RESPONSE OF AFRICAN-AMERICAN AN D CAUCASIAN WOMEN IN A RURAL SETTING TO A LIFESTYLE IN TERVENTION FOR OBESITY By KATIE ALLISON RICKEL 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 2008
2007 Katie Allison Rickel
To my mother, my father, and my sister
iv ACKNOWLEDGMENTS I would like to acknowledge my mother, Jody Rickel, for her extraordinary levels of patience, love, and dedication to my h ealth and well-being. I would also like to acknowledge my father, David Rickel, for instil ling in me a passion to thrive and succeed in whatever I choose to do. Additionally, I w ould like to thank my sister, Emily Rickel, for her constant support of, and enthus iasm about, all of my endeavors. Finally, this dissertation would not have been possible without the guidance provided by Dr. Michael Perri, who served as a knowledgeable and caring mentor throughout my gradua te school training.
v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION......................................................................................................11 Relationship between Obesity, Level of Urbanization, and Race/Ethnicity..............12 Consequences of Obesity in Af rican Americans and Caucasians..............................13 Lifestyle Interventions: Impact of Behavioral Factors...............................................13 Dietary Intake......................................................................................................13 Physical Activity.................................................................................................18 Physiological Responses to Lifestyle Treatment........................................................22 Weight loss..........................................................................................................22 Improvements in Metabolic Risk Factors............................................................23 Improvements in Cardiorespiratory Fitness........................................................24 Role of Social Support in Weight-loss Interventions.................................................25 Racial/ethnic Differences in Patterns of Weight Change...........................................25 Aims and Hypotheses of Current Study.....................................................................28 2 RESEARCH METHODS AND PROCEDURES......................................................30 Parent Study: Treatment of Obesity in Underserved Rural Settings (TOURS).........30 Participating County Cooperative Extension Service (CES) Offices.........................30 Participants.................................................................................................................31 Procedures...................................................................................................................32 Recruitment and Screening..................................................................................32 Data Collection....................................................................................................32 Intervention..........................................................................................................34 Statistical Procedures and Analyses....................................................................36
vi 3 RESULTS...................................................................................................................39 Recruitment and Screening.........................................................................................39 Participants Starting a nd Completing Phase I............................................................39 Baseline Characteristics..............................................................................................40 Primary Aim: Changes in Body Weight.....................................................................40 Secondary Aim: Effect of Extended Care Program within each Racial/Ethnic Group......................................................................................................................41 Tertiary Aim: Changes in Selected Outcomes............................................................42 Changes during lifestyle in tervention (Months 0 6)..................................42 Outcomes during extended care (Months 6 18).........................................46 4 DISCUSSION.............................................................................................................64 Main Findings.............................................................................................................64 Comparison to Previous Studies.................................................................................65 Change in Energy Balance as a Mediat or for Discrepant Weight Outcome..............67 Implications for Clinical Services..............................................................................69 Limitations and Strengths...........................................................................................70 Future Directions........................................................................................................73 Summary and Conclusions.........................................................................................74 LIST OF REFERENCES...................................................................................................75 BIOGRAPHICAL SKETCH.............................................................................................84
vii LIST OF TABLES Table page 3-1 Reasons for exclusi on by racial/ethnic group...........................................................50 3-2 Participant demographics at baseline.......................................................................51 3-3 Weight-related outcomes of African-A merican and Caucasian participants...........52 3-4 Percentage of weight change from baseline.............................................................52 3-5 Systolic and diastolic blood pre ssure during lifestyle intervention..........................53 3-6 Triglycerides and LDL-choleste rol during lifestyle intervention............................53 3-7 Hemoglobin A1c and C-Reactive pr otein during lifestyle intervention..................53 3-8 Performance on 6-minute walk test (6MWT) and weekly energy expenditure during lifestyle intervention.....................................................................................54 3-9 Daily caloric and saturated fat intake during lifestyle intervention.........................54 3-10 Systolic and diastolic bl ood pressure during extended care.....................................54 3-11 Triglycerides and LDL-chol esterol during extended care........................................55 3-12 Hemoglobin A1c and C-Reactiv e protein during extended care..............................55 3-13 Performance on 6-minute walk test (6MWT) and weekly energy expenditure during extended care................................................................................................56 3-14 Daily caloric and saturated fat intake during extended care....................................56
viii LIST OF FIGURES Figure page 3-1 Screening and assessment of i ndividuals for the parent study.................................57 3-2 Categories of weight lost from baseline...................................................................58 3-3 Weight regain of African-American a nd Caucasian participants assigned to control versus inte rvention conditions (EMM + SE)............................58 3-4 Systolic blood pressure at Months 0, 6, and 18 (EMM + SE)..................................59 3-5 Diastolic blood pressure at Months 0, 6, and 18 (EMM + SE)................................59 3-6 Triglycerides at Months 0, 6, and 18 (EMM + SE)..................................................60 3-7 LDL-cholesterol at Months 0, 6, and 18 (EMM + SE)............................................60 3-8 Hemoglobin A1c at Mont hs 0, 6, and 18 (EMM + SE)...........................................61 3-9 C-reactive protein at Mo nths 0, 6, and 18 (EMM + SE)..........................................61 3-10 Distance covered on 6-minute walk te st at Months 0, 6, and 18 (EMM + SE).......62 3-11 Weekly caloric expenditure in activities of at least moderate intensity at Months 0, 6, and 18 (EMM + SE)........................................................................................62 3-12 Self-report caloric intake at Months 0, 6, and 18 (EMM + SE)...............................63 3-13 Self-reported saturate d fat intake at Months 0, 6, and 18 (EMM + SE)..................63
ix 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 RESPONSE OF AFRICAN-AMERICAN AN D CAUCASIAN WOMEN IN A RURAL SETTING TO A LIFESTYLE IN TERVENTION FOR OBESITY By Katie Allison Rickel August 2008 Chair: Michael G. Perri Major: Psychology African-American and Caucasian individuals exhibit differential patterns of weight loss and regain, with African-Americans ach ieving smaller initial losses but better maintenance of lost weight th an their Caucasian counterparts. However, this disparity has not been studied in medically unders erved rural populations. The current study examined the response of obese, rural wome n aged between 50 and 75 years (AfricanAmerican, n = 43; Caucasian, n = 181) to a 6-month lifesty le intervention for weight management followed by randomization to one of two 12-month personal extended care programs (group office-based sessions or indivi dual telephone-based co unseling) or to an educational control condition. M ean weight losses immediatel y following initial lifestyle treatment were smaller for African-American participants than for Caucasian participants ( M s = -6.99 + 4.30 kg, -10.67 + 5.14 kg, respectively, p < .001), but weight regain during extended care was comparable between the two groups. In addition, although Caucasians randomized to either of the extended care progr ams regained less weight than Caucasians
x assigned to the control condition ( M s = 1.03 + 6.14 kg vs. 4.23 + 6.45 kg, for the extended care and control groups, respectively, p = .002), the provision of extended care did not influence weight rega in among African-Americans ( M s = 1.67 + 5.58 kg vs. 1.34 + 4.68 kg for the extended care a nd control groups respectively, p = .85). Exploratory analyses revealed that both African-Ame ricans and Caucasians made significant improvements in various cardiovascular/me tabolic (e.g., systolic and diastolic blood pressure, hemoglobin A1c, C-reactive protein, fitness) and behavioral (e.g., caloric and saturated fat intake, physical act ivity) factors during the life style intervention. Further, African-Americans generally exhibited superi or maintenance of these changes during the extended care period. Collectively, these finding s suggest that a) raci al/ethnic differences in weight loss during lifestyle intervention ar e observed in rural settings, b) despite disparate weight loss outcomes, African-Ame ricans are able to achieve and maintain significant physiological and be havioral improvements, and c) the effectiveness of and necessity for extended care programs among we ight-reduced African-Americans deserve increased attention in future research.
11 CHAPTER 1 INTRODUCTION Obesity, defined as a Body Mass Index, BMI, > 30 kg/m2, is a leading risk factor for the development of a vast array of hea lth conditions. Moreover, it has been wellestablished that lifestyle in terventions can produce clinica lly significant weight reduction, leading to health benefits acr oss a variety of these domains (Klein et al., 2004; Wadden, Butryn, & Byrne, 2004; Wing & Gorin, 2003). However, the existing research has largely ignored African-Americans and reside nts of rural areas, two populations with perhaps the greatest need for eff ective obesity treatments. The few studies that have compared trea tment response of African-Americans to that of Caucasians have found mixed out comes. Although some interventions have produced similar weight reductions and hea lth improvements in both groups (Gower, Weinsier, Jordan, Hunter, & Desmond, 2002; Weinsier et al., 2000; Weinsier, Hunter, Schultz, Zuckerman, & Darnell, 2002), others seem to be less effective in AfricanAmerican than in Caucasian participants (Darga, Holden, Olson, & Lucas, 1994; Foster, Wadden, Swain, Anderson, & Vogt, 1999). Further, the vast majority of these data come from efficacy trials delivered to urban populat ions within academic medical settings. Thus, the current study attempte d to fill this gap in the literature by evaluating how race/ethnicity is related to response of rura l participants to a lif estyle intervention for obesity. Specifically, we examined racial/eth nic differences in 1) response immediately following lifestyle intervention, 2) maintena nce of response during a follow-up period, and 3) the influence of m odality of extended care.
12 Relationship between Obesity, Level of Urbanization, and Race/Ethnicity Obesity is generally more prevalent in rural communities than in urban communities (23% vs. 16%) (Eberhardt et al., 2001). However, trends appear to vary by geographic region. Specifically, in the west, urban communities have the highest rates of obesity, whereas in the South, obesity is more prevalent in rural than in urban settings (Eberhardt et al., 2001). One manifestation of this latter trend is the presence of the stroke belt, the largely rural southeastern po rtion of the United States in which stroke mortality (to which obesity contributes) is particularly high. Differences between Midwest and Southern rural populations across a variety of factors (e.g., diet composition, level of occupational physical activity, racial/ethnic make-up, mean household income level, access to health care services) have been cited as possible explanations for their oppos ite patterns of obesity pr evalence (Martin et al., 2005). The relationship between race/ethnicity and prevalence of obesity has not been well-studied in rural populati ons. Across the United States, the prevalence of obesity is higher among African-Americans adults than among Caucasians adults (45.0% vs. 30.6%, Ogden et al., 2006). However, some i nvestigators have dem onstrated that this racial/ethnic difference in obesity prevalence does not replicate when assessed in rural communities (e.g., Appel, Harrell, & De ng, 2002; Hayward, Pienta, & McLaughlin, 1997; Patterson, Moore, Probst, & Shinogle, 2004) These researchers postulate that this difference in obesity prevalence disappears because there may be fewer racial/ethnic differences in the presence of certain risk factors for obe sity (e.g., physical inactivity, high-fat diet) within rural communities. Studies that have examined this hypoth esis explicitly have found that physical activity patterns are more similar among Af rican-Americans and Caucasians in rural
13 areas than in urban communities (e.g., Parks, Housemann, & Brownson, 2003). By contrast, qualitative differences in diet are found in both rural and urban communities (e.g., Champagne et al., 2004), with African-A mericans consuming less optimal diets (i.e., higher fat, fewer fruits and vegetables) than Caucasians. Thus, it is possible that homogeneity across other domains (e.g., educa tional attainment, income level) may be contributing to the absence of a racial/et hnic disparity in obesity in some rural communities. Consequences of Obesity in Af rican Americans and Caucasians Interestingly, obesity may exert a diffe rential impact on mortality in AfricanAmerican compared to Caucasian populations. For example, Calle and colleagues (Calle, Thun, Petrelli, Rodriguez, & Heath, 1999) f ound a 20 to 30% increase in all-cause mortality in black women with BMIs exceeding 35; however, white women in this obesity category exhibited an elevated morta lity risk of 75 to 100%. Further, although Caucasians in the highest category of obesity status exhibited greater mortality risk than did Caucasians at a healthy body weight, this relation was not found among AfricanAmericans. Others have obser ved that these relations persis t even after controlling for differences in age, income, or gender (D urazo-Arvizu et al., 1997; Sanchez, Reed, & Price, 2000). Nonetheless, treating obes ity in both racial/ethnic groups leads to significant health benefits. As previously mentioned, lifestyle in tervention has emerged as an effective means to improve health via weight reduction. Lifestyle Interventions: Impa ct of Behavioral Factors Dietary Intake Traditional racial/ethnic dietary practices. Arguably the most potent ingredient of lifestyle interventions, diet ary modification is an important determinant of success in
14 obesity treatment programs. Typically, par ticipants follow a low-calorie diet (i.e., a deficit of 500 to 1000 kilocalories below stea dy state) with some attention to improving quality of diet composition (e.g., reducing fa t intake, increasing fruit and vegetable consumption). Given that community-dwell ing African-Americans often fail to meet national dietary guidelines with respect to fruit, vegetable, and fat intake, the dietary prescriptions given in lifestyle interventions may differ substantially from what AfricanAmericans are accustomed to consuming (Gary et al., 2004). Indeed, the average African-American woman eats proportionately more high-fat meats and fried foods, as well as fewer fruits, vegetables, and low-fat dairy products when compared to a typical Caucasian woman (Ard, Skinner, Chen, Aickin, & Svetkey, 2005). Thus, African Americans participating in lifestyle interventi ons are likely asked to make more dramatic changes in their eating habits than are their Caucasian counterparts. Further, some research has suggested that there are specific qu alitative differences between the African-American and the Caucasian diet. For example, in terms of source of dietary fat intake, African-Americans c onsume dietary fat primarily from red meat, poultry, and fish. In contrast, Caucasian partic ipants tend to consume fat primarily in the forms of added fat, dairy foods, baked goods, mixed dishes, and vegetables/salads (Kristal, Shattuck, & Patt erson, 1999). This distinction becomes important when considering the ease with which dietary modifications might be made; AfricanAmericans have to alter what is likely the main dish, whereas Caucasians would alter foods that supplement a meal. Collectivel y, these factors sugges t that the AfricanAmerican diet may be less amenable to traditional caloric restriction techniques than is the Caucasian diet.
