Validation of the Jackson Heart Study Physical Activity Survey in African Americans

Transcription

1 Journal of Physical Activity and Health, 2009(Suppl 1), 6, S124 S Human Kinetics, Inc. Validation of the Jackson Heart Study Physical Activity Survey in African Americans Todd A. Smitherman, Patricia M. Dubbert, Karen B. Grothe, Jung Hye Sung, Darla E. Kendzor, Jared P. Reis, Barbara E. Ainsworth, Robert L. Newton Jr., Karen T. Lesniak, and Herman A. Taylor Jr. Background: Physical inactivity has been consistently linked to cardiovascular disease, yet few instruments have been validated for assessment of physical activity in African Americans, a group particularly vulnerable to heart disease. The current study aimed to establish the psychometric properties of the activity survey used in the Jackson Heart Study (JHS) among African Americans, the JHS Physical Activity Cohort survey (JPAC). Methods: Test-retest reliability over 2 weeks was assessed using a convenience sample of 40 African Americans. Convergent validity with accelerometer and pedometer data were assessed in 2 samples from the JHS (N = 404 and 294, respectively). Results: Test-retest reliability was excellent, with intraclass correlations =.99 for the JPAC total and index scores. Higher JPAC total scores were significantly associated with higher raw accelerometer and pedometer counts. Spearman correlations between JPAC total scores and accelerometer (rho =.24) and pedometer counts (rho =.32) were consistent with these results. Most subscales were significantly correlated with the objective measures. The JPAC total score was most strongly associated with objectively-measured activity. Conclusion: This study provides support for the Smitherman, Grothe, and Taylor Jr. are with the University of Mississippi Medical Center, Jackson MS, Dubbert is with the G. V. (Sonny) Montgomery VAMC, Jackson, MS Sung is with Jackson State University, Jackson MS Kendzor is with the University of Texas M. D. Anderson Cancer Center, Houston, TX Reis is with San Diego State University, San Diego CA, Ainsworth is with Arizona State University, Mesa AZ Newton Jr. is with Pennington Biomedical Research Center, Baton Rouge, LA Lesniak is with Loma Linda University, Loma Linda, CA reliability and validity of the JPAC as a tool for assessing physical activity among African Americans across a variety of domains. Keywords: physical activity, accelerometry, pedometry, cardiovascular health African Americans experience disparate rates of cardiovascular disease (CVD), and contributing factors to these disparities are now being identified. 1 Lower levels of leisure time physical activity for African Americans may contribute to higher rates of CVD in this group. African American women, especially those living in the Southeast, bear the highest CVD burden with regard to specific risk factors, including physical inactivity. 2,3 Until recently, many activity studies used measures not validated for ethnic minority groups or that lacked items specific to activities often performed by women. 4 The measurement of physical activity has improved vastly over the past 15 years. The advent of objective monitoring methods such as pedometers and accelerometers and the improvement of self-report measures have increased measurement accuracy. 5 Despite these advances, however, the literature on measures for assessing activity in African Americans remains sparse. The Coronary Artery Risk Development in Young Adults (CARDIA) and the Atherosclerosis Risk in Communities (ARIC) studies, 2 large-scale epidemiologic studies of cardiovascular disease in African Americans and Caucasians, examined self-reported physical activity, but noted limitations of their assessment instruments. 4,6 The ARIC study used the Baecke physical activity survey, a well-validated instrument in Caucasians, but did not examine its psychometric properties in the ethnically-diverse sample and noted potential concerns with work-related activity perceptions. 4 The Jackson Heart Study (JHS), the largest epidemiological study of CVD in African Americans to date, included 5302 men and women in the baseline S124

2 JHS Survey Validation S125 examination. In this observational study, a variety of methods were used to measure physical activity, including self-reported physical activity using the Jackson Heart Physical Activity Cohort instrument, or JPAC, as well as combinations of self-report and objective methods (ie, accelerometer, pedometer) for subsamples. As described by Dubbert et al, 7 the JPAC instrument was derived from modifications to the Baecke physical activity survey 8 used in the ARIC study. Retaining similarities to ARIC measures was desirable because 22% of JHS participants were also part of the Jackson, MS ARIC sample. The JPAC instrument is most similar to a recent major modification of the Baecke/ARIC instrument, the Kaiser Physical Activity Survey (KPAS). 