Do improvements in participan ts problem-solving abilities from 6 to 18 months or baseline to 18 months impact weight change over the trial? (5) Do participants problem-solving ab ilities improve from 6 to 18 months or baseline to 18 months differently in the te lephone or in-person groups as compared to the mail group?
19 (6) Do improvements in participants problem -solving abilities from 6 to 18 months or baseline to 18 months mediate the impact of Phase 2 assignment on weight change? (7) Does Phase 1 adherence to self-monitoring medi ate the effect of problem-solving abilities on Phase 1 weight change? and Does Phase 2 adherence to self-monitoring mediate the effect of improvements in problem solvi ng on weight maintenance during Phase 2? We hypothesized: 1. Higher pre-intervention problem-solving ab ilities, as measured by the SPSI-R, would be significantly associated with gr eater total percent b ody weight change at 6 months. 2. The weight-loss intervention would result in increased problem-solving abilities from baseline to 6 months. 3. Greater improvements in problem-solving abilities from baseline to 6 months would be associated with greater total percent body weight change at 6 months. 4a. Greater improvements in problem-solving abilities from 6 months to 18 months would be related to greater total percen t body weight change from 6 months to 18 months. 4b Greater improvements in problem-solving abilities from baseline to 18 months would be related to greater total percen t body weight change from baseline to 18 months. 5a. Problem-solving abilities would increase fr om 6 months to 18 months in the phone or in-person Phase 2 groups but not in the mail group. 5b. Problem-solving abilities would increase from baseline to 18 months in the phone or in-person Phase 2 groups but not in the mail group. 6a. The effect of extended care conditions on weight change from months 6 to 18 would be mediated by changes in problem -solving skills from months 6 to 18. 6b. The effect of extended care conditions on weight change from baseline to 18 months would be mediated by changes in problem-solving skills from baseline to 18 months. 7a. The effect of changes in problem-sol ving skills from baseline to 6 months on weight change from baseline to 6 months would be mediated by Phase 1 adherence.
20 7b. The effect of changes in problem-sol ving skills from baseline to 18 months on weight change from baseline to 18 m onths would be mediated by Phase 2 adherence.
21 Table 1-1. Overview of the Problem-Solving Process ______________________________________________________________________ Problem Orientation Problem perception recognizing rather than denying or avoiding a problem Problem attribution attributing th e problem to a modifiable cause Problem appraisal viewing the problem as an opportunity for self-improvement rather than a threat Perceived control viewing oneself as cap able of changing the problem situation Time /effort commitment recognizing that effective problem solving requires time and effort Problem Definition and Formulation Gathering relevant factual information separating objective facts from assumptions or misconceptions Understanding the problem understanding what factors are responsible for the problem (i.e., its causes) Setting a realistic problem-solving goal sett ing specific, concrete, attainable goals Re-appraising the problem understanding the benefits of change versus status quo Generation of Alternative Solutions Quantity principle the more alternatives that are consid ered, the greater the odds of identifying a good one Deferment of judgment principle susp ending judgment help s generate creative solutions Variety principle the greater the variety of solutions, the greater the chances of developing a solution Decision Making Anticipating solution outcomes identifyi ng the expected positive and negative consequences of a solution Evaluating solution outcomes judging the likelihood that a solution will resolve a problem Preparing a solution plan identifying e ither a single solution or a solution combination Solution Implementation and Verification Carrying out the solution plan actual implementation of the specific plan for solving the problem Self-monitoring self-observation of the im plementation of the solution plan and its outcome Self-evaluation judging whether the so lution plan has resolved the problem Self-reinforcement recognizing successful problem resolution with positive self statement Troubleshooting and recyclingif the probl em is not resolve d, returning to the problem-solving process ________________________________________________________________________
22 CHAPTER 2 MATERIALS AND METHODS Participants Participants were healthy but sedentar y women from medically underserved rural areas who volunteered to take part in an 18-month study examining the effects of a lifestyle weight-loss in tervention on obesity and health. Three cohorts of participants were recruited and randomized at 6-month in tervals. Eligibility requirements included being female, age between 50 75 years, body mass index (BMI) of 30 kg/m2 and above, and weight less than 350 pounds (so as to allow for measurement on a balance beam scale). Potential participants were excluded if, at screening, thei r medical history, clinical examination, or laboratory results revealed underlying diseases likely to limit lifespan and/or increase risk of inte rvention, such as: cancer requiri ng treatment in the past 5 years, serious infectious diseases, myocardi al infarction or cere brovascular accident within the last 6 months, unstable angina within the past 6 months, congestive heart failure, chronic hepatitis, cirrhosis, chronic ma labsorption syndrome, chronic pancreatitis, irritable bowel syndrome, previous bari atric surgery, history of solid organ transplantation, history of musculo-skeletal conditions that limit walking, chronic lung diseases limiting physical activity, serum cr eatinine > 1.5 mg/dl, anemia (hemoglobin < 10 g/dl), and any other condition likely to limit five-year life expectancy. Potential participants were also deemed ineligible based on the following criteria: metabolic values out of range, use of certa in medications, and conditions or behaviors
23 likely to affect the conduct of the trial. Metabolic excl usions included: fasting blood glucose > 125 mg/dl at screen ing if not known to be diab etic (diabetic patients under active treatment were enrolled if appr oved by primary provider), fasting serum triglycerides > 400 mg/dl at sc reening despite appr opriate drug treatment, and resting blood pressure > 140 / 90 mmHg despite ap propriate drug treatment. Medication exclusions included: antipsychotic agents monoamine oxidase inhibitors, systemic corticosteroids, antibiotics for HIV or TB, chemotherapeutic drugs, or current use of prescription weight-loss drugs. Conditions or behaviors likely to affect the conduct of the trial included: unwilling or unabl e to give informed consent, unable to read English at the 5th grade level, unwilling to accept random assi gnment, unable to travel to Extension Office for intervention sessions, participation in another rand omized research project, weight loss > 10 pounds in past 6 months, likel y to move out of the county in next 2 years, major psychiatric disorder, excessi ve alcohol intake, b ody weight > 159 kg, and other conditions which in the opinion of the staff would advers ely affect participation in the intervention. Women who were pregnant or planning to become pregnant during the course of the study were also excluded. Measures Weight. Weight was measured to the nearest 0.1 kilogram using a calib rated and certified balance beam scale. Weights were used to calculate total percen t body weight change (e.g., percent weight change for Phase 1 wa s calculated: 6-month weight baseline weight / baseline weight) for each participant across the trial. Total percent body weight change was utilized instead of net weight change so as to acc ount for participants
24 differing baseline or starting wei ghts. Percent weight change from baseline to 6 months, 6 months to 18 months, and baseline to 18 m onths were treatment outcomes in this study. Social Problem Solvi ng Inventory-Revised. Problem-solving abilities were assessed with the Social Problem Solving Inventory-Revised (SPSI-R; Maydeu-Olivares & DZurilla, 1996). The SPSI-R is a 52item, self-report measure that asked participan ts to rate statements on a 5-point Likert scale ranging from 0 (not very true of me) to 4 (extremely true of me). The SPSI-R is based upon the five-component model of social problem solving and has five scales. Problem orientation was measured on two scales: Positive Problem Orientation and Negative Problem Orientation. Problemsolving style was assessed on the three remaining scales, which include: Rational Problem Solving, Impul sivity/Carelessness Style, and Avoidance Style. The scores are inter-correlated such that a positive problem orientation is most highly correlated w ith a rational problem solving style ( r = .6 to .7). A negative problem orientation is most highl y correlated with the dysfunctional problem solving styles of impulsivity/carelessness ( r = .5 to .6) and avoidance ( r = .6 to .7). Furthermore, the relationships between SPSI-R subscales have been shown to be stable across different populations, varying in both ag e and psychiatric stat us (Maydeu-Olivares & DZurilla, 1996). These scores were combined to create a composite score, Social Problem Solving Summary Score, which was used in this st udy as both a predictor of treatment outcome and as a treatment outcome. This Social Problem Solving Su mmary Score, as well as the other 5 component scale scores, have a m ean of 100 and a standard deviation of + 15 points. The guidelines for interpreting scores are such that 0-70 is considered extremely to very much below normal, 71-85 is cons idered below normal, 86-114 is considered
25 normal, 115-129 is considered above normal, 130 and above is considered very much to extremely above average (Maydeu-Olivares & DZurilla, 1996). Therefore, improvements in problem-solving scores can be interpreted by assessing upward movement (i.e., increases in scores) towa rds more higher functioning or more above average levels of problem-solving ability. One study examining the effects of PST on depression showed statistically significant improvements of 3-5 points in mean problem-solving ability (specifically on scales: problem definition and formulation, generation of alternatives, and decisionmaking) from pre to post-treatment in a sample of depressed older adults was associated with clinically significant changes (i.e., impr ovements) in depression (Arean et al., 1993). The SPSI-R has sound psychometric properties with an internal consistency of .76 to .92 and test-retest reliabil ity of .72 to .88 (DZurilla, Ne zu, & Maydeu-Olivares, 2002). The measure is also particularly notewo rthy, as it has been shown to possess the sensitivity necessary to assess the effects of problem-solving skills training in clinical settings (Nezu, Nezu, Friedman, Faddis, & Houts, 1998). Self-Monitoring Records. Self-monitoring of foods consumed was a key intervention tool. These records allowed participants to compare their act ual daily eating behavi ors with their goal behaviors. Therefore, these records facilitated an awarene ss in the participants that fostered initiation of corrective actions as n eeded when goals were not being met. Thus, it was the process of utilizing self-monitoring on a daily basis that was meaningful, rather than the absolute accuracy of the reco rds that was important in this study. It is important to note that self-monito ring is just one tech nique that denotes adherence to an intervention. Because self-mon itoring is supported in obesity research as
26 an important predictor of adherence a nd treatment outcomes (Sarwer & Wadden, 1999), completion of self-monitoring records was used as our proxy of adherence for this study. Specifically, only self-monitoring records that had total daily caloric intake calculated were used (i.e., partially completed records, which would not allow for meaningful selfmonitoring, were not utilized in calculating adherence). Self-monitoring was measured over the 24 sessions of Phase 1. Thus, there was a possibility of having a maximum total of 161 records during initial treatment. Selfmonitoring was also calculated during Phas e 2, with the possibility of completing a maximum total of 364 records. Phase 2 record s were either given to the interventionist by the in-person group members at session or mailed in by the telephone-based and mailcontact group participants. Procedure Recruitment. Participants were recruited through a vari ety of methods, including media articles, direct mailings, newspaper announcements, a nd presentations to community groups. Recruitment specifically target ed rural counties, utilizing the OMB classification of rural (Ricketts, Johnson-Webb, & Tayl or, 1998). Following telephone screening, potentially eligible persons were invited to attend an assessment during which the purposes and procedures of the study were e xplained and informed consent was obtained. Initial Assessment. After providing consent, each individual underwent an electrocardiogram (ECG), a blood-draw, and a 6-minute walk test. Measurem ents of height, weight, and girth, and of resting heart rate and blood pressure were taken for each participant. Participants completed medical history forms that incl uded questions concerning current and past
27 illnesses, past surgeries, and current medicati on use. They also completed other various questionnaires assessing dietar y intake, physical activity, heal th-related quality of life, depressive symptoms, and problem solving. Following the initial assessment, potential participants results were reviewed to determin e if participation crit eria had been met. Participants were then contact ed via the telephone to inform them of their eligibility. Schedule of Assessment. Measures were administered at a ssessments occurring at baseline (preintervention), 6 months (at the completion of Phase 1), and 18 months (at the completion of Phase 2). Phase 1. Phase 1 occurred during months 1-6 after in itial randomization. Participants met in groups for 24 weekly sessions, each lasting approximately 90-minutes, in their local County Extension Offices. Interventionists, who were either Family and Consumer Sciences Agents (FCS) or graduate student s with a bachelors or masters degree in behavioral or exercise scie nce, led the weekly groups. All interventionists underwent extensive training, monitoring, and supervision throughout the full delivery of treatment. The individual sessions in Phase 1 ha d three main components. The first component was a check-in segment, where part icipants had a private weigh-in and then met as a group to discuss progress made over the past week. Each participant had an opportunity during this segment of the sessi on to discuss any problems encountered or successes achieved since the previous sessi on. During this group check-in, participants provided support and suggestions for each other as problems or challenges were reported, while descriptions of achievements were me t with an abundance of positive feedback. Skill training was the second major component of the Phase 1 sessions. During this