15 Barriers to dietary modification. Assuming that African-Americans need to make more drastic dietary changes than do th eir Caucasian counterpa rts, it would follow that African-Americans also may face more barriers when attempting these dietary modifications. Kumanyika (2002) has describe d a number of structural (e.g., targeted marketing of high-calorie foods on black te levision channels, abundance of fast food establishments and shortage of supermarkets in minority communities, food choice at social gatherings), economic (e.g., reliance on homemade foods, lower family income), and sociocultural (e.g., traditiona l high-fat preparation of f oods, food insecurity, distrust of recommendations from medical field) challenges with which African-Americans must grapple. Existing research on dietary modification. Very few studies have examined racial/ethnic differences with regard to ch anges in dietary intake during and/or after participation in a lifestyle change program. In the analyses that have been conducted, African-American participants do appear to be able to ma ke similar and significant dietary improvements when compared to Ca ucasians participants (Glass, Miller, Szymanski, Fernhall, & Durnstine, 2002; Kr istal et al., 1999; Svet key et al., 2005). For example, after completing a nutrition educat ion intervention focused on adopting a lowfat diet, reductions in fat from dairy foods, baked goods, poul try, and fish actually were greater in African-American th an in Caucasian participants (Kristal et al., 1999). Racial/ethnic differences in we ight change outcome were not assessed in this research. Further, in a study that did target weight loss, both African-American and Caucasian participants achieved comparable reductions in saturated fat and comparable increases in fruits, vegetables, and low-fa t dairy products (Svetkey et al., 2005). In spite of these
16 comparable behavioral changes, however, Af rican-American women in this study lost significantly less weight than Caucasian women. Intervention strategies. Several program-related factors seem to be especially important when implementing dietary modifi cations in a multi-ethnic population. The inclusion of clear and simple nutrition edu cation, the use of indi vidualized counseling, and instruction in estimating caloric and fat c ontent appear to be particularly beneficial for African-American women (Bronner & B oyington, 2002). Additionally, the use of ethnic recipes (modified to be lo wer in calories and fat) is an effective way to facilitate dietary change in minority populations (Agurs-Collins, Kumanyika, Ten Have, & Adams-Campbell, 1997). Assessment of dietary intake. The present study used a food frequency questionnaire to assess changes in dietary in take. However, a number of researchers (e.g., Tucker et al., 2005) have posited that th e surveys and questionnai res used to collect information about dietary patterns may not be valid for use in multicultural samples. Nonetheless, many studies have found suppor t for the cross-cultural validity of food frequency questionnaires by comparing these da ta to those derived from food diaries, 24hour dietary recalls, and/or bi ological markers (e.g., Knut sen, Fraser, Linsted, Beeson, & Shavlik, 2001; Resnicow et al., 2000). Kris tal and colleagues (Kristal et al., 2000) found similar associations among food frequency questionnaires, 24-hour recalls, and food diaries across racial/ethnic groups. Simila rly, Knutsen and colleagues (Knutsen et al., 2001) found consistent associations between food frequency questionnaires, serum nutrient levels, and 24-hour recalls when data derived from African-American participants were compared to data derived from Caucasian participants.
17 However, others (e.g., Tucker et al., 2005) still argue that traditional assessments (e.g., food frequency questionnaires, telephone surv eys) need to be tailored because they often fail to include questions about specific ethnic foods that make up large parts of the diets of minority populations. Some research has experimented with this concept by modifying measures for use in populations of interest. For example, Mayer-Davis and colleagues (Mayer-Davis et al., 1999) designed a food frequency interview for an epidemiologi cal study conducted in the southern U.S. (Insulin Resistance Arterioscler osis Study, IRAS). After co nsultation with nutritionists who were familiar with participants traditio nal diets, modifications to the assessment included adding regional food choices, such as ham hock and red chili. When compared to nutrient intake data derived from eight 24-hour dietary reca lls, the food frequency interview had comparable validity across racial/ethnic groups (rural non-Hispanic Caucasians, r = 0.61; non-Hispanic African Americans, r = .50). This study was limited by its failure to compare the validity of this tailored measure to standard food frequency questionnaires. Subsequently, another group (Tucker et al., 2005) completed a comparative analysis. Specifically, a similar modifica tion technique was us ed to tailor a food frequency questionnaire for use in the impoveri shed rural region of the Lower Mississippi Delta. To this end, Tucker and colleagues (Tucker et al., 2005) added numerous regional foods (e.g. turnip greens, chitterlings, squirr el) and modified suggest ed portion sizes to customize a standardized food frequency in strument (Block FFQ, Block et al., 1986) for use in their study. Using 24-hour dietar y recall to assess c onvergent validity, the modified questionnaire appear ed to yield more accurate data than the standard food
18 frequency questionnaire. Evaluati on of these tailored assessment tools continue, and these investigations may provide a means toward mo re valid data about dietary practices of minority groups. Nonetheless, despite the possible limitations of standardized dietary assessment instruments, the Block FFQs cross-cultural validity in previous work justifies the use of this tool in the current study. Physical Activity Patterns and predictors of physical activity. As with their dietary practices, community-dwelling African-American and Caucasian participants differ in their usual patterns of leisure time physical activity. McGuire and colleagues (McGuire, Ahluwalia, & Strine, 2006) found that Caucasians accrue higher amounts of leisure time physical activity than do African-Americans. This pattern is particularly pronounced in the southern part of the countr y, even after controlling for age, income, and BMI (Washburn, Kline, Lackland, & Wheeler, 1992). Some authors have suggested that AfricanAmericans lower levels of leisure time physi cal activity might be offset by higher levels of occupational physical activ ity (e.g., Kumanyika, 2002). Ho wever, when data from doubly-labeled water experiments are examin ed (which do not differentiate between leisure-time or occupational activities), African-Americans still exhibit lower energy expenditure, largely due to ove rall lower levels of physical activity (Carpenter et al., 1998; Starling, Toth, Matthews, & Poehlman, 1998; Tuten, Petosa, Sargent, & Weston, 1995). In addition, the correlates of participation in physical activity in African-American women deserve attention. Ransdell and Wells (1998) found marital status to be a reliable correlate of physical activity in Africa n-American women, with married women participating in higher levels of activity than those who are single, widowed, or divorced.
19 The increasing prevalence of single-parent families (usually consisting of a mother and her children) in the African-A merican community suggests that this relationship may be a particularly important one for future study (U.S. Census Bureau, 2006). Further, Eyler and colleagues (Eyler et al., 2003) suggested that observing others who are physically active appears to be a significant predictor of activity for rural African-Americans but not for rural Caucasians. Because rural settings often are not conducive to physical activity participation in public places (i.e., due lack of sidewalks, poor street lighting, dangerous traffic patterns), African-Americans are not lik ely to see others in the community engage in physical activity. Barriers to increasing physical activity. When attempting to increase physical activity as part of a weight-loss program African-Americans face a number of barriers usually not addressed in trad itional lifestyle interventions. For example, AfricanAmericans endorse a preference for lifestyl e activities (e.g., walking, gardening) over formal exercise programs (Wanko et al., 2004). Although these life style activities may provide health benefits, the re latively modest amount of en ergy expended is unlikely to result in weight loss. However, this caveat is not always made clear to participants. In addition, almost half of African -American patients in an urba n diabetes clinic indicated that pain was a substantial barrier to exer cise (Wanko et al., 2004). Although Caucasians also cite pain as an exercise deterrent, Afri can-Americans seem to suffer greater levels of disability (in terms of mob ility and functionality) associ ated with obesity (Houston, Stevens, & Cai, 2005). It is possible that th is differential burden in disability serves to further compound weight management difficu lties for African-Americans; that is, reductions in mobility will likely result in lo wer overall energy expenditure, which helps
20 maintain obesity. Lastly, Kumanyika (2002) identified unique ec onomic (e.g., lack of discretionary time and money) and sociocu ltural (e.g., prestige of living a comfortable sedentary lifestyle) barriers African Am ericans face when attempting to adopt and maintain a program of physical activity. Existing research on physical activity modification. Research evaluating racial/ethnic differences in physical activity durin g and after participation in a weightloss intervention is even scarcer than the data comparing dietary modi fications. Weinsier and colleagues (Weinsier et al., 2002) f ound that, after weight loss, Caucasian participants became more physically active, whereas African-American participants became less physically active. A parallel pa ttern emerged when comparing changes in weight-adjusted aerobic capacity (ml/kg/min), a likely correlate of changes in physical activity. African-Americans appeared to exhib it decreased aerobic ca pacity after weight loss, whereas Caucasians made increases in aerobic capacity. (Data concerning aerobic capacity unadjusted for body weight were not presented in the study. Thus, these outcomes are difficult to interp ret.) Others in this gr oup (Walsh, Hunter, Sirikul, & Gower, 2004) have hypothesized that physical activity is more aversive to AfricanAmericans after weight loss, due to decr eases in aerobic capa city, strength, and flexibility. It is important to note that thes e data have not been replicated independently in weight-reduced African-American participan ts. Thus, caution is warranted again when interpreting these observations. Intervention strategies. Nonetheless, increases in phys ical activity are possible for African-Americans and are facilitated when weight-loss programs utilize culturallysensitive programs. Incorporating religi on and spirituality (Underwood & Powell, 2006),
21 offering exercise programs and equipment at residential sites (Belza et al., 2004), building relationships among pa rticipants and emphasizing so cial support for exercise (Young & Stewart, 2006), and teaching pain management techniques (Golightly & Dominick, 2005), all have been effective in helping African-Ameri cans increase physical activity during and/or after at tempts at weight reduction. Assessment of physical activity. Daily physical activity patterns have been assessed in a variety of wa ys, including self-report measures and intervieweradministered questionnaires, but the validity of methods of physical activity assessment has been a matter of controversy. One wa y in which researchers have attempted to establish validity of these methods is to co rrelate these data with physiological criteria, such as changes in cardiova scular fitness, heart rate, adiposity, or blood pressure. However, measures often have been valid ated in predominately Caucasian samples (Singh, Fraser, Knutsen, Lindset, & Bennett., 2001). Therefore, it remains unclear whether these instruments are equally valid for use in minority populations. Resnicow and colleagues (Resnicow et al., 2003) sought to examine the validity of the Community Healthy Activities Mode l Program for Seniors (CHAMPS) physical activity questionnair e in a sample of African-American adults. Prior to their study, the CHAMPS had only been validated in Ca ucasian and Asian-American populations (Harada, Chiu, King, & Stewart, 2001). Re snicows group compared CHAMPS data to the physiological measures of estimated VO2 max, blood pressure, and BMI. The study revealed only a moderate correlation between CHAMPS indices and estimated VO2 max. However, this study was limited in its appro ach to validity testi ng. Changes in the physiological measures assessed usually re quire significant increases in vigorous
22 intensity activity. Thus, because the CHAM PS questionnaire emphasizes activities of moderate intensity, it is perhaps not surpri sing that few significant findings emerged. The use of accelerometry might have provided the authors with a more sensitive measure of change in overall physical activity level. Physiological Responses to Lifestyle Treatment A computerized literature search of PubMed and PsychInfo using the terms obesity combined with African-Ameri can, black, ethnic differences, Caucasian, white, weight control, wei ght loss, or obesity treatment uncovered 16 independent intervention studies that have compared African-American and Caucasian participants physiological response to lifestyle interventions. Because this research includes data covering a variety of outcome measures, only severa l key physiological responses will be described in detail below. Weight loss Relatively few studies have directly co mpared weight loss achieved by AfricanAmericans to that achieved by Caucasians in the context of the same weight-loss trial (Kumanyika, 2002). In the small body of existing literature, none of the data suggest that African-Americans lose more we ight than Caucasians. In fact, in half of the sixteen studies that have examined ra cial/ethnic differences in weig ht loss, (Darga et al., 1994; Foster et al., 1999; Kumanyi ka et al, 2002; Kumanyika, Doroskeka, Self, Bahnson, & Robertson, 1997; Kumanyika, Obarzanek, Stevens, Hebert, & Whelton,1991; Wing & Anglin, 1996; Wylie Rosett et al., 1993; Yanovski, Gormally, Lesser, Gwirtsman, & Yanovski, 1994), Caucasian participants clear ly achieved weight losses of greater magnitude than did African-American particip ants. Of the remaining studies, 4 do not suggest racial/ethnic differences in amount of weight lost dur ing treatment (Glass, Miller,
23 Szymanski, Fernhall, & Durstine, 2002; Hong, Li, Wang, Elashoff, & Heber, 2005; Nicklas et al, 2003; Weinsier et al., 2002) a nd 4 yielded more complicated patterns (e.g., interactions between weight category a nd race/ethnicity, interactions between race/ethnicity and time). In one of these latter st udies, Djuric and colleague s (Djuric et al., 2002) found a racial/ethnic difference only among the heav iest women in their study, with AfricanAmericans losing less weight than Caucasians In another study, analyzing data from the Diabetes Prevention Program (DPP), a greate r proportion of Caucasians met the studys weight-loss goal than did African-American s by the end of the 6 month intensive treatment period (57% vs. 36%, respectively, Wing et al., 2004). However, after a variable follow-up period ( M time to follow-up = 3.2 years) this difference disappeared (38% vs. 32%). Similar trends were noted by Wassertheil-Smoller and colleagues (Wassertheil-Smoller et al., 1985) and Steven s and colleagues (Stevens et al., 2001) in other multi-center weight-loss trials. These latter three studies highlight the possibility that African-Americans may experience a differe nt rate and pattern of weight change during and after weight-loss attempts. This issue will be discussed in a later section. Improvements in Metabolic Risk Factors Although weight-loss outcomes are largely mixed, African-American and Caucasian women appear to reap similar metabol ic benefits from weight loss. In studies wherein African-Americans and Caucasians lo st similar amounts of weight, both Djuric and colleagues (Djuric et al., 2002) and Nick las and colleagues (Nicklas et al., 2003) observed similar and significant improvements in both groups in HDL-cholesterol, LDLcholesterol, triglyceri des, systolic and diastolic blood pr essure, fasting blood glucose, and insulin sensitivity.
24 These findings also have been replicated in studies where racial/ethnic differences in weight change existed. For example, Darga and his researchers (Darga et al., 1994) found that, even though Caucasians lost mo re weight than African-Americans, both groups experienced similar and significant im provements in blood pressure and in lipid profiles. One explanation of this surprising outcome ma y be that African-Americans exhibit a lower threshold at which weight lo ss produces physiological changes (Nicklas et al., 2003). Another equally plausible explanatio n is that the threshold at which weight loss confers metabolic benefits is low for a ll individuals, regardless of race/ethnicity. Therefore, even though weight losses were smaller for African-Americans in Dargass research, the losses were sufficient to produce metabolic improvements. Improvements in Cardiorespiratory Fitness In contrast to this similarity in meta bolic response, some studies suggest robust racial/ethnic differences in cardiorespiratory response to weight loss. Weinsier and his group (Weinsier et al., 2002) observed that, after weight loss, African-American women became less fit (i.e., exhibited increases in weight-adjusted VO2 max) whereas Caucasian women became more fit (i.e., exhibite d decreases in weight-adjusted VO2 max). Further, both Nicklas and colleagues (N icklas et al., 1999) and Hunt er and colleagues (Hunter, Weinsier, Zuckerman, & Darnell, 2004) found increases in weight-adjusted VO2 max to be greater in Caucasian than in African-Am erican participants. The latter group also determined that African-Americans experien ced significantly more physiologic difficulty during exercise, as evidenced by higher heart rates, lower ventilation, and higher ratings of perceived exertion. It may be important to note that the particip ants in these studies did not exercise during their weight-loss at tempts. Thus, these outcomes may not be representative of outcomes associated with more comprehensive lifestyle treatments.
