QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Davis, R. E. Articles by Costello, R. B. Search for Related Content PubMed PubMed Citation Articles by Davis, R. E. Articles by Costello, R. B. Pubmed/NCBI databases Medline Plus Health Information Dietary Supplements Healthy Living

---

Supplement: The Examination of Two Short Dietary Assessment Methods, within the Context of Multiple Behavioral Change Interventions in Adult Populations

Use of Signal Detection Methodology to Identify Subgroups of Dietary Supplement Use in Diverse Populations^1,2

³ Department of Health Behavior and Health Education, School of Public Health, University of Michigan, Ann Arbor, MI 48109-2029; ⁴ Behavioral Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD 20892-7344; ⁵ Program in Public Health Nutrition, Department of Nutrition, and Department of Society, Human Development, and Health, Harvard School of Public Health, Boston, MA 02115; ⁶ Department of Clinical Nutrition, Rush University Medical Center, Chicago, IL 60612; ⁷ Hunt Consulting Associates, Consultant to Harvard School of Public Health, Program in Public Health Nutrition, Lyme, NH 03768; ⁸ Cancer Prevention and Control Program, Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, Columbia, SC 29208; ⁹ Department of Nutrition and Food Sciences, University of Rhode Island, Kingston, RI 02881; ¹⁰ Illinois Institute of Technology, Institute of Psychology, Chicago, IL 60616; ¹¹ Departments of Medicine, Clinical and Social Sciences in Psychology, Psychiatry, University of Rochester, Rochester, NY 14642; ¹² Applied Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD 20892-7344; ¹³ Oregon Research Institute, Eugene, OR 97403; ¹⁴ Division of Health Promotion and Sports Medicine, Oregon Health & Science University, Portland, OR 97239; and ¹⁵ Office of Dietary Supplements, NlH, Bethesda, MD 20892-7517

^* To whom correspondence should be addressed. E-mail: reda{at}umich.edu.

	ABSTRACT

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

 
Despite widespread use of dietary supplements, little is known about correlates and determinants of their use. Using a diverse sample from 7 interventions participating in the Behavior Change Consortium (n = 2539), signal detection methodology (SDM) demonstrated a method for identifying subgroups with varying supplement use. An SDM model was explored with an exploratory half of the entire sample (n = 1268) and used 5 variables to predict dietary supplement use: cigarette smoking, fruit and vegetable intake, dietary fat consumption, BMI, and stage of change for physical activity. A comparison of rates of supplement use between the exploratory model groups and comparably identified groups in the reserved, confirmatory sample (n = 1271) indicates that these analyses may be generalizable. Significant indicators of any supplement use included smoking status, percentage of energy from fat, and fruit and vegetable consumption. Although higher supplement use was associated with healthy behaviors overall, many of the identified groups exhibited mixed combinations of healthy and unhealthy behaviors. The results of this study suggest that patterns of dietary supplement use are complex and support the use of SDM to identify possible population characteristics for targeted and tailored health communication interventions.

	Introduction

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

The efficacy of dietary supplements in the prevention and treatment of disease remains equivocal (11). However, their widespread use warrants a deeper understanding of which population subgroups are more or less likely to use them. This information can be used either to encourage use of supplements found to be safe and effective or discourage use of products found to be ineffective and/or harmful. Studies are therefore needed to establish how dietary supplement use covaries with multiple health behaviors and how these patterns vary among population subgroups defined by sociodemographics or other characteristics that may be targeted by behavioral interventions.

The purpose of this article is to explore population subgroups that are more and less likely to report dietary supplement use using a relatively novel statistical technique, signal detection methodology (SDM).¹⁶ SDM can be a useful means for exploring relations among variables in large datasets such as that used for the present analyses. The population used for this study draws on 7 diverse samples participating in interventions associated with the Behavior Change Consortium (BCC), which is a collaborative group of 15 different behavior change intervention studies (12). None of the 7 studies represented in these analyses specifically encouraged or discouraged dietary supplement use.