28 segment of the session, the inte rventionist utilized session-b y-session intervention plans, contained in the TOURS treatment manual, which provided specific objectives, activities, and handouts for the participants. These session plans targeted skill building associated with cognitive-behavioral se lf-management skills (e.g., se lf-monitoring, goal-setting), proper exercise technique (e.g., importance of stretching), and nutritional training (e.g., accurate calculation of caloric intake). A cooking demonstration and/or food tasting of low-fat, low-calorie foods occurred during this segment of the session during certain designated weeks of the trial. Lastly, the final segment of the weekly Phase 1 sessions was a check-out discussion, where each particip ant identified specific behavioral goals (e.g., calorie goals, step goals ) for the week. The session ended with feedback and encouragement from the interventionist and group members regarding the goals for the upcoming week. During the Phase 1 sessions, the same trea tment manual was utilized in all groups and all counties. The main treatment objec tives for Phase 1 were the same for all participants, including: decreasing calori c intake in a nutritionally sound manner and increasing moderate intensity exercise (primari ly in the form of walking) to 180 minutes per week or 30 minutes a day/ 6 days a wee k, which translates to an extra 3000 steps per day (with approximately 100 steps equaling 1 mi nute of walking), so as to facilitate a weight loss of approximately 0.4 kg per week. Other objectives for Phase 1 included: an increase in knowledge about healthy weight management (e .g., benefits of weight loss, basics of energy balance, and proper weight -loss methods), improvement in quality of diet (e.g., reduction in total fat consump tion with a specific emphasis on decreasing saturated fat consumption, increase in consum ption of fruits and vegetables, increase in
29 consumption of whole grains), increase in phys ical activity (e.g., gradua l increase in daily steps as monitored by a pedometer), and acquisition of behavioral self-management skills (e.g., stimulus control, social support, and c ognitive restructuring). No formal problemsolving therapy occurred during Phase 1. Phase 2. Phase 2 occurred during months 7-18 of th e intervention. Main objectives for Phase 2 included: maintenance of healthy eating beha viors (e.g., isocaloric intake tailored to maintain lost weight, maintenance of low c onsumption of fats), maintenance of healthy physical activity levels (e.g., maintenance of 180 minutes of moderate intensity exercise per week), and maintenance of behavioral skills aimed at weight management (e.g., ongoing or intermittent self-monitoring of weight and health behaviors, developing selfreliance skills helpful in sustaini ng long-term weight management). In Phase 2, participants were randomly assigned one of three follow-up program conditions: (1) office-based (in-person) ma intenance program, (2) telephone-based maintenance program, or (3) mail-contact. A Latin Square design was employed, such that county, Phase 1 session time, a nd Phase 2 condition assignment were counterbalanced across the trial. Individuals were initially randomly assigned to Phase 1 groups, with schedule availability as th e only limitation, and then each group was randomly assigned to a Phase 2 condition. All groups were informed of their assignments at the conclusion of Phase 1 and were strongly encouraged to maintain the healthy lifestyle changes in diet and physical activity they accomplished during Phase 1. Participants assigned to the office-based maintenance program continued to meet at the County Extension Office for on-site group sessions twice a month. Phase 2 sessions were comprised of three main com ponents. The first component was again a
30 check-in segment, similar to that of Phase 1, where participants were weighed and then met as a group for each participant to discuss what challenges (if any) were encountered, what went well during the past two weeks, and what problem or issue (if any) would she desire group assistance wit h. Next, the key segment of the session involved the interventionist, who had been trained in PST, leading the group through the problemsolving process. The goal was to choose one group members problem and ultimately generate a solution plan for that participant to implement. The interventionist used the 5stage problem-solving model to guide the participants through the problem-solving process. The last component of the Phas e 2 session was simila r to the check-out segment of Phase 1, where participants id entified specific behavi oral goals for the upcoming 2-week period and then provided f eedback and encouragement to each other regarding the goals. With consent of the participants, randomly selected sessions were video-recorded for quality control to ensure protocol adherenc e. Interventionists adherence to the PST protocol during the Phase 2 in-person groups was checked routinely by the projects treatment fidelity manager. Participants assigned to the telephone-bas ed maintenance program were provided with a 15-page, detailed review describing the various problem-solving steps. They also had two telephone contacts per month with their interventionist The phone-contact session had three main objectives: (1) prompt th e participant to continue utilizing weightmanagement skills learned during Phase 1, (2) us e problem solving to identify challenges to successful weight maintenance and create plans to overcome those challenges, and (3) provide support and reinforcement for conti nued efforts at weight management. These
31 phone sessions had three components, similar to those in the office-based program. First, the interventionist checked-in with the pa rticipant. The participant was asked to provide her weight as measur ed that week and how often she met her behavioral goals within the past two weeks. The interventi onist then asked what went well in the participants eating and activity and also what problems or challenges the participant encountered within the past two weeks. Positive feedback was provided by the interventionist in response to participants achievemen t of goals and successful maintenance of healthy lifestyle behaviors. In response to problems encountered by the participant, the interventionist guided the participant through the same 5-step problemsolving process as implemented in the office-based maintenance program. Again, the ultimate goal was for the interventionist to a ssist the participant in generating a solution plan to implement within the ensuing 2-week period be tween phone contacts. The conversation ended with the familiar check-o ut, where the participant reviewed her goals for the next two weeks and was provi ded with feedback a nd encouragement from the interventionist. Phone conversations were approximat ely 10 15 minutes in length. All participants had a spec ified time for their phone appointme nt with their interventionist. To maximize the likelihood of reaching partic ipants by phone, up to five callbacks were made to individuals who could not be reach ed initially. All telephone contacts were logged and a check-sheet was completed by the in terventionist to reco rd: the length of the phone session, the progress reported by the pa rticipant, a description of the problemsolving process implemented, and the solu tion that was to be implemented by the participant before the next phone contact. Lastly, with consent of the participant,
32 randomly selected phone calls were audio-record ed for quality control to ensure protocol adherence. Interventi onists adherence to the PST prot ocol during the Phase 2 telephone calls was checked routinely by the projects treatment fidelity manager. Participants assigned to the mail-contact group were provided with a 15-page, detailed review describing the various pr oblem-solving steps. They also received biweekly newsletters providing psycho-educ ational information about nutrition and exercise. The newsletters provided low-fat, low-calorie recipes and offered tips about healthy lifestyle behaviors that promote ma intenance of lost weight. The mail-based maintenance condition served as a control or comparison group to the effects of the PST implemented in the office-based and te lephone-based maintenance programs. Analyses Hypothesis 1: Higher pre-intervention pr oblem-solving abilities, as measured by the SPSI-R, would be significantly associated with greater total percent body weight change at 6 months. This hypothesis wa s examined by utilizing a Pearson productmoment correlation analysis to assess if pre-intervention SPSI-R scores were significantly correlated with total percent body weight chan ge from baseline to 6 months. Hypothesis 2: The weight-loss interven tion would result in increased problemsolving abilities from baseline to 6 months. This hypothesis was investigated utilizing a within-group repeated measures analysis of variance to assess if SPSI-R scores changed significantly from baseline to 6 months. Hypothesis 3: Greater improvements in pr oblem-solving abilities from baseline to 6 months would be associated with greater total percent body weight change at 6 months. This hypothesis was examined with a Pears on product-moment correl ation analysis to
33 assess if SPSI-R change scores from baseline to 6 months were sign ificantly correlated to total percent body weight change from baseline to 6 months. Hypothesis 4a: Greater improvements in problem-solving abili ties from 6 months to 18 months would be related to greater to tal percent body weight ch ange from 6 months to 18 months. This hypothesis was inve stigated with a P earson product-moment correlation analysis to assess if changes in problem solving from 6 months to 18 months were significantly correlated with total per cent body weight change during Phase 2 of the trial. Hypothesis 4b: Greater improvements in pr oblem-solving abilities from baseline to 18 months would be related to greater total percent body weight change from baseline to 18 months. This hypothesis was investigated with Pearson product-moment correlation analyses to assess if SPSI-R change scores from baseline to 18 months were significantly correlated with total percent body weight change over the entire the trial. Hypothesis 5a: Problem-sol ving abilities would increas e from 6 months to 18 months in the phone or in-person Phase 2 groups, but not in the mail group. This was examined with a 2 (time) x 3 (condition) re peated measures analysis of variance. Hypothesis 5b: Problem-solving abilitie s would increase from baseline to 18 months in the phone or in-person Phase 2 groups, but not in the mail group. This hypothesis was examined with a 2 (time) x 3 (condition) repeated measures analysis of variance. Hypothesis 6a: The effect of extended care conditions on weight change from months 6 to 18 would be mediated by change s in problem-solving skills from months 6 to 18. This hypothesis was assessed was examined using a series of analyses. For change
34 in problem-solving abilities from 6 months to 18 months to act as a mediator of the effect of Phase 2 assignment on total percent body weight change from 6 to 18 months, the following conditions would need to be met: (1) an ANOVA would have to show that Phase 2 group assignment impacted change in problem-solving abilities from 6 to 18 months such that the telephone group as compared to the mail group or the in-person group as compared to the mail group significantl y differed in changes in problem solving from 6 to 18 months, (2) t-tests would have to show Phase 2 assignment (i.e., to the telephone vs. mail group or in-person vs. mail group) produced signifi cant differences in weight change from 6 to 18 months, (3) a regression analysis w ould have to show improvements in problem solving from 6 to 18 months significantly predicted weight change from 6 to 18 months, and (4) when both Phase 2 treatment condition assignment and improvements in problem solving from 6 to 18 months were in a regression model, such that the effect of the change in probl em solving was controlled for, the impact of Phase 2 assignment on weight change from 6 to 18 months would have to be lessened (Baron & Kenny, 1986). Hypothesis 6b: The effect of extended care conditions on weight change from baseline to 18 months would be mediated by changes in proble m-solving skills from baseline to 18 months. This hypothesis was assessed was examined using a series of analyses. For change in problem-solving abilit ies from baseline to 18 months to act as a mediator of the effect of Phase 2 assignme nt on total percent body weight change from baseline to 18 months, the following conditi ons would need to be met: (1) an ANOVA would have to show that Phase 2 group a ssignment impacted change in problem-solving abilities from baseline to 18 months such that the telephone group as compared to the
35 mail group or the in-person group as compared to the mail group significantly differed in changes in problem solving from baseline to 18 months, (2) t-tests would have to show Phase 2 assignment (i.e., to the telephone vs. mail group or in-person vs. mail group) produced significant differences in weight change from baseline to 18 months, (3) a regression analysis would have to show im provements in problem solving from baseline to 18 months significantly predicted weight change from baseline to 18 months, and (4) when both Phase 2 treatment condition assignment and improvements in problem solving from baseline to 18 months were in a regression model, such that the effect of the change in problem solving was controlled for, the impact of Phase 2 assignment on weight change from baseline to 18 months would have to be lessened (Baron & Kenny, 1986). Hypothesis 7a: The effect of changes in problem-solving skills from baseline to 6 months on weight change from baseline to 6 months would be mediated by Phase 1 adherence. This hypothesis was assessed with a series of re gression analyses. For Phase 1 adherence to act as a mediator of the effect of changes in problem-solving skills from baseline to 6 months on total percent body wei ght change from baseline to 6 months, the following conditions would need to be met: (1 ) a regression analysis would have to show that the change in problem solving from baseline to 6 m onths significantly predicted Phase 1 adherence, (2) a regression analysis would have to show that Phase 1 adherence significantly predicted weight change from baseline to 6 months, (3) a regression would have to show change in problem solving from baseline to 6 months significantly predicted weight change from baseline to 6 months and (4) when both change in problem solving from baseline to 6 months and Phas e 1 adherence were in a regression model,
36 such that the effect of Phase 1 adherence was controlled for, the impact of change in problem solving on weight change would have to be lessened (Baron & Kenny, 1986). Hypothesis 7b: The effect of changes in problem-solving skills from baseline to 18 months on weight change from baseline to 18 months would be mediated by Phase 2 adherence. For Phase 2 adhere nce to act as a mediator of the effect of change in problem-solving skills from baseline to 18 m onths on total percent body weight change from baseline to 18 months, the following c onditions would need to be met: (1) a regression analysis would have to show that the change in problem solving from baseline to 18 months significantly predicted Phase 2 adherence, (2) a regression analysis would have to show that Phase 2 adherence signifi cantly predicted weight change from baseline to 18 months, (3) a regression would have to show change in problem solving from baseline to 18 months significantly predic ted weight change from baseline to 18 months and (4) when both change in problem solvi ng from baseline to 18 months and Phase 2 adherence were in a regression model, such that the effect of Phase 2 adherence was controlled for, the impact of change in pr oblem solving from baseline to 18 months on weight change would have to be lessened (Baron & Kenny, 1986).