9 The KPAS was designed to improve physical activity assessment in women by including assessment of physical activity resulting from care giving, housework, and other homeand yard-based activities. Variants of the KPAS have repeatedly demonstrated adequate test-retest reliability (index r =.79 to.91) and validity as related to other subjective and objective indices of physical activity and fitness among ethnically-diverse samples of women However, the JPAC improved upon the ARIC instrument by providing a more comprehensive measure of physical activity, updating occupational activity coding and sports activity choices, and using MET (multiples of resting metabolic rate, 1 MET = 1 kcal/kg/hr) values of energy expenditure from the 2000 Compendium of Physical Activity 12 to classify sports-related activity. For the JPAC, several items of the KPAS were modified to be applicable to men as well as women and some items were omitted to meet the time constraints of the JHS. The primary objectives of this study were to establish the reliability and validity of the JPAC instrument in African American men and women. This is the first study to validate the JPAC. Specific goals were 1) to estimate the 2-week test-retest reliability of the JPAC total and index scores; 2) to estimate the validity of the JPAC total and index scores using 2 objective comparison measures of physical activity (Actigraph accelerometer and Yamax pedometer) in 2 separate samples; and 3) to estimate the validity of the JPAC measure using a subjective comparison of physical activity (past week physical activity recall). We hypothesized that test-retest reliability would be satisfactory over the 2-week interval and that self-reported activity from the JPAC would be moderately correlated with objective activity data gathered from accelerometers and pedometers and with subjective activity reported in the past week. Methods The research was approved by the Institutional Review Board of the University of Mississippi Medical Center, with ratification by the boards of the Jackson State University and Tougaloo College study partners, and written informed consent was obtained from all participants. Measures The 40-item Jackson Heart Physical Activity Cohort (JPAC) survey is an interviewer-administered instrument that assesses physical activity over the past 12 months and typically takes approximately 15 minutes to administer. The JPAC provides 4 index (subscale) scores as well as a total score, calculated as the sum of the 4 index scores. The Active Living Index contains 7 items that assess activity during leisure time (similar to the 4 counterpart questions on the KPAS). The Work Index includes 8 items assessing work-related physical activity and is similar to counterpart sections on the KPAS and ARIC measures, with the difference that the JPAC asks about both paid and volunteer work activity. The Work Index included a score for the respondent s occupational intensity (1 = low, 3 = moderate, 5 = high) also, based on the Department of Labor Occupational Codes and consistent with the KPAS. The Home, Family, Yard, and Garden Index ( Home/Life ) contains 7 items that assess activities that occur around the home, including care giving, cooking, major and minor cleaning, home repairs, and outdoor work. This Index combined some items from the counterpart section of the KPAS, adapted others to include activities often reported by men, 13 and deleted references to shoveling snow. The Sport and Exercise Index contains 18 items and is identical to counterpart items on the KPAS. Most JPAC items ask for 5-level, categorical Likertstyle responses, although items on the Sports Index require specification of particular activities. Scoring for the JPAC generally follows procedures described for the KPAS 9 and ARIC 14 instruments. Likert-scale responses for each item are assigned a number (ranging from 1 to 5, with 1 indicating lowest activity) and summed within each index. The resulting summed score is then divided by the number of index items to generate an index score ranging from 1 to 5, with lower scores representing lower relative physical activity. The Work Index score was an average of the 8 Work items as well as the respondent s occupational intensity code. The Sports Index is scored differently from the other 3: Activities are coded by MET intensity, multiplied by the time engaged in each activity, and then mapped onto an ordinal (1 5) scale, with lower scores indicating lower activity. Index scores were not calculated if any items were missing from the respective index. To limit psychometric assumptions, and in contrast to scoring of the KPAS, scores for the Work Index were assigned a weight of 0 if the respondent reported not working (including volunteer work). Total scores are calculated by summing the 4 index scores, and can thus range from 3 to 20 for those respondents with complete data. The JPAC is available for public use at the JHS website: jhsinfo/forresearchers/formsmanuals/visit1forms/ tabid/109/default.aspx (labeled PACA ).