25 Role of Social Support in Weight-loss Interventions Caucasian women receive more naturalistic social support for weight loss than do African-Americans (Nothwehr, 2004). The lack of social support fo r weight loss in the African-American community is related to multiple factors. For one, the AfricanAmerican ideal body type is indeed larger than the Caucasian ideal (Kumanyika, Wilson, & Guilford-Davenport, 1993). African-Americans tend to endorse more positive body image and to experience less cultural pressure to lose we ight, even after controlling for age, BMI, and education (Smith, Thom pson, Raczynski, & Hilner, 1999). In addition, Kumanyika (2002) points out that in many minority cultures, f ood serves as an important form of social currency, such that individuals show appreciation of each other through the serving and eating of large quant ities of traditional (often hi gh-calorie, high-fat) foods. Thus, restricting ones calo ries would likely not be supported among ones family and friends. The general lack of support for weight loss within the African-American community suggests that a social support ne twork built into th e intervention itself may be particularly important for African-American participants. Indeed, African-American women seem to particularly benefit from tr eatments that are deliver ed in a group-session format (Bronner & Boyington, 2002). Enhanced perceived control and bolstered selfefficacy have been suggested as the variable s mediating the relationship between social support and success in group weight-loss inte rventions among African-American women (Wolfe, 2004). Racial/ethnic Differences in Patterns of Weight Change As previously mentioned, African-American s and Caucasians may lose and regain lost weight at different rates. Two indepe ndent multi-center trials (Wassertheil-Smoller
26 et al., 1985; Stevens et al., 2001) found that African-Americans lost less weight during initial treatment than did Caucasians; howev er, at long-term follo w-up, net weight losses became similar. In the DISH (Dietary Intervention Study of Hypertension) trial, Caucasians and African-Americans achiev ed weight changes of -2.5 and -1.7 kg, respectively, at 8 weeks; how ever, at 56 weeks, the net ch anges were -4.1 and -5.0 kg for Caucasians and African-Americans, respec tively (Wassertheil-Smoller et al., 1985). Similarly, in the TOHP II (Tri als of Hypertension Preven tion, Phase II) study, Stevens and colleagues (Stevens et al., 2001) found that Caucasians exhibited greater net changes than African-Americans at 6 mont hs (-2.4 kg vs. .0 kg, respectively, p < .01) and at 18 months (-1.0 kg vs. 0.2 kg, respectively, p < .03) but that groups exhibited similar net change by 36 months (0.6 kg vs. 1.4 kg, respectively, p > 0.2). It is possible, however, that the relatively small weight changes reporte d in these studies limit the utility of these data in illustrating racial/ethnic disparities in response to lifestyle interventions. These studies were not designed to evaluate such differences, and they likely lacked the power necessary to detect meaningful patterns. Further, the interpretation of these data is potentially flawed. In both studies, the authors suggest that weight loss achieved by African-Americans is comparable to that achieved by Caucasians when participants are assessed long-term. This interpretation is based upon the observation that the differe nces between Caucasian and AfricanAmerican losses are no longer significant at follow-up. Ho wever, the authors overlook an important distinction be tween their findings, lack of evidence for a difference, versus their interpretation of the data, evidence for lack of a difference (Gomberg-
27 Maitland, Frison, & Halperin, 2003). In other wo rds, being unable to detect a difference between values does not imply that the values are equivalent. A statistical procedure called an equiva lence trial, however, can provide this evidence for lack of a difference. To employ this method, an equivalence margin the smallest difference between values that woul d be clinically meani ngful is chosen a priori on the basis of both clin ical judgment and existing data in the literature. In an equivalence trial, the null hypothesis states th at the difference betwee n the values exceeds the equivalence margin. The objective of the an alysis is to gather evidence that rejects this hypothesis, thereby supporting the altern ative hypothesis that the two values are clinically equivalent (Tamayo-Sarver, Albert, Tamayo-Sarver, & Cydulka., 2005). Nonetheless, there are a number of f actors that could indeed produce a true convergence in these data. For one, it is possi ble that African-Americans continue to lose weight after the initial treatment period e nds, whereas Caucasians only maintain lost weight during follow-up. Alternatively, Af rican-Americans might be better able to sustain their behavioral cha nges and weight losses because they both are generally of smaller magnitude than the change and loss achieved by Caucasians after initial treatment. In the Trial of Nonpharmacologi c Interventions in the Elderly, Kumanyika and colleages (Kumanyika et al., 2002) found support for both explanations. Their data revealed that, among participants who failed to lose 3.6 kg or more at 6 months, AfricanAmericans were more successful than Caucasia ns in continuing to lose weight and in maintaining lost weight during the follow-up period. A more recent study (Kumanyika et al ., 2005) examined program-related factors that might influence maintenance of lost weig ht in African-American participants. After
28 completion of an initial weight-loss program, participants were randomized to (a) further group sessions, (b) staff-facilita ted self-help program, or (c) a clinic visit only control condition. The authors were surprised to find that neither of the maintenance programs was superior to clinic visits only. In fact, none of the gr oups exhibited significant weight changes during the 10-month follow-up period. This is the only study to our knowledge that has studied the effectiven ess of extended care approaches in an exclusively AfricanAmerican population; thus, it is possible that aspects of the content and/or delivery specific to these programs were responsible for their ineffectiveness. Further, no study to date has directly compared the responses of African-Ameri can and Caucasian individuals to different extended care programs in the cont ext of a single trial. Given the established importance of social support for African -Americans attempting to manage weight (Boyington & Bronner, 2002), group support ma y be superior to individualized counseling for this population. Aims and Hypotheses of Current Study There were three aims of the current study. The primary aim was to evaluate racial/ethnic differences in weight change (a) immediately after completing a 6-month lifestyle intervention (Month 0 to Month 6), and (b) duri ng a 12-month extended care period (Month 6 to Month 18). It was hypothesi zed that (a) Caucasians would lose more weight than African-Americans between M onth 0 and Month 6, (b) Caucasians would exhibit greater weight regain during the 12-month extended care period, between Month 6 and Month 18. The secondary aim was to compare the effects of two extended care programs (office-based group sessions vs. telephone-based individual counseling) on changes in weight within each racial/ethnic group. Afri can-Americans, perhaps even more so than
29 Caucasians, may derive benefit from social support built into the in tervention. Thus, it was hypothesized that African-Americans assi gned to the office-based group extended care program would exhibit smaller weight re gain than African-Americans assigned to the telephone-based individualized extended care program. The effect was not expected to occur among Caucasian participants. The tertiary aim was to evaluate raci al/ethnic differences in changes along a variety of physiological and be havioral domains, including (a) systolic and diastolic blood pressure, (b) LDL-cholesterol, (c) trigly cerides, (d) HbA1c, (e) C-reactive protein, (f) distance covered duri ng 6-minute walk test (a measure of fitness), (g) daily caloric intake, (h) daily saturated fat intake, a nd (i) energy expended during activities of moderate or greater intensit y. In addition, racial/ethnic differences in adherence and program satisfaction were explored. Because these tertiary aims were exploratory, no a priori hypotheses were postulated.
30 CHAPTER 2 RESEARCH METHODS AND PROCEDURES Parent Study: Treatment of Obesity in Underserved Rural Settings (TOURS) TOURS was conducted as a three-arm randomiz ed controlled trial to evaluate the effectiveness of interventions designed to improve long-term weight management in obese women from medically underserved rura l counties. Particip ants were randomized to one of three 18-month treatment progr ams delivered in rural counties through Cooperative Extension Service offices. Each treatment program began with an identical 6-month lifestyle intervention for weight loss (Phase I), and treatment conditions differed in the extended care program delivered dur ing the 12 months following the lifestyle intervention (Phase II). The three extended care progr ams included two intervention conditions, (a) continued office-based gr oup sessions, (b) in dividual telephone counseling, and (c) a control c ondition providing health educa tion via mail. Assessments were conducted at baseline (Month 0), postlifestyle interventi on (Month 6), and postextended care period (Month 18). Participating County Cooperative Extension Service (CES) Offices All assessments and interventions were carried out in CES offices of the six participating rural counties in north Florida. All six counties have been designated in whole or in part as Health Professional Shortage Areas by the U.S. Department of Health and Human Services. The demographic characteristics of th e six counties (based on the 2000 Census) indicate that the population is largely white (80.2%), with AfricanAmericans (15.8%) making up the next larges t racial/ethnic categ ory. As a group, the
31 participating counties are characterized by relatively low levels of educational attainment. The mean percentage of adults who are hi gh school graduates is 63.4%, compared with the U.S. national average of 84.1%. Compared to the U.S. population as a whole, the participating counties have relatively lo w household income levels ($26, 517 vs. $50,890) and relatively high levels of poverty (21.3% vs. 11.3%). Participants Participants were 224 obese women (African-American, n = 43; Caucasian, n = 181) from medically underserved rural counties who volunteered to take part in an 18month lifestyle obesity treatment program. For the parent study, three cohorts of participants were recruited and randomized at 6-month intervals. Eligibility requirements included age between 50 and 75 year s, body mass index (BMI) of 30 kg/m2 or greater, and current residence in a rural county in No rth Central Florida. Individuals who were unwilling or unable to give informed consent, who were unwilling to accept random assignment, or who were participating in another research project were not accepted. Potential participants were excluded if, at screening, thei r medical history, clinical examination, or laboratory results revealed underlying disease likely to limit lifespan and/or increase risk of inte rventions (e.g., cancer within 5 years, chronic or recurrent respiratory or gastrointestinal conditions, hist ory of musculo-skeletal conditions) or that revealed metabolic abnormalities despite appropriate treatment (e.g., fasting blood glucose > 125 mg/dl, fasting serum triglycerides > 400 mg/dl, resting blood pressure > 140/90). Also, participants were excluded if th ey were unable to travel to their local CES office for intervention sessions, were diagnosed with a major psychiatric disorder, or had lost more than 10 pounds in the 6 months prior to screening.
32 Procedures Recruitment and Screening Participants were recruited through a vari ety of methods, including direct mailings, media articles, radio announcements, and co mmunity presentations. Both culturallysensitive approaches and dire ct solicitation to community gr oups were used to attract a demographically representative sample. Fo r example, African-American members of our recruitment staff gave presentations to chur ches with predominately African-American congregations. Following a preliminary telephone screening, prospective participants were invited to an in-person assessment at their local CES office. This assessment included completion of informed consent and a medical evaluation conducted by a mobile clinical assessment team led by a licensed registered nurse. Data Collection Race/ethnicity. As part of a demographic questionna ire, participants were asked to report the race/ethnicity with which they most strongly identified. Only participants identifying as African-American or as Caucas ian were used for anal yses in the present study. Weight. Weight (assessed in light clothing and without shoes) was measured to the nearest 0.1 kilogram with a calibrate d and certified balance beam scale. Blood pressure and blood analysis. Systolic and diastolic blood pressure were measured by a registered nurse, who used a st andardized protocol. Three readings were taken, spaced one minute apart, and the last tw o were averaged. The registered nurse or a study phlebotomist drew 22 ml of blood, and Quest Diagnostics Clinical Laboratories analyzed the sample for a metabolic and lipid profile. Of the levels assessed, only LDL-
33 cholesterol, triglycerides, C-reactive protei n, and HbA1c were considered in the current study. Dietary intake. The Block 95 Food Questionnair e (Block et al., 1986) is a validated self-report instrument that asks res pondents to estimate th eir consumption of a wide variety of foods over the past year. Scoring yields estimate s of macroand micronutrient intake, as well as intake by sp ecific food group. A recent validational study showed a strong correlation between the Block 95 Food Frequency Questionnaire and four-day food records (S ubar et al., 2001). Physical activity. The CHAMPS Physical Activity Qu estionnaire (Stewart et al., 2001) is a 41-item self-report measure deve loped specifically for the assessment of physical activity in adults aged 50 years and older. The CHAMPS measure yields estimates of energy expended per week in all physical activities and in activities of at least moderate intensity. Two-week test-retes t reliability coefficients ranged from .62 to .76, and discriminant validity has been establis hed in groups of older adults with varied physical activity levels (Stewart et al., 2001). Physical fitness. The 6-Minute Walk Test (6MWT) is a commonly used assessment procedure designed to evaluate chan ges in fitness in people with low exercise capacity. Participants walk along an indoor course, trying to cover as much ground as possible in 6 minutes. Distance walked is m easured and recorded to the nearest foot. The reliability of the 6MWT in low-fit popul ations is high (Kervio, Ville, Leclercq, Daubert, & Carre, 2004), and its validity as a measure of fitness is supported by its high correlation ( r = 0.68) with peak oxygen uptake during ma ximal exercise te sting (Zugck et al., 2002).
34 Adherence to behavioral modification. Participants were supplied with food and exercise diaries in which they were instruct ed to record information about daily eating (type, amount, and caloric content of foods c onsumed) and daily physical activity (steps measured by pedometer). During the lifestyle intervention, diaries were to be completed each day, and participants were instructed to submit the previous weeks food and exercise diaries at each group meeting. During the extended care period, diaries were to be completed at least three tim es each week, and participants were encouraged to submit diaries every two weeks during th is 12-month extended care period. Satisfaction with lif estyle intervention. Participants also completed a 36-item questionnaire (specifically designed for the present study) that yiel ds composite scores assessing their (a) overall sa tisfaction with the lifestyl e intervention, (b) perceived usefulness of strategies taught during the lifestyle intervention, and (c) current use of strategies taught during the life style intervention. It is im portant to note that this questionnaire has not been validated. Intervention During the lifestyle intervention (Month 0 Month 6), all participants completed the same obesity treatment program. The in tervention consisted of 24 weekly 90-minute group (n = 10-14 participants/group) sessions led by bachelors and/or masters level interventionists who followed a structured protocol. During sessions, participants (a) reported on, and received feedback about, th eir previous weeks progress in reaching eating and physical activity goa ls; (b) learned skills related to cognitive-behavioral selfmanagement, healthful eating behaviors, and phy sical activity practice; and (c) identified specific behavioral goals for the coming w eek, receiving encouragement from fellow group participants.