	Methods

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

 
    Sample population. The data for this study are drawn from 2887 participants from 7 NIH-funded health behavior intervention research sites participating in the Nutrition Working Group (NWG) of the BCC (12). The 7 sites include Emory University, Harvard School of Public Health (HSPH), the Illinois Institute of Technology (IIT), Oregon Health & Science University (OHSU), the Oregon Research Institute (ORI), the University of Rochester (ROC), and the University of Rhode Island (URI). Details about the BCC, the individual study populations, recruitment, and the designs of the 7 intervention trials are described elsewhere in this supplement (13). The starting sample for the present analyses comprised 2539 participants with complete baseline data for any dietary supplement use, gender, and age.

    Dietary supplement measures. Selection of measures of dietary supplement behavior varied by site; choice of measure depended on the study population, goals, setting, and the feasibility of including additional measures in existing evaluation protocols. HSPH queried supplement use via a single item: "On average, how many days a week do you take multivitamins?" Responses were closed-ended and ranged from "Never" to "7 d a week." Emory also used a single item: "During the past year, have you taken any vitamins or mineral supplements?" Response options consisted of "Yes, fairly regularly," "Yes, but not regularly," and "No." The remaining 5 of the 7 sites collected dietary supplement data using 1 of 2 longer questionnaire formats, Form A and Form B. Both forms contained questions about multivitamin use, questions about 3 specific vitamins/minerals/nutrients, a checklist of 19 additional supplements, and 1 open-ended item about dietary supplement use. Form A also included 3 additional questions about specific vitamins/minerals/nutrients and an open-ended vitamin supplement question. Form B, which was validated in the IIT population (14), included 9 additional questions about specific vitamins/minerals/nutrients and assessed usual dose per day. ROC and URI used Form A, and OHSU and IIT used Form B. ORI used a modified version of Form B and assessed number of pills per week and dose by interview. With the exception of 1 vitamin, which was queried as "take niacin/nicotinic acid such as niacin or Niaspan?" all vitamin, mineral, and nutrient use was queried in the longer forms either specifically (e.g., "Do you currently take a β-carotene supplement?") or generally (e.g., "Do you currently take other nonvitamin supplement?"). Participants were coded as having no dietary supplement use if they answered "No" to the single global item for HSPH and Emory or to all of the dietary supplement use questions on Forms A or Form B for ROC and URI and OHSU, IIT, and ORI, respectively. If participants answered "Yes" to any of the dietary supplement questions, they were coded as using dietary supplements.

    Other measures. Other measures include a 19-item National Cancer Institute (NCI) fruit and vegetable screener (FVS) (15) used in all 7 sites and a 16-item NCI percentage energy from fat screener (PFat) (16) administered in 6 sites. The FVS queried frequency of usual consumption of 10 categories of fruits and vegetables over the prior month (17). Portion sizes were asked for 9 items: 100% juice, fruit, lettuce salad, French fries/fried potatoes, other white potatoes, cooked dried beans, other vegetables, tomato sauce, and vegetable soups. A single item asking the frequency of consuming "mixtures that included vegetables" was not included in calculating fruit and vegetable intake levels. The FVS screener estimates daily servings of fruits and vegetables using the 1998 USDA Food Guide Pyramid defined servings (18). The PFat screener asks about usual consumption practices in the past 12 mo. Frequency of intake is asked for 15 food groups that were selected in earlier analyses to optimally predict intake of percentage energy from fat. Portion size is not explicitly asked as part of the PFat screener; however, the scoring algorithms assign gender- and age-specific median portion sizes in grams to each food group asked. Individual percentage energy from fat is estimated based on frequency responses, assigned portion size, and gender-specific regression coefficients relating intake of each food group to percentage energy from fat. The performance of both the FVS and PFat screeners at baseline varied by site and gender. Deattenuated Pearson correlation coefficients for the PFat and true intake (as estimated from 24-h dietary recall assessments (24HR) using a measurement error model) were significantly different from 0 (P < 0.05) for men and women in all sites, ranging from 0.52 to 0.77 among men and 0.36 to 0.59 among women. At baseline, the FVS significantly (P < 0.05) overestimated intake at 2 of 4 sites for men and all 4 sites for women. Correlations between the 24HR data and the FVS by site ranged from 0.31 (P = 0.07) to 0.47 (P < 0.01) in men and from 0.43 to 0.63 (P < 0.01) in women. Additional details on the development, scoring, and testing of the FVS and PFat screeners can be found elsewhere (15,17,19,20).