3 S126 Smitherman et al Test-Retest Reliability For the purposes of evaluating the test-retest reliability of the JPAC, a convenience sample of adult African Americans was recruited from a college campus and local churches via advertisements and word-of-mouth during Efforts were made to match pairs of participants on relevant demographic variables (ie, age, gender). Participants of the JHS were excluded. The JPAC was first administered during a face-to-face interview and subsequently readministered via phone 2 weeks later by a trained African American interviewer. Because occupational information was not available for this sample, the intensity value of the respondent s given occupation was not included in computations of testretest coefficients for the Work Index. Validation with Accelerometer Data A trained African American interviewer administered the JPAC to each JHS participant at his/her home before the initial clinic examination. A subset of JHS participants agreed to wear a 24-hour blood pressure monitor and an accelerometer for 1 day following the clinic examination 1 to 3 months later. At the close of the clinic examination, these participants were provided with a Manufacturing Technology, Inc (Model 7164 Actigraph, MTI, Fort Walton Beach, FL) accelerometer to objectively quantify their free-living physical activity for a continuous 24-hour period. The accelerometer monitoring period commenced immediately following the clinic examination and occurred simultaneously with 24-hour ambulatory blood pressure monitoring. JHS study personnel retrieved both instruments the following day. The Actigraph accelerometer is a small ( cm), lightweight (42 g), battery-powered device that measures acceleration and deceleration of movement (at a magnitude of 0.05 to 2G and frequency of 0.25 to 2.5 Hz) in a single vertical plane over a userdefined time interval (epoch). The Actigraph accelerometer has been shown to be a valid objective measure of moderate-level energy expenditure in both laboratory 15 and field settings, 16 and for these reasons it is often used for estimating the validity of newly-developed self-report instruments. Before distribution, each accelerometer was initialized to sample movement data at 1-minute epochs; after each epoch, the accumulated count was stored in memory and reset to zero automatically. Participants were instructed to affix the accelerometer in a provided waist pouch, worn at belt level. They were further instructed to remove the device only while bathing, swimming, or sleeping and not to alter their typical activity patterns. After the 24-hour monitoring period concluded, data for each accelerometer were downloaded to a digital file for subsequent analysis. In addition to raw accelerometer counts, previously established and widely-used cutpoint values (Hendelman et al, 16 Freedson et al, 17 and Swartz et al 18 ) were used to deter- mine the amount of time (in minutes/day) spent performing moderate intensity physical activity ( 3.00 METs) during the monitoring period. Validation with Pedometer Data Another subset of individuals from the JHS participated in the JHS Diet and Physical Activity Substudy (DPASS, N = 499), which was developed to provide more detailed information about the dietary and activity habits of JHS participants (see Carithers et al 19 ). Like other JHS participants, DPASS participants were administered the JPAC at their home before their JHS clinic examination. Two to 3 months later, at the first DPASS visit, they completed a 7-day physical activity recall (the physical activity module of the Behavioral Risk Factor Surveillance Survey, or BRFSS). 20 Before their next DPASS clinic visit, they were mailed a Yamax SW-200 pedometer (Yamax Corporation, Tokyo, Japan) to wear and record steps for 3 consecutive days during the week before their return clinic visits. The Yamax brand has been shown to accurately quantify steps taken in both laboratory 21 and free-living conditions. 22 Data from the Yamax SW-200 pedometer also correlate strongly with those obtained from the Actigraph accelerometer. 