35 Certain procedures of this initial lifestyle treatment were culturally tailored to suit the special needs and issues of rural wo men. These included provision of a recipe booklet, Down Home Healthy Cookin(Natio nal Institutes of Health, 2000), which provides recipes and cooking tips for prep aring traditional Af rican-American and Southern dishes with lower fat, lower calor ie, and lower sodium ingredients. Cooking demonstrations and food tastings were incorpor ated into selected in tervention sessions in order to expose participants to healthful menu items. In a ddition, a simplified version of the self-monitoring materials was made avai lable to those in need of assistance. At the conclusion of the lifestyle interven tion (Month 6), participants were notified of their randomization assignment to one of the three 12-month extended care conditions: an office-based extended care program, a te lephone-based extended care program, or an educational control condition. Office-based extended care program. In this condition, part icipants continued to meet at their CES office for on-site group sessions every other week. At each session, participants submitted any eating or exercise diaries completed since the last session and were weighed. Session format and conten t were guided by social problem solving theory, which utilizes a 5-stage problem solving model consisting of (a) problem orientation, (b) problem definition and formul ation, (c) generation of alternatives, (d) decision making, and (e) solution implementa tion. During the sessi on, interventionists led the participants in group problem solving with the goal of generating a solution plan for dealing with one of the problem situati ons described by a part icular group member. Sessions concluded with each participant iden tifying specific behavioral goals for the
36 following two-week period and receiving feedback and support from fellow group members. Telephone-based extended care program. Participants randomized to this condition received two scheduled telephone c ontacts each month fr om their lifestyle intervention leader. During the telephone call, leaders attempted to (a) prompt participants to continue usi ng key weight-management strate gies, (b) use the previously described problem-solving model to identify sp ecific problems and to generate a plan, (c) provide support and reinforcement for conti nued efforts at weight management, and (d) collect participants self-repor ted weight as well as inform ation recorded on any eating and exercise diaries complete d. (Participants were then encouraged to submit their diaries and self-reported weight by mail.) Interventionists followed a specific protocol when interacting with participants. Education control condition. Participants randomized to this condition received biweekly newsletters with educational in formation about proper eating and physical activity. The newsletters included low-fat, low-calorie recipes, as well as tips for leading a lifestyle conducive to weight maintenance. In additi on, a summary of the problemsolving model was provided. Participants were asked to mail their self-reported weights, as well as any completed eating and exercise diaries, to our laborat ory every two weeks. Statistical Procedures and Analyses Only Phase I treatment completers (i.e., part icipants who had atte nded at least half of the lifestyle intervention sessions includi ng at least one session between weeks 18 and 24) were included in the statistical analyses. Further, because the l iterature suggests that African-American and Caucasian individuals exhibit differential patterns of weight change during and after lifes tyle intervention (Kumanyika 2002), it would have been
37 imprudent to impute values for missing weight data. Thus, when no data were available for a participant on a certain outcome measure, the participant was excluded from that analysis. This is reflected in the n s reported in the tables at the end of chapter 3. Preliminary analyses using ANOVA proce dures were carried out to determine whether the African-American and Caucasian participants differed with respect to baseline demographic (e.g., age, weight, e ducation) characteristics. Differences uncovered were included as covariates in the subsequent analyses. Primary aim. To address the studys primary aim, weight changes achieved between (a) Month 0 and Month 6 and be tween (b) Month 6 and Month 18 were calculated. We expected to observe (a) a gr eater magnitude of wei ght loss by Caucasians between Months 0 and 6 and (b) a greater ma gnitude of weight regain by Caucasians between Months 6 and 18. An analysis of variance (ANOVA), with race/ethnicity as the categorizing variable, was used to examine racial/ethnic differences in weight change between Months 0 and 6, and a parall el ANOVA was conducted with respect to racial/ethnic differences in weight change between Months 6 and 18. Analyses were also conducted for body mass index and pe rcentage of initial weight lost. Secondary aim. For the secondary aim, in whic h the influence of the type of extended care condition was evaluated with re spect to weight change, African-American participants and Caucasian participants were examined separately. For each racial/ethnic group, a 2 X 2 repeated measures ANOVA wa s conducted, with two time points (Month 6 and Month 18) and two categories of extended care condition (office-based group sessions and telephone-based individual counseling).
38 Tertiary aim. Lastly, the tertiary aim entailed examination of racial/ethnic differences along a variety of other parameters Changes in systolic and diastolic blood pressure, LDL-cholesterol, triglycerides, HbA1c, distance covered during 6MWT, daily caloric intake, daily saturated fat intake and energy expended during activities of moderate intensity were assessed as outco me measures. Adherence measures (i.e., attendance at group sessions, completion of food and exercise logs) and program satisfaction were also considered. For each tertiary aim outcome measure, two 2 x 2 repeated measures ANOVA were employed, one for changes during the lifestyle intervention and one for changes during the extended care program. Thus, the first repeated measures ANOVA used two assessmen t points (Month 0 and Month 6) and two racial/ethnic categories (African-American and Caucasian), and the second used two time periods (Month 6 and Month 18) and two raci al/ethnic categories (African-American and Caucasian). Estimated marginal means (usi ng Bonferroni correcti ons) were calculated and used to decompose significant interactions For all analyses, significance testing was conducted at = 05.
39 CHAPTER 3 RESULTS Recruitment and Screening A consort statement summarizing the re sults of the screening and assessment process is displayed in Figure 3-1. The r easons for exclusion by racial/ethnic group (for exclusions verified during the telephone screen or at the in -person assessment visit) are presented in Table 3-1. Ther e were no significant differe nces in the proportion of African-American compared to Caucasian participants who were excluded for a nonmusculoskeletal medical condition [ 2 (1) = 2.07, p = .15] or on the basis of screening visit results (e.g., elevated blood pr essure, abnormal lipid panel) [ 2 (1) = 1.94, p = .16]. However, a marginally greater proportion of Caucasians (relative to African-Americans) were excluded for a musculoskeletal condition [ 2 (1) = 3.55, p = .06]. Participants Starting and Completing Phase I Of the African-American and Caucasian pa rticipants eligible for randomization, 3 African-Americans and 26 Caucasians were not randomized. Of the African-American individuals not randomized, one participant reported a work conflict, another no longer met the studys weight criteria, and a third wa s lost to follow-up. Caucasians who were not randomized had endorsed a variety of r easons (e.g., conflict with the group meeting time, incipient illness/injury, lost to followup). As described prev iously, participants were categorized as having completed Phase I of the study if they a) a ttended at least half of the Phase I group sessions and b) attended at least one session between weeks 18 and 24. Proportion of participants classified as Phase I completers did not differ by
40 racial/ethnic group [2 (1) = 2.80, p = .094]. Of the 61 African-Americans who began the study, 43 (70.5%) were classified as Phase I completers. Of the 225 Caucasians who began the study, 181 (80.4%) were classified as Phase I completers. Baseline Characteristics Table 3-2 displays the baseline demogr aphic characteristic s of the AfricanAmerican ( n = 43) and Caucasian ( n = 181) women who completed Phase I of the trial. African-American participants were significantly younger than Caucasian participants (for African-American and Ca ucasian participants, 58.01 + 5.71, 60.06 + 6.07 years, respectively, p < .05). Thus, age was used as a covari ate in all subseque nt analyses that included both African-American and Caucasian pa rticipants. Conseque ntly, all such data presented are estimated marginal means (EMM) w ith standard errors (SE). In all but one analysis (i.e., change in distance covered dur ing the 6-minute walk test between Months 0 and 6), the effect of this cova riate was not significant. Over all, groups were similar with regard to baseline body weight and BMI, em ployment status, marita l status, education level, and annual household income ( p s > .05). Primary Aim: Changes in Body Weight As shown in Table 3-3, Caucasians lost more weight than African-Americans between Months 0 and 6 [ F (1, 221) = 20.70, p < .001]. Racial/ethnic differences in weight change between Months 6 and 18 were not significant ( p = .535). Within-group analyses revealed that both groups exhibited significant re ductions in weight between Months 0 and 6 ( p s < .001), whereas only Caucasian pa rticipants exhibited significant regain between Months 6 and 18 ( p < .001). Weight regain by African-Americans did approach significance, however ( p = .066).
41 Further (also described in Table 3-3), pa rallel findings emerged for percentage weight lost from baseline and for changes in BMI, with one exception. Namely, although African-Americans did not exhibit significant increases in weight gain or BMI between Months 6 and 18, this group did exhibit a signi ficant percent weight change during that period ( p = .040). Table 3-4 and Figure 3-2 further describe weight losses achieved at Month 6 and at Month 18 by African-American and Caucasia n participants. Chi-square analyses revealed that a greater proportion of Cau casians (compared to African-Americans) achieved weight losses of at least 10% of initial body weight at both Month 6 and Month 18 (Month 6, 2(1) = 16.21, p < .001; Month 18, 2(1) = 11.77, p = .001). The modal loss for African-Americans at Months 6 and 18 was between 5.00 and 9.99%, whereas the modal loss for Caucasians at Months 6 and 18 was greater than 10.0%. In addition to these analyses of thos e participants who completed Phase I, additional analyses were conducted using all av ailable data from both Phase I completers and Phase I non-completers (African-Americans, n = 50, Caucasians, n = 197). Identical findings emerged, with one exception. A lthough African-Americans weight regain between Months 6 and 18 only approached sign ificance in the completers analyses, their regain did reach significance when Ph ase I non-completers were included ( p = .006). Secondary Aim: Effect of Extended Care Pr ogram within each Racial/Ethnic Group The studys original secondary aim entailed comparing, within each racial/ethnic group, the weight changes of participants in the two ex tended care programs (officebased group sessions versus telephone-based i ndividual counseling). These analyses did not reveal a significant intera ction between time and type of extended care program assignment in either the African-American or Caucasian participants. African-American
42 participants assigned to the offi ce-based group sessions regained 3.02 + 4.38 kg, whereas those assigned to the telephone-based individual counseling regained 1.03 + 6.07 kg ( p = .35). Although this difference is not statistically significan t, it may have clinically meaningful implications. However, Caucasia n participants assigned to the office-based group sessions exhibited leve ls of weight regain ( M = 0.76 + 6.19 kg) that were neither statistically nor clinically diffe rent from weight regain ex hibited by those assigned to the telephone-based individual counseling (1.39 + 6.12 kg, p = .55). Further consideration prompted two additi onal post hoc analyses, however. Again for each racial group, those assigned to e ither of the extended care programs were compared to those assigned to the contro l condition. As displayed in Figure 3-3, participation in a extended care program was unrelated to weight regain for AfricanAmericans (extended care program, 1.67 + 1.03 kg; control condition, 1.34 + 1.56 kg, p = .85). However, Caucasians assigned to an extended care program regained significantly less weight than those assigned to the c ontrol condition (extended care program, 1.03 + 0.58 kg; control condition, 4.23 + 0.83 kg, p = .002). Parallel analys es using all available data (i.e., data for both Phase I completers and Phase I non-completers ) revealed the same pattern of findings. Tertiary Aim: Changes in Selected Outcomes Changes during lifestyle in tervention (Months 0 6) Systolic blood pressure. As shown in Table 3-5 and Figure 3-4, repeated measures ANCOVA revealed significant main effects for time [ F (1, 221) =58.51, p < .001] and ethnicity [ F (1, 221) = 3.88, p = .050] between Month 0 and Month 6, but the interaction effect was not significant ( p = .554). There were no racial/ethnic differences in systolic blood pressure at either Month 0 or Month 6. Further, at Month 6, both
43 African-Americans and Caucasians had exhi bited a significant decrease from Month 0 ( p s < .001). Diastolic blood pressure. Significant main effects for time [ F (1, 221) = 20.97, p < .001] and ethnicity [ F (1, 221) = 7.97, p = .005] emerged with resp ect to diastolic blood pressure during the lifestyle intervention (see Table 3-5 and Figure 3-5). However, the interaction between time and race/eth nicity did not reach significance ( p = .441). African-Americans were found to have significan tly higher diastolic blood pressure than Caucasians at Month 0 ( p = .043) and at Month 6 ( p = .007), and both groups exhibited significant decreases in dias tolic blood pressure during this period (for AfricanAmericans and Caucasians, p = .036 and p < .001, respectively). It should be noted that 69.8% of African -Americans and 44.2% of Caucasians were being treated with hypertensive medications over the course of the study. However, including medication status as a covariate in these analyses had no apparent effect on the outcomes just described. Triglycerides. Repeated measures ANCOVA unc overed significant main effects for time [ F (1, 217) = 10.06, p = .002] and race/ethnicity [ F (1, 217) = 14.16, p < .001], but no interaction was apparent between tim e and race/ethnicity with respect to triglycerides ( p = .183, see Table 3-6 and Figure 3-6). Caucasian participants presented with higher levels of triglycerides at both Month 0 (p < .001) and at Month 6 ( p = .009) when compared to African-Americans. Further, only Caucasians demonstrated significant decreases in triglycerides over th e course of the lifestyle intervention ( p < .001).
44 LDL-cholesterol. Table 3-6 and Figure 3-7 illu strate that repeated measures ANCOVA did not find main ef fects for either time ( p = .236) or race/ethnicity (p = .589) with regard to LDL-cholesterol. A marginally significant interaction between time and race/ethnicity did emerge, however [ F (1, 215) = 3.34, p = .069]. African-American and Caucasian participants did not differ in LDL-c holesterol at either M onth 0 or at Month 6, and only Caucasians achieved significant decr eases in LDL-cholesterol between those two assessment points ( p = .001). Although 66.7% and 40.2% of African-Ame ricans and Caucasians, respectively, were being treated with lipid-lowering medica tions, inclusion of medi cation status in the analyses did not influence the outcomes presented. Hemoglobin A1c. As shown in Table 3-7 and Figure 3-8, there were main effects for time [ F (1, 218) = 53.30, p < .001] and race/ethnicity [ F (1, 218) = 36.68, p < .001] with regard to HbA1c assessed at Months 0 and 6. The intera ction between time and race/ethnicity was not significant ( p = .295). African-Americans exhibited higher HbA1c at both assessment points when compared to Caucasians ( p s < .001), but both groups made significant decreases in HbA1c during the lifestyle intervention ( p s < .001). Moreover, 45.6 % of African-Americans and 28.7% of Caucasians were being treated with diabetic medications. However, includi ng diabetic medication st atus as a covariate did not affect these results. C-reactive protein. Repeated measures ANCOVA revealed only a main effect for time with respect to C-reactive protein [ F (1, 218) = 43.07, p < .001, see Table 3-8 and Figure 3-9]. Neither the main effect for ra ce/ethnicity nor the interaction effect was significant (for the main effect of race/ ethnicity and the interaction effect, p = .618 and p
45 = .811, respectively). African-American and Caucasians did not differ in terms of Creactive protein levels at M onth 0 or at Month 6, but bot h groups achieved significant decreases during the lifestyle interventi on (for African-American and Caucasian participants, p = .001 and p < .0001, respectively). Six-minute walk test. Table 3-8 and Figure 3-10 show significant main effects for time [ F (1, 215) = 38.21, p < .001] and race/ethnicity [ F (1, 215) = 10.71, p = .001], and the interaction between time and race/ethnicity approached significance [ F (1, 215) = 2.86, p = .092] with respect to distance covered during the 6-minute walk test. In terms of statistical significance, Caucasian particip ants performance was superior to AfricanAmerican participants perf ormance at both Month 0 ( p = .024) and at Month 6 ( p < .001), but these differences were not clinically meaningful. Further, both groups made significant increases in distance covered betw een Months 0 and 6 (for African-American and Caucasian participants, p = .014 and p < .001, respectively). Energy expenditure during activities of at least moderate intensity. Repeated measures ANCOVA uncovered main effects for time [ F (1, 214) = 47.65, p < .001] and race/ethnicity [ F (1, 214) = 3.95, p = .048] with regard to self-reported weekly energy expenditure during activi ties of at least moderate intensity, but there was no interaction effect ( p = .235). As shown in Table 3-8 and Figu re 3-11, African-American participants reported lower energy expenditure than Caucasians at Month 0 ( p = .014) but not at Month 6. Further, both racial/ethnic groups reported incr eased weekly energy expenditure over the course of the lifest yle intervention (for African-American and Caucasian participants, p s < .001).