Other variables assessed across all 7 sites included race and ethnicity, BMI, smoking status (smoker/nonsmoker), age, educational status, income level, employment status, gender, and marital status. Most of these measures are described elsewhere in this supplement (13). In addition, physical activity was measured using a stage of change measure (21) across all 7 sites. Participants were coded as being in 1 of 5 categories: precontemplation, contemplation, preparation, action, or maintenance stages of change for physical activity.

    Statistical procedures. SDM (22) was used to identify the 2 groups of dietary supplement use (any supplement use vs. no supplement use). SDM has been previously used to delineate groups for cardiovascular disease risk (23,24), physical activity (25,26), weight status (27), smoking cessation (28–30), participation in an alcohol treatment program (31), and achieving a low-fat diet (32). SDM relies on receiver operating characteristic curves to identify nonoverlapping, homogeneous, and maximally differentiated groups on a designated dichotomous outcome (22,27). The groups are defined through an iterative process of selecting the most discriminating cutpoint of the most discriminating indicator variable until preset stopping rules are satisfied (22). SDM is an exploratory method, the results of which should be empirically explored through confirmatory procedures. However, groups identified using SDM have been validated in prior research (26,33). SDM is capable of identifying higher-order interactions that might go unnoticed or be uninterpretable via other methods (22,24,27). SDM is also a nonparametric procedure and, as such, does not require linearity between variables or a normal distribution and is only minimally impacted by multicollinearity among indicator variables (22). The datasets and variables used in the SDM analyses were prepared using SAS 9.1 for Windows (SAS Institute, Cary, NC, 2002–2003).

An exploratory SDM model analyzed use of any supplements among a randomly selected subsample comprising half of the original sample (n = 1268). The exploratory model included the following indicator variables: physical activity stage of change, dietary fat intake, fruit and vegetable consumption, smoking status, and BMI. These predictors represented all available baseline indicators of health behaviors from the cross-site BCC dataset. The exploratory model was prepared as single datasets in SAS and converted to text files for use with the ROC4 SDM software (Department of Veterans Affairs, Mental Illness Research, Education, and Clinical Centers, 2002). The sensitivity and specificity of an SDM test (signified as r) range from 0 to 1 and can be adjusted to match the goals of the analysis by prioritizing the minimization of false positives (r = 0), prioritizing the minimization of false negatives (r = 1), or balancing the risk of incurring either type of error for maximum efficiency (r = 0.5) (23,28). The sensitivity/specificity ratio was set to r = 0.5 for these analyses to evenly balance the risk of incurring type I or type II errors and maximize efficiency. Because the ROC4 program produces a maximum of 3 levels of splitting per run, all models were run using subsets of the data, as necessary, until terminal groups were achieved. Terminal groups were reached when a group contained fewer than 50 subjects and/or the significance level of the multiple testing increased above P < 0.01.

In replicating procedures conducted by Atienza et al. (26) in prior research, the parameters characterizing the exploratory groups were applied to the reserved, confirmatory sample (n = 1271) to create groups in the confirmatory sample with the same characteristics as those identified in the exploratory sample. Any supplement use was calculated for the confirmatory groups, and the percentages of any supplement use were compared between comparable exploratory and confirmatory groups using ² analyses to explore the generalizability of the exploratory model.

Using the group parameters identified in the exploratory SDM analysis, sociodemographic and select 24HR variables were compared across comparably defined groups from the entire sample (n = 2539) to further define potential group differences. The ² tests of homogeneity and 1-way ANOVA analyses were used to determine whether significant differences existed among groups.

	Results

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

 
    Sample population. The study sample consisted of 2539 participants and is described in Table 1. The overall rate of dietary supplement use in the sample was 62.9% (not shown in Table 1). Most participants were middle-aged, female, married or living with a partner, employed full-time, white or African American, and had a high school education or higher.

View larger version (21K): [in this window] [in a new window]   FIGURE 1  SDM results for any supplements exploratory split-sample model (n = 1268).