37 CHAPTER 3 RESULTS Participant Characteristics Of 1,350 people who made telephone inquiries concerning the study, 559 underwent medical screening for eligibility. Of the 559 who were assessed for eligibility, 261 potential participants were screened out for the following reasons: BP > 140/90 mmHg (35.5%); abnormal lipid values (16.9%); other screen ing visit results (13%); medical conditions, such as self-reported ca rdiac or musculoskeletal problems (12.5%); out of range values on screeni ng visit results, such as gluc ose > 125 mg/dl (10%); out of range BMI (7.3%); and non-medical issues, such as not being able to attend meetings (4.7%). Ultimately, 298 were deemed eligib le to participate and accepted randomization to one of three treatment conditions. Table 31 presents the baseline characteristics of the 298 individuals who began the treatment program. Preliminary Analyses Of the 298 participants w ho were randomized, 297 comple ted the baseline SPSI-R. Analyses showed baseline problem-solving sc ores did not significantly differ by level of income, years of education, or ethnicity ( p s > .05). Of the 298, 278 participants (94%) completed the 6-month evaluation, with 273 of the 278 completing a full evaluation including the SPSI-R (see Table 3 for SPSI-R scale abbreviations). The mean weight loss from baseline to 6 months was 8.3 + 6.0 kg (baseline values were carried forward for missing data in an intent-to-treat, ITT, anal ysis). The mean adherence for Phase 1 was 74%, with participants completing a mean of 119 Phase 1 self-mon itoring records out of
38 a possible 161 records. The mean percent reduction in weight was 8.6 + 6.1%. Percent weight change achieved from baseline to 6 months was > 5% for the majority of the sample (68%), with (41%) of the sample achieving > 10% (see Figure 3-1 for proportions of the sample who achieved weight reductions of various percentages during Phase 1). At this time, 2 of 3 cohorts of particip ants have completed both Phase 1 and Phase 2 of the trial, with the final cohort scheduled to finish in August 2006. Of the 200 participants in cohorts 1 and 2, 173 (87%) returned for the 18-month evaluation. Of those returning for the 18-month evalua tion, 148 were categorized as Phase 2 starters/Non-Phase 1 drop-outs. Phase 2 star ters were defined as participants who did not withdraw from Phase 1, completed more th an half of Phase 1 sessions, and attended at least 1 session during weeks 18-24 of Phase 1 (Note: Information about Phase 2 treatment was discussed during sessions 18-24 of Phase 1). Of the 148 Phase 2 starters, 144 completed a full 18-month evaluation incl uding the SPSI-R. The other 4 Phase 2 starter participants were weighed at 18 months but did not co mplete all necessary questionnaires for full completion of the 18-m onth assessment. Analyses showed that these 144 Phase 2 completers (i.e., those w ho were Phase 2 starters and full 18-month assessment completers) did not differ significan tly in baseline demographics or problem solving from the other 56 participants of cohorts 1 and 2 ( p s > .05). Also, analyses showed no significant differences between Ph ase 2 treatment conditions in regards to demographics, attrition rates, baseline weights, weights at 6 months, baseline problem solving, or problem solving at 6 months (s ee Table 3-3 for characteristics of Phase 2 completers by treatment condition and Table 3-4 for weights of Phase 2 completers across the trial by treatment condition). Thus analyses regarding Phase 2 outcomes will
39 be reported only for those 144 participants w ho actually received the Phase 2 treatment and completed a full evaluation at 18 months. The mean adherence for Phase 2 was 28%, with participants completing a mean of 101 Phase 1 self-monitoring reco rds out of a possible 364 records. The overall mean weight regain for the Phase 2 completers of cohorts 1 and 2 during Phase 2 was 2.0 + 7.5 kg with an overall mean percent loss from baseline to 18 months of 8.6 + 8.8 %. In terms of total percent body weight regain, an ANOV A showed percent weight regain from months 6 to 18 was significan t by Phase 2 group assignment ( F = 3.64, p = .029). Posthoc tests with Sidak corrections showed signifi cantly greater percent weight regain in the mail group as compared to the in-person group ( p = .027), such that from 6 to 18 months participants in the mail group regained over 4% of their initial 6-month weight while participants in the in-pers on group regained less than 0.3%. See Table 6 for percent weight change during Phase 2 by treatment c ondition and see Figure 2 for percentages of Phase 1 weight loss maintained during Phas e 2 as a function of Phase 2 treatment condition. Primary Analyses To address hypothesis 1, (i.e., higher pr e-intervention proble m-solving abilities would be significantly associated with great er total percent body weight change at 6 months), a Pearson product-moment correlation analysis was utilized to examine if preintervention SPSI-R scores were significantly correlated to total percent body weight change from baseline to 6 months. Results showed greater tota l percent body weight change at 6 months was significantly rela ted to baseline Social Problem Solving Summary Score ( n = 277, r = -.133, p = .027). This shows a low (< 2% of variance
40 accounted for), significant association betw een higher pre-intervention problem-solving abilities and greater total percent body weight change during Phase 1. Because baseline Social Problem Solving Summary Score was significantly related to greater total percent body weight change at 6 months, individual SPSI-R scale scores were then assessed to provide further detail into component problem -solving abilities at baseline. Specifically, greater total per cent body weight change at 6 months was significantly related to baseline scor es on Positive Problem Orientation ( n = 277, r = .118, p = .05), Impulsivity/Carelessness Style ( n = 277, r = .123, p = .042), and Avoidance Style ( n = 277, r = .139, p = .021). Lower scores on Impulsivity/ Carelessness Style and Avoidance Style and higher scores on Positive Problem Orientation were associated with greater decr eases in weight from baseline to 6 months. These results revealed certain pre-interventi on problem-solving abilities seem relevant to initial Phase 1 weight change. To address hypothesis 2, (i.e., the wei ght-loss interventi on would result in increased problem-solving abilities from base line to 6 months), a within-group repeated measures analysis of variance (ANOVA) was used to assess if SPSI-R scores changed significantly from baseline to 6 months. We found that there were no significant changes in problem solving from baseline to 6 months ( F = .95, p = .33). See Table 3-6 for SPSIR means and standard deviations for c ohorts 1-3 at baseline and 6 months. It is important to note that a ceiling effect with rega rds to problem-solving scores may have occurred. There were high levels of problem-solving ab ilities within the sample at baseline; thus, little room for impr ovement in SPSI-R scores from baseline to 6 months or baseline to 18 months existed. The mean and standard deviation for baseline
41 Social Problem Solving Summar y Score found in our sample ( n = 297, M = 106.5, SD = 12.9) were somewhat higher as comp ared to a normative sample ( M = 100, SD = 15) for the SPSI-R (DZurilla, Nezu, & Maydeu-O livares, 2002) and the distribution was significantly negatively skewed ( skewness statistic = -.444, SE = .141), such that the majority of the sample had above averag e problem-solving abilities at baseline (see Figure 3-3 for graph of negatively skewed di stribution of baseline SPSIS scores). To address hypothesis 3, (i.e., greater im provements in problem-solving abilities from baseline to 6 months would be associat ed with greater tota l percent body weight change at 6 months), a Pearson product-mome nt correlation analysis was used to assess the relationship between SPSI-R change scor es from baseline to 6 months and total percent body weight change fr om baseline to 6 months. Re sults showed total percent body weight change from baseline to 6 months was significantly re lated to change from baseline to 6 months in Social Problem Solving Summary Score ( n = 273, r = -.223, p = .001). This shows a modest (> 5% of variance accounted fo r), significant association between improvements in problem-solving ab ilities during Phase 1 and greater total percent body weight change during Phase 1. Furthermore, because change from baseline to 6 months in Social Problem Solving Summary Score was significantl y related to greater total pe rcent body weight change at 6 months, individual SPSI-R scale scores were then assessed, revealing that improvements in Positive Problem Orientation, Nega tive Problem Orientation, Decision Making, Solution Implementation and Verifica tion, Rational Problem Solving, and Impulsivity/Carelessness Style we re all significantly related to greater total percent body weight change at 6 months. Results showed that from baseline to 6 months, as scores
42 increased on the following scales weight decreased: Positive Problem Orientation ( n = 273, r = -.152, p = .012) and Rational Problem Solving ( n = 273, r = -.143, p = .018). Results showed that from baseline to 6 m onths, as scores decreased on the following scales weight also decreased: Negative Problem Orientation ( n = 273, r = .155, p = .01) and Impulsivity/Carelessness Style ( n = 273, r = .144, p = .017). This finding that improvements in problem -solving abilities over the first 6 months of treatment were significantly related to Ph ase 1 weight change, suggested the need for further exploration because previous findings from the analyses examining hypothesis 2 revealed no significant changes in SPSI-R sc ores from baseline to 6 months. Thus, analyses were utilized to examine the differe nces in changes in problem-solving abilities from baseline to 6 months in participants who achieved a c linically signif icant amount of weight loss (> 5%) as compared to those who did not achieve a clinically significant amount of weight loss, which may account fo r the discrepancy in findings between the analyses in regards to hypothesis 2 and hypot hesis 3. Indeed, two within-group t-tests showed significant improvements from baselin e to 6 months in problem-solving abilities (SPSIS) in participants who lost 5% or greater of their tota l body weight from baseline to 6 months ( p < .001), while those who lost less than 5% showed a significant decrease ( p < .001) in problem-solving abilities from ba seline to 6 months (see Table 3-7 for SPSIS means and standard deviations by percent weight loss). Hypothesis 4(a), (i.e., greater improvements in problem-solving abilities from 6 to 18 months would be related to greater tota l percent body weight change from 6 to 18 months), was assessed using a Pearson produc t-moment correlation analysis. Change in
43 Social Problem Solving Summ ary Score from 6 to 18 months was not significantly related to total percent body weight change from 6 to 18 months ( p = .23). Hypothesis 4(b), (i.e., greater improveme nts in problem-solving abilities from baseline to 18 months would be related to greater total percent body weight change from baseline to 18 months), was assessed using a Pearson product-moment correlation analysis, which showed change in Social Problem Solving Summary Score from baseline to 18 months was significantly related to tota l percent body weight ch ange from baseline to 18 months ( n = 144, r = -.171, p = .04). This shows a small to modest (< 3% of variance accounted for), signi ficant association between improvements in overall problem-solving abilities over the intervention a nd greater total pe rcent body weight change over the course of the trial. Follo w-up analyses showed weight change from baseline to 18 months was not significantly related to improvements from baseline to 18 months in any other individual SPSI-R scales ( p s > .05). To address hypothesis 5(a), (i.e., problemsolving abilities w ould increase from 6 months to 18 months in the phone Phase 2 group or in-person group but not in the mail group), a 2 (time) x 3 (condition) repeated measures ANOVA was used to assess if Phase 2 treatment condition significantly impacted ch ange in problem-solving (SPSIS) scores from 6 months to 18 months. Results showed no effect for time ( p = .81) or for Phase 2 condition ( p = .64). Thus, this suggests that prob lem solving did not significantly change from 6 to 18 for the sample as a whole or between specific Phase 2 treatment conditions. Furthermore, the Phase 2 treatment conditions did not differ in SPSIS scores at baseline ( p = .35), 6 months ( p = .60), or 18 months ( p = .27). See Table 3-8 for SPSIS means and standard deviations by Phase 2 condition.
