Harsohena Kaur. Nicole Nollen, Christie Befort, Kim Pulvers, and Aimee S. James. Jasjit S. Ahluwalia. Matthew S. Mayo and Qingjang Hou

Transcription

1 Health Psychology Copyright 2008 by the American Psychological Association 2008, Vol. 27, No. 3(Suppl.), S252 S /08/$12.00 DOI: / (Suppl.).S252 Demographic and Psychosocial Factors Associated With Increased Fruit and Vegetable Consumption Among Smokers in Public Housing Enrolled in a Randomized Trial Nicole Nollen, Christie Befort, Kim Pulvers, and Aimee S. James University of Kansas Medical Center Matthew S. Mayo and Qingjang Hou University of Kansas Medical Center Harsohena Kaur University of Minnesota School of Medicine Jasjit S. Ahluwalia University of Minnesota School of Medicine Objectives: To examine the demographic and psychosocial factors associated with increased fruit and vegetable (FV) consumption among smokers residing in public housing. Design: Cluster randomized trial of 20 public housing developments (HDs). Ten housing developments were randomly assigned to a FV intervention and 10 to a smoking cessation intervention. Primary Outcome: Change in daily FV intake over the past 7 days at 8 weeks postbaseline. Results: Above the effect of treatment, baseline confidence for vegetable consumption (model coefficient 0.19, SE 0.07, p.01), decreased barriers (model coefficient 0.12, SE 0.04, p.002) and increased agency (model coefficient 0.08, SE 0.04, p.03) were significantly associated with week 8 FV consumption. Conclusions: Although the intervention produced significant change in FV consumption, the majority of individual psychosocial factors were not associated with this change. Future studies examining the dynamic interaction between interventions and individual, social, and environmental factors are needed to more fully explain dietary change among public housing residents. Keywords: fruits, vegetables, dietary change, low income Fruit and vegetable (FV) consumption is protective against several chronic diseases, including cardiovascular disease (Ascherio, Katan, Zock, Stampfer, & Willett, 1999; Hooper et al., 2001) and certain cancers (Block, Patterson, & Subar, 1992; Gandini, Merzenich, Robertson, & Boyle, 2000; Wargovich, 1997). In 1991, the National Cancer Institute (NCI) initiated the 5 A Day for Better Health Program as a national effort to increase FV consumption. Since then, several behavioral interventions aimed at increasing FV consumption have been examined (Langenberg et al., 2000; Resnicow et al., 2004; Steptoe, Perkins-Porras, Rink, Hilton, & Cappuccio, 2004). In a recent meta-analysis, Ammerman et al. (2002) found that among 22 behavioral interventions for improving FV intake, 17 reported significant increases, with an Nicole Nollen, Christie Befort, Kim Pulvers, and Aimee S. James, Department of Preventive Medicine and Public Health and the Kansas Masonic Cancer Research Institute, University of Kansas Medical Center; Harsohena Kaur, Department of Pediatrics, University of Minnesota School of Medicine; Matthew S. Mayo and Qingjang Hou, Department of Preventive Medicine and Public Health and the Kansas Masonic Cancer Research Institute, University of Kansas Medical Center; Jasjit S. Ahluwalia, Department of Medicine and Office of Clinical Research, University of Minnesota School of Medicine. This research was supported by NIH R01 CA Correspondence concerning this article should be addressed to Nicole L. Nollen, Ph.D., University of Kansas Medical Center, 3901 Rainbow Boulevard, Mail Stop 1008, KS City, KS nnollen@kumc.edu average increase of 0.6 servings per day (Ammerman, Lindquist, Lohr, & Hersey, 2002). Despite these efforts, FV consumption among the United States population remains well below recommended guidelines; in 2000, only one fourth of the U.S. population consumed at least 5 FV servings per day (Serdula et al., 2004). Low socioeconomic groups have even lower FV intake compared to their more affluent counterparts (Serdula et al., 2004) and are at increased risk for many cancers and chronic illnesses associated with diet (Ward et al., 2004). Individuals of lower socioeconomic status are also more likely to smoke (Paavola, Vartiainen, & Haukkala, 2004), and in turn, smokers report consuming fewer FV compared to nonsmokers (Genkinger, Platz, Hoffman, Comstock, & Helzlsouer, 2004). Taken together, smokers of low socioeconomic status are among those at highest risk for inadequate FV consumption and for multiple negative health consequences. This group, therefore, may derive substantial health benefit from increasing their FV consumption (Hung et al., 2004). Understanding factors that determine FV consumption is important for improving the effectiveness of dietary interventions. Among low-income populations, several psychosocial factors have been identified as cross-sectional correlates of FV consumption, including financial hardship, time constraints, knowledge of recommended intake, perceived barriers and benefits, self-efficacy, social support, and the perceived importance of increased FV intake (Trudeau, Kristal, Li, & Patterson, 1998; Van Duyn et al., 2001). There is evidence that positive dispositional traits, such as S252