23 The step count data were typically collected 3 to 6 months after administration of the JPAC, depending on when the participant entered the DPASS protocol. Figure 1 details the time sequencing of data collection for both the accelerometer and DPASS (pedometer) subsamples. Both the JHS and the DPASS were observational studies with no interventional components. Participants were instructed to leave their pedometer affixed to their waist and go about their typical activities, removing the pedometer only while bathing, swimming, or sleeping. They were asked to record their daily steps each evening for the 3 days of monitoring in a provided step log and to reset the pedometer to zero. Resulting daily pedometer step counts were averaged across all available days. Data Analysis Statistical analyses were conducted using the Statistical Analysis System (SAS Institute, Cary, NC) and the Statistical Package for the Social Sciences (SPSS, SPSS Inc., Chicago, IL) version To assess test-retest reliability of the JPAC subscales and total scale, intraclass correlation coefficients (ICCs) between the first and second JPAC administrations were calculated within the reliability sample. A 2-way, mixed effects model was used (person effects random, measure effects fixed) to designate the ratio of between-groups variance to total variance. Convergent validity with objectively-measured physical activity (24-hour accelerometer data and 3-day pedometer data) was assessed in 2 ways: 1) Participants were separated into tertiles based on their JPAC total scores, and objective physical activity data were compared across tertiles using linear trend analyses and

4 JHS Survey Validation S127 Figure 1 Sequencing of JHS Accelerometer and DPASS (pedometer) Subsamples. 2) Spearman correlations between accelerometer/ pedometer values and JPAC total and index scores were calculated. (Spearman correlations were used because the activity scores were not normally distributed.) Within the pedometer sample, Spearman correlations between JPAC scores and self-reported variables from the BRFSS 7-day physical activity module were also computed. Results Test-Retest Reliability Sample Complete test-retest data were available on 40 African Americans. The demographic characteristics and mean JPAC scores for these individuals are presented in Table 1. The intraclass correlations for the 2-week test-retest reliability of the JPAC were very high, despite the fact that the 2 administrations used different formats (faceto-face interview vs. phone interview). Reliability for the JPAC total score was.99, and index score reliability coefficients were all.99, confirming that scores for the JPAC 12-month recall were virtually invariant during the 2-week administration interval. Validation with Accelerometer Sample Four hundred and four (404) JHS participants participated in 24-hr ambulatory monitoring using an accelerometer. Table 1 presents the demographic characteristics of the accelerometer sample, as well as their mean scores on the JPAC. As is evident from the table, the majority of participants in this sample were middleaged, female, and overweight/obese. Minutes per day spent in activities of moderate intensity varied considerably depending upon the cutpoint used to classify raw accelerometer counts (Table 2), a finding similar to that reported in other studies of accelerometer-measured activity. 24 Specifically, the Freedson et al cutpoint 17 resulted in the most conservative estimate of moderate + vigorous activity, as even the highest JPAC tertile engaged in only minutes (SD = 21.61) of moderate/vigorous activity per day. The cutpoint recommended by Hendelman et al 16 yielded the most liberal estimate (mean for highest JPAC tertile = minutes/day ± ), while the cutpoint of Swartz and colleagues 18 produced estimates that were between those of Freedson and Hendelman (mean for highest JPAC tertile = ± 62.99). To assess the validity of the JPAC in relation to accelerometer-measured activity, participants first were separated into tertiles based on their JPAC total scores. Linear trend analyses were then conducted to determine whether the JPAC group tertiles evidenced true differences in objectively-measured physical activity. As can be seen in Table 2, raw accelerometer counts/min increased significantly as JPAC total scores increased. Likewise, as JPAC total scores increased, minutes of moderate intensity activity increased significantly for 2 of the 3 most common recommended cutpoints (the third trending toward significance at P =.08). As a second means of establishing the JPAC as a valid indicator of objectively-measured physical activity, Spearman correlations between JPAC scores and accelerometer values were calculated. Table 3 displays the correlations between JPAC scores and accelerometer counts and minutes of various-intensity activity. The JPAC total (rho =.24, ) and 3 of the 4 index scores (rho =.11 to.17) were significantly and positively correlated with mean raw accelerometer counts/ min. JPAC total scores also were significantly correlated with minutes of moderate-intensity activity regardless