44 To address hypothesis 5(b), (i.e., proble m-solving abilities would increase from baseline to 18 months in the phone or in-person Phase 2 groups but not in the mail group), a 2 (time) x 3 (condition) repeated measures ANOVA was used to assess if Phase 2 treatment condition significantly impacted ch ange in problem-solving scores from baseline to 18 months. Results s howed there was no effect by condition ( p = .53) but there was an effect for time ( F = 6.05, p = .015). Thus, these results suggest that improvements in Social Problem Solving Summa ry Score were not significantly different by treatment condition, but rather the whole sample improved in problem-solving ability (see Table 3-9 for means and standard devi ations for the total sample of Phase 2 completers SPSI-R scores across the trial). Follow-up t-tests showed that in addition to improvements in the Social Problem Solving Summary Score, participants scores also improved from baseline to 18 months on Avoidance Style ( n = 144, t = 3.25, p = .001) and Impulsivity/Carelessness ( n = 144, t = 2.02, p = .045). A follow-up 2 (time) x 1 (total sample) repeated measures ANCOVA showed that the significant time effect for change in problem solving (SPSIS) from baseline to 18 months became non-significant when controlli ng for change in problem solving from baseline to 6 months ( p = .235). Thus, the significant improvements in problem solving over the entire trial were acc ounted for by improvements made during the initial 6-month intervention. To address hypothesis 6(a), (i.e., the eff ect of extended care conditions on weight change from months 6 to 18 would be medi ated by changes in problem-solving skills from months 6 to 18), a 2 (time) x 3 (c ondition) repeated measures ANOVA showed the Phase 2 groups did not significan tly differ in change in prob lem-solving abilities from 6
45 to 18 months ( F = .45, p = .64), as shown in the analyses for hypothesis 5a. Therefore, because Phase 2 group assignment (i.e., neither assignment to telephone vs. mail nor inperson vs. mail) did not significantly impact change in problem-solving abilities over Phase 2, Phase 2 improvements in problem solving could not have mediated the impact of Phase 2 assignment on Phase 2 weight cha nge (i.e., in-person ach ieving significantly better Phase 2 weight maintenance as compared to mail). To address hypothesis 6(b), (i.e., the eff ect of extended care conditions on weight change from baseline to 18 months would be mediated by changes in problem-solving skills from baseline to 18 months), a 2 (time) x 3 (condition) repeated measures ANOVA showed the Phase 2 groups did not significan tly differ in change in problem solving abilities from baseline to 18 months ( F = .64, p = .53), as shown in the analyses for hypothesis 5b. Therefore, because Phase 2 gr oup assignment (i.e., neither assignment to telephone vs. mail nor in-person vs. mail) did not significantly impact change in problemsolving abilities from base line to 18 months, improvements in problem solving from baseline to 18 months could not have medi ated the effect of Phase 2 assignment on weight change from baseline to 18 months. To address hypothesis 7(a), (i.e., the eff ect of changes in problem-solving skills from baseline to 6 months on weight change from baseline to 6 months would be mediated by Phase 1 adherence), a series of regression analyses were used. A linear regression showed change in Social Problem Solving Summary Score from baseline to 6 months significantly pred icted Phase 1 adherence ( n = 273, r = .223, p < .001). A subsequent linear regression showed adherence to Phase 1 self-monitoring significantly predicted weight loss from baseline to 6 months ( n = 273, r = .540, p < .001). Next, a
46 linear regression showed change in problem solving (SPSIS) from baseline to 6 months significantly predicted wei ght loss over Phase 1 ( n = 273, r = .223, p < .001). Lastly, a hierarchical regression showed that in a re gression model with onl y change in problem solving as a predictor, the model signi ficantly predicted Phase 1 weight loss ( r = .223, p < .001), such that change in problem solving fr om baseline to 6 mont hs alone significantly predicted Phase 1 weight loss ( Standardized Beta = -.223 p < .001). Next, in a regression model that significantly predicted Phase 1 weight loss ( r = .550, p < .001), which had both change in problem solving from baseline to 6 months and Phase 1 adherence to self-monitoring both predicting to tal percent body weight loss from baseline to 6 months, such that Phase 1 adherence was controlled for, Phase 1 adherence remained a significant predictor of Phase 1 weight loss ( Standardized Beta = -.515 p < .001), while the effect of change in problem solving on Phase 1 weight loss weakened ( Standardized Beta = -.108 p = .038). Thus, the association between improvements in problem solving and total percent body weight loss during Phase 1 was partially mediated by adherence to self-monitoring. Lastly, to address hypothesis 7(b), (i.e., the effect of changes in problem-solving skills from baseline to 18 months on weight change from baseline to 18 months would be mediated by Phase 2 adherence), a regression showed changes in problem solving from baseline to 18 months did not significantly predict Phase 2 adherence to self-monitoring ( r = .147, p = .08), although marginal significance was noted. Therefore, because improvements in problem solving did not sign ificantly predict Phase 2 adherence, Phase 2 adherence could not have mediated the eff ect of improvements in problem solving from baseline to 18 months on weight ch ange from baseline to 18 months.
47 Secondary Analyses Secondary analyses were utilized to furt her assess the relationship between Phase 2 adherence and Phase 2 weight loss. An ANOVA showed Phase 1 adherence did not differ by Phase 2 group assignment ( p = .93). However, a follow-up ANOVA showed Phase 2 adherence did differ by Phase 2 group assignment ( F = 3.08, p = .049). Post-hoc analyses with Sidak corrections showed the telephone group completed significantly more self-monitoring records during Phase 2 as compared to the mail group ( p = .048). The telephone vs. in-person ( p = .76) and the mail vs. in-person groups ( p = .29) did not significantly differ in Phase 2 adherence (see Table 3-10 for number of records completed during Phase 1 and Phase 2 by treatm ent condition). Therefore, because Phase 2 adherence did not differ significantly between the two groups that differed in Phase 2 weight change (i.e., the mail and in-person groups), Phase 2 adherence did not mediate the impact of Phase 2 assignment to mail or in-person conditions on Phase 2 weight change. Follow-up t-test analyses, how ever, showed that assignment to an active personal contact extended care group (telephone or in-p erson) as compared to the mail control group, (who had no personal contact with a health care professional), produced significantly different Phase 2 adherence ( t = 2.41, p = .018), such that the active groups completed more self-monitoring records ( M = 114.4, SD = 98.2) than the mail group ( M = 76.0, SD = 87.1). The analyses also showed that assignment to treatment in one of the active groups produced significan tly different percent body weight change from 6 to 18 months as compared to assignment to the mail group ( t = -2.50, p = .014), such that the active groups experienced si gnificantly less percent body we ight regain from 6 to 18 months ( M = .834, SD = 7.04) than the mail group ( M = 4.15, SD = 7.84). Furthermore, a
48 linear regression analysis showed Phase 2 adherence significantly predicted percent weight change from 6 to 18 months ( n = 144, r = -.416, p < .001). Lastl y, a hierarchical regression showed that in a regression mode l with only Phase 2 assignment to an active or control condition as a pred ictor, the model significantl y predicted Phase 2 weight change ( r = .212, p < .05), such that Phase 2 assignme nt alone significantly predicted weight change from 6 to 18 months ( Standardized Beta = .212 p = .011). Next, in a regression model that significantly pr edicted Phase 2 weight change ( r = .438, p < .001), which had both Phase 2 assignment and Ph ase 2 adherence to self-monitoring both predicting total percen t body weight change from 6 to 18 months, such that Phase 2 adherence was controlled for, Phase 2 adhe rence remained a significant predictor of Phase 2 weight change ( Standardized Beta = -.390 p < .001), while the effect of Phase 2 assignment on Phase 2 weight change weakened such that it became non-significant ( Standardized Beta = .138 p = .077). Thus, the effect of Phase 2 assignment on weight change from 6 to 18 months was mediated by adherence.
