Share this article on:

00005768-199608000-0001300005768_1996_28_1026_singh_adventists_8miscellaneous-article< 123_0_11_13 >Medicine & Science in Sports & Exercise©1996The American College of Sports MedicineVolume 28(8)August 1996pp 1026-1037Validity of selected physical activity questions in white Seventh-day Adventists and non-Adventists[Epidemiology]SINGH, PRAMIL N.; TONSTAD, SERENA; ABBEY, DAVID E.; FRASER, GARY E.Center for Health Research, Loma Linda University, Nichol Hall,#2008, Loma Linda, CA 92350Submitted for publication May 1995.Accepted for publication January 1996.We gratefully acknowledge the effort and skills of secretary Cathy Provencio in typing this manuscript and the expert advice of Dr. Ralph Paffenbarger in helping to design the study. This research was supported by the National Heart, Lung, and Blood Institute, grant HL34174, and the National Cancer Institute, grant 5-R01-CA14703.Address for correspondence: Pramil N. Singh, Center for Health Research, Loma Linda University, Nichol Hall, #2008, Loma Linda, CA 92350.Present address for Dr. Tonstad: Solstad Terrasse 34, N-1364 Hvalstad, Norway.ABSTRACTThe validity and reliability of selected physical activity questions were assessed in both Seventh-day Adventist (N = 131) and non-Adventist(N = 101) study groups. Vigorous activity questions similar to those used by others and new questions that measured moderate and light activities were included. Validation was external, comparing questionnaire data with treadmill exercise time, resting heart rate, and body mass index(kg·m-2), and internal, comparing data with other similar questions. Both Adventist and non-Adventist males showed significant age-adjusted correlations between treadmill time and a “Run-Walk-Jog Index” (R = 0.28, R = 0.48, respectively). These correlations increased substantially when restricting analysis to exercise speeds exceeding 3 mph (R= 0.39, R = 0.71, respectively). Frequency of sweating and a vigorous physical activity index also correlated significantly with treadmill time in males. Correlations were generally weaker in females. Moderate- and light-intensity questions were not correlated with physical fitness. Internal correlations (R= 0.50-0.78) between the above three vigorous activity questions were significant in all groups, and correlations (R = 0.14-0.60) for light and moderate activity questions were also documented. Test-retest reliability coefficients were high for vigorous activity questions (R = 0.48-0.85) and for one set of moderate activity questions (R = 0.43-0.75). No important differences in validity and reliability were found between Adventists and non-Adventists, but the validity of vigorous activity measures was generally weaker in females.Numerous studies have shown that higher levels of physical activity are associated with a decreased risk of ischemic heart disease and diabetes melitus and an increased life expectancy(5,15,19,20,22,24). In addition, increased physical fitness as measured by time to exhaustion on a treadmill test is also protective for ischemic heart disease and all-cause mortality (7,8). Although fitness measures(treadmill, cycle performance) are objective and useful when validating an assessment of physical activity, they are not practical or cost-effective in epidemiologic investigations of large populations (13). Most population studies examining the relationship between physical activity and coronary heart disease (CHD) have used physical activity questionnaires that have been validated against such physiologic measures.Seventh-day Adventists rarely use tobacco and alcohol and often subscribe to a vegetarian diet pattern. Thus, they are well suited for investigations designed to distinguish the effects of certain lifestyle practices on risk of disease. In a recent study of over 27,000 California Seventh-day Adventists who responded to a lifestyle and demographic questionnaire in 1974, increased exercise was associated with a decrease in fatal CHD events of up to 50%(10). In this study, physical activity was assessed by three items on the questionnaire that asked for frequency of occupational and leisure time activity. These items specifically measured vigorous activities and could not be used to examine the effects of moderate- and low-intensity activities. Furthermore, the validity of these questions has not been evaluated.The purpose of this study was to investigate the validity and reliability of a physical activity questionnaire for use in future epidemiologic investigations of Adventists. Questionnaire items of known validity in non-Adventists and newly designed questions measuring moderate- and low-intensity activities were used. The questionnaire was administered to both Adventists and non-Adventists, and the results obtained in the two groups are compared. The results of this study are also compared to findings from other populations.MATERIALS AND METHODSWe measured physical activity and fitness levels during “health testing” programs where subjects underwent treadmill testing and completed a short questionnaire. The validity of the questionnaire was studied by comparison of responses to questionnaire items with direct (treadmill time) and indirect (resting heart rate, body mass index) measures of physical fitness. There is an established relation between physical fitness and both health outcome and actual physical activity (12). In determining correlations between self-reported physical activity measures and fitness indicators, we tested the validity of using simple, self-administered questionnaire items to measure physical fitness and, by this, to indirectly estimate physical activity. Internal validity of self-reported physical activity measures was also studied by comparison of responses to similar items. Reliability was evaluated by comparing responses to questions on the initial questionnaire with responses to the identical questions on a follow-up questionnaire mailed to subjects 6 wk after the initial