46 Caloric intake. As displayed in Table 3-9 and Figure 3-12, repeated measures ANCOVA uncovered a significant main effect for time [ F (1, 199) = 32.64, p < .001] as well as an interaction between time and race/ethnicity [ F (1, 199) = 4.44, p = .036] for self-reported daily caloric intake. Caucasians reported higher daily caloric intake than African-Americans at Month 0 ( p = .040) but not at Month 6. Further, both racial/ethnic groups reported significant decreases in calor ic intake between Month 0 and Month 6 (for African-American and Caucasians, p = .050 and p < .001, respectively). Saturated fat intake. There emerged significant main effects for time [ F (1, 199) = 70.59, p < .001] and race/ethnicity [ F (1,199) = 6.78, p = .010], as well as a significant interaction between time and race/ethnicity [ F (1, 199) = 10.14, p = .002], in terms of selfreported saturated fat intake during the lifes tyle intervention (see Table 3-9 and Figure 313). Caucasians reported higher intakes of satu rated fat than African -Americans at Month 0 ( p = .002) but intakes were co mparable at Month 6 ( p = .621). Nonetheless, both groups reported significant d ecreases in saturated fat inta ke from Month 0 to Month 6 (for African-American and Caucasian participants, p = .005 and p < .001, respectively). Outcomes during extended care (Months 6 18) Systolic blood pressure. Repeated measures ANCO VA uncovered only a main effect for time [ F (1, 203) = 4.24, p = .041] for systolic blood pressure during the extended care period. As shown in Table 3-10 and Figure 3-4, there were no racial/ethnic differences at either Month 6 or at Month 18, and only Caucas ian participants exhibited a significant increase in systolic blood pressure between Months 6 and 18 ( p = .001). Diastolic blood pressure. Table 3-10 and Figure 3-5 display main effects for both time [ F (1, 203) = 5.85, p = .016] and race/ethnicity [ F (1, 203) = 7.91, p = .005] with regard to diastolic blood pressure from Month 6 to Month 18. African-Americans
47 exhibited higher diastolic blood pressure when compared to Caucas ians at both Month 6 ( p = .014) and at Month 18 ( p =.030). However, only Cauc asian participants exhibited significant increases in diastolic blood pr essure during the extended care period ( p = .001). Triglycerides. As shown in Table 3-11 and Fi gure 3-6, only a main effect for race/ethnicity [ F (1, 198) = 10.49, p = .001] emerged in a repeated measures ANCOVA examining triglyceride level dur ing the extended care period. Neither the main effect for time ( p = .466) nor the interaction effect ( p = .183) was significant. Caucasians had higher triglyceride levels than Af rican-Americans at both Month 6 ( p = .006) and at Month 18 (p = .002), and neither group demonstr ated significant cha nges in triglyceride level between Months 6 and 18. LDL-cholesterol. There were no significant main effects for time ( p = .311) or race/ethnicity ( p = .589) with respect to LDL-chol esterol during extended care, but a significant interaction between time and race/ethnicity emerged [ F (1, 195) = 6.11, p = .014]. As Table 3-11 and Figure 3-7 illustrate, there were no racial/ethnic differences in LDL-cholesterol at Month 6, but Caucasians exhibited marginally higher levels than African-Americans at Month 18 ( p = .067). Further, only Caucasian participants showed significant increases in LDL-chol esterol during extended care ( p < .001). Hemoglobin A1c. Repeated measures ANCOVA revealed main effects for time [ F (1, 200) = 4.01, p = .047], race/ethnicity [ F (1, 200) = 48.14, p < .001], and a marginally significant interaction between time and race/ethnicity [ F (1, 200) = 3.59, p = .060] for HbA1c during extended care (see Table 3-12 and Figure 3-8). AfricanAmericans exhibited higher HbA1c compared to Caucasians at both Month 6 and 18 ( p s
48 < .001), and only African-American participants showed a si gnificant increase in HbA1c ( p = .033) during extended care. C-reactive protein. As displayed in Table 3-12 and Figure 3-9, there was a main effect for time [ F (1, 188) = 4.99, p = .027] and a marginally si gnificant main effect for race/ethnicity [ F (1, 188) = 2.69, p = .102] for C-reactive prot ein during extended care. The interaction between time and race/ ethnicity did not reach significance ( p = .814). African-Americans and Caucasia ns did not differ in C-reactiv e protein levels at Month 6 or at Month 18. Further, only Caucasians demonstrated significant reductions in CReactive protein during extended care ( p = .015). Six-minute walk test. Table 3-13 and Figure 3-10 indica te that a main effect for race/ethnicity [ F (1, 182) = 8.57, p = .004] emerged in a repeated measures ANCOVA conducted for distance covered during 6-mi nute walk test. There was neither a significant main effect for time ( p = .946) nor a significant in teraction between time and race/ethnicity ( p = .707). African-Americans covered less distance than Caucasians at both Month 6 ( p = .024) and Month 18 ( p = .009) assessments, and neither group exhibited significant changes in perform ance between Month 6 and Month 18. Energy expenditure during activities of at least moderate intensity. Repeated measures ANCOVA resulted in a marginally significant main effect for time [ F (1, 197) = 2.61, p = .108] during extended care (see Table 313 and Figure 3-11). However, neither the main effect for race/ethnicity nor the inte raction effect reached statistical significance (for main effect of race /ethnicity and interaction, p = .379 and p = .945, respectively). There were no racial/ethnic differences in se lf-reported energy expenditure at Month 6 or
49 at Month 18, and only Caucasians exhibited marginally significant decreases in selfreported energy expenditure during extended care ( p = .077). Caloric intake. As displayed in Table 3-14 a nd Figure 3-12, a main effect for race/ethnicity [ F (1, 187) = 6.37, p = .012] and an interaction between time and race/ethnicity [ F (1,187) = 6.34, p = .013] were observed for se lf-reported cal oric intake during extended care. The main effect for time was not significant ( p = .158). AfricanAmericans and Caucasians did not report signif icantly different caloric intakes at Month 6, but Caucasians reported signi ficantly higher caloric intake than African-Americans at Month 18 ( p = .001). Further, only African-America ns reported significant decreases in caloric intake during extended care ( p = .031). Saturated fat intake. A significant main effect for race/ethnicity [ F (1, 187) = 7.79, p = .006] and a significant interaction between time and race/ethnicity [ F (1,187) = 11.17, p = .001] emerged for self-reported satura ted fat intake between Months 6 and 18 (see Table 3-14 and Figure 3-13). No main effect for time was observed ( p = .879). African-American and Caucasian participan ts did not report si gnificantly different intakes of saturated fat when assessed at Month 6, but Ca ucasians endorsed significantly higher intakes of saturated fat when comp ared to African-Americans at Month 18 ( p < .001). Further, Caucasians reported significant increases in saturate d fat during extended care ( p < .001), whereas the increases reporte d by African-Americans only approached significance ( p = .079). Adherence and program satisfaction. The percentage of lifestyle intervention sessions attended by African-American and Cauc asian participants were not significantly different ( M s = 89.70 + 8.91, 89.51 + 10.33%, respectively, p = 0.914). In addition, both
50 groups completed similar percentages of the requested food ( M s = 85.45 + 17.60, 89.95 + 15.23%, respectively, p = .237) and exercise ( M s = 83.35 + 18.31, 85.15 + 18.38%, respectively, p = 0.605) records between Months 0 and 6. Of those participants assigned to either of the extended car e interventions (i.e., office -based groups sessions or telephone-based individual c ounseling), racial/ethnic diffe rences in percentage of sessions attended approached significance (for African-Americans and Caucasians, M s = 53.40 + 37.12, 42.20 + 41.01%, respectively, p = .086), with African-Americans attending a greater pe rcentage of sessions. In addition, subscale totals derived from responses on a questionnaire designed to assess participants satisfaction with the progr am were analyzed via independent t-tests, with race/ethnicity as the cat egorizing variable. These anal yses revealed that AfricanAmericans expressed higher levels of overall satisfaction [ t (203) = 3.09, p = .003], perceived the strategies ta ught to be more useful [ t (203) = 2.74, p = .007] and endorsed higher levels of current use of the strategies [ t (203) = 2.40, p = .017] when compared to their Caucasian counterparts. Further, Af rican-American and Caucasian participants expressed similar satisfacti on with their group leaders [ t (218) = 0.06, p = .954] Because this questionnaire has not been validated, results should be interpreted with caution. Table 3-1. Reasons for excl usion by racial/ethnic group. African-AmericanCaucasian (N = 148 )(N = 779) Exclusionary Criteria Number (%)Number (%) <50 or >75 years old 19 (12.8) 30 (3.9) <30 or >50 BMI 16 (10.8) 121 (15.5) Non-medical reason 16 (10.8) 137 (17.6) Did not complete assessment 26 (17.6)100 (12.8) Medical condition (not musculoskeletal) 20 (13.5) 134 (17.2) Musculoskelatal condition 16 (10.8) 124 (15.9) Screening visit result 35 (23.6) 133 (17.1)
51 Table 3-2. Participant dem ographics at baseline. African-AmericanCaucasian (N = 43 )(N = 181 ) Characteristic Mean + SD or n (%) Mean + SD or n (%) Body weight (kg) 99.87 + 16.79 95.82 + 14.46 Body Mass Index (kg/m2) 38.13 + 6.16 36.49 + 4.52 Age (yrs) 58.01 + 5.71 60.06 + 6.07a Employed (fullor parttime, %) 21 (48.8) 86 (47.5) Marital Status Never married 1 (2.3) 2(1.1) Divorced or separated 9 (20.9) 16 (8.8) Widowed 6 (14.0) 15 (8.3) Presently married 27 (62.8) 142 (78.5) Cohabitating with partner 0 (0.0) 6 (3.3) Annual Income < $19,000 8 (18.6) 38 (21.0) $20,0004,000 13 (30.2) 40 (22.1) $35,0009,000 9 (20.9) 37 (20.4) $50,0004,000 8 (18.6) 38 (21.0) $75,0009,000 5 (11.6) 17 (9.4) >$100,000 0 (0.0) 7 (3.9) Dont know 0 (0.0) 4 (2.2) Education < High School 3 (7.0) 9 (5.0) GED 1 (2.3) 10 (5.5) High School diploma 6 (14.0) 41 (22.7) Vocational training 3 (7.0) 33 (18.2) Some college/Associates degree 14 (32.6) 55 (30.4) Bachelors degree 5 (11.6) 14 (7.7) Post-bachelors study or degree 11 (25.6) 19 (10.5) a p <.05 for between-group differences
52 Table 3-3. Weight-related outcomes of Afri can-American and Caucasian participants. African-American Caucasian (N = 43) (N = 181) Outcome Estimated Mean + SE Estimated Mean + SE Body weight (kg) Month 0 99.87 + 16.79 95.82 + 14.46 Change between Month 0 and 6 -6.83 + 0.76 b -10.10 + 0.37ab Change between Month 6 and 18 1.49 + 0.99 2.18 + 0.48b % weight change from baseline Change between Month 0 and 6 -6.90 + 0.80b -11.20 + 0.40ab Change between Month 6 and 18 1.70 + 1.00b 2.30 + 0.50b Body Mass Index (kg/m 2) Month 0 38.00+ 0.75 36.52 + 0.36 Change between Month 0 and 6 -2.60 + 0.29b -4.08 + 0.14ab Change between Month 6 and 18 0.55 + 0.38 0.82 + 0.18b ap < .05 for racial/ethnic difference, bp < .05 for within-group difference Table 3-4. Percentage of weight change from baseline. African-American Caucasian (N = 43) (N = 181) Assessment time Count (%) Count (%) Month 6 Weight gain 1 (2.3) 3 (1.7) 0.00-4.99% loss 10 (23.3) 18 (9.9) 5.00-9.99% loss 21 (48.8) 52 (28.7) > 10.00% loss a 11 (25.6) 108 (59.7) Month 18 Weight gain 5 (12.5) 23 (13.5) 0.00-4.99% loss 14 (35.0) 45 (26.3) 5.00-9.99% loss 15 (37.5) 27 (15.8) > 10.00% loss b 6 (15.0) 76 (44.4) a 2(1) = 16.21, p < .001, b 2(1) = 11.77, p = .001
53 Table 3-5. Systolic and diastolic blood pressure during lifestyle intervention. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Systolic blood pressure (mm Hg)ab N 43 181 Month 0 127.70 + 1.37 125.57 + 0.67 Month 6 120.53 + 1.71e 117.18 + 0.83e Diastolic blood pressure (mm Hg)ab N 43 181 Month 0d 77.23 + 1.16 74.61 + 0.56 Month 6d 74.55 + 1.23e 70.83 + 0.59e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 0. Table 3-6. Triglycerides and LDL-choles terol during lifestyle intervention African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Triglycerides (mg/dL)ab N 40 180 Month 0d 110.33 + 9.49 152.67 + 4.46 Month 6d 101.18 + 9.87 130.12 + 4.63e LDL-cholesterol (mg/dL) N 40 178 Month 0 116.33 + 4.69 123.47 + 2.21 Month 6 117.92 + 5.08 115.99 + 2.40e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 0. Table 3-7. Hemoglobin A1c and C-Reactive protein during life style intervention. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Hemoglobin A1c (%)ab N 40 181 Month 0d 6.48 + 0.10 5.86 + 0.05 Month 6d 6.15 + 0.08e 5.61 + 0.04e C-Reactive protein (mg/dL)a N 40 181 Month 0 6.23 + 0.95 5.94 + 0.44 Month 6 3.82 + 0.54e 3.34 + 0.25e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 0.
54 Table 3-8. Performance on 6-minute walk test (6MWT) and weekly energy expenditure during lifestyle intervention. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Distance covered on 6MWT (ft)ab N 42 176 Month 0d 1351.55 + 27.15 1420.48 + 13.19 Month 6d 1407.24 + 27.09e 1518.50 + 13.16e Weekly energy expenditure in activities of at least moderate intensity (kcal/wk)ab N 43 174 Month 0d 619.64 + 207.93 1195.25 + 102.88 Month 6 1629.54 + 235.47e 1906.58 + 116.51e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 0. Table 3-9. Daily caloric and saturated fa t intake during lifestyle intervention. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Daily caloric intake (kcal/day)ac N 37 165 Month 0d 1506.39 + 104.33 1745.61 + 48.92 Month 6 1315.10 + 69.14e 1327.32 + 32.42e Daily saturated fat intake (g/day)abc N 37 165 Month 0d 17.28 + 1.70 23.25 + 0.80 Month 6 12.86 + 0.91e 13.37 + 0.43e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 0. Table 3-10. Systolic and diastolic blood pressure during extended care. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Systolic blood pressure (mm Hg)a N 40 166 Month 6 120.22 + 1.75 116.70 + 0.86 Month 18 121.50 + 1.90 119.80 + 0.93e Diastolic blood pressure (mm Hg)ab N 40 166 Month 6d 74.14 + 1.28 70.62 + 0.61 Month 18d 75.60 + 1.15 72.81 + 0.56e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 6.