49 Table 3-1. Baseline Characteristics ( n = 298) ________________________________________________________________________ Variable M SD % Weight (kg) 96.5 14.9 Age (years) 59.3 6.3 Ethnicity Caucasian 75 African American 21 Hispanic 2 Native American 2 Education < 12 years 37 13-15 years 43 > 16 years 20 Income < $10,000 6 $10,000-$34,999 41 $35,000-$75,000 40 > $75,000 11 unknown 2 ________________________________________________________________________
50 Table 3-2. SPSI-R Scales and Abbreviations ________________________________________________________________________ Scale Abbreviation Positive Problem Orientation (PPO) Negative Problem Orientation (NPO) Problem Definition and Formulation (PDF) (Subscale of RPS) Generation of Alternative Solutions (GAS) (Subscale of RPS) Decision Making (DM) (Subscale of RPS) Solution Implementation and Verification (SIV) (Subscale of RPS) Rational Problem Solving (RPS) Impulsivity/Carelessness Style (ICS) Avoidance Style (AS) Social Problem Solving Su mmary Score (SPSIS) ________________________________________________________________________
51 Table 3-3. Characteristics of Phase 2 Completers by Treatment Condition ( n = 144) ________________________________________________________________________ Telephone In-Person Mail ( n = 43) ( n = 51) ( n = 50) Variable % % % Ethnicity Caucasian 65 82 76 African American 33 16 18 Hispanic 2 2 4 Native American 0 0 2 Education < 12 years 42 43 36 13-15 years 42 51 54 > 16 years 16 6 10 Income < $10,000 5 4 4 $10,000-$34,999 40 47 23 $35,000-$75,000 46 45 51 > $75,000 9 4 20 unknown 0 0 2 ________________________________________________________________________
52 Table 3-4. Weights (kg) of Phase 2 Completers by Treatment Condition ( n = 144) ________________________________________________________________________ Telephone In-Person Mail (n = 43) (n = 51) (n = 50) Weights M SD M SD M SD Baseline 96.3 17.2 98.8 14.9 95.7 13.9 6-Month 87.2 16.7 88.4 14.2 85.2 13.4 18-Month 88.5 18.1 88.8 16.9 88.8 15.2 ________________________________________________________________________
53 Table 3-5. Percent Body Weight Change from 6 to 18 Months by Phase 2 Condition ( n = 144) ________________________________________________________________________ Telephone In-Person Mail (n = 43) (n = 51) (n = 50) M SD M SD M SD Percent Weight Change 1.48 6.78 .292 a 7.27 4.15 b 7.84 ________________________________________________________________________ Note. Means with different subscripts indi cate significant between-groups differences ( p < .05).
54 Table 3-6. Cohorts 1-3 SPSI-R Scores Across Phase 1 ( n = 273) ________________________________________________________________________ Baseline 6 Months SPSI-R Scales M SD M SD PPO 101.8 14.9 101.4 15.8 NPO 92.1 11.1 91.7 11.5 PDF 99.0 15.2 100.2 15.3 GAS 102.0 15.0 102.3 15.9 DM 98.0 14.4 99.0 15.7 SIV 100.1 14.9 100.1 15.1 RPS 99.6 15.0 100.4 15.8 ICS 92.2 13.9 91.0 14.0 AS 93.3 11.8 93.1 12.2 SPSIS 106.5 12.9 107.1 13.4 ________________________________________________________________________
55 Table 3-7. SPSI-S Scores Across Phase 1 by Category of Percent Weight Loss ( n = 273) ________________________________________________________________________ < 5% Wt Loss > 5% Wt Loss ( n = 69) ( n = 204) SPSIS Scores M SD M SD Baseline 109.3 11.5 105.6 13.2 6-Month 106.7 13.1 107.2 13.5 ________________________________________________________________________
56 Table 3-8. SPSI-S Scores Across Tr ial by Phase 2 Treatment Condition ( n = 144) ________________________________________________________________________ Telephone In-Person Mail ( n = 43) ( n = 51) ( n = 50) SPSIS Scores M SD M SD M SD Baseline 107.9 15.3 104.0 11.6 104.9 13.0 6-Month 109.1 13.9 107.1 11.9 106.2 14.6 18-Month 110.2 14.7 107.0 11.3 105.7 14.6 ________________________________________________________________________
57 Table 3-9. Phase 2 Completers SPSI-R Scores Across Trial (n = 144) ________________________________________________________________________ Baseline 6 Months 18 Months SPSI-R Scales M SD M SD M SD PPO 101.1 15.1 101.6 16.2 101.0 16.6 NPO 92.7 10.2 91.2 10.0 91.3 11.0 PDF 98.2 16.1 99.9 16.4 99.7 15.7 GAS 101.2 15.6 102.4 17.1 102.0 16.1 DM 97.3 15.5 99.2 16.7 98.2 16.2 SIV 99.0 15.9 100.4 16.1 99.5 14.6 RPS 98.8 16.1 100.5 17.1 99.8 16.0 ICS 93.5 a 14.5 90.9 14.1 91.4 b 12.9 AS 94.0 a 11.6 92.9 11.4 91.1 b 10.3 SPSIS 105.4 a 13.3 107.4 13.5 107.5 b 13.6 ________________________________________________________________________ Note. Means with different subscripts indi cate significant differences across time periods ( p s < .05).
58 Table 3-10. Phase 2 Completers Self -monitoring by Treatment Condition ( n = 144) ________________________________________________________________________ Telephone In-Person Mail ( n = 43) ( n = 51) ( n = 50) Self-monitoring M SD M SD M SD Phase 1 records 139.4 31.7 138.3 36.1 136.7 30.6 Phase 2 records 123.8 a 103.1 106.4 94.1 76.0 b 87.1 ________________________________________________________________________ Note. Means with different subscripts indi cate significant between-groups differences ( p < .05).
59 Figure 3-1. Proportion of Sample with Differe nt Levels of Phase 1 Weight Change Gained 0 -4.9% 5.0 -9.9 % 10% or g reate r Wei g ht Chan g e Cate g or y 0 10 20 30 40 50 % Pro p ortion of Sam p le with Different Levels of Phase 1 Wei g ht Chan g e 27 27 41 4
60 Percentage of Phase 1 Weight Loss Maintained During Months 6-18 as a Function of Extended Care Condition0 20 40 60 80 100MailTelephoneIn-Person Phase 2 Treatment Condition Percent % Figure 3-2. Percentage of Ph ase 1 Weight Loss Maintain ed During Months 6-18 as a Function of Extended Care Condition 66 86 95
61 Distribution of Baseline Problem-Solving Scores 140 120 100 80 60 40 SPSI: Total Score baseline 40 30 20 10 0 Frequency Mean = 106.54 Std. Dev. = 12.907 N = 297 Figure 3-3. Distribution of Ba seline Problem-Solving Scores Cohorts 1-3
62 CHAPTER 4 DISCUSSION In the present study, we found small e ffects showing baseline problem-solving abilities and change in problem-solving ab ilities from baseline to 6 months were significantly related to total percent body weight loss from baseline to 6 months, and Phase 1 adherence partially mediated the e ffect of improvements in problem solving on Phase 1 weight loss. We found problem-solvi ng scores did not significantly change from baseline to 6 months or 6 to 18 months. Sc ores from baseline to 18 months significantly improved in overall problem solving, accounted for by significant changes made from baseline to 6 months, and showed a small ye t significant association with weight loss over the entire trial. Lastl y, improvements in problem-solving skills from 6 to 18 months or baseline to 18 months di d not mediate the effect of Phase 2 group assignment on weight change. However, we found Phase 2 adherence significantly mediated the effect of Phase 2 assignment on weight change from 6 to 18 months. Some limitations to this study should be noted. First, in terms of problem solving, the ceiling effect due to partic ipants above average levels of problem-solving abilities at baseline may have caused a restriction of ch ange. Thus, this may have limited possible increases in scores across the trial and can perhaps explain the findings that problemsolving abilities did not signifi cantly increase from baseline to 6 months or 6 months to 18 months. The above average problem-sol ving skills at baseline may possibly be accounted for by above average levels of edu cation attainment within the sample, with the majority of the 298 participants at ba seline (63%) having achie ved education levels
63 greater than 12 years. Also, only 2 of 3 c ohorts were utilized in analyses examining Phase 2 outcomes. Therefore, this reduced number of participants could have made it harder to detect significant relationships. Next, although all interventionists participated in regularly scheduled training sessions so as to help ensure standardization of Phase 2 treatment, Phase 2 treatment fidelity was not assessed quantitatively. Thus, complete adherence to Phase 2 problem-solving protocol by each interventionist cannot be confirmed. Lastly, we only assessed adhe rence through one technique: self-monitoring. There are many other self-regulatory skills that are taught to improve adherence to an intervention (e.g., stimulus control, self-re inforcement, and cognitive restructuring), which were not assessed in this study. Furt hermore, some participants records may have been more detailed and accurate than othe rs. Therefore, although research has shown adherence behaviors tend to cluster, such that successful adhere nce to one type of behavior is typically associ ated with successful adheren ce to another (The Diabetes Prevention Program Research Group, 2004), our utilization of onl y self-reported food monitoring records could have limited the full insight that may have been attained by using multiple, quality-controlled measures of adherence. While recognizing the limitations of the study, it is important to emphasize that this studys findings add to the literature in a va riety of ways. First, this study provides knowledge about obesity treatm ent in a rural setting, which is important for a population who are medically underserved, understudied, and impacted by obesity-related diseases (Eberhardt et al. 2001; Economic Research Services, 1993). For example, this study shows that barriers to obesity treatment in rural settings can be overcome such that successful weight losses can be achieved. Therefore, wh ile many high-cost obesity
64 treatment interventions have shown the efficacy of their interventions in university-based settings (Diabetes Preventi on Program Research Group, 2004) this study showed that utilizing County Cooperative Extension Servi ce Offices was an economical and effective way to provide obesity treatment to a popul ation who often have difficulties finding transportation such that they can obtain medical care or preventative services (Economic Research Services, 1993). Furthermore, as detailed below, insights from this study regarding the role of psychologi cal factors (i.e., problem solv ing skills) and adherence to behavioral strategies (i.e., adherence to self-monitoring) in the process of weight management, are helpful in that they may ai d in the development of effective lifestyle interventions targeting rura l populations in the future. Specifically, this study provides insight into psychological factors impacting weight loss in lifestyle interventions, whic h few studies have examined prospectively (Byrne, 2002). Our findings showed that baseline overall problem-solving abilities (SPSIS) and change in overall problem-solvi ng abilities from baseline to 6 months were related to greater initial Phase 1 weight lo sses. Although the findings revealed only a low to modest, positive association between initial problem-solving abilities and initial weight loss, it adds to the scant literature on ps ychological contributors to early success in lifestyle interventions. Furthermore, our findings showed that particular pre-in tervention problemsolving abilities, including Positive Problem Orientation, Impulsivity/Carelessness Style, and Avoidance Style, appear to be related to initial weight loss. This suggests that participants who entered the lifestyle in tervention with a more positive outlook or orientation to problems obtaine d better weight losses from baseline to 6 months. This
65 may have occurred because participants with a greater positive problem orientation may have had a greater tendency to expect a pos itive resolution to their challenges and possessed a greater self-efficacy, which may have enabled them to more actively participate in effective problem-solvi ng strategies (DZur illa & Nezu, 1999). For example, participants with a higher baseli ne positive problem or ientation may have recognized that they would need to put a si gnificant amount of time and effort into the intervention if they wanted to be successful in their weight loss. Thus, they may have put more time into looking at nutri tion labels, scheduling in ex ercise, and completing selfmonitoring records, all of wh ich may have helped them achieve successful weight loss over Phase 1. Furthermore, participants who report ed less impulsivity or carelessness in their baseline problem-solving st rategies obtained be tter initial weight losses. This may be explained in that those participants who were less impulsive/careless, problem solved in a less hurried and more complete, eff ective manner (DZurilla & Nezu, 1999). For example, participants who were able to be less impulsive or carele ss may have been able to make healthier food choices when dining at restaurants or at par ties because they could carefully assess which food choices would be mo st nutritious and fit within their calorie goals, rather than impulsively choosing what initially appealed to them the most (e.g., high-fat, high-calorie foods). Thus, their ab ility to not act impulsively or carelessly in their decision-making while eating out may have significantly impacted their caloric intake, such that they consumed many less ca lories, which may have ultimately aided in their successful weight loss over time. Lastly, our findings suggest that those participants with lower baseline avoidance scores, w ho entered the weight-loss program reporting they were less likely to avoid problems or challenges, were achieved greater weight