2 FACTORS ASSOCIATED WITH INCREASED FV CONSUMPTION S253 goal-directedness/hope (Snyder, 2002), are associated with preventive health behaviors such as physical activity (Harney, 1990), protective sexual practices (Floyd & McDermott, 1998), and adherence to medical regimes (Moon, Snyder, & Rapoff, 2001). The connection between goal-directedness/hope and FV consumption has never been examined. Furthermore, FV intervention studies with low income populations have demonstrated that changes in psychosocial factors predict changes in FV intake longitudinally. For example, Langenberg et al. (2000) found that improvements in self-efficacy, social support, knowledge, perceived importance, and personal responsibility for FV consumption accounted for most of the intervention effect among nearly 1500 low-income women enrolled in a dietary intervention program (Langenberg et al., 2000). Steptoe et al. (2004) likewise found that short-term changes in psychosocial factors after an 8-week intervention, including improved social support, self-efficacy, and perceived benefits, significantly predicted 12-month changes in FV consumption among a low-income population (Steptoe et al., 2004). Our group recently completed a randomized clinical trial with smokers residing in public housing developments. A FV intervention was used as the comparison group, in place of a placebocontrolled group for the smoking cessation arm. The results of the trial, described in detail in elsewhere (Ahluwalia et al., 2007), found significantly greater increases in fruits and vegetables in the FV group compared to the cessation group. The purpose of the present analysis was to further our understanding of the factors that led to increased FV consumption in this high-risk group. In particular, our objectives were the following: (1) examine the baseline demographic and psychosocial factors associated with baseline FV consumption, (2) examine what baseline demographic and psychosocial factors predicted change in FV consumption at week 8, and (3) assess whether changes in psychosocial factors from baseline to week 8 were associated with changes in consumption. Based on previous research and the theoretical frameworks guiding the intervention, we were particularly interested in the impact of demographic variables, perceived barriers, social support, autonomous regulation, motivation, confidence, and goal-directedness on dietary change. Method, Setting, and Design Study design, methods, and inclusion/exclusion criteria have been described in detail elsewhere (Jeffries, Choi, Butler, Harris, & Ahluwalia, 2005). In brief, this study was a cluster-randomized trial in which 20 public and section 8 housing developments (HDs) were randomly assigned to a fruit and vegetable (FV) or smoking cessation intervention. All HD residents were invited to attend a community health fair held at their development. Health fair attendees completed a questionnaire assessing smoking status. From this, smokers were identified for potential inclusion in the randomized trial. All participants provided written informed consent. The trial procedures were approved by the University of Kansas Medical Center s human subjects committee. Randomization occurred at the HD level at the conclusion of each health fair. Block randomization (with a block size of 4) was used to stratify by elderly and family HDs because age is associated with key smoking variables. Treatment assignment was revealed to the study coordinator and research staff only after each health fair was complete. Sequential enrollment continued until 20 HDs were randomized; 10 to the smoking cessation arm and 10 to the comparison arm. Recruitment occurred between October 2001 and May 2003; participants were followed for 6 months. Intervention The FV intervention included three components: MI counseling sessions, educational materials, and resources that included a bag of fresh FV and a cookbook of healthy recipes containing FV. MI was used as a counseling technique to move participants toward commitment to change through the use of reflective listening and exploration of their ambivalence about change (Miller & Rollnick, 2002). Each participant received five individual MI sessions with a trained master s level counselor at baseline, Day 10, and Weeks 3, 5, and 20. in the FV arm received standard FV educational materials at baseline and two videos addressing how to buy FV on a budget and how to incorporate FV at every meal and snack. in the smoking cessation arm received an 8-week supply of 4-mg nicotine gum, instructions for using the gum, educational materials related to quitting, and five sessions of motivational interviewing (MI) counseling aimed at smoking cessation. Theoretical Framework Selection of the intervention components was guided by Social Cognitive Theory (SCT), the Health Belief Model (HBM), and Two-Dimensional Model of Cultural Sensitivity. SCT posits that self-efficacy (an individual s belief in their ability to make a change) and outcome expectations (belief that making a change will lead to a desired outcome) are important determinants of health behavior change (Maibach & Cotton, 1995). Accordingly, the intervention addressed the importance of and self-efficacy to change and positive expectations for cooking healthy, such as the nutritional value and health benefits of commonly available FV. HBM suggests that behavior change is most likely to occur when an individual believes her/she is at risk (perceived susceptibility), that the consequences are serious (perceived severity), and that the benefits (perceived benefits) outweigh the barriers (perceived barriers; Becker, 1977). Following from the HBM, a portion of each MI session focused on an activity where the benefits and barriers to increasing FV consumption were explored. Another exercise encouraged participants to consider the reason for and risks and consequences of maintaining their current level of intake. Finally, the Two-Dimensional Model of Cultural Sensitivity guided the cultural tailoring of the intervention (Resnicow, Soler, Braithwaite, Ahluwalia, & Butler, 2000). Surface structures were incorporated by featuring predominantly African American images and graphics in the materials and by developing a cookbook incorporating recipes and foods common among the African American community. Deep structures were incorporated through the print materials and counseling, which addressed unique barriers (i.e., poverty, dietary prefer-