55 Table 3-11. Triglycerides and LDL-c holesterol during extended care. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Triglycerides (mg/dL)b N 36 165 Month 6d 99.76 + 10.64 132.17 + 4.95 Month 18d 101.86 + 10.27 137.21 + 4.78 LDL-cholesterol (mg/dL)c N 36 162 Month 6 116.08 + 5.02 115.57 + 2.44 Month 18 112.53 + 5.65 124.04 + 2.65e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 6. Table 3-12. Hemoglobin A1c and C-Reac tive protein during extended care. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Hemoglobin A1c (%)ab N 37 166 Month 6d 6.18 + 0.08 5.59 + 0.04 Month 18d 6.36 + 0.11e 5.59 + 0.05 C-Reactive protein (mg/dL)a N 32 159 Month 6 4.11 + 0.62 3.28 + 0.28 Month 18 3.16 + 0.45 2.51 + 0.20e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 6.
56 Table 3-13. Performance on 6-minute walk test (6MWT) and weekly energy expenditure during extended care. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Distance covered on 6MWT (ft)b N 37 148 Month 6d 1428.23 + 29.21 1525.17 + 14.49 Month 18d 1431.21 + 30.20 1520.88 + 14.98 Weekly energy expenditure in activities of at least moderate intensity (kcal/wk) N 38 162 Month 6 1682.98 + 254.33 1894.05 + 122.66 Month 18 1410.69 + 283.74 1644.24 + 136.84 ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 6. Table 3-14. Daily caloric and saturate d fat intake during extended care. African-American Caucasian Outcome Estimated Mean + SE Estimated Mean + SE Daily caloric intake (kcal/day)bc N 34 156 Month 6 1254.61 + 71.97 1344.57 + 33.22 Month 18d 1101.10 + 80.18e 1388.27 + 37.01 Daily saturated fat intake (g/day)bc N 34 156 Month 6 12.18 + 0.97 13.41 + 0.45 Month 18d 10.62 + 1.08 15.12 + 0.50e ap < .05 for overall main effect of time, bp < .05 for overall main effect of race/ethnicity, cp < .05 for overall interaction effect, dp < .05 for time-specific r acial/ethnic difference, ep < .05 for change from Month 6.
57 Figure 3-1. Screening and assessment of individuals for the parent study. 328 individuals eligible for randomization 30 individuals were not randomized: -15 (50.0%) no longer met study requirements -10 (33.3%) were unable to attend session time assigned 5 (17.7%) were lost to follow-up 298 participants randomized -61 African-Americans -225 Caucasians -12 other racial/ethnic g rou p 1350 individuals completed telephone screening (TS) 659 individuals excluded at TS: -234 (29.6%) did not meet age, BMI, or residence criteria -265 (33.5%) endorsed a non-medi cal exclusion criteria -292 (36.9%) endorsed an exclusionary medical condition 691 individuals invited to inperson assessment visit (AV) 363 individuals excluded during or after AV: -132 (36.0%) did not attend AV -17 (7.3%) exhibited BMI <30 or >50 -11 (4.7%) endorsed non-medical exclusion criteria at AV -29 (12.5%) endorsed exclusiona ry medical condition at AV -23 (12.5%) exhibited glucose > 125 mg/dL -39 (16.9%) exhibited ab normal lipid profile -82 (35.5%) exhibited BP > 140/90 mm Hg -30 (13.0) revealed other metabolic abnormality 64 participants did not complete Phase I -18 African-Americans (28.1%) -44 Caucasians (68.8%) -2 other racial/ethnic group (3.1%) 234 participants completed Phase I 43 African-Americans -181 Caucasians -10 other racial/ethnic group
58 Figure 3-2. Categories of wei ght lost from baseline. Figure 3-3. Weight regain of African-Ameri can and Caucasian participants assigned to control versus intervention conditions (EMM + SE). p < .001 ns p = .001 ns ns ns ns ns 0 10 20 30 40 50 60 70 Gain 0-4.99% 5.00-9.99% >10.00% Gain 0-4.99% 5.00-9.99% >10.00% Month 6Month 18 African-American Caucasian -1 0 1 2 3 4 5 6 African-AmericanCaucasianWeight Regained (kg) Control Condition Intervention Conditions ns p < .002
59 105 110 115 120 125 130 135 Month 0Month 6Month 18mm Hg AfricanAmerican Caucasian Figure 3-4. Systolic blood pressure at Months 0, 6, and 18 (EMM + SE) Figure 3-5. Diastolic blood pressure at Months 0, 6, and 18 (EMM + SE). 66 68 70 72 74 76 78 80 Month 0Month 6Month 18 AfricanAmerican Caucasianmm Hg
60 Figure 3-6. Triglycerides at M onths 0, 6, and 18 (EMM + SE). Figure 3-7. LDL-cholesterol at Months 0, 6, and 18 (EMM + SE). 60 80 100 120 140 160 180 Month 0 Month 6Month 18mg/dL African-American Caucasia n 90 95 100 105 110 115 120 125 130 135 Month 0Month 6 Month 18mg/dL African-American Caucasian
61 Figure 3-8. Hemoglobin A1c at Months 0, 6, and 18 (EMM + SE). Figure 3-9. C-reactive protein at Months 0, 6, and 18 (EMM + SE). 5 5.2 5.4 5.6 5.8 6 6.2 6.4 6.6 6.8 Month 0Month 6Month 18% African-American Caucasian 0 1 2 3 4 5 6 7 8 Month 0 Month 6Month 18mg/dL African-American Caucasian
62 Figure 3-10. Distance covered on 6-minute wa lk test at Months 0, 6, and 18 (EMM + SE). Figure 3-11. Weekly caloric expenditure in act ivities of at least moderate intensity at Months 0, 6, and 18 (EMM + SE). 0 500 1000 1500 2000 2500 Month 0 Month 6Month 18kcal/w k African-American Caucasian 1200 1250 1300 1350 1400 1450 1500 1550 1600 Month 0Month 6Month 18f t African-American Caucasian
63 Figure 3-12. Self-report cal oric intake at Months 0, 6, and 18 (EMM + SE). Figure 3-13. Self-reported satu rated fat intake at Months 0, 6, and 18 (EMM + SE). 0 5 10 15 20 25 30 Month 0 Month 6Month 18grams African-American Caucasia n 800 1000 1200 1400 1600 1800 2000 Month 0 Month 6Month 18calories African-American Caucasia n
64 CHAPTER 4 DISCUSSION Main Findings The primary aim of this study was to exam ine racial/ethnic differences in weight changes achieved (a) between Month 0 and M onth 6 and (b) between Month 6 and Month 18. Consistent with our hypothesis, the magnitude of weight loss exhibited by Caucasians was greater than that achieve d by African-Americans between Months 0 and 6. Contrary to our hypothesis, however, there was no statistically si gnificant racial/ethnic difference in weight change achieved between Months 6 and 18. Both groups exhibited significant weight reductions between Months 0 and 6, and in Phase I completer analyses, only Caucasians exhibited significant we ight regain between Months 6 and 18. The secondary aim of this study was to i nvestigate whether African-American and Caucasian participants responded differentiall y to the types of extended care programs offered. Initial plans were to compare the effectiveness of the office-based group condition to the individualiz ed telephone counseling c ondition. For both AfricanAmericans and Caucasians, the weight regain exhibited by those assi gned to the officebased condition was not significan tly different than the weight regain exhibited by those assigned to the telephone-based counseling condition. However, this difference may have been clinically meaningful among African-American participants, with those assigned to the telephone-based counseling exhi biting smaller regain than those assigned to the office-based group condition.
65 An additional analysis collapsed data across extended care programs and compared these to outcomes achieved by those in the e ducational control condi tion. We found that assignment to an extended care program was a ssociated with smaller weight regain in Caucasian, but not African-American, individual s. In fact, African -Americans exhibited only small and non-significant weight regain whether assigned to an extended care program or to the control condition. The tertiary aim of this study inspected a variety of other outcome measures for racial/ethnic differences. African-American and Caucasian participants exhibited comparable attendance at group sessions and similar completion of self-monitoring instruments. Between Months 0 and 6, both African-Americans and Caucasians exhibited significant improvements in systo lic blood pressure, diastolic blood pressure, hemoglobin A1c, C-reactive protein, and pe rformance on the 6-minute walk test. Further, only Caucasians exhibited signi ficant reductions in triglycerides and LDLcholesterol between Mo nths 0 and 6. Both groups reported increased weekly energy expenditure and decreased caloric and sa turated fat intake during the lifestyle intervention phase. Between Month 6 and 18, Caucasians ( but not African-Americans) exhibited significant increases in systolic blood pr essure, diastolic blood pressure, and LDLcholesterol, as well as decr eases in C-reactive protein. In addition, African-Americans reported significant decreases in caloric intake between Months 6 and 18, whereas Caucasians reported significant increases in saturated fat inta ke during the same period. Comparison to Previous Studies Overall, our findings concur with exis ting literature that suggests AfricanAmericans lose smaller amounts of weight during lifestyle intervention relative to
66 Caucasians. The two most recent clinical tria ls that have examined this pattern are the Trials for Hypertension Prevention (T OHP) and the Trial of Nonpharmacologic Interventions in the Elderly (TONE). In T OHP, Stevens and colleagues (Stevens et al., 2001) found that Caucasians exhibited greater weight reductions relative to AfricanAmericans at 6 months (from base line, -2.4 kg vs. .0 kg, respectively, p < .01) and at 18 months (from baseline, -1.0 kg vs. 0.2 kg, respectively, p < .03). These reductions are smaller but consistent with the weight losses achieved by our Caucasian and AfricanAmerican participants at 6 months (fro m baseline, -10.7 kg vs. -6.9 kg, respectively, p < .001) and at 18 months (from baselin e, -8.6 kg vs. -5.2 kg, respectively, p = .009). It is interesting to note that the magnitude of rega in between Months 6 and 18 were similar in these two studies, despite TOHP s continuing intensive inte rvention versus the present studys transitioning into a less intensive extended care period. In TONE, Kumanyika and colleagues (Kuma nyika et al., 2002) also observed that Caucasian participants lost significantly more weight than African-American participants at 6 months (-5.9 kg vs. -2.7 kg, respectively, p < .001) and at the end of a variable follow-up period ( M time to follow-up = 27 weeks; -4.9 kg vs. -2.0 kg, respectively, p < .01). It is difficult to compare magnitude of regain in TONE with that in TOURS due to the latters earlier endpoint. Nonetheless, it is noteworthy that the present effectiveness study was able to replicate the patterns obs erved in TOHP and TONE, given that the latter two were conducted in an academic medi cal setting. These parallels lend support to the contention that raci al/ethnic discrepancies in weight loss during lifestyle intervention are robust across study setting and sample.