66 losses at 6 months. This lack of avoidance may have enabled greater initial weight losses because they may have been less likely to procrastinate, wait passively, or depend on others for a solution when problems arose. Fu rthermore, less avoidant participants may have been more likely to acknowledge and act ively try to resolve problems as necessary (DZurilla & Nezu, 1999). For example, part icipants with low baseline avoidance style scores may have been more likely throughout the intervention to se lf-monitor to try to identify and redress problems that were o ccurring rather than tryi ng to ignore or not actively seek information and solutions to certain problems (e.g., high caloric intake, low caloric expenditure) that may ha ve been hindering their success in weight loss. Thus, our findings may suggest that screening partic ipants problem-solving skills prior to acceptance into the intervention, in terms of their level of overall problem-solving ability, positive problem orientation, impulsivity/carelessness, and avoidance, may provide an indication of who may be more likely to do well in the interventi on and could add in small part to the decision-making process rela ted to who is chosen to participate in the intervention. Next, our findings showed improvements ma de during the initial treatment from baseline to 6 months in Positive Problem Orientation, Negative Problem Orientation, Impulsivity/Carelessness Style, and Rational Pr oblem Solving Style with its subscales of Decision Making and Solution Implementation a nd Verification, were also significantly associated with initial weight loss. Th ese improvements achieved during the lifestyle intervention, which had no formal problem -solving therapy training component, could have been accounted for by the numerous cogniti ve-behavioral strategi es the participants were taught (e.g., how to talk back to negative thoughts, how to break dysfunctional
67 behavior chains, how to make less impulsi ve/more healthy food choices, and how to overcome barriers to particip ating in regular physical activity) that may have subsequently helped them improve in their problem-solving abilities. In addition, it is important to note that while these fi ndings may represent actual improvements in problem-solvi ng abilities and their significant impact on weight change, there is also the possibility that the significan t findings are an artifact of participants biased opinions of their pr oblem-solving skills based on their magnitude of Phase 1 weight loss, which may have impacted their self-report of changes in problem-solving. Thus, participants who lost weight during Phase 1 may have self-reported they had improved in their problem-solving, whereas a pa rticipant who did not lose weight during Phase 1 may have been discouraged by her performance and thus rated herself as not improving or actually decreasing in probl em-solving ability during Phase 1. In terms of the impact of Phase 1 improvements in problem solving on Phase 1 weight change, our findings s uggest that those participants who increased their likelihood to perceive a problem in a positive manner (i.e., perceive the problem as a challenge rather than threat) and decreased their nega tive orientation towards problems (i.e., their likelihood to perceive problems as threats th at are insurmountable) obtained greater weight losses at 6 months. This may have occurred because participants who had a higher positive problem orientation and a lo wer negative problem orientation may have been more likely to appraise a problem as so mething that is normal and solvable and were more likely to put forth the necessary efforts to resolve the problem (DZurilla & Nezu, 1999).
68 Levels of impulsivity and carelessness were again shown to be important in relation to initial weight loss. Our findings suggest that those part icipants who decreased their likelihood towards impulsive or careless problem solving over the first 6 months of treatment obtained better weight losses. Ag ain, this most likely occurred because they were able to cope with problems in a mo re functional, less na rrow or hasty manner. Lastly, our findings suggest that thos e participants who im proved their rational problem solving, particularly their decisi on-making and solution implementation and verification, obtained better weig ht losses at 6 months. This may be explained in that those participants who improved their ability to problem solve in a deliberate, organized manner were better able to a ppropriately identify and employ effective solutions to their problems (DZurilla & Nezu, 1999). Thus, our st udy showed the importance of problemsolving abilities on initia l weight-loss during a lifestyle intervention. Again it is important to note that im provements in problem-solving abilities during initial treatment were small (i.e, ga ins of only a few points) and accounted for only 5% of the variance in terms of weight lo ss over Phase 1. Thus, 95% of the variance was not accounted for by changes in problem solving but rather other factors that impacted weight change. Factors that may have been relevant in influencing weight change include: adherence to calorie and phys ical activity goals, ut ilization of social support, utilization of stress management skills, and utilization of other cognitivebehavioral strategies taught over the course of Phase 1. Furthermore, we found that problem-solv ing abilities improved from baseline to 18 months in overall problem solving ( SPSIS), Avoidance Style, and Impulsivity/ Carelessness. Although the significance was accounted for by Phase 1 improvements,
69 these findings again suggested a relation be tween problem-solving abilities and weight loss. There was a significant association between improvements in overall problem solving over the course of the trial and weight loss from baseline to 18 months. Thus, participants who improved in their general problem solving during the trial, and most importantly during Phase 1, obtaine d greater weight losses over the course of the trial. This may be explained in that those part icipants who improved their overall problemsolving may have engaged in more construc tive or effective problem solving, which may have enabled them to overcome barriers to long-term weight management. Next, our findings showed that the imp act of improvements in problem-solving abilities on greater Phase 1 weight loss was partially mediated by adherence to Phase 1 self-monitoring. Therefore, this suggests th at the benefit of part icipants improvements in Phase 1 problem solving was partially acc ounted for by their bette r adherence to selfmonitoring. This finding, which highlight s the importance of adherence to selfmonitoring to successful outcomes during weig ht-loss interventions, has been strongly supported by the weight-loss literature that emphasizes the value of self-monitoring on weight management (Sarwer & Wadden, 1999). Self-monitoring has even been referred to as the single most effective technique in behavioral treatment (Perri, Nezu, & Viegener, 1992), in that it f acilitates a heightened sense of awareness about daily food consumption, which is essential when trying to decrease caloric c onsumption so as to initiate or maintain weight-loss. Furthermore, self-regulation theory (Kan fer, 1970) provides addi tional insight into the benefits of completing daily self-monitori ng records when attempting to lose weight. Self-regulation, which refers to the process by which an individual tries to exert control
70 over his/her behavior/cogniti ons/environment, consists of three stages: (1) selfmonitoring, (2) self-evaluation, and (3) self -reinforcement. The first stage, selfmonitoring, entails deliberately and carefully attending to ones ow n behavior (Kanfer & Goldstein, 1980). This is particularly important in obesity treatment, as selfmonitoring of eating enable s participants to evaluate a vari ety of issues related to caloric intake that they may never had been awar e of previously, including: how many calories they are consuming each day, what types of foods they are consuming, and what quantities of food they are consuming. Next, through the self-evaluation stage, participants are able to utilize the informati on they learned about themselves in the selfmonitoring stage and judge how their behaviors compare to the behaviors they think they ought to be doing, which are also known as performance criteria or standards. In terms of obesity treatment, self-monitoring reco rds allow participants to recognize if they are meeting their individual daily calorie goals if they are coming close to meeting their goals, or if they are consuming many more calories than their daily goals allow. Furthermore, they can identify patterns of eating that may be bene ficial (i.e., eating a healthy breakfast) or that may be hindering (i.e., consuming large amounts of saturated fats each day) their expected weight loss. Fi nally, the third stage of self-reinforcement is the individuals reaction to their self-observatio n and self-judgment. In relation to weight loss, this may mean that if a participant se lf-monitors her eating, id entifies that she is almost always eating within her daily calorie goal limit, and subsequently loses weight, she will be more likely to respond positively and be motivated to continue with her current behaviors. If a participant recogni zes from her self-monitoring that she is not meeting her goals regularly and is subseque ntly gaining weight, she may use this to
71 change her behaviors such that she may be more likely to meet her expected standards and lose or maintain her weight. This exemplifies the importance of daily selfmonitoring in that eating behavi ors can not be judged and modi fied, so as to stop weight regain, if deliberate efforts are not first made to vigilantly monitor eating habits. Thus, continued self-monitoring, which facilitates self-regulation of eating behaviors, can be a powerful strategy for long-term weight management. Our findings again showed the importance of self-monitoring, such that Phase 2 adherence to self-monitoring mediated the imp act of Phase 2 assignment (to an active vs. control condition) on Phase 2 weight change. The great importance of self-monitoring during Phase 2 on weight management during mont hs 6 to 18 of the trial can be explained once more within the context of self-regulation theory. Du ring Phase 2, participants who continued to monitor their eating behavior were able to continue judging their eating habits against the standards they were taught during the intervention (i.e., proper nutrition, appropriate caloric in take vs. caloric expenditure), and modify their behaviors as necessary to fight weight regain and con tinue losing or maintain their weights. A participant who successfully lost weight dur ing Phase 1 with self-monitoring but chose not to self-monitor her eating behaviors dur ing Phase 2 when trying to maintain her weight, most likely had a much more difficu lt time being aware of her actual caloric intake and thus would have been unable to as accurately judge how closely her eating behaviors matched her standards and would have become much more susceptible to post-treatment weight regain. Therefore, our findings suggest that greater adherence to Phase 2 self-monitoring, which may have help ed keep participants on track for
72 effective self-regulation of caloric intake, predicted less weight regain from 6 to 18 months. Thus, the question becomes: Why would participants have chosen to stop utilizing self-monitoring skills that were so beneficial to their initial weight loss during Phase 1? It is here that the maintenance problem of obesity treatment becomes relevant and the reason why many researchers believe in a mo re continuous-care approach to obesity treatment becomes explicated. First, participants perceptions of the benefits of continued practice of behavioral strategies learned during the initial treatment often becomes lessened during the weight maintena nce period because goals shift from losing weight, which is often very motivating for part icipants, to then trying to only maintain or stay at the same weight. Furthermore, staying at the sa me weight, or battling weight regain, after initial treatment actually beco mes harder due to numerous physiological (e.g., decreased energy needs, decreased meta bolic rate), psychological (e.g., unrealistic expectations, lack of reinfo rcement for maintenance), and environmental factors (e.g., constant exposure to low-co st, highly palatable, energy-de nse foods) that actually prime participants for weight regain after initial tr eatment (Perri & Corsica, 2002). Therefore, it appears that while teaching participants ski lls such as self-mon itoring, goal-setting, and problem solving to overcome barriers is impor tant for initial weight loss, it seems it becomes imperative that partic ipants then be continually gui ded and encouraged to utilize those skills, especially when they may be losing motivation and perceiving higher costs as opposed to benefits in regards to maintain ing healthy behaviors, so as to combat the almost inevitable post-treatment weight regain.