5 S128 Smitherman et al Table 1 Characteristics of the 3 Validation Samples Sample source Test-retest reliability sample Non-JHS African American sample Accelerometer sample JHS Accelerometer Subsample participants Pedometer sample JHS Diet and Physical Activity Substudy (DPASS) participants Sample size N = 40 N = 404 N = 294 Mean age (SD) years (15.7) years (11.54) years (9.55) Gender 50.00% female 68.81% female 62.24% female Mean education (SD) N/A years (3.78) years (3.88) % married 57.50% 57.25% N/A Mean BMI (SD) (4.54) (5.75) (6.59) Mean weight (SD) kg (11.16) kg (17.51) kg (20.20) JPAC mean scores Mean total score 8.97 (3.20) 8.32 (2.56) 8.22 (2.66) Mean active living score 1.60 (0.49) 2.13 (0.83) 2.14 (0.86) Mean sports index score 2.72 (1.37) 2.19 (1.23) 2.13 (1.25) Mean home/life score 2.20 (0.56) 2.28 (0.63) 2.32 (0.61) Mean work score 2.45 (1.54) 2.60 (0.66) 2.62 (0.64) Note. Mean JPAC scores for the reliability sample are based on the first of 2 administrations so as to negate the potential influence of practice effects. of the cutpoint used (rho =.17 to.25, all P <.001). With the exception of the Work Index scores, minutes of moderate activity were significantly correlated with the various JPAC subscales, although there was some variability in these correlation coefficients depending upon the index and the cutpoint used. Validation with Pedometer Sample Pedometer data were obtained for 382 of the 499 participants enrolled in the DPASS. To maintain consistency with previous literature suggesting that 3 days of pedometer data are needed to provide a sufficient estimate, 25 participants with less than 3 days of complete data were excluded from data analysis; those with extremely low (<1,000 mean daily steps) or extremely high (>20,000 mean daily steps) step averages were likewise excluded to remove outliers and suspect data. After these exclusions (n = 88), 294 participants remained with viable pedometer data, 55 of whom were also part of the accelerometer sample. Although the 88 excluded participants were older (), less educated (), and had a higher proportion of males (P =.03) than those retained, there were no differences between those excluded versus those retained on weight, BMI, marital status, or any JPAC scores. Demographic and psychometric information for the pedometer sample is shown in Table 1, indicating that the pedometer sample was quite similar to the accelerometer sample. As was the case with linear trend analyses with the raw accelerometer counts, raw pedometer counts/day increased significantly () from the lowest to highest JPAC total score tertiles (see Table 2). Although increases in self-reported minutes of activity (moderate minutes, vigorous minutes, moderate + vigorous minutes, minutes walked) from the 7-day recall were noted with higher JPAC total scores, this trend was significant only for self-reported minutes walked. In this case, participants in the highest JPAC tertile reported walking approximately 60 minutes/week more than the lowest JPAC tertile and 25 minutes/week more than the middle tertile. Spearman correlations between the JPAC total and index scores and pedometer counts/day and self-reported activity from the 7-day recall were calculated as they were with the accelerometer sample (see Table 3). The JPAC total score correlated most strongly with daily pedometer counts (rho =.32, ), while the Sports (rho =.29, ) and Home/Life (rho =.15, ) Index scores were also significantly correlated with daily counts. The Active Living and Work Index scores did not correlate significantly with daily pedometer counts. Regarding self-reported activity in the pedometer sample, the majority of JPAC scores (Total and Index scores) were significantly correlated with minutes of moderate activity, minutes of vigorous activity, minutes of moderate + vigorous activity, and minutes walked over the past week. The consistent exception was that Work Index scores did not correlate significantly with other self-reported physical activity, a finding similar to that obtained regarding the relationship between the Work Index and the objective measures.