67 Some of the racial/ethnic patterns obs erved in metabolic changes were also consistent with those found in the existing li terature. For example, our data regarding similar improvements in blood pressure c oncur with the findings of Nicklas and colleagues (Nicklas et al., 2003) and those of Stevens and colleagues (Stevens et al., 2001). However, the racial/ethnic differences that emerged with respect to changes achieved in LDL-cholesterol and triglyceride are discrepant with those of previous studies. In their lifesty le interventions, Gower and associ ates (Gower et al., 2002) did not find racial/ethnic differences in changes in LDL-cholesterol, and Nicklas and colleagues (Nicklas et al., 2003) did not find racial/ethnic differences with respect to improvements in triglycerides. It should be noted, however, that African-Americans lost similar amounts of weight (but less body fat) when co mpared to their Caucasian counterparts in each of these studies. Thus, it is possible that the smaller weight changes achieved by African-American participants in our study we re responsible for th e observed pattern of outcomes. Change in Energy Balance as a Medi ator for Discrepant Weight Outcome There are a variety of possible mechanisms that may contribute to the discrepant weight losses achieved by African-American and Caucasian participants during lifestyle interventions. The most obvious possible medi ator would be differen tial changes in diet and exercise, thus resulting in different ma gnitude of weight change. Unfortunately, however, the cost and labor of utilizing unbiased biomarkers of energy intake (such as doubly labeled water, Trabulsi & Schoeller, 2001) is usually not feasible for use in lifestyle intervention. Self-report measures of dietary in take and energy expenditure are notoriously inaccurate (e.g., Hill & Davies, 2001), and their validity may be especially
68 suspect when used in diverse populations (e.g., Resnicow et al., 2003; Tucker et al., 2005). Nonetheless, the current study did collect se lf-report data related to caloric intake (via the Block Food Frequency questionnaire) and to energy expe nditure during daily activities (via the CHAMPS questionnaire). As previously described, there were no racial/ethnic differences in changes in ener gy expenditure, but Caucas ian participants did report larger decreases in cal oric intake between Months 0 and 6 and between Months 0 and 18. Thus, it is possible that Caucasians greater caloric reducti on contributed to this groups superior weight losses. To examine this possibility quantitatively, projected weight loss can be calculated by combining a) changes in self-reported caloric intake on the Block Food Frequency questionnaire and b) changes in self-reported energy expenditure during all activities on the CHAMPS questionnaire. Such analyses revealed that African-A mericans projected weight loss of 7.96 kg approximated their actual weight loss of 6.99 kg between Months 0 and 6. Similarly, the projected weight loss of 10.99 kg for Caucasia ns approached their actual loss of 10.67 kg during the same time interval. It may be important to note that the Block 95 food frequency questionnaire asks participants to consider their intake patterns over the previous year when making responses. However, the agreement between projected and actual losses suggests that participants were reporting thei r most recent patterns of intake at Month 6 (i.e., rather than averaging their intake over the past year, which would have included a period prior to the beginning of the study). There are two major implications of these analyses. First, it does appear that differential reduction in caloric intake may have been responsibl e for the differential
69 weight outcomes observed. Caucasians reporte d greater caloric reductions and lost more weight between Months 0 and 6. The other im portant implication is that projected and actual losses appear to be comparably correla ted in each racial/ethni c group. Thus, these outcomes lend support to the cross-cultura l validity of the instruments. Implications for Clinical Services The design of this study, consisting of an intensive weekly intervention followed by a less demanding extended care program, allows for examination of th e possibility that the intensity of an intervention may have di fferential impact for African-American and Caucasian participants. On the basis of we ight outcome and dietary change, Caucasian participants appear to have benefited more from the intensive portion. On the other hand, African-Americans superior maintenance of metabolic improvements and continued dietary changes during the follow-up period sugges ts that they fared better in the program of lower intensity. African-Americans commonl y cite transportation difficulties and lack of time to attend meetings as barriers to successful participation in weight loss groups (Maillet, DEramo, & Spollett, 1996). Thus, the increased convenience of the follow-up programs (i.e., attending in-person sessions only every other wee k, receiving telephone calls in their own homes) may have been pa rticularly beneficial to African-American participants. This pattern bears implications for the types of weight loss programs that might best be recommended to African-American indi viduals. For instance, the primary care clinic has been suggested as a possible venue in which to deliver weight loss interventions (Rippe, 1998; Simkin-Silverma n & Wing, 1997). Primary care offers the potential convenience (i.e., indi viduals can receive services during normal clinic visits) and the semi-frequent contact that may have been helpful to the American-American
70 participants in the current study. In addition, accessing se rvices via primary care may allow individuals to circumvent the stig ma that the African-American community commonly associates with formal weight lo ss programs. Martin and colleagues (Martin et al., 2006) found that six monthly outpatien t visits (each lasting approximately 15 minutes) administered to African-American patients in their primary care clinics produced a mean 2.0 kg weight loss by Month 6. This reduction is comparable to losses achieved during a group lifes tyle intervention (designed for African-American participants) requiring 10 to 16 hours of participant contact (Kanders, Ullman-Joy, Foreyt, 1994). Thus, although lif estyle intervention remains the gold standard, primary care weight management may be a promisi ng alternative, especially for AfricanAmericans. Limitations and Strengths In considering the implications of these findings, a few potential limitations should be considered. Because African-Americans a nd Caucasians were not recruited in equal numbers by the parent study, all of the raci al/ethnic comparisons utilized a smaller number of African-American than Caucasian participants. Unequal groups lead to a variety of potential statisti cal confounds (e.g., violations of homogeneity of variance); however, these were considered and corrected for in the an alyses (e.g., using Levenes statistic). Additionally, larger sample si zes (as opposed to smaller ones) are generally more representative of the population from which the sample is drawn. Thus, it is possible that our results bette r represent the referent Cau casian population than they do the referent African-American population. Further, the preponderance of Caucasian pa rticipants in the li festyle intervention groups (as compared to the minority of African -American participants ) represents another
71 potential problem. Although the interventi on was tailored for use in a multi-ethnic sample, our African-American participants ma y have faced more barriers developing a sense of belonging or common experience with the others in the group. However, this potential limitation is mitigated by our finding that African-Americans were highly satisfied with the program. Moreover, the Afri can-Americans in our study were able to achieve weight losses comparable to or better than those pa rticipating in studies with exclusively African-American samples (e.g., Kumanyika, 2002). Additional limitations of the study relate to the statistica l analyses and procedures employed. For one, we were unable to comple te a full intent-to-treat analysis because much of the data related to Phase I n on-completers were missing. As mentioned previously, common methods of imputing we ight data (e.g., carry ing last observation forward, substituting baseline weights for missing values) were deemed inappropriate due to the inconsistent patterns of weight lo ss and regain observed wh en comparing AfricanAmerican and Caucasian participants. To acco unt for this deficiency, analyses using all available data (which included both Phase I completers and Phase I non-completers) were conducted, and these revealed a similar patter n of outcomes overall. The only difference lay in the significant within-group weight regain exhibited by African-Americans between Months 6 and 18. This discrepancy, coupled with the finding that a marginally higher percentage ( p = .09) of African-A mericans (compared to Caucasians) were classified as Phase I non-completers, suggest s that the outcomes for African-Americans may be limited in generalizability. The other limitation related to our statistic al analyses is the sheer number of tests conducted, especially with respect to our tert iary aims. However, we considered these
72 analyses exploratory rather than definitive, as we were interested in exploring trends that might help explain findings related to our primary and secondary aims. Despite these limitations, the current study ha s a number of impor tant strengths. A major limitation of many other studies in th is area is the conf ounding of race/ethnicity and socioeconomic status (e.g., Juhaeri et al., 2003; Kumanyika et al., 1991), wherein African-American participants are of lower in come, employment, and educational status. However, the African-American and Caucasian participants in our sample were comparable on variables related to socioeconomic status. Sim ilar patterns of employment were observed in each racial/ethnic group, with 48.8% of African-American and 47.5% of Caucasian participants holding a fullor pa rt-time job. In addition, similar percentages of African-American and Caucasian partic ipants reported annual incomes between $50,000 and $99,000 (30.2 and 30.4%, respectively). In terms of education, approximately twice as many African-Americans had earned at least a bachelors degree as had Caucasian participants (37.2 and 18.2%, re spectively). Thus, it is unlikely that the racial/ethnic differences in we ight change observed for our participants were driven by differences in resources. Another strength of the pres ent study is that our lifest yle intervention was useful and satisfying for our African-American partic ipants, per their self -report on our program satisfaction questionnaire. A great deal of research is currently being conducted to design weight loss programs that are acceptable to Af rican-American participants (e.g., Kim et al., 2006; Smith West, Dilillo, Bursac, Gore, & Greene, 2007). The majority of these new programs target specific aspects of Af rican-American culture (e.g., spirituality, community) and build weight loss interventi ons into pre-existing infrastructures (e.g.,
73 churches, community groups). Although this approach has mer it, it is possible that such programs discard elements of lifestyle intervention that have proven efficacious in the established literatu re. The TOURS program, on the other hand, was able to accommodate African-American participants with only minor adjustments to a protocol that has produced successful out comes in previous research. Future Directions To our knowledge, the current study is the first to directly compare AfricanAmerican and Caucasian response to program s of extended care. The extended care programs utilized in this study differed in modality (i.e., in-person group counseling vs. individualized telephone based counseling) but were similar in their theoretical underpinnings (i.e., social problem solving). Our finding that extended care did not influence weight change of African-American participants can be interpreted in two ways. One interpretation is that African-American participants are able to maintain weight losses without the support of a personal extended care program. Our AfricanAmerican participants did e xhibit superior maintenance of metabolic improvements and behavior change when compared to their Caucasian counterparts. Further, some have argued that African-Americans smaller cha nges may simply be easier to maintain biologically (e.g., Kumanyika, 2002). However, an alternative interpretation is that an extended care program grounded in social problem solving is not effective for African-American participants. To this end, future research should examine how ex tended care programs governed by other theoretical orientations might influence we ight maintenance among African-Americans. For example, the use of peer educators has em erged as a powerful tool in the treatment of obesity with African-American individuals (e.g., Auslander, Haire-Joshu, Houston, Rhee,
74 & Williams, 2002; Kennedy et al., 2005; Williams et al., 2006). The majority of this research has been based in pre-existing social networks, such as churches or community centers, and has focused on initiation of weight loss. However, this strategy has not been evaluated within a program of extended care following weight loss. The use of peer support as a maintenance strategy has demons trated mixed effectiveness in samples comprised of Caucasian and minority indivi duals (e.g, Perri et al., 1987; Verheijden, Bakx, van Weel, Koelen, & van Staveren, 2005), but benefits specific to AfricanAmerican participants may have been obscu red by the unfavorable responses of other racial/ethnic groups. Summary and Conclusions The present study found that a) Caucasia ns lost more weight than AfricanAmericans during lifestyle intervention, b) Caucasians, but not African-AfricanAmericans, benefited from extended care progr ams utilizing social problem solving, and c) despite disparate weight losses, simila r metabolic and behavi oral (i.e., diet and physical activity) improvements were achieve d by African-American and Caucasians participating in lifestyle interv ention. These findings can be util ized in future clinical and research endeavors that strive to improve the treatment of obesity for African-Americans. Given the dearth of research targeting th ese underserved and high-risk individuals, our pattern of findings might inform ways in whic h future treatment can be more effective, more convenient, and better suited to the goals of this population.
75 LIST OF REFERENCES Agurs-Collins, T.D., Kumanyika, S.K., Te n Have, T.R., & Adams-Campbell, L.L. (1997). A randomized controlled trial of we ight reduction and exer cise for diabetes management in older African-American subjects. Diabetes Care, 20 1503-1511. Appel, S.J., Harrell, J.S., & Deng, S. (2002). Racial and socioeconomic differences in risk factors for cardiovascular di sease among Southern rural women. Nursing Research, 51 140-147. Ard, J.D., Skinner, C.S., Chen, C., Aickin, M., & Svetkey, L.P. (2005). Informing cancer prevention strategies for African American s: the relationship of African American acculturation to fruit, vegetable, and fat intake Journal of Behavioral Medicine, 28 239-247. Auslander, W., Haire-Joshu, D., Houston, C., Rhee, C.W., & Williams, J.H. (2002). A controlled evaluation of staging dietary patt erns to reduce the risk of diabetes in African-American women. Diabetes Care, 25, 809-814. Belza, B., Walwick, J., Shiu-Thornton, S., Schwartz, S., Taylor, M., & LoGerfo, J. (2004). Older adult perspec tives on physical activity and exercise: voices from multiple cultures. Prevention of Chronic Diseases, 1 A09. Block, G., Hartman, A.M., Dresser, C.M., Carroll, M.D., Gannon, J., & Gardner, L. (1986). A databased approach to diet questionnaire design and testing. American Journal of Epidemiology, 124 453-469. Bronner, Y. & Boyington, J.E. (2002). Developi ng weight loss interventions for African American women: elements of successful models Journal of the National Medical Association, 94 224-235. Carpenter, W.H., Fonong, T., Toth, M.J., et al (1998). Total daily energy expenditure in free-living older African Americans and Caucasians American Journal of Physiology, 274 E96E101. Calle, E.E., Thun, M.J., Petrelli, J.M., Rodr iguez, C., & Heath, C.W. Jr. (1999). Body mass index and mortality in a pros pective cohort of U.S. adults. New England Journal of Medicine, 341 1097-1105.
76 Champagne, C.M., Bogle, M.L., McGee, B.B., Yadrick, K., Allen, H.R., Kramer, T.R. et al. (2004). Dietary intake in the lower Mi ssissippi delta region : results from the Foods of our Delta Study. Journal of the American Dietetic Association, 104 199207. Darga, L.L., Holden, J.H., Olson, S.M., & Lucas, C.P. (1994). Comparison of cardiovascular risk factors in obese blacks and whites. Obesity Research, 2 239245. Diabetes Prevention Program Research Gr oup. (2004). Achieving weight and activity goals among Diabetes Prevention Program lifestyle participants. Obesity Research, 12 1426-1435. Djuric, Z., Lababidi, S., Heilbrun, L.K., De pper, J.B., Poore, K.M., & Uhley, V.E. (2002). Effect of low-fat a nd/or low-energy diets on anthropometric measures in participants of the Womens Diet Study. Journal of the American College of Nutrition, 21 38-46. Durazo-Arvizu, R., Cooper, R.S., Luke, A., Prewitt, T.E., Liao, Y., & McGee, D.L. (1997). Relative weight and mortality in U.S. blacks and whites: findings from representative national population samples. Annals of Epidemiology, 7 383-395. Eberhardt, M.S., Ingram, D.D ., Makuc, D.M., et al. (2001). Urban and rural health chartbook. Health, United States, 2001. Hyattsville, MD: National Center for Health Statistics. Eyler, A.A., Matson-Koffman, D., Young, D.R ., Wilcox, S., Wilbur, J., Thompson, J.L. et al. (2003). Quantitative study of correlate s of physical activity in women from diverse racial/ethnic groups. The Womens Cardiovascular Health Network Project: summary and conclusions. American Journal of Preventive Medicine, 25, 93-103. Foster, G.D., Wadden, T.A., Swain, R.M ., Anderson, D.A., & Vogt, R.A. (1999). Changes in resting energy expenditure af ter weight loss in obese African-American and white women. American Journal of Clinical Nutrition, 69 13-17. Gary, T.L., Baptiste-Roberts, K., Gregg, E.W., Williams, D.E., Beckles, G.L., Miller, E.J. III, et al. (2004). Fruit, vegetable and fat intake in a population-based sample of African-Americans. Journal of the National Medical Association, 96 1599-1605. Glass, J.N., Miller, W.C., Szymanski, L. M., Fernhall, B., & Durstine, J.L. (2002). Physiological responses to weight-loss intervention in inactive obese AfricanAmerican and Caucasian women. Journal of Sports, Medicine, & Physical Fitness, 42 56-64. Golightly, Y.M. & Dominick, K.L. (2005). Racial variations in self-reported osteoarthritis symptom severity among veterans. Aging in Clinical and Experimental Research, 17, 264-269.
77 Gomberg-Maitland, M., Frison, L., & Halperin, J.L. (2003). Activ e-control clinical trials to establish equivalence or noninferiority : methodological and st atistical concepts linked to quality. American Heart Journal, 146 398-403. Gower, B.A., Weinsier, R.L., Jordan, J.M ., Hunter, G.R., & Desmond, R. (2002). Effects of weight loss on changes in insulin se nsitivity and lipid concentrations in premenopausal African-American and white women. American Journal of Clinical Nutrition, 76 923-927. Harada, N.D., Chiu, V.,King, A.C., Stewart, A.L. (2001) An evaluation of three self report physical activity inst ruments for older adults. Medicine & Science in Sports & Exercise, 33, 962-970. Hayward, M.D., Pienta, A.M., & McLaughli n, D.K. (1997). Inequality in men's mortality: the socio-economic status gradient and geographic context. Journal of Health and Social Behavior, 38, 313-330. Hill, R.G. & Davies, P.S.W. (2001). The va lidity of self-reported energy intake as determined using the doubly labeled water technique. British Journal of Nutrition, 85 415. Hong, K., Li, Z., Wang, H.J., Elashoff, R., & He ber, D. (2005). Analysis of weight loss outcomes using VLCD in black and white overweight and obese women with and without metabolic syndrome. International Journal of Obesity, 29 436-442. Houtson, D.K., Stevens, J., & Cai, J. (2005) Abdominal fat distri bution and functional limitations and disability in a biracial cohort: the Ar therosclerosis Risk in Communities study. International Journal of Obesity, 29 1457-1463. Hunter, G.R., Weinsier, R.L., Zuckerman, P.A ., & Darnell, B.E. (2004). Aerobic fitness, physiologic difficulty and physical ac tivity in black and white women. International Journal of Obesity, 28 1111-1117. Juhaeri, B., Stevens, J., Chambless, L.E., Ni eto, F.J., Jones, D., Schreiner, P. et al. (2003). Associations of weight loss and changes in fat distribution with the remission of hypertension in a bi-ethnic cohort: the Artherosclerosis Risk in Communities Study. Preventive Medicine, 36 330-339. Kanders, B., Ullman-Joy, P., & Foreyt, J.(1994) The Black American Lifestyle Intervention (BALI): the design of a wei ght loss program for working class African American women. Journal of the American Dietetic Association, 94 310-312. Kennedy, B.M., Paeratakul, S., Champagne, C.M ., Ryan, D.H., Harsha, D.W., McGee, B. et al. (2005). A pilot church-based wei ght loss program for African-American adults using church members as health e ducators: a comparison of individual and group intervention. Ethnicity and Disease, 15 373-378.