	Discussion

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

Prior studies have documented associations among healthy behaviors such as not smoking, low fat intake, high fruit and vegetable consumption, high fiber intake, and regular exercise with higher dietary supplement use (1,3,5–9). In these analyses, dietary supplement intake is generally related to engagement in other healthy behaviors, but this is not always the case. Inconsistent relations were found among these variables when they were combined and studied in tandem as indicators of dietary supplement use. These findings suggest that, for some participants, supplement use may represent a compensatory behavior for other, less healthy lifestyle choices. A supplement use rate of over 45% was associated with a mixed profile of healthy and unhealthy behaviors for 2 of the 6 groups (Groups D and F). And, even among the groups evincing the least healthy behavioral combination, Group E, over 23% of participants reported dietary supplement use. The members of these groups may be consciously compensating for engaging in unhealthy behaviors by actively pursuing other, healthier behaviors such as higher fruit and vegetable intake, regular physical activity, or dietary supplementation. In a cluster analysis of health lifestyle behaviors, Maibach et al. (34) identified a cluster of smokers who, aside from their smoking, appear to be consciously trying to live a healthy lifestyle. In our analyses, some participants may be including supplement use to compensate for other less healthy behaviors such as smoking, lack of exercise, or low fruit and vegetable consumption. This compensation dynamic may be influenced by the fact that our sample was drawn entirely from participants enrolling in health promotion interventions. Thus, this sample may be more predisposed to healthy behaviors, even among smokers, for example, who might be more ready to quit (and therefore more likely to engage in other health-promoting behaviors) than smokers from the general population. Thus, our results may be more applicable to individuals interested in changing health behaviors.

This study has several limitations. For one, dietary supplement use was inconsistently measured across the 7 study sites. Two sites used single-item measures, and 5 sites used checklist forms. Within these 2 formats, there was additional variability in the wording and format of the items. The fact that dietary supplement use was reduced to a single binary variable may have reduced some of the measurement problems. However, by collapsing use into 2 levels, we also lost variability that may have enhanced our analyses. In addition, as noted earlier, SDM is an exploratory method, and any findings should be substantiated using confirmatory analytic procedures. SDM analyses can also produce small sample sizes when compared with starting sample sizes, so interpretations based on smaller subgroups in these analyses should be made with caution. Our findings may be confounded by the significant sample differences across study sites. For example, the lack of ROC and IIT participants from certain groups likely reflects the focus of the behavioral interventions from these sites on smoking cessation and cardiac risk reduction, respectively. Several of the sites consisted of behavior change interventions targeting populations with known health risks such as older age (URI), smoking (ROC), coronary artery disease (IIT), and diabetes with sedentary lifestyles and high-fat diets (ORI). Two of the sites focused on relatively healthy populations such as African Americans from local churches (Emory) and firefighters (OHSU). These differences may limit the generalizability of our findings.

Another weakness of this study is that analyses included behavioral data that were acquired through self-report. Further, this study also lacked a measure of actual physical activity and relied instead on physical activity stage of change. As a consequence, our findings may be subject to social desirability and measurement error biases. Another limitation is the absence of psychosocial variables such as social support, self-efficacy, or intrinsic motivation. Inclusion of these variables may have provided more informative findings related to future intervention development. Last, the SDM results presented use a stopping rule of P < 0.01, but some SDM practitioners recommend a more stringent stopping rule of P < 0.001. Because many of our cuts would have disappeared with a stopping rule of P < 0.001, we opted to use a stopping rule of P < 0.01 as a better demonstration of the types of results that can be found through an SDM analysis.

Despite these limitations, this study also has several strengths. First, the inclusion of a large and sociodemographically diverse sample may have enriched the findings. Second, this study exhibits the use of a relatively novel, empirical, hierarchical method for identifying groups with varying health behaviors. Although the efficacy of dietary supplement use remains equivocal, our findings demonstrate that SDM may be a useful method for generating hypotheses for future research on dietary supplement use and other nutrition-related behaviors. For example, the findings from this article may stimulate research on psychological predictors of supplement use. SDM may also be useful in identifying groups that may serve as audience segments for targeted or tailored health communication interventions either to encourage or to discourage dietary supplement use. For instance, if supplement use is found to be beneficial, low-use populations such as those represented in Group E might be targeted by an intervention encouraging supplement use. In addition, behavioral and sociodemographic group characteristics generated through SDM procedures may be utilized to increase message salience and credibility when designing behavioral intervention materials. Health communication specialists may use SDM-generated data to craft individualized messages for tailored interventions. A tailored message that increases a participant's awareness of success in one health behavior, for instance, may assist him or her in building self-efficacy to address challenges in implementing another health behavior.