3 S254 NOLLEN ET AL. ences) and core cultural values (i.e., religion/spiritualism, collectivism, and family) of our study sample. Outcome Variable Fruit and vegetable intake. The primary outcome for the present analyses was change in FV consumption at week 8, assessed using a 2-question measure adapted from the Health Habits and History Questionnaire (HHHQ; Block, Hartman, & Naughton, 1990; Block, Thompson, Hartman, Larkin, & Guire, 1992). Using an open-ended response format, items ask, Over the past 7 days, how many times did you eat fruit, not including vegetables and Over the past 7 days, how many times did you eat vegetables, not including fruit. A follow-up question asked if this was times per day or per week. All responses were converted to a times per day estimate. Items were then summed to derive an index of daily FV intake. The reliability of this measure has been documented in previous studies (Resnicow et al., 2004, 2001). Additionally, validity data indicates small to moderate correlations with 24-hr dietary recalls and total serum carotenoids (Resnicow, Odom, et al., 2000). Predictor Variables Demographic characteristics. Baseline assessment of demographic information included age, race/ethnicity, marital and employment status, income, and education. Psychosocial factors. Perceived barriers to FV consumption were assessed with 13 items deemed most relevant by expert consultants from a 27-item scale (Resnicow et al., 2000; Resnicow & Wallace, 2000). Items used a 4-point continuum ranging from Doesn t affect me at all to Makes it very difficult. Evidence of construct validity is derived from significant associations found between barriers and FV consumption (r , p.01) in previous studies (Resnicow et al., 2000) and the significant relationship found between decreased barriers and FV change (model coefficient 0.12, SE 0.04, p.002) in the present study. Internal consistency for the items used in this study was Social support for eating FV was assessed by 2 items derived from the work of Sallis (1987; Sallis, Grossman, Pinski, Patterson, & Nader, 1987) and Baranowski et al. (1995). Items included, How much encouragement do you get from your family to eat more FV? and How much encouragement do you get from your friends or work colleagues to eat more FV? Response categories were on a 4-point continuum ranging from None to A lot. Evidence of predictive validity is derived from the scale s ability to predict dietary change found in previous studies (Fuemmeler et al., 2006). Internal consistency for the items used in this study was Motivation and confidence were assessed separately for fruits and vegetables using single items found to be related to dietary and smoking cessation outcomes in previous studies (Resnicow et al., 2000; Resnicow & Wallace, 2000; Rollnick, Butler, & Stott, 1997). Questions were phrased based on the program s predetermined goal of encouraging a daily intake of up to four servings of fruits and five vegetables. Fruit items included On a scale from 0 10, with 0 representing not at all motivated and 10 representing very motivated, how motivated are you to increase/maintain your eating of 4 fruits/day? and Assuming that you wanted to, how confident (0 [not at all confident] and 10 [very confident]) are you that you can increase/ maintain your eating of 4 fruits/day? Items assessing motivation and confidence for increasing vegetable were phrased similarly. Autonomous regulation, or the extent to which participants increased their FV consumption because it held personal importance for them (vs. doing so because of external pressure), was assessed with three items from the 6-item Autonomous Motivation subscale of the Treatment Self-Regulation Questionnaire (TSRQ; Williams, McGregor, Zeldman, Freedman, & Deci, 2004). Items selected were based on those with the highest factor loadings from a similar study of dietary change (Resnicow et al., 2005). A sample item is: The reason I eat fruits and vegetables is because I personally believe it is a good thing for my health. Responses are on a 10-point scale ranging from Not at all true to Very true. The TSRQ has demonstrated adequate reliability and validity (Williams, Grow, Freedman, Ryan, & Deci, 1996; Williams, McGregor, Zeldman, Freedman, & Deci, 2004; Williams, Rodin, Ryan, Grolnick, & Deci, 1998). Internal consistency for the items used in this study was Goal-directedness was assessed using the Hope Scale (Snyder et al., 1991). This scale consists of 8 items from which a summary score and two subscale scores (Pathways and Agency) are derived. The Pathways subscale taps the perceived ability to find routes to reach desired goals and the Agency subscale measures the perceived motivation to sustain goal-relevant pursuits. Sample items include: There are lots of ways around any problem (Pathways) and I energetically pursue my goals (Agency). Responses fall along an 8-point continuum ranging from Definitely false to Definitely true. The Hope Scale has demonstrated adequate reliability ( ) and validity (Snyder et al., 1991) and, in the current study, was found to be internally consistent (.80, Agency;.82, Pathways). Procedures Trained research staff explained the purpose of the study to participants and obtained informed consent. Upon enrollment, research staff and participants were blind to treatment assignment (e.g., smoking cessation or FV consumption). learned of their treatment assignment at their first in-person counseling visit (Week 0). A baseline assessment was conducted at the health fair. Outcome assessments were conducted at Weeks 8 and 26. A trained research staff member read all survey items aloud to participants. were compensated $40 at the health fair and at Weeks 8 and 26. To minimize attrition, participants received phone-calls and postcards reminding them of each upcoming visit or telephone counseling session, and additional incentives (e.g., movie tickets, T-shirt, water bottle, or tote bag) were provided at interim in-person visits. Statistical Analyses Categorical variables were summarized with frequencies and percentages and continuous variables were summarized with means and standard deviations. Change scores were compared between the two treatment groups using a mixed model to account for the intraclass correlation. To minimize the possibility of a Type I error, the individual impact of each predictor on baseline FV

4 FACTORS ASSOCIATED WITH INCREASED FV CONSUMPTION S255 consumption and FV change was assessed using a mixed model adjusted for treatment and clustering (intraclass correlations) within housing developments. Only significant variables were included in the final mixed model, which examined the factors jointly associated with FV change. For all mixed models, the Proc Mixed approach was used. This approach assumes a compound symmetric correlation to estimate the clustering, or intraclass correlations, within housing developments. All analyses were performed on an intent to treat basis. Missing values for FV intake were imputed as no change. Results On average, 21% (range 8% 66%) of adult residents attended a health fair at their HD. Of the 813 health fair attendees, 273 smokers (33.6%) were identified and screened, 204 met screening criteria, and 173 (84.8% of those eligible) were randomized, of whom 107 were assigned to the FV arm and 66 were assigned to the cessation arm (see Figure 1). Of the 173 participants, 151 (87.3%) attended their Week 8 visit. No significant differences were found between groups in the rate of missing data ( p.22) or in baseline FV intake among those who dropped out and those who returned ( p.87). Baseline characteristics of the study sample are summarized in Table 1. Means, standard deviations, and change scores by treatment group for FV intake and the psychosocial variables at baseline and Week 8 are summarized in Table 2. Significant differences were found in FV change scores between the FV and cessation groups. Specifically, at 8 weeks postbaseline, the increase in FVs reported by the intervention group was 1.58 ( p.001) servings greater than the increase reported in the cessation group (see Table 2). No other significant between group differences were found in psychosocial change scores. Baseline Factors Associated With Baseline FV Consumption At baseline, motivation (r s 0.18, p.02) and confidence (r s 0.22, p.004) for eating more fruit, motivation (r s 0.18, 20 developments approached and assessed for eligibility Randomized 20 developments, mean # part./hd=8.65; range= participants screened 204 participants eligible 173 randomized Excluded 0 developments 57 participants did not meet inclusion criteria 7 participants declined participation 11 other reasons Allocated to FV intervention 10 developments, mean # part/hd=10.7; range= enrolled Allocated to smoking cessation intervention 10 developments, mean # part/hd=6.6; range= enrolled Lost to follow-up at Week 8 0 developments 11/107 (10.3%) did not return at Week 8 Lost to follow-up at Week 8 0 developments 11/66 (16.7%) did not return at Week 8 Figure 1. Flow diagram of PATH cluster randomized trial.

5 S256 Table 1 Baseline Characteristics by Treatment Group NOLLEN ET AL..03) were significantly associated with increased FV consumption at week 8. p.02) and confidence (r s 0.19, p.01) for eating more vegetables, autonomous regulation (r s 0.21, p.005), perceived ability to find routes/alternatives to reach desired goals (i.e., pathways) (r s 0.19, p.01), perceived motivation to sustain pursuit of desired goals (i.e., agency) (r s 0.18, p.02), and goal-directedness (r s 0.18, p.02) were significantly related to baseline FV consumption. Baseline Predictors of FV Change FV (n 107) A series of t tests were run to assess whether demographic and baseline psychosocial factors predicted change in FV consumption at Week 8 after accounting for treatment group. Baseline confidence for eating more vegetables was a significant predictor of change in FV consumption at Week 8 (see Table 3). For every unit increase in confidence, FV consumption increased by 0.19 servings per day (SE 0.07, p.01). Psychosocial Change as Correlates of FV Change Cessation (n 66) Demographic characteristics Age (in years), M (SD) 48 (13.1) 43 (14.3) Female, no. (%) 68 (63.6%) 53 (80.3%) Race/ethnicity, no. (%) African American 93 (86.9%) 50 (75.8%) White 7 (6.5%) 11 (16.7%) Hispanic 2 (1.9%) 3 (4.6%) Other 4 (3.7%) 2 (3.0%) Married or living with a partner, 10 (9.4%) 5 (7.6%) no. (%) HS education, no. (%) 46 (43.0%) 22 (33.3%) Unemployed, no. (%) 42 (39.3%) 30 (45.5%) Monthly income $ (72.9%) 49 (74.2%) Insured, no. (%) 85 (79.4%) 49 (74.2%) Smoking-related variables Cigarettes smoked per day, M (SD) 16 (9.2) 19 (13.0) Age of initiation (in years), M (SD) 19 (6.6) 18 (5.2) To examine the individual associations between psychosocial and FV change scores from baseline to Week 8, a series of mixed models were run. Each psychosocial change score was analyzed in a separate model along with treatment. The results of these analyses are presented in Table 3. After accounting for the effect of treatment, increases in FV intake at week 8 were associated with a decrease in barriers and an increase in agency. For every unit decrease in barriers and increase in agency, FV consumption increased by 0.12 (SE 0.04, p.001) and 0.08 (SE 0.04, p.03) servings per day, respectively. Barriers and agency were then analyzed together, along with treatment, in a final mixed model to examine if changes in barriers and agency were jointly associated with change in FV consumption at week 8. Results of the final model, displayed in Table 4 indicate that, above the effect of treatment, decreased barriers (model coefficient 0.12, SE 0.04, p.002) and increased agency (model coefficient 0.08, SE 0.04, p Discussion Consistent with other studies conducted among low-income populations (Langenberg et al., 2000; Steptoe, Perkins-Porras, Rink, Hilton, & Cappuccio, 2004), this secondary analysis of a randomized trial found small but significant cross-sectional associations between baseline FV consumption, motivation, and confidence for fruits and vegetables, autonomous regulation, and goal-directedness, indicating that individuals scoring higher on these domains were consuming a greater number of fruits and vegetables at the onset of the trial. Of these baseline factors, only confidence for vegetables remained a significant predictor of FV intake at Week 8; however, it is possible that this relationship may be due to chance (i.e., Type I error), given the number of baseline factors examined. Consistent with the literature, which has found limited predictive utility of baseline psychosocial factors (Baranowski, Cullen, & Baranowski, 1999; Steptoe et al., 2004), none of the other baseline factors significantly predicted 8-week FV consumption, calling into question the usefulness of baseline characteristics in predicting longitudinal dietary change among a low-income sample. Although the reasons for these findings are unknown, it is possible that participants struggled to accurately estimate psychosocial factors at baseline causing a disconnect between initial attitudes/intentions and the ability to carry out dietary change (Steptoe, Perkins-Porras, Rink, Hilton, & Cappuccio, 2004). For example, participants may have overestimated their baseline confidence to eat more fruits and vegetables due to inexperience with change attempts and overcoming barriers. Barriers displayed an increasing trend from baseline to Week 8, indicating that barriers outside of their control (e.g., fresh fruits and vegetables were too costly or not readily available at their grocery stores) may have diminished any effect of initial perceived ability to increase FV intake. Contrary to many of the current health behavior change models (Montano & Kasprzyk, 2002), it is also possible that attitudes are an artifact of behavior, not the cause, thus limiting their predictive utility. The notion of behavior-shaping attitude is not new to the psychology literature (Bem, 1972), and a recent meta-analysis of physical activity change supports this idea. Specifically, although attitude, intention, and self-efficacy accounted for much of the variance in activity level across studies, these relationships were attenuated when prior behavior entered the model, suggesting that prior behavior may be more important in influencing behavior change than current attitudes (Haggar, Chetzisarantis, & Biddle, 2002). Future research is needed to further examine and explore the causal relationships between past and/or present behavior and current attitudes in shaping behavior, particularly as it relates to dietary patterns. Our findings and those of others (Langenberg et al., 2000; Steptoe et al., 2004), indicate that changes in psychosocial factors may play a more important role in dietary change than baseline values. Specifically, our results suggest that interventions aimed at decreasing FV-specific barriers and increasing agency-related goal-directedness may be particularly important for improving FV intake in this low-income population. Inter-

6 FACTORS ASSOCIATED WITH INCREASED FV CONSUMPTION S257 Table 2 Factor Means, Standard Deviations, and Change Scores by Treatment Group Baseline Week 8 Change score Factor Group M (SD) M (SD) M (SD) p a FV FV 2.06 (1.73) 3.88 (3.02) 1.82 (2.68) Cessation 2.17 (1.63) 2.41 (1.94) 0.24 (1.73) Perceived barriers FV 4.71 (4.90) 5.42 (5.02) 0.71 (5.17) Cessation 5.38 (5.69) 4.68 (4.76) 0.69 (4.38) 0.07 Social support: family FV 1.40 (1.30) 1.50 (1.26) 0.10 (1.14) Cessation 1.12 (1.33) 1.19 (1.35) 0.08 (1.06) Social support: friends FV 0.78 (1.07) 1.03 (1.21) 0.25 (1.33) Cessation 0.74 (1.15) 0.92 (1.26) 0.18 (0.99) Motivation for fruit FV 7.74 (2.38) 8.55 (2.35) 0.81 (2.27) Cessation 7.24 (3.09) 8.18 (3.01) 0.94 (2.30) Motivation for vegetables FV 7.88 (2.49) 8.49 (2.34) 0.62 (2.36) 0.75 Cessation 7.59 (2.92) 7.76 (2.95) 0.17 (2.89) Confidence for fruit FV 8.31 (2.07) 9.07 (1.98) 0.76 (2.49) 0.27 Cessation 8.11 (2.74) 8.53 (2.34) 0.42 (2.61) Confidence for vegetables FV 7.93 (2.62) 8.72 (2.19) 0.79 (2.59) 0.39 Cessation 8.38 (2.29) 8.48 (2.53) 