78 Kervio, G., Ville, N.S., Leclercq, C., Daubert, J.C., & Carre, F. (2004). Intensity and daily reliability of the six-minute walk test in moderate chronic heart failure patients. Archives of Physical Me dicine and Rehabilitation, 85 1513-1518. Kim, K.H., Linnan, L., Campbell, M.K., Brooks, C., Koenig, H.G., & Wiesen, C. (2006). The WORD (wholeness, oneness, righte ousness, deliverance): A faith-based weight-loss program utilizing a community-b ased participantory research approach. Health Education and Behavior, Epub Dec 15 Klein, S., Sheard, N.F., Pi-Sunyer, X., Daly, A ., Wylie-Rosett, J., Kulkarni, K., & Clark, N.G. (2004). Weight management through lif estyle modification for the prevention and management of type 2 diabetes: rati onale and strategies. A statement of the American Diabetes Associat ion, the North American A ssociation for the Study of Obesity, and the American Society for Clinical Nutrition. American Journal of Clinical Nutrition, 80, 257-263. Knutsen, S.F., Fraser, G.E., Linsted, K.D ., Beeson, W.L., & Shavlik, D.J. (2001). Comparing biological measurements of v itamin C, folate, alpha-tocopherol and carotene with 24-hour dietary recall informa tion in nonhispanic blacks and whites. Annals of Epidemiology, 11 406-416. Kristal, A.R., Shattuck, A.L, & Patterson, R.E. (1999). Differences in fat-related dietary patterns between black, Hispanic, and wh ite women: results from the Womens Health Trial Feasibility St udy in Minority Populations. Public Health Nutrition, 2 253-262. Kristal, A.R., Vizenor, N.C., Patterson, R.E., Neuhouser, M.L., Shattuck, A.L., & McLerran, D. (2000). Precision and bias of food frequencybased measures of fruit and vegetable intakes. Cancer Epidemiology, Biomarkers, & Prevention, 9, 939944. Kumanyika, S.K. (2002). Obesity treatment in minorities. In T.A. Wadden and A.J. Stunkard (Eds.), Handbook of obesity treatment (pp. 416-446). New York: Guilford Press. Kumanyika, S.K., Dorosheko, L.H., Self, M.S., Bahnson, J., & Robertson, J. (1997). Nutritional safety of weight and sodium reduction in hypertensive older adults. Canadian Jouranl of Cardiology, 13, 51B. Kumanyika, S.K., Espeland, M.A., Bahnson, J.L ., Bottom, J.B., Charleston, J.B., Folmar, S., Wilson, A.C., & Whelton, P. K. (2002). Ethnic Comparison of weight loss in the trial of nonpharmacologic in terventions in the elderly. Obesity Research, 10, 96-106. Kumanyika, S.K., Obarzanek, E., Stevens, V. J., Hebert, P.R., & Whelton, P.K. (1991). Weight-loss experiences of black and white participants in NHLBI-sponsored clinical trials. American Journal of Clinical Nutrition, 53 1631S-1638S.
79 Kumanyika, S.K., Shults, J., Fassbender, J., Whitt, M.C., Brake, V., Kallan, M.J. et al. (2005). Outpatient weight management in African-Americans: the Healthy Eating and Lifestyle Program (HELP) study Preventive Medicine, 41 488-502. Kumanyika, S., Wilson, J.F., & Guilford-Davenpor t, M. (1993). Weight -related attitudes and behaviors of black women. Journal of the American Dietetic Association, 93 416-422. Maillet, N.A., DEramo, M.G., & Spollett, G. (1996). Using focus groups to characterize the health beliefs and practices of black women with non-insulin-dependent diabetes. Diabetes Education, 22 39-46. Martin, S.L., Kirkner, G.J., Mayo, K., Matthew s, C.E., Durstine, J.L., & Hebert, J.R. (2005). Urban, rural, and regional variations in physical activity. Journal of Rural Health, 21 239-244. Mayer-Davis, E.J., Vitolins, M.Z., Carmichael, S.L., Hemphill, S., Tsaroucha, G., Rushing, J. et al. (1999). Validity an d Reproducibility of a food frequency questionnaire interview in a multi-cultural epidemiologic study. Annals of Epidemiology, 9, 314-324. McGuire, L.C., Ahluwalia, I.B., & Strine T. W. (2006). Chronic disease-related behaviors in U.S. older women: Behavioral Risk Factor Surveillance System 2003. Journal of Womens Health, 15, 3-7. 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. National Institutes of Health. (2000). Down Home Healthy Cookin (NIH Publication No. 00-3408SV). Nicklas, B.J., Dennis, K.E., Berman, D.M ., Sorkin, J., Ryan, A.S., & Goldberg, A.P (2003). Lifestyle intervention of hypocal oric dieting and walking reduces abdominal obesity and improves coronary heart disease risk factors in obese, postmenopausal, African-Ameri can and Caucasian women. Journal of Gerontology, 58A 181-189. Nothwehr, F. (2004). Attitudes and behaviors related to weig ht control in two diverse populations. Preventive Medicine, 39 647-680. Ogden, C.L., Carroll, M.D., Curtin, L.R., Mc Dowell, M.A., Tabak, C.J., & Flegal, K.M. (2006). Prevalence of over weight and obesity in the United States, 1999-2004. Journal of the American Medical Association, 295 1549-1555. Parks, S.E., Housemann, R.A., & Brownson, R.C. (2003). Differential correlates of physical activity in urban and rural adults of various socioeconomic backgrounds in the United States. Journal of Epidemiology and Community Health, 57 29-35.
80 Patterson, P.D., Moore, C.G., Probst, J.C., & Shinogle, J.A. (2004). Obesity and physical inactivity in rural America. Journal of Rural Health, 20 151-159. Perri, M.G. & Corsica, J.A. (2002). Impr oving the maintenance of weight lost in behavioral treatment of obesity. In T. A. Wadden and A. J. Stunkard (Eds.), Handbook of obesity treatment (pp. 357-379). New York: Guilford Press. Perri, M.G., McAdoo, W.G., McAllister, D.A., La uer, J.B., Jordan, R.C., Yancey, D.Z. et al. (1987). Effects of peer support and ther apist contact on longterm weight loss. Journal of Consulting and Clinical Psychology, 55 615-617. Pollock, M.L., & Wilmore, J.H. (1990). Exercise in health and disease: evaluation and prescription for prevention and rehabilitation. 2nd ed. Philadelphia, PA: W.B. Saunders Co. Ransdell, L.B. & Wells, C.L. (1998). Physical activity in urban white, African-American, and Mexican-American women. Medicine & Science in Sports & Exercise, 30, 1608-1615. Resnicow, J., Odom, E., Wang, T., Dudley, W.N., Mitchell, D., Vaughan, R. et al. (2000). Validation of three food frequency questionnaires and 24-hour recalls with serum carotenoid levels in a sa mple of African-American adults American Journal of Epidemiology, 152, 1072-1080. Resnicow, K., McCarty, F., Blissett, D., Wang, T., Heitzler, C., & Lee, R.E. (2003). Validity of a modified C HAMPS physical activity questionnaire among AfricanAmericans. Medicine & Science in Sports & Exercise, 35 1537-1545. Rippe, J.M. (1998). The case for medical mana gement of obesity: a call for increased physician involvement. Obesity Research, 6 23S-33S. Sanchez, A.M., Reed, D.R., & Price, R.A. (2000). Reduced mortality associated with body mass index (BMI) in African Am ericans relative to Caucasians. Ethnic Disparities, 10 24-30. Simkin-Silverman, L., & Wing, R. (1997). Ma nagement of obesity in primary care. Obesity Research, 5 603-612. Singh, P.N., Fraser, G.E., Knutsen, S.F., Lindsted, K.D., & Bennett, H.W. (2001). Validity of a physical activity questionna ire among African-American Seventh-day Adventists. Medicine & Science in Sports & Medicine, 33 468-475. Smith, D.E., Thompson, J.K., Raczynski, J.M ., & Hilner, J.E. (1999). Body image among men and women in a biracial cohort: The CARDIA study International Journal of Eating Disorders, 25 71-82.
81 Smith West, D., Dilillo, V., Bursac, Z., Gore S.A., & Greene, P.G. (2007). Motivational interviewing improves weight loss in women with Type 2 diabetes. Diabetes Care, Epub Mar 2. Starling, R.D., Toth, M.J., Matthews, D.E., Poehlman, E.T. (1998). Energy requirements and physical activity of older free-liv ing African Americans: a doubly-labeled water study. Journal of Clinical Endoc rinology and Metabolism, 83 1529. Stevens, V.J., Obarzanek, E., Cook, N.R., Lee, I., Appel, L.J., West, D.S., Milas, N.C., Mattfeldt-Beman, M., Beldens, L., Bragg, C., Millstone, M., Raczynski, J., Brewer, A., Singh, B., & Cohen, J. (2001). Long-term weight loss and changes in blood pressure: Results of the trials of hypertension prevention, phase II. Annals of Internal Medicine, 134 1-11. Stewart, A.L., Mills, K.M., King, A.C., Haske ll, W.L., Gillis, D., & Ritter, P.L. (2001). CHAMPS physical activity questionnai re for older adults: outcomes for interventions. Medicine and Science in Sports and Exercise, 33 1126-1141. Subar, A.F., Thompson, F.E., Kipnis, V., Midthune, D., Hurwitz, P., McNutt, S., McIntosh, A., & Rosenfeld, S. (2001). Comp arative validation of the Block, Willet, and National Cancer Institute F ood Frequency Questionnaires. American Journal of Epidemiology, 154 1089-1099. Svetkey, L.P., Erlinger, T.P., Vollmer, W.M., Fe ldstein, A., Cooper, L.S., Appel, L.J. et al. (2005). Effect of lifestyle modifi cations on blood pressure by race, sex, hypertension status, and age. Journal of Human Hypertension, 19, 21-31. Tamayo-Sarver, J.H., Albert, J.M., Tama yo-Sarver, M., & Cydulka, R.K. (2005). Advanced statistics : how to determine whether your intervention is different, at least as effective as, or e quivalent: a basic introduction. Academic Emergency Medicine, 12 536-542. Trabulsi, J. & Schoell er, D.A. (2001). Evaluation of dietary assessment instruments against doubly labeled water, a biom arker of habitual energy intake. American Journal of Physiology: Endocrinology and Metabolism, 281 E891-E899. Tucker, K.L., Maras, J., Champagne, C., Connell, C., Goolsby, S., Weber, J. et al. (2005). A regional food-frequency questionnai re for the US Mississippi Delta. Public Health Nutrition, 8 87-96. Tuten, C., Petosa, R., Sargent, R., & Weston, A. (1995). Biracial differences in physical activity and body composition among women. Obesity Research, 3 313. Underwood, S.M. & Powell, R.L. (2006). Re ligion and spirituality: influence on health/risk behavior and cancer screeni ng behavior of African Americans. ABNF, 17 20-31.
82 U.S. Census Bureau. (2006). Americas Families and Living Arrangements: 2005. Last Revised September 21, 2006. Retrieved 15 March 2007 from U.S. Census http://www.census.gov/population/www/socdemo/hh-fam/cps2005.html Verheijden, M.W., Bakx, J.C., van Weel, C., Koelen, M.A., & van Staveren, W.A. (2005). Role of social support in lif estyle-focused weight management interventions. European Journal of Clinical Nutrition, 59 S179-S186. Wadden, T.A., Butryn, M.L., & Byrne, K.J. (2004) Efficacy of lifestyle modification for long-term weight control Obesity Research, 12 151S-162S. Walsh, M.C., Hunter, G.R., Sirikul, B., & Gower, B.A. (2004). Comparison of self reported with objectively assessed energy expenditure in black and white women before and after weight loss. American Journal of Clinical Nutrition, 79 10131019. Wanko, N.S., Brazier, C.W., Young-Rogers, D., Dunbar, V.G., Boyd, B., George, C.D. et al. (2004) Exercise preferences and barrier s in urban African Americans with type 2 diabetes. Diabetes Education, 30 502-513. Wassertheil-Smoller, S., Langford, H.G., Bl aufox, M.D., Oberman, A., Hawkins, M., Levine, B. et al. (1985). E ffective dietary intervention in hypertensives: sodium restriction and weight reduction Journal of the American Dietetic Assocation, 85, 423-430. Washburn, R.A., Kline, G., Lackland, D. T., & Wheeler, F.C.(1992). Leisure time physical activity: are there black/white differences? Preventive Medicine, 21, 127135. Weinsier, R.L., Hunter, G.R., Desmond, R.A., Byrne, N.M, Zuckerman, P.A., & Darnell, B.E. (2002). Free-living activity energy expenditure in women successful and unsuccessful at maintaining a normal body weight. American Journal of Clinical Nutrition, 75 499-504. Weinsier, R.L., Hunter, G.R., Schultz, Y., Zuckerman, P.A., & Darnell, B.E. (2002). Physical activity in free-living, overwei ght white and black women: divergent responses by race to diet-induced weight loss. American Journal of Clinical Nutrition, 76 736-742. Weinsier, R.L., Hunter, G.R., Zuckerman, P.A., Redden, D.T., Darnell, B.E., Larson, D.E. et al. (2000). Energy expenditure and free-living physical act ivity in black and white women: comparison before and after weight loss. American Journal of Clinical Nutrition, 71 1138-1146. William, J.H., Auslander, W.F., de Groot M., Robinson, A.D., Houston, C., & HaireJoshu, D. (2006). Cultural relevancy of a diabetes prevention nutrition program for African American women. Health Promotion and Practice, 7 56-67
83 Wing, R.R. (2004). Behavioral approaches to the treatment of obesity. In: Bray, G.A. & Bouchard, C., eds. Handbook of Obesity: C linical Applications 2nd ed., New York: Marcel Dekker. Wing, R.R. & Anglin, K. (1996). Effectiveness of a behavioral wei ght control program for blacks and whites with NIDDM. Diabetes Care, 19, 409-413. Wing, R.R. & Gorin, A.A. (2003). Behavioral techniques for treating the obese patient. Primary Care, 30 375-371. Wing, R.R., Hamman, R.F., Bray, G.A. et al. (2004). Achieving weight and activity goals among diabetes prevention progr am lifestyle participants. Obesity Research, 12, 1426-1434. Wylie-Rosett, J., Wassertheil-Smoller, S., Blaufox, S., Davis, B.R., Langford, H.G., Oberman, A., et al. (1993). Trial of anti hypertensive intervention and management: greater efficacy with weight reduction than with a sodi um-potassium intervention. Journal of the American Dietetic Association, 93 408-415. Yanovski, S.Z., Gormally, J.F., Lesser, M.S., Gwirtsman, H.E., & Yanovski, J.A. (1994). Binge eating disorder affects outcome of comprehensive very -low-calorie-diet treatment. Obesity Research, 2 205-212. Young, D.R. & Stewart, K.J. (2006). A church-based physical activity intervention for African American women. Family and Community Health, 29, 103-117. Zugck, C., Kruger, C., Durr, S. et al. (2002). Is the 6-minute walk test a reliable substitute for peak oxygen uptake in patient s with dilated cardiomyopathy? European Heart Journal 21, 540-549.
84 BIOGRAPHICAL SKETCH Katie Allison Rickel was born Decemb er 21, 1980 to Jody Rickel and David Rickel. She was raised in El kins Park, Pennsylvania, with her younger sister, Emily. She graduated from Cheltenham High School in 1999, and she earned a bachelors degree in psychology from Duke University in 2003. Since 2003, Katie has been a doctoral student at the University of Florida in the Department of Clinical and Health Ps ychology, specializing in medical psychology. Katie will return to Duke University to co mplete her predoctoral internship at Duke University Medical Center, from 2007-2008, after which she will have fulfilled all requirements for her doctorate. Her ultimate goal is to be a psychologist in a medical setting.