Our article contributes new knowledge about the relations among dietary supplement use, health behaviors, and sociodemographic characteristics not just by identifying correlates of supplement use, as has been done in numerous prior studies, but by identifying specific constellations of behaviors that exist among participants in health intervention studies. Prior studies on dietary supplement use have focused on a more limited set of determinants and do not account for multiple interactions among predictors.

The results of this study indicate that SDM may be a useful tool for exploring dietary data. These findings also suggest that dietary supplement users are heterogeneous and that different health communications messages may be required to reach and motivate different groups. A general motivation toward healthy living may also be capitalized on to support additional behavioral achievements. Health behaviors do not occur in isolation from one another, and this interconnectedness makes the task of impacting one behavior infinitely complex. But, by embracing this complexity, intervention researchers may find an unexpected ability to influence health behavior change as a synergistic system of interconnected acts and inspirations.

	FOOTNOTES

 
¹ Published in a supplement to The Journal of Nutrition. This effort was organized by the National Cancer Institute (NCI) and the Behavior Change Consortium (BCC) Nutrition Working Group (NWG) to present the outcome data from a multidisciplinary collaboration from 7 BCC sites and 2 federal agencies: University of Rhode Island, Harvard School of Public Health, Oregon Health & Science University, Oregon Research Institute, Illinois Institute of Technology, Emory University, University of Rochester, the NCI, and the Office of Dietary Supplements. The BCC NWG was supported by National Institutes of Health funding initiative RFA OD-93-002 with additional NCI supplemental funding R01AG16588, R01HD37368, R01AR45901, R01HL62156, R01HL62158, R01HL64959, and R01MH59594. The opinions or assertions contained herein are the private ones of the authors and are not to be considered as official or reflecting the views of the National Institutes of Health. Guest Editors: Shirley A. A. Beresford, University of Washington, Seattle, WA and Lisa M. Klesges, St. Jude Children's Research Hospital, Memphis, TN, and Helaine R. H. Rockett, Harvard Medical School and Brigham and Women's Hospital, Boston, MA. Guest Editor disclosure: S. A. A. Beresford, L. M. Klesges, and H. R. H. Rockett will receive compensation from NCI, DCCPS, BRP for editorial services provided for this supplement publication; L. M. Klesges was a member of the BCC.

² Author disclosures: R. E. Davis, K. Resnicow, A. A. Atienza, K. E. Peterson, A. Domas, A. Hunt, T. G. Hurley, A. L. Yaroch, G. W. Greene, T. Goldman Sher, G. C. Williams, J. R. Hebert, L. Nebeling, F. E. Thompson, D. J. Toobert, D. L. Elliot, C. DeFrancesco, and R. B. Costello, no conflicts of interest.

¹⁶ Abbreviations used: 24HR, 24-h recall; BCC, Behavior Change Consortium; FVS, fruit and vegetable screener; HSPH, Harvard School of Public Health; IIT, Illinois Institute of Technology; NCI, National Cancer Institute; NWG, Nutrition Working Group; OHSU, Oregon Health & Science University; ORI, Oregon Research Institute; PFat, NCI percentage energy from fat screener; ROC, University of Rochester; SDM, signal detection methodology; URI, University of Rhode Island.

	LITERATURE CITED

TOP ABSTRACT Introduction Methods Results Discussion LITERATURE CITED

 

1. Balluz LS, Okoro CA, Bowman BA, Serdula MK, Mokdad AH. Vitamin or supplement use among adults, Behavioral Risk Factor Surveillance System, 13 states, 2001. Public Health Rep. 2005;120:117–23.[Medline]

2. Timbo BB, Ross MP, McCarthy PV, Lin CTJ. Dietary supplements in a national survey: Prevalence of use and reports of adverse events. J Am Diet Assoc. 2006;106:1966–74.[CrossRef][Medline]

3. Millen AE, Dodd KW, Subar AF. Use of vitamin, mineral, nonvitamin, and nonmineral supplements in the United States: The 1987, 1992, and 2000 National Health Interview Survey Results. J Am Diet Assoc. 2004;104:942–50.[CrossRef][Medline]

4. Kelly JP, Kaufman DW, Kelley K, Rosenberg L, Anderson TE, Mitchell AA. Recent trends in use of herbal and other national products. Arch Intern Med. 2005;165:281–6.[Abstract/Free Full Text]