73 Thus, this need for assisti ng participants in the conti nued utilization of important behavioral strategies, lends further creden ce to the idea of deve loping a continuous-care approach to weight manageme nt, where participants have extended contact with health care providers so as to facil itate more effective long-term weight management. This idea has also been supported by research that has shown extended care models of weight management are in fact more successful than short-term approaches (Wadden, Butryn, & Byrne, 2004; Wadden, Crerand, & Brook, 2005; Wing, 2002). Specifically, one study showed that participants in an extended care problem-solving therapy condition attained significantly better weight lo sses than a behavioral treat ment group who had no follow-up after initial treatment. Their study also show ed an association between problem solving and adherence such that problem solving impacted adherence, which was a partial mediator of long-term weight main tenance (Perri et al., 2001). Our findings also showed that Phas e 2 group assignment did not impact participants problem-solving ab ilities from 6 to 18 months. Furthermore, our findings showed the entire sample did not significan tly change or improve in problem-solving during Phase 2, which may be accounted for by the above average baseline problemsolving scores and the ceiling effect, which was discussed earlier in this section, that left little room for additional improvement during Phase 2 treatment. Phase 2 group assignment, however, did impact weight chan ge from 6 to 18 months, such that those participants in the active conditions had si gnificantly less weight regain than those participants in the mail-control group. Sp ecifically, the mechanism by which the active conditions impacted participants, such that th ey achieved better weig ht maintenance, was through enhanced Phase 2 self-monitoring. Th erefore, this suggest s that the regular
74 contact with a health care professional, w ho helped participants problem solve to overcome barriers so as to maintain healthy lifestyle behaviors such as self-monitoring, significantly impacted weight change from 6 to 18 months for the participants in the active conditions such that they had less rega in as compared to those in the mail group who had no continued contact with a health care professional. This finding is supported by research that has shown therapist-led follow-up care enhances weight maintenance (Perri, McAllister, Gange, Jordan, McAdoo, & Nezu, 1988). It should be noted that due to the desi gn of this study, which lacked a control group who had contact with a health care pr ofessional but received no PST, our findings cannot distinguish that the significant compone nt of the active groups was the PST rather contact with an interventioni st. Therefore, our findings suggest that participants improvements in problem-solv ing abilities over a follow-up period was not as important as being in an active follow-up condition where they had regular contact with an interventionist, whose training in how to utilize the problem-solving process appropriately so as to aid, guide, motivat e, and encourage participants may have impacted the participants such that they ma intained better conti nued adherence to selfmonitoring, as compared to those in the control group. In conclusion, the results of this study suggest that in a rural population, baseline problem-solving abilities and improvements in problem-solving abilities over the first 6 months of treatment are signi ficantly related to initial weight loss in a lifestyle intervention, accounting for a small (2-5%) amount of the variance. Furthermore, adherence to self-monitoring over the first 6 months partially medi ates the effect of improvements in initial problem-solving abil ities on initial weight loss. Adherence
75 during Phase 2 of the intervention is esp ecially important fo r long-term weight management from months 6 to 18 of the interv ention as it mediates the effect of Phase 2 assignment on Phase 2 weight change. Therefor e, findings from this study suggest that continued contact after initial treatment, so as to facilitate continued self-regulation through maintenance of healthy lifestyle behavi ors/behavioral strate gies such as selfmonitoring, is helpful for successful long-te rm weight management. Thus, future interventions may benefit by utilizing a more continuous-care problem-solving approach to obesity treatment, which w ould utilize problem-solving based extended treatment, where participants have personal contact with health care professionals who have been trained in PST facilitation. The trained health care professionals could then best assist participants in overcoming the numerous and chronic challenges associated with weight management and target important factors such as self-monitoring and other behavioral strategies associated with succe ssful long-term weight management such as: regular weighing, eating a low-fa t, low-calorie diet and part icipating in regular daily physical activity (Hill, Wyatt, Phelan, & Wi ng, 2005). Furthermore, important goals for future studies in obesity treatment may be to try to identify what dose of extended care treatment is necessary for optimal maintenance of lost weight, while also considering cost-effectiveness (Wing, 2003). Ultimately, th e goal is to develop the most effective, economical weight management intervention, with the potential fo r wide dissemination, to successfully combat the epidemic of obesity in this country.
76 LIST OF REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions Thousand Oaks: Sage. Arean, P. A., Perri, M. G., Nezu, A. M., Sche in, R. L., Christopher, F., & Joseph, T. X. (1993). Comparative effectiveness of social problem-solving therapy and reminiscence therapy as treatments for depression in older adults. Journal of Consulting and Clinical Psychology, 61 1003-1010. Arterburn, D. E., Maciejewski, M. L., & Tsev at, J. (2005). Impact of morbid obesity on medical expenditures in adults. International Journal of Obesity, 29 334-339. Baron, R. M., & Kenny, D. A. (1986). The mode rator-mediator variab le distinction in social psychological research: Conceptual, strategic, and statis tical considerations. Journal of Personality and Social Psychology, 51 1173-1182. Bond, D. S., Lyle, R. M., Tappe, M. K., Seehafer, R. S., & DZurilla, T. J. (2002). Moderate aerobic exercise, TAi Chi, and social problem-solving ability in relation to psychological stress. International Journal of Stress Management, 9 329-343. Bray, G. A. (2002). Drug treatment of obesity. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 317-338). New York: Guilford Press. Byrne, S. M. (2002). Psychological aspects of weight maintenance and relapse in obesity. Journal of Psychosomatic Research, 53 1029-1036. Cooper, R., Cutler, J., Desvigne-Nickens, P., Fortmann, S. P., Friedman, L., Havlik, R., et al. (2000). Trends and disparities in cor onary heart disease, stroke, and other cardiovascular diseases in the United Stat es: Findings of the National Conference on Cardiovascular Disease Prevention. Circulation, 102 3137-3147. Cooper, Z., & Fairburn, C. G. (2001). A ne w cognitive behavioral approach to the treatment of obesity. Behavior Research and Therapy, 39 499-511. Davey, G. C. L., & Levy, S. (1999). Internal statements associated with catastrophic worrying. Personality and Individual Differences, 26 21-32. Douglass, H. M. (2000). Problem-solving abi lity, parasuicidal behavior and borderline traits in women receiving skills-based therapy. Dissertation Abstracts International: Section B: The Sciences & Engineering, 61 2197.
77 Dudley-Brown, S. (2002). Prevention of psychological distress in persons with inflammatory bowel disease. Issues in Mental Health Nursing, 23 403-422. Dugas, M. J., Gagnon, F., Ladouceur, R., & Fr eeston, M. H. (1998). Generalized anxiety disorder: A preliminary test of a conceptual model. Behaviour Research and Therapy, 36 215-226. DZurilla T. J., & Goldfried, M. R. (1971). Problem solving and be havior modification. Journal of Abnormal Psychology, 78 107-126. DZurilla T. J., & Nezu, A. M. (1982). Social problem solving in adults. In P. C. Kendall (Ed.), Advances in cognitive-be havioral research and therapy (pp. 201-274). New York: Academic Press. DZurilla T. J., & Nezu, A. M. (1999). Problem-solving therapy: A social competence approach to clinical intervention (2nd ed.). New York: Springer. DZurilla T. J., Nezu, A. M., & Maydeu-Olivares, A. (2002). Manual for the social problem-solving inventory-revised. (SPSI-R) North Tonawanda, NY: Multi-Health Systems, Inc. Eberhardt, M. S., Ingram, D. D., & Makuc, D. M. (2001). Urban and rural chartbook Health, United States, 2001. Hyattsville, MD : National Center for Health Statistics. Economic Research Services. (1993). Rural conditions and trends Washington, D. C.: U.S. Department of Agriculture. Elliot, T. R., Shewchuck, R. M., Miller, D. M., & Richards, J. S. (2001). Profiles in problem-solving: Psychological well-b eing and distress among persons with diabetes mellitus. Journal of Clinical Psychol ogy in Medical Settings, 8 283-291. Field, A. E., Barnoya, J., & Colditz, G. A. (2002). Epidemiology and health and economic consequences of obesity. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 3-18). New York: Guilford Press. Findeis, J. L., Henry, M., Hirschl, T. A., Le wis, W., Ortega-Sanchez, I., & Peine, E. (2001). Welfare reform in rural America: A review of current research Columbia, MO: Rural Policy Research Institute. Flegal, K. M., Graubard, B. I., Williamson, D. F., & Gail, M. H. (2005). Excess deaths associated with underweight, overweight, and obesity. JAMA, 293 1861-1867. Foster, G. D., Wadden, T. A., Vogt, R. A., & Brewer, G. (1997). What is a reasonable weight loss? Patients expectations and ev aluations of obesity treatment outcomes. Journal of Consulting and Clinical Psychology, 65 79-85. Frye, A. A., & Goodman, S. H. (2000). Whic h social problem-solving components buffer depression in adolescent girls? Cognitive Therapy and Research, 24 637-650.
78 Hauptman, J., Lucas, C., Boldrin, M. N., Collins, H., & Segal, K. R. (2000). Orlistat in the long-term treatment of obes ity in primary care settings. Archives of Family Medicine, 9 160-167. Hill, J. O., Wyatt, H., Phelan, S., & Wing, R. R. (2005). The National Weight Control Registry: Is it useful in helping deal with our obesity epidemic? Journal of Nutrition Education and Behavior, 37 206-210. Hill-Briggs, F. (2003). Problem-solving in diab etes self-management: A model of chronic illness self-management behavior. Annals of Behavioral Medicine, 25 182-193. Institute of Medicine. (1995). Weighing the options: Criteri a for evaluating weightmanagement programs Washington, D. C.: National Academy Press. Jeffery, R. W., Wing, R. R., Thorson, C., & Bu rton, L. R. (1998). Use of personal trainers and financial incentives to increase exerci se in a behavioral weight-loss program. Journal of Consulting and Clinical Psychology, 66 777-783. Kanfer, F. H. (1970). Self-regulation: Researc h, issues, and speculati ons. In C. Neuringer & J. L. Michael (Eds.), Behavior modification in clinical psychology New York: Appleton-Century-Crofts. Kanfer, F. H. (1980). Self-management methods. In F. H. Kanfer & A. P. Goldstein (Eds.), Helping People Change (pp. 334-389). New York: Pergamon Press. Kayman, S., Bruvold, W., & Stern, J. S. ( 1990). Maintenance and relapse after weight loss in women: Behavioral aspects. American Journal of Clinical Nutrition, 52 800-807. Kerns, R. D., Rosenberg, R., & Otis, J. D. (2002). Self-appraised problem solving and pain relevant social support as predic tors of the experience of chronic pain. Annals of Behavioral Medicine, 24 100-105. Klein, S., Burke, L. E., Bray, G. A., Blair, S., Allison, D. B., Pi-Sunyer, X., et al. (2004). Clinical implications of obesity with specific focus on cardiovascular disease: A statement for professionals from the American Heart Association Council on Nutrition, Physical Activity, and Metabolism. Circulation, 110 2952-2967. Kneebone, I. I., & Martin, P. R. (2003). Copi ng and caregivers of people with dementia. British Journal of Health Psychology, 8 1-18. Kramer, F. M., Jeffery, R. W., Forster, J. L., & Snell, M. K. (1989). Long-term follow-up of behavioral treatment for obesity: Patterns of weight gain among men and women. International Journal of Obesity, 13 124-136. Latner, J. D., Stunkard, A. J., Wilson, G. T., Zelitch, D. S., & Labouvie, E. (2000). Effective long-term treatment of obesity: A continuing care model. International Journal of Obesity, 24 893-898.