0.11 (2.37) Autonomous regulation FV 28.5 (3.38) 28.9 (3.45) 0.41 (3.64) 0.08 Cessation 26.2 (7.01) 27.2 (6.30) 0.95 (4.54) Goal-directedness: total FV 49.8 (11.8) 50.2 (11.1) 0.38 (10.8) 0.41 Cessation 50.9 (12.5) 50.9 (10.5) 0.02 (7.83) 0.80 Goal-directedness: pathways FV 24.9 (5.92) 25.1 (5.92) 0.19 (5.58) Cessation 25.7 (5.87) 26.1 (5.02) 0.29 (3.67) 0.89 Goal-directedness: agency FV 25.2 (6.15) 25.2 (5.80) 0.07 (5.18) Cessation 25.6 (6.25) 24.9 (6.10) 0.74 (4.43) 0.29 Note. a Comparison of change scores between groups. estingly, agency reflects a general cognitive disposition rather than a FV-specific construct and suggests that individuals who believe they have the ability to reach desired goals, and are able to maintain this positive disposition in the face of challenges, are more likely to experience success (Snyder, 2002). Results highlight the importance of addressing both construct- and trait-specific factors in eliciting dietary change and suggest that agency may be a necessary component to compliment behavior change by providing individuals the willpower to carry out construct-specific changes. However, these results should not be overinterpreted. No significant relationships were found between treatment and psychosocial change scores, suggesting that decreased barriers and increased agency were brought about by factors outside of the intervention. It may be that the intervention impacted psychosocial factors (e.g., knowledge, intentions, social norms) not examined in the current study. Alternatively, the lack of relationships between treatment and psychosocial change may be an artifact of our cluster randomized design, which utilized housing developments as the primary unit of analysis, thereby limiting the degrees of freedom and decreasing the likelihood of finding significant relationships. It is also important to note that the majority of baseline and psychosocial change factors were not significantly associated with behavior change. This lack of a relationship is consistent with the literature in this area (Baranowski, Cullen, & Baranowski, 1999) and may be due, in part, to our assumption of linear relationships. As suggested by Resnicow and Vaughn (2006), the inability of current behavior change models to adequately predict behavior change may be related to an assumption that change is a rational and gradual process under an individual s volitional control. In contrast, some literature suggests that behavior change may also be related to teachable moments (e.g., being diagnosed with a chronic health condition) or spontaneous, unplanned, or random decisions that are difficult to predict but may serve as tipping points (Larabie, 2005; Matzger, Kaskutas, & Weisner, 2005; West & Sohal, 2006). In addition, the inability of the majority of psychosocial factors to predict FV consumption may be due to our exploration of individual-level variables to the exclusion of broader social and environmental factors (Baranowski, Cullen, Nicklas, Thompson, & Baranowski, 2003). Future research in the area could be improved by innovative methods that invoke and test the random events hypothesis, by implementation of multilevel interventions addressing the complex interactions between individual, social, and environmental factors, and a greater focus on disentangling not only if interventions work but how and for whom they work. Several limitations should be noted. Our findings are based on a comparison group of smokers recruited from health fairs at their HD to receive a dietary intervention as an attention control in a smoking cessation trial. Our results, therefore, may reflect self-selection bias. The external validity of our findings is limited by the fact that participants were not representative of the entire HD community and caution should also be taken in generalizing the findings to nonsmokers or other low income populations. The assessment of FV consumption was based on self-report via a brief screener and is, therefore, subject to response and recall bias; however, our randomized design should have distributed these potential biases equally across both study groups and data from other trials has found consis-

7 S258 NOLLEN ET AL. Table 3 Predictors of Change in FV Consumption at Week 8 Predictor variable at baseline 8-week change in predictor variable Predictor Estimate a (SE) p Estimate a (SE) p Demographic Age 0.03 (0.01) 0.06 Race/ethnicity 0.36 (0.49) 0.46 Education 0.28 (0.38) 0.46 Marital status 0.58 (0.65) 0.38 Employment (other vs. employed) 0.85 (0.77) 0.26 Income 0.51 (0.42) 0.23 Psychosocial Motivation for fruit 0.03 (0.07) (0.07) 0.45 Motivation for vegetable 0.11 (0.07) (0.07) 0.94 Confidence for fruit 0.06 (0.08) (0.07) 0.51 Confidence for vegetable 0.19 (0.07) (0.08) 0.47 Autonomous regulation 0.01 (0.04) (0.05) 0.07 Social support: family 0.02 (0.17) (0.16) 0.89 Social support: friends 0.18 (0.15) (0.15) 0.23 Perceived barriers 0.03 (0.04) (0.04) Hope: pathways 0.03 (0.03) (0.04) 0.58 Hope: agency 0.03 (0.03) (0.04) 0.03 Hope: total 0.01 (0.02) (0.02) 0.68 Note. a Coefficient for each individual predictor after taking into account the effect of treatment. tency between self-reported FV intake and relevant biomarkers (Steptoe et al., 2003). Additionally, validity data on the screener indicates small to moderate correlations with 24-hr dietary recalls and total serum carotenoids (Resnicow et al., 2000). Finally, caution should be used in inferring causality in our change analyses. We examined the associations between changes in psychosocial factors with FV intake