6 JHS Survey Validation S129 Table 2 Accelerometer and Pedometer Values Across Tertiles of Total Activity Based on JPAC Total Scores Lowest JPAC tertile mean (SD) Middle JPAC tertile mean (SD) Highest JPAC tertile mean (SD) Linear trend P-value Accelerometer sample Sample size Mean raw counts/min (67.57) (65.39) (87.38) <.001 Min/day moderate activity Freedson ( 1952) 6.20 (10.20) 8.38 (11.03) (21.61) <.001 Hendelman ( 191) (116.99) (107.18) (103.30).08 Swartz ( 574) (63.81) (63.83) (62.99) <.01 Pedometer sample Sample size Mean daily counts ( ) ( ) ( ) <.0001 Sample size PWA min moderate activity past 7 days (199.74) (244.80) (193.21).20 Sample size PWA min vigorous activity past 7 days (97.93) (125.88) (116.52).23 Sample size PWA min moderate + vigorous past 7 days (242.02) (299.34) (244.75).16 Sample size PWA min walked past 7 days (111.63) (181.76) (164.99).01 Abbreviations: PWA, past week activity recall from the Behavioral Risk Factor Surveillance Survey 20 (self-reported recall of minutes of various intensity activity within pedometer sample). Note. Sample sizes across columns of JPAC tertile total scores add up to less than the total sample size for the accelerometer and pedometer samples, as total scores were not calculated for those participants who had 1 or more missing index scores. Discussion This study evaluated the reliability and validity of the instrument used to assess physical activity among African Americans of the Jackson Heart Study cohort (JPAC). Test-retest reliability was assessed using a convenience sample of 40 African Americans. Results indicate excellent test-retest reliability across a 2-week interval for both the total scale and the 4 subscales (Active Living, Sports, Home/Life, Work). In each case, the ICC well exceeded.80, the value considered adequate in research on physical activity. 26 Reliability coefficients may be particularly high because of the nature of the broad response categories on the JPAC and brief test-retest interval. Researchers wishing to use the JPAC can thus be confident that this measure is stable over time. Validity using objectively-measured accelerometer and pedometer data were assessed using 2 large subsamples from the Jackson Heart Study. Both subsamples were descriptively similar to the entire JHS sample of 5302 African Americans, in which the majority of JHS participants are middle-aged (M = 54.9, SD = 12.9), female (64.03%), and overweight/obese (M BMI = 31.9, SD = 7.6). 1 Linear trend analyses using tertiles of JPAC total scores confirmed that accelerometer and pedometer counts increased significantly with increased self- reported activity on the JPAC. Validity coefficients for the JPAC total score were typically higher than those of the index scores for both accelerometer and pedometer measures (.24 and.32, respectively). Compared with other work of this genre, these correlations are generally comparable although slightly smaller in size. 24,27 The more modest correlations obtained in the current study may be explained in part by the facts that the accelerometer was worn for only 24 hours and simultaneously with an ambulatory blood pressure monitor, the pedometer was worn for 3 consecutive days, and both objective measures occurred at least 1 to 3 months after the JPAC interview. These methodological limitations were dictated by concerns about overall JHS participant burden and imply that the obtained validity coefficients may be underestimates. The correlations and their significance levels thus are relatively impressive given these limitations. With the exception of the Work Index, relationships between the JPAC subscales and objective activity data were typically significant but varied. Consistent with validation studies of the Baecke-type instruments, the Sports Index correlated most strongly with other measures. Despite excellent test-retest reliability, the Work Index correlated poorly with objective measures of physical activity. Other investigators have also reported similarly high reliability but poor validity for the work

7 S130 Smitherman et al Table 3 Spearman Correlations Between JPAC Scores and Accelerometer/Pedometer Values Accelerometer Mean raw counts/min.24 Min/day of moderate activity Freedson ( 1952).25 Hendelman ( 191).17 P <.001 Swartz ( 574).24 Pedometer Mean counts/day.32 PWA min moderate activity past 7 days.17 PWA min vigorous activity past 7 days.19 PWA min moderate + vigorous past 7 days.17 PWA min walked past 7 days.18 JPAC index scores Total score Active living Sports Home/Life Work P = P = P < P < P = P < P = P = P = P = P = P = P = P = P = P = P = P = P = P = P =.74 Abbreviations: PWA, past week activity recall from the Behavioral Risk Factor Surveillance Survey 20 (self-reported recall of minutes of various intensity activity within pedometer sample). subscale of the Baecke physical activity survey compared with other types of physical activity, 28,29 possibly because of inaccurate reporting and inability of accelerometers to detect upper-body activity contributing to work scores. 