5. Fennell D. Determinants of supplement usage. Prev Med. 2004;39:932–9.[CrossRef][Medline]

6. Lyle BJ, Mares-Perlman JA, Klein BEK, Klein R, Greger JL. Supplement users differ from nonusers in demographic, lifestyle, dietary and health characteristics. J Nutr. 1998;128:2355–62.[Abstract/Free Full Text]

7. White E, Patterson RE, Kristal AR, Thornquist M, King I, Shattuck AL, Evans I, Satia-Abouta J, Littman AJ, Potter JD. VITamins And Lifestyle cohort study: Study design and characteristics of supplement users. Am J Epidemiol. 2004;159:83–93.[Abstract/Free Full Text]

8. Slesinski MJ, Subar AF, Kahle LL. Dietary intake of fat, fiber, and other nutrients is related to the use of vitamin and mineral supplements in the United States: The 1992 National Health Interview Survey. J Nutr. 1996;126:3001–8.[Abstract/Free Full Text]

9. Radimer K, Bindewald B, Hughes J, Ervin B, Swanson C, Picciano MF. Dietary supplement use by US adults: Data from the National Health and Nutrition Examination Survey, 1999–2000. Am J Epidemiol. 2004;160:339–49.[Abstract/Free Full Text]

10. Dwyer JD, Costello RB. Assessment of dietary supplement use. In: Coulston AM, Boushey CJ, editors. Nutrition in the prevention and treatment of disease. 2nd ed. Elsevier; In press.

11. Huang HY, Caballero B, Chang S, Alberg AJ, Semba RD, Schneyer CR, Wilson RF, Cheng TY, Vassy J, et al. The efficacy and safety of multivitamin and mineral supplement use to prevent cancer and chronic disease in adults: A systematic review for a National Institutes of Health State-of-the-Science conference. Ann Intern Med. 2006;145:372–85.[Abstract/Free Full Text]

12. Greene GW, Nebeling LC, Greaney ML, Lindsay AC, Hardwick CK, Toobert DJ, Resnicow K, Williams GC, Elliot DL, et al. A qualitative study of a nutrition working group. Health Promot Pract. 2007;8:299–306.[Abstract/Free Full Text]

13. Yaroch AL, Nebeling L, Thompson FE, Hurley TG, Hebert JR, Toobert DJ, Resnicow K, Greene GW, Williams GC, et al. Baseline design elements and sample characteristics for seven sites participating in the Nutrition Working Group of the Behavior Change Consortium. J Nutr. 2008;138:185S–192S.[Abstract/Free Full Text]

14. Allender JC. Validation of the dietary supplement questionnaire used in the Couples Intervention for Cardiac Risk Reduction Study. Chicago: Rush University Medical Center; 2004.

15. Greene GW, Resnicow K, Thompson FE, Peterson K, Hurley TG, Hebert JR, Toobert DJ, Williams GC, Elliot DL, et al. Correspondence of the NCI Fruit and Vegetable Screener to repeat 24-H recalls and serum carotenoids in behavioral intervention trials. J Nutr. 2008;138:200S–204S.[Abstract/Free Full Text]

16. Thompson FE, Midthune D, Williams GC, Yaroch AL, Hurley TG, Resnicow K, Hebert JR, Toobert DJ, Greene GW. et al. Evaluation of a short dietary assessment instrument for percentage energy from fat in an intervention study. J Nutr. 2008;138:193S–199S.[Abstract/Free Full Text]

17. Thompson FE, Subar AF, Smith AF, Midthune D, Radimer KL, Kahle LL, Kipnis V. Fruit and vegetable assessment: performance of 2 new short instruments and a food frequency questionnaire. J Am Diet Assoc. 2002;102:1764–72.[CrossRef][Medline]

18. United States Department of Agriculture Center for Nutrition Policy and Promotion. The Food Guide Pyramid. Home and Garden Bulletin No. 252. Washington, DC: USDA; 1996.

19. Thompson FE, Midthune D, Subar AF, Kipnis V, Kahle LL, Schatzkin A. Development and evaluation of a short instrument to estimate usual dietary intake of percentage energy from fat. J Am Diet Assoc. 2007;107:760–7.[CrossRef][Medline]

20. Applied Research Program NCI, National Institutes of Health. Percent energy from fat screener. Available from: http://riskfactor.cancer.gov/diet/screeners/fat/ (accessed December 1, 2006).