measured at the same time point; therefore, it is not possible to determine whether psychosocial changes preceded increased FV consumption or were secondary consequences or correlates to it. In conclusion, despite significant cross-sectional associations found at baseline, the majority of psychosocial factors were not associated with dietary change. While the intervention was successful in increasing FV consumption among smokers residing in public housing, the individual-level psychosocial factors examined in the current study were not robust for explaining this change. Dietary interventions among low-income populations are critical to improving the health of this underserved and high risk group. Future studies examining the dynamic interaction between modifications in dietary change, individual-level, and environmental factors are needed, as are dismantling studies that provide insight into how, why, and for whom interventions work. Table 4 Final Model of the Factors Associated With FV Change Scores Variable Estimate (SE) p Effect size estimate Treatment (FV 1) 1.70 (0.37) Barriers 0.12 (0.04) Goal-directedness: agency 0.08 (0.03) References Ahluwalia, J. S., Nollen, N., Kaur, H., James, A. S., Mayo, M. S., & Resnicow, K. (2007). Pathway to Health: Cluster-randomized trial to increase fruit and vegetable consumption among smokers in public housing. Health Psychology, 26, Ammerman, A. S., Lindquist, C. H., Lohr, K. N., & Hersey, J. (2002). The efficacy of behavioral interventions to modify dietary fat and fruit and vegetable intake: A review of the evidence. Preventive Medicine, 35, Ascherio, A., Katan, M. B., Zock, P. L., Stampfer, M. J., & Willett, W. C. (1999). Trans fatty acids and coronary heart disease. New England Journal of Medicine, 340, Baranowski, T., Cullen, K. W., & Baranowski, J. (1999). Psychosocial correlates of dietary intake: Advancing dietary intervention. Annual Review of Nutrition, 19, Baranowski, T., Cullen, K. W., Nicklas, T., Thompson, D., & Baranowski, J. (2003). Are current health behavioral change models helpful in guiding prevention of weight gain efforts? Obesity Research, 11, 23S 43S. Baranowski, T., Hearn, M., Baranowski, J., Lin, L., Doyle, C., Wahlay, N., et al. (1995). TEACH WELL: The relation of teacher wellness to elementary student health and behavior outcomes: Baseline subgroup comparisons. Journal of Health Education, 26, S61 S71. Becker, M. (1977). Selected psychosocial models and correlates of individual health-related behaviors. Medical Care, 5, Bem, D.J. Self-perception theory. (1972). In L. Berkowitz (ed.) Advances in experimental social psychology. Vol 6. (1 62). New York: Academic Press. Block, G., Hartman, A. M., & Naughton, D. (1990). A reduced dietary questionnaire: Development and validation. Epidemiology, 1, Block, G., Patterson, B., & Subar, A. (1992). Fruit, vegetables, and cancer prevention: A review of the epidemiological evidence. Nutrition and Cancer, 18, Block, G., Thompson, F. E., Hartman, A. M., Larkin, F. A., & Guire, K. E. (1992). Comparison of two dietary questionnaires validated against multiple dietary records collected during a 1-year period. Journal of the American Dietetic Association, 92, Floyd, R., & McDermott, D. (1998, Aug). Hope and sexual risk-taking in gay men. Paper presented at the American Psychological Association, San Francisco, CA.

8 FACTORS ASSOCIATED WITH INCREASED FV CONSUMPTION S259 Fuemmeler, B. F., Masse, L. C., Yaroch, A. L., Resnicow, K., Campbell, M. K., Carr, C., et al. (2006). Psychosocial mediation of fruit and vegetable consumption in the body and soul effectiveness trial. Health Psychology, 25, Gandini, S., Merzenich, H., Robertson, C., & Boyle, P. (2000). Metaanalysis of studies on breast cancer risk and diet: The role of fruit and vegetable consumption and the intake of associated micronutrients. European Journal of Cancer, 36, Genkinger, J. M., Platz, E. A., Hoffman, S. C., Comstock, G. W., & Helzlsouer, K. J. (2004). Fruit, vegetable, and antioxidant intake and all-cause, cancer, and cardiovascular disease mortality in a communitydwelling population in Washington County, MD. American Journal of Epidemiology, 160, Haggar, M., Chetzisarantis, N., & Biddle, S. (2002). A meta-analytic review of the theories of reasoned action and planned behavior in physical activity: Predictive validity and the contribution of additional variables. Journal of Sport and Exercise Psychology, 24, Harney, P. (1990). The Hope Scale: Exploration of construct validity and its influlence on health. Lawrence, KS: University of Kansas. Hooper, L., Summerbell, C. D., Higgins, J. P., Thompson, R. L., Capps, N. E., Smith, G. D., et al. (2001). Dietary fat intake and prevention of cardiovascular disease: Systematic review. British Medical Journal, 322, Hung, H. C., Joshipura, K. J., Jiang, R., Hu, F. B., Hunter, D., Smith-Warner, S. A., et al. (2004). Fruit and vegetable intake and risk of major chronic disease. Journal of the National Cancer Institute, 96, Jeffries, S. K., Choi, W., Butler, J., Harris, K. J., & Ahluwalia, J. S. (2005). Strategies for recruiting African-American residents of public housing developments into a randomized controlled trial. Ethnicity and Disease, 15, Langenberg, P., Ballesteros, M., Feldman, R., Damron, D., Anliker, J., & Havas, S. (2000). Psychosocial factors and intervention-associated changes in those factors as correlates of change in fruit and vegetable consumption in the Maryland WIC 5 A Day Promotion Program. Annals of Behavioral Medicine, 22, Larabie, L. C. (2005). To what extent do