24 Work physical activity in ARIC was inversely related to sports and other exercise, 30 rendering it most useful as one component of total activity rather than as an individual predictor per se. The poor validity of this index despite the strong validity of the JPAC total appears to support this notion. Disagreement exists about which accelerometer cutpoints are most appropriate for quantifying minutes of activity in various types of physical activity research. 31 The 3 most common cutpoints for minutes of moderate or higher activity were employed in the current study, and our data confirm the considerable discrepancies between these 3 metrics. 9 The most liberal cutpoint of Hendelman et al 16 was less strongly correlated with JPAC total scores and resulted in estimates that even the lowest JPAC tertile engaged in nearly 4 hours of moderate + vigorous activity per day. Thus, it is very likely that the moderate-activity cutpoint of Hendelman et al is too overinclusive. By comparison, the more conservative cutpoints of Freedson et al 17 and Swartz et al 18 provided estimates that the middle JPAC tertile engaged in 8.38 and minutes of moderate + vigorous activity daily, respectively. These estimates were also significantly associated with JPAC total scores in the linear trend analyses, unlike those of Hendelman et al Thus, the moderate intensity estimates of Freedson and Swartz are likely most informative for the current study, particularly in view of high rates of inactivity among African Americans in the southeastern United States. This study has several strengths that contribute to our understanding of physical activity assessment. This is one of the very first physical activity assessment instruments to be validated specifically for use with African Americans. The methods used in the current study also permit comparisons between self-reported activity and 2 objective measures of physical activity in minimally overlapping African American samples, samples which are considerably larger than those typically employed in research on physical activity assessment. Thirdly, our study analyzed the validity of the JPAC using 3 different accelerometer cutpoints in addition to raw counts; this decision was intended to inform not only statements about the validity of our measure but also to spark further dialogue about the relative discrepancies between these 3 commonly-used cutpoints. Limitations described above (eg, small convenience sample for reliability analyses, collection of accelerometer data in conjunction with wearing an ambulatory blood pressure monitor, time interval between the objective measures and the JPAC interview) were dictated by concerns about participant burden but merit attention in future studies. In summary, our results found that the JPAC is a highly reliable measure of physical activity in African Americans and reflects objectively-derived activity data

8 JHS Survey Validation S131 relatively well. Modest but statistically significant correlations were obtained between the JPAC Total Score and raw accelerometer counts, minutes of moderate + vigorous accelerometer-measured activity, mean daily pedometer counts, and self-reported activity minutes over the past week. Smaller but significant associations were found also between 3 of the 4 JPAC index scores and the various indices of objectively-measured activity, the one consistent exception being poor validity for the Work Index. In conjunction with limitations of the current study, these findings suggest that the JPAC is an appropriate measure to assess physical activity of African Americans that occurs across multiple domains (active living, sports/exercise, home life, and paid or volunteer work). The consistently strong associations between the JPAC total score and the various objective metrics indicate that researchers can use the JPAC total score as a single variable measure of physical activity for a variety of analyses. Evaluating the JPAC with objective measures over longer periods of time and with African Americans from geographic areas other than the southeastern United States will be valuable for future research. Acknowledgments This research from the Jackson Heart Study was supported by NIH contracts N01-HC-95170, N01-HC-95171, and N01-HC that were provided by the National Heart, Lung, and Blood Institute and the National Center for Minority Health and Health Disparities. The authors wish to thank William D. Johnson and Tanya Taylor for their help with preliminary data analysis and collection, respectively, as well as the participants and data collection staff of the Jackson Heart Study. References 1. Taylor HA, Wilson JG, Jones DW, et al Towards a resolution of cardiovascular health disparities in African Americans: design and methods of the Jackson Heart Study. Ethnicity & Disease. 