21. Prochaska JO, Redding CA, Harlow LL, Rossi JS, Velicer WF. The transtheoretical model of change and HIV prevention: A review. Health Educ Q. 1994;21:471–86.[Medline]

22. Kraemer HC. Evaluating medical tests: Objective and quantitative guidelines. Newbury Park: Sage Publications; 1992.

23. Williams JE, Flora JA. Health behavior segmentation and campaign planning to reduce cardiovascular disease risk among Hispanics. Health Educ Q. 1995;22:36–48.[Medline]

24. Winkleby MA, Flora JA, Kraemer HC. A community-based heart disease intervention: Predictors of change. Am J Public Health. 1994;84:767–72.[Abstract/Free Full Text]

25. King AC, Kiernan M, Oman RF, Kraemer HC, Hull M, Ahn D. Can we identify who will adhere to long-term physical activity? Signal detection methodology as a potential aid to clinical decision making. Health Psychol. 1997;16:380–9.[CrossRef][Medline]

26. Atienza AA, Yaroch AL, Masse LC, Moser RP, Hesse BW, King AC. Identifying sedentary subgroups: The National Cancer Institute's Health Information National Trends Survey. Am J Prev Med. 2006;31:383–90.[CrossRef][Medline]

27. Kiernan M, Kraemer HC, Winkleby MA, King AC, Taylor CB. Do logistic regression and signal detection identify different subgroups at risk? Implications for the design of tailored interventions. Psychol Methods. 2001;6:35–48.[CrossRef][Medline]

28. Killen JD, Fortmann SP, Kraemer HC, Varady A, Newman B. Who will relapse? Symptoms of nicotine dependence predict long-term relapse after smoking cessation. J Consult Clin Psychol. 1992;60:797–801.[CrossRef][Medline]

29. Smith PM, Kraemer HC, Miller NH, DeBusk RF, Taylor CB. In-hospital smoking cessation programs: Who responds, who doesn't? J Consult Clin Psychol. 1999;67:19–27.[CrossRef][Medline]

30. Norman SB, Norman GJ, Rossi JS, Prochaska JO. Identifying high- and low-success smoking cessation subgroups using signal detection analysis. Addict Behav. 2006;31:31–41.[Medline]

31. O'Connor SM, Davies JB, Heffernan DD, Van Euk R. An alternative method for predicting attrition from an alcohol treatment programme. Alcohol Alcohol. 2003;38:568–73.[Abstract/Free Full Text]

32. Winkleby MA, Howard-Pitney B, Albright CA, Bruce B, Kraemer HC, Fortmann SP. Predicting achievement of a low-fat diet: A nutrition intervention for adults with low literacy skills. Prev Med. 1997;26:874–82.[CrossRef][Medline]

33. James KE. White RF, Kraemer HC. Repeated split sample validation to assess logistic regression and recursive partitioning: An application to the prediction of cognitive impairment. Stat Med. 2005;24:3019–35.[Medline]

34. Maibach EW, Maxfield A, Ladin K, Slater M. Translating health psychology into effective health communication: The American Healthstyles Audience Segmentation Project. J Health Psychol. 1996;1:261–77.[Abstract]

This article has been cited by other articles:

				S. A. A. Beresford, L. M. Klesges, and H. R. H. Rockett Introduction J. Nutr., January 1, 2008; 138(1): 183S - 184S. [Full Text] [PDF]

				A. L. Yaroch, L. Nebeling, F. E. Thompson, T. G. Hurley, J. R. Hebert, D. J. Toobert, K. Resnicow, G. W. Greene, G. C. Williams, D. L. Elliot, et al. Baseline Design Elements and Sample Characteristics for Seven Sites Participating in the Nutrition Working Group of the Behavior Change Consortium J. Nutr., January 1, 2008; 138(1): 185S - 192S. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Davis, R. E. Articles by Costello, R. B. Search for Related Content PubMed PubMed Citation Articles by Davis, R. E. Articles by Costello, R. B. Pubmed/NCBI databases Medline Plus Health Information Dietary Supplements Healthy Living