79 Marx, E. M., & Schulze, C. C. (1991). Inte rpersonal problem-solving in depressed students. Journal of Clinical Psychology, 47 361-367. Maydeu-Olivares, A., & DZurilla, T. J. ( 1996). A factor analytic study of the social problem solving inventory: An in tegration of theory and data. Cognitive Therapy and Research, 20 115-133. Miner, R. C., & Dowd, E. T. (1996). An empi rical test of the problem solving model of depression and its application to th e prediction of anxiety and anger. Counselling Psychology Quarterly, 9 163-176. Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes of death in the United States, 2000. JAMA, 291 1238-1245. National Heart, Lung, and Blood Institute (1998). Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: the evidence report. Obesity Research, 6 51S-209S. Nezu, A. M. (1985). Differences in psyc hological distress between effective and ineffective problem solvers. Journal of Counse ling Psychology, 32 135-138. Nezu, A. M. (2004). Problem solvi ng and behavior therapy revisited. Behavior Therapy, 35 1-33. Nezu, A. M., & DZurilla, T. J. (1989). So cial problem solving and negative affective conditions. In P.C. Kenda ll, & D. Watson (Eds.), Anxiety and depression: Distinctive and o verlapping features San Diego, CA: Academic Press, Inc. Nezu, A. M., & Nezu, C. M. (2001). Problem Solving Therapy. Journal of Psychotherapy Integration, 11 187-205. Nezu, A. M., Nezu, C. M., Friedman, S. H., Faddis, S., & Houts, P. S. (1998). Coping with cancer: A problem solving approach. Washington, DC: American Psychological Association. Nezu, A. M., Nezu, C. M., McClure, K. S., Fe lgoise, S. H., & Houts, P. S. (2003). Project Genesis: Assessing the efficacy of problem -solving therapy for distressed adult cancer patients. Journal of Consulting and Clinical Psychology, 71 1036-1048. Nezu, A. M., Nezu, C. M., & Perri, M. G. (1989). Problem-solving therapy for Depression New York: Wiley. Nezu, A. M., & Perri, M. G. (1989). Soci al problem-solving therapy for unipolar depression: An initial dismantling investigation. Journal of Consulting & Clinical Psychology, 57 408-413.
80 Ogden, C. L., Carroll, M. D., Curtin, L. R., Mc Dowell, M. A., Tabak, C. J., & Flegal, K. M. (2002). Prevalence of overweight and obesity in the United States, 1999-2004. Journal of the American Medical Association, 295 1549-1555. Pakenham, K. I. (2001). Coping with multiple sclerosis: Development of a measure. Psychology, Health, & Medicine, 6 411-428. Pearson, T. A. (1996). Socioeconomic stat us and cardiovascular disease in rural populations. In: Report of the Conference on Socioeconomic Status and Cardiovascular Disease and Health (pp. 101-108). Bethesda, MD: National Institutes of Health, National Heart, Lung, and Blood Institute. Pearson, T. A., & Lewis, C. (1998). Rural ep idemiology: Insights from a rural population laboratory. American Journal of Epidemiology, 148 949-957. Perri, M. G., & Corsica, J. A. (2002). Impr oving the maintenance of weight lost in behavioral treatment of obesity. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 357-379). New York: Guilford Press. Perri, M. G., McAllister, D. A., Gange, J. J., Jordan, R. C., McAdoo, W. G., & Nezu, A. M. (1988). Effects of four maintenance programs on the long-term management of obesity. Journal of Consulting and Clinical Psychology, 56 629-634. Perri, M. G., Nezu, A. M., McKelvey, W. F., Shermer, R. L., Renjilian, D. A., & Viegener, B. J. (2001). Relapse prevention training and problem-solving therapy in the long-term management of obesity. Journal of Consulting and Clinical Psychology, 69 722-726. Perri, M. G., Nezu, A. M., & Viegener, B. J. (1992). Improving the long-term management of obesity: Theory, re search, and clinical guidelines New York: Wiley. Perri, M. G., Shapiro, R. M., Ludwig, W. W., Twentyman, C. T., & McAdoo, W. G. (1984). Maintenance strategies for the tr eatment of obesity: An evaluation of relapse prevention training and post tr eatment contact by mail and telephone. Journal of Consulting and Clinical Psychology, 52 404-413. Platt, J. J., Scura, W. C., & Hannon, J. R. (1973). Problem-solving thinking of youthful incarcerated heroin addicts. Journal of Community Psychology, 1 278-281. Platt, J. J., & Spivack, G. (1972). Social competence and effective problem solving in psychiatric patients. Journal of Clinical Psychology, 28 3-5. Platt, J. J., & Spivack, G. (1973). Studies in problem-solv ing thinking of psychiatric patients: Patient-control differences a nd factorial structur e of problem-solving thinking. Proceedings, 81st Annual Convention of the American Psychological Association, 8 461-462.
81 Platt, J. J., & Spivack, G. (1975). Manual for the means-ends problem-solving procedures (MEPS): A measure of interp ersonal cognitive prob lem-solving skills Philadelphia: Hahnemann Community Me ntal Health Retardation Center. Platt, J. J., Spivack, G., Altman, N., Altm an, D., & Peizer, S. B. (1974). Adolescent problem solving thinking. Journal of Consulting and Clinical Psychology, 42 787793. Podnos, Y. D., Jimenez, J. C., Wilson, S. E., Stevens, C. M., & Nguyen, N. T. (2003). Complications after laparoscopic ga stric bypass: A review of 3464 cases. Archives of Surgery, 138 957-961. Priester, M. J., & Clum, G. A. (1993). Percei ved problem-solving ability as a predictor of depression, hopelessness, and suicide ideation in a co llege population. Journal of Counseling Psychology, 40 79-85. Puhl, R. & Brownell, K. D. (2001). Bias, discrimination, and obesity. Obesity Research, 9 788-805. Ricketts, T. C., Johnson-Webb, K. D., & Taylor, P. (1998). Definitions of rural: A handbook for health policy makers and researchers Chapel Hill, NC: Federal Office of Rural Health Policy. Sarwer, D. B., & Wadden, T. A. (1999). The tr eatment of obesity: Whats new, whats recommended. Journal of Womens Health & Gender-Based Medicine, 8 483-493. Schootman, M. & Fuortes, L. J. (1999). Breas t and cervical carcinoma: The correlation of activity limitations and rurality with scr eening, disease incidence, and mortality. Cancer, 86 1087-1094. Sheehy, N., & OConnor, R. C. (2002). Cognitive style and suicidal behavior: Implications for therapeutic interventi on, research lacunae, and priorities. British Journal of Guidance & Counseling, 30 353-362. Sjostrom, C. D., Lissner, L., Wedel, H., & Sj ostrom, L. (1999). Reduction in incidence of diabetes, hypertension, and lipid disturbances after intentional we ight loss induced by bariatric surgery: Th e SOS Intervention Study. Obesity Research, 7 477-484. Sjostrom, L., Lindroos, A. K., Peltonen, M., To rgerson, J., Bouchard, C., Carlsson, B., et al. (2004). Lifestyle, diabet es, and cardiovascular ri sk factors 10 years after bariatric surgery. The New England Journal of Medicine, 351 2683-2693. Smith, I. G., & Goulder, M. A. (2001). Random ized placebo-controlled trial of long-term treatment with sibutramine in mild to moderate obesity. Journal of Family Practice, 50 505-512. Spivack, G., Platt, J. J., & Shure, M. B. (1976). The problem-solving approach to adjustment San Francisco: Jossey-Bass.
82 Stein, C. J., & Colditz, G. A. (2004). The epidemic of obesity. The Journal of Clinical Endocrinology& Metabolism, 89 2522-2525. The Diabetes Prevention Program Resear ch Group. (2004). Achieving weight and activity goals among Diabet es Prevention Program lif estyle participants. Obesity Research, 12 1426-1435. U.S. Census Bureau. (2002). Census 2000 redistricting data (Public law 94-171) summary file: 2000 census of population and housing Washington, D.C.: U.S. Department of Commerce. U.S. Department of Health and Human Services. (2001). The Surgeon Generals call to action to prevent and decrease overweight and obesity Rockville, MD: U.S. Department of Health and Human Services Public Health Service, Office of the Surgeon General. Available from: U.S. GPO, Washington. U.S. Department of Health and Human Se rvices. (2002). List of designated primary medical care, mental health, and dental health professional shortage areas. Federal Register, Vol. 67 No. 34, February 20, 2002. Wadden, T. A., Butryn, M. L., & Byrne, K. J. (2004). Efficacy of lifestyle modification for long-term weight control. Obesity Research, 12 151S-162S. Wadden, T. A., Crerand, C. E., & Brock, J. (2005). Behavioral tr eatment of obesity. Psychiatric Clinics of North America 28 51-70. Wadden, T. A., Womble, L. G., Stunkard, A. J., & Anderson, D. A. (2002). Psychosocial consequences of obesity and weight loss. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 144-169). New York: Guilford Press. Williamson, D. F. (1999). The prevention of obesity. The New England Journal of Medicine, 341 1140-1141. Wilson, G. T. (1994). Behavioral treatmen t of obesity: Thirty years and counting. Advances in Behaviour Research and Therapy, 16 31-75. Wing, R. R. (2002). Behavioral we ight control. In T. A. Wadden & A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 301-316). New York: Guilford Press. Wing, R. R. (2003). Behavioral interventions for obesity: Recognizing our progress and future challenges. Obesity Research, 11 3S-6S. Wing, R. R., & Jeffery, R. W. (1999). Benefits of recruiting participan ts with friends and increasing social support for we ight loss and maintenance. Journal of Consulting and Clinical Psychology, 67 132-138. Zebb, B. J., & Beck, J. G. (1998). Worry versus anxiety: Is there really a difference? Behavior Modification, 22 45-61.
83 BIOGRAPHICAL SKETCH Mary Murawski graduated summa cum laude from Saint Anselm College in 2002, where she received a Bachelor of Arts in psyc hology. She received a Master of Sciences in clinical psychology from the University of Florida in 2004 and is currently working at the University of Florida towards her Ph.D. in clinical psychology, focusing in the area of health psychology. She will begin her clinical internship with the Boston VA Consortium in September 2006.