2005;15:S6-4 S Mensah GA, Mokdad AH, Ford ES, Greenlund KJ, Croft JB. State of disparities in cardiovascular health in the United States. Circulation. 2005;111: Whitt M, Kumanyika S, Bellamy S. Amount and bouts of physical activity in a sample of African American women. Med Sci Sports Exerc. 2003;35: Pereira MA, Folsom AR, McGovern PG, et al for the Atherosclerosis Risk in Communities Study Investigators. Physical activity and incident hypertension in Black and White adults: The Atherosclerosis Risk in Communities Study. Prev Med. 1999;28: Dubbert PM. Physical activity and exercise: Recent advances and current challenges. J Consult Clin Psychol. 2002;70: Anderssen N, Jacobs DR, Jr, Sidney S, et al Change and secular trends in physical activity patterns in young adults: a seven-year longitudinal follow-up in the Coronary Artery Risk Development in Young Adults Study (CARDIA). Am J Epidemiol. 1996;143: Dubbert PM, Carithers T, Ainsworth BE, Taylor HA, Jr, Wilson G, Wyatt SB. Physical activity assessment methods in the Jackson Heart Study. Ethn Dis. 2005;15(Suppl 6):S56 S Baecke JA, Burema J, Fritjers JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36: Ainsworth BE, Sternfeld B, Richardson MT, Jackson K. Evaluation of the Kaiser Physical Activity Survey in women. Med Sci Sports Exerc. 2000;32: Sternfeld B, Ainsworth BE, Quesenberry CP, Jr. Physical activity patterns in a diverse population of women. Prev Med. 1999;28: Sternfeld B, Ngo L, Satariano WA, Tager IB. Associations of body composition with physical performance and self-reported functional limitation in elderly men and women. Am J Epidemiol. 2002;156: Ainsworth BE, Haskell WL, Whitt MC, et al Compendium of Physical Activities: An update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;32(Suppl 9):S498 S Dong L, Block G, Mandel S. Activities contributing to total energy expenditure in the United States: Results from the NHAPS study. Int J Behav Nutr Phys Act. 2004;1: Richardson MT, Ainsworth BE, Huei-Ching W, Jacobs DR, Leon AS. Ability of the Atherosclerosis Risk in Communities (ARIC)/Baecke Questionnaire to assess leisuretime physical activity. Int J Epidemiol. 1995;24: Melanson EL, Jr. Validity of the Computer Science and Applications, Inc. (CSA) activity monitor. Med Sci Sports Exerc. 1995;27: Hendelman D, Miller K, Baggett C, Debold E, Freedson P. Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Med Sci Sports Exerc. 2000;32(Suppl 9):S442 S Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Med Sci Sports Exerc. 1998;30: Swartz AM, Strath SJ, Bassett DR, Jr, O Brien WL, King GA, Ainsworth BE. Estimation of energy expenditure using CSA accelerometers at hip and wrist sites. Med Sci Sports Exerc. 2000;32(Suppl 9):S450 S Carithers T, Dubbert PM, Crook ER, et al Diet assessment in African Americans: Methods used in the Jackson Heart Study. Ethn Dis. 2005;15(Suppl 6):S49 S Yore MM, Ham SA, Ainsworth BE, et al Reliability and validity of the instrument used in BRFSS to assess physical activity. Med Sci Sports Exerc. 2007;39: Crouter SC, Schneider PL, Karabulut M, Bassett DR. Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Med Sci Sports Exerc. 2003;35: Schneider PL, Crouter SE, Lukajic O, Bassett DR. Accuracy and reliability of 10 pedometers for measuring steps over a 400-m walk. Med Sci Sports Exerc. 2003;35:

9 S132 Smitherman et al 23. Tudor-Locke C, Ainsworth BE, Thompson RW, Matthews CE. Comparison of pedometer and accelerometer measures of free-living physical activity. Med Sci Sports Exerc. 2002;34: Schmidt MD, Freedson PS, Pekow P, Roberts D, Sternfeld B, Chasan-Taber L. Validation of the Kaiser Physical Activity Survey in pregnant women. Med Sci Sports Exerc. 2006;38: Tudor-Locke C, Burkett L, Reis JP, Ainsworth BE, Macera CA, Wilson DK. How many days of pedometer monitoring predict weekly physical activity in adults? Prev Med. 2005;40: Dale D, Welk GJ, Matthews CE. Methods for assessing physical activity and challenges for research. In: Welk GJ, ed. Physical Activity Assessment for Health-Related Research. Champaign, IL: Human Kinetics; 2002: Chasan-Tabar L, Schmidt MD, Roberts DE, Hosmer D, Markenson G, Freedson PS. Development and validation of a Pregnancy Physical Activity Questionnaire. Med Sci Sports Exerc. 2004;36: Ainsworth BE, Jacobs DR, Jr, Leon AS, Richardson MT, Montoye HJ. Assessment of the accuracy of physical activity questionnaire occupational data. J Occup Med. 1993;35: Jacobs DR, Jr, Ainsworth BE, Hartman TJ, Leon AS. A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med Sci Sports Exerc. 1993;25: Evenson KR, Rosamond WD, Cai J, Pereira MA, Ainsworth BE. Occupational physical activity in the Atherosclerosis Risk in Communities study. Ann Epidemiol. 2003;13: Ainsworth BE, Bassett DR, Jr, Strath SJ, et al Comparison of three methods for measuring time spent in physical activity. Med Sci Sports Exerc. 2000;32(Suppl 9):S457 S464.