smokers plan quit attempts? Tobacco Control, 14, Maibach, E., & Cotton, D. (1995). Moving people to behavior change: A staged social cognitive approach to message design. In E. W. Maibach & R. L. Parrot (Eds.), Designing health messages: Approaches from communication theory and public health practice. Thousand Oaks, CA: Sage Matzger, H., Kaskutas, L. A., & Weisner, C. (2005). Reasons for drinking less and their relationship to sustained remission from problem drinking. Addiction, 100, Miller, W. R., & Rollnick, S. (2002). Motivational interviewing: Preparing people for change (2nd ed.). New York, NY: Guilford Press. Montano, D., Kasprzyk, D. The theory of reasoned action and the theory of planned behavior. (2002). In K. Glanz B. Rimer, F. Lewis, (Eds) Health behavior and health education: Theory, research, and practice. (67 98) San Francisco, CA: Jossey-Bass. Moon, C., Snyder, C., & Rapoff, M. (2001). The relationship of hope to children s asthma treatment adherence. Lawrence, KS: University of Kansas. Paavola, M., Vartiainen, E., & Haukkala, A. (2004). Smoking from adolescence to adulthood: The effects of parental and own socioeconomic status. European Journal of Public Health, 14, Resnicow, K., Campbell, M. K., Carr, C., McCarty, F., Wang, T., Periasamy, S., et al. (2004). Body and soul. A dietary intervention conducted through African-American churches. American Journal of Preventive Medicine, 27, Resnicow, K., Jackson, A., Blissett, D., Wang, T., McCarty, F., Rahotep, S., et al. (2005). Results of the healthy body healthy spirit trial. Health Psychology, 24, Resnicow, K., Jackson, A., Wang, T., De, A. K., McCarty, F., Dudley, W. N., et al. (2001). A motivational interviewing intervention to increase fruit and vegetable intake through Black churches: Results of the Eat for Life trial. American Journal of Public Health, 91, Resnicow, K., 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 sample of African- American adults. American Journal of Epidemiology, 152, Resnicow, K., Soler, R., Braithwaite, R., Ahluwalia, J. S., & Butler, J. (2000). Cultural sensitivity in substance use prevention. Journal of Community Psychology, 28, Resnicow, K., & Vaughan, R. (2006). A chaotic view of behavior change: A quantum leap for health promotion. International Journal of Behavioral Nutrition and Phsycial Activity, 3, 25. Resnicow, K., Wallace, D., Jackson, A., DiGirolamo, A., Odom, E., Wang, T., et al. (2000). Dietary change through African-American churches: Baseline results and program description of the Eat for Life Trial. Journal of Cancer Education, 15, Rollnick, S., Butler, C., & Stott, N. (1997). Helping smokers make decisions: The enhancement of brief interventions for general medical practice. Patient Education and Counseling, 31, Sallis, J., Grossman, R., Pinski, R., Patterson, T., & Nader, P. (1987). The development of scales to measure social support for diet and exercise behaviors. Preventive Medicine, 16, Serdula, M. K., Gillespie, C., Kettel-Khan, L., Farris, R., Seymour, J., & Denny, C. (2004). Trends in fruit and vegetable consumption among adults in the United States: Behavioral risk factor surveillance system, American Journal of Public Health, 94, Snyder, C. (2002). Hope theory: Rainbows in the mind. Psychological Inquiry, 13, Snyder, C., Harris, C., Anderson, J., Holleran, S., Irving, L., Sigmon, S., et al. (1991). The will and the ways: Development and validation of an individual differences measure of hope. Journal of Personality and Social Psychology, 60, Steptoe, A., Perkins-Porras, L., McKay, C., Rink, E., Hilton, S., & Cappuccio, F. (2003). Psychological factors associated with fruit and vegetable intake and with biomarkers in adults from a low-income neighborhood. Health Psychology, 22, Steptoe, A., Perkins-Porras, L., Rink, E., Hilton, S., & Cappuccio, F. P. (2004). Psychological and social predictors of changes in fruit and vegetable consumption over 12 months following behavioral and nutrition education counseling. Health Psychology, 23, Trudeau, E., Kristal, A. R., Li, S., & Patterson, R. E. (1998). Demographic and psychosocial predictors of fruit and vegetable intakes differ: Implications for dietary interventions. Journal of the American Dietetic Association, 98, Van Duyn, M. A., Kristal, A. R., Dodd, K., Campbell, M. K., Subar, A. F., Stables, G., et al. (2001). Association of awareness, intrapersonal and interpersonal factors, and stage of dietary change with fruit and vegetable consumption: A national survey. American Journal of Health Promotion, 16, Ward, E., Jemal, A., Cokkinides, V., Singh, G. K., Cardinez, C., Ghafoor, A., et al. (2004). Cancer disparities by race/ethnicity and socioeconomic status. Ca: A Cancer Journal For Clinicians, 54, Wargovich, M. J. (1997). Experimental evidence for cancer preventive elements in foods. Cancer Letters, 114, West, R., & Sohal, T. (2006). Catastrophic pathways to smoking cessation: Findings from national survey. British Medical Journal, 332, Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L. (1996). Motivational predictors of weight loss and weight-loss maintenance. Journal of Personality and Social Psychology, 70, Williams, G. C., McGregor, H. A., Zeldman, A., Freedman, Z. R., & Deci, E. L. (2004). Testing a self-determination theory process model for promoting glycemic control through diabetes self-management. Health Psychology, 23, Williams, G. C., Rodin, G. C., Ryan, R. M., Grolnick, W. S., & Deci, E. L. (1998). Autonomous regulation and long-term medication adherence in adult outpatients. Health Psychology, 17,