questionnaire administration and treadmill test.SubjectsSeventh-day Adventist subjects were recruited through church announcements for a “health testing” program in a community with a large proportion of Seventh-day Adventists. Adventist participants in this program were offered a free assessment of physical fitness level when they responded to the questionnaire and agreed to complete a follow-up questionnaire 6 wk later. A group of non-Adventist subjects was recruited through community announcements for a health testing program. A standard protocol for initial questionnaire administration, treadmill testing, and 6-wk follow-up questionnaire mailing was also used in this group. Written informed consent was obtained from all subjects.Selection criteria used to enroll the Adventist and non-Adventist subjects required that subjects be 25 yr or older and that individuals with a history of clinical illness affecting their ability to exercise be excluded from participation. Overall, 232 subjects completed the initial questionnaire and treadmill test. Twenty-seven subjects with a positive treadmill test(development of typical angina or >1.25 mm of ST depression on the ECG[electrocardiogram]) or bundle branch block on the ECG were excluded from the validity analyses involving comparisons between fitness indicators (treadmill time, resting heart rate, body mass) and questionnaire data but were not excluded from other analyses.The response rate to the 6-wk follow-up questionnaire was 79% (183/232). However, only 112 of these 183 responders indicated on their follow-up questionnaire that they did not change their exercise habits since administration of the initial questionnaire and treadmill test. Hence, these 112 were the subjects included in the reliability study.The total number of subjects enrolled in the validity and reliability studies is shown in Table 1.TABLE 1. Total subjects enrolled in validity and reliability studies.MeasurementsPhysical activity questionnaire. The physical activity questionnaire obtained information about subjects' usual activity levels over the previous 3 months. These questions included some items previously validated and used by others, as well as some additional questions that we designed to measure vigorous and moderate- and light-intensity activities. A description of the variables measured and indices derived from the questionnaire are found in the Appendix(Table A1).TABLE A1. Variables and indices on the Physical Activity Questionnaire.The first 10 items were modified from the questionnaires used in the Harvard College Alumni Study (17) and a substudy of the Aerobics Center Longitudinal Study (12). These items included stair-climbing, city blocks walked in a day (1 block = 1/12 mile), pace of walking, activities with an intensity sufficient to “work up a sweat” (frequency, duration), and run-walk-jog workouts (frequency, duration, distance). In addition, we designed a 10-item “vigorous activity” question group to obtain a report of the frequency (times per week) and duration (minutes per session) of 1) leisure-time vigorous activities (nine items: walking, running, or jogging; aerobics; bicycle or stationary cycling; swimming; sports; moderate to heavy labor; yard work; and two additional “write-in” vigorous leisure time activities) and 2) occupational vigorous activity (one item).Following data collection from non-Adventists and the first group of Adventists, we designed questions to measure moderate- and light-intensity activities. These were included as six additional items (two sets of questions) on the questionnaires administered to the final 60 Adventist subjects (52 with a valid treadmill test). These questions are shown in theAppendix (Table A2).TABLE A2. Two sets of moderate- and light-intensity question items administered to 60 Adventist subjects.The self-administered 6-wk follow-up questionnaire was identical to the initial self-administered questionnaire except for one additional question asking if they had changed their physical activity patterns during the 6-wk interval following the treadmill test. If so, they were excluded from reliability analyses.Treadmill test. Prior to the treadmill testing, resting blood pressure, pulse rate, height, and weight were measured by trained personnel. Assessment of physical fitness was performed through maximal exercise treadmill testing using either the Bruce or the modified Bruce protocol. Treadmill duration using the modified Bruce protocol was adjusted in all analyses to correspond to the Bruce protocol.Data analysis. Responses to certain questionnaire items were used to compute pertinent physical activity indices. A “Run-Walk-Jog Index” was computed using the responses to questions asking for time(minutes) per run-walk-jog workout (excluding rest stages), distance traveled(miles) per workout, and frequency (workouts per week) of run-walk-jog workouts. Using the Compendium of Physical Activities from the SAFE study(3), each subject was assigned an MET value (metabolic equivalent of the task) based on the calculated speed (distance per workout divided by time per workout) of their run-walk-jog workout (walking, up to 4.5 mph = 2.0-4.5 MET; running, 5-11 mph = 8.0-18.0 MET). Subjects indicating no walking, running, or jogging in their exercise program were assigned a zero run-walk-jog MET value. The final index value was computed using the following formula: EquationEquation 1AAn index of energy expenditure from up to nine vigorous “sports and recreation” activities (walking, running, and jogging; aerobics; cycling; swimming; sports; moderate to heavy labor; yard work; and two additional vigorous leisure time activities) was determined as the sum of energy expended in each of these activities using the above formula with appropriate MET values (3).An aerobic “Physical Activity Index” was calculated by determining the total energy expended (MET min·wk-1) from stair-climbing, walking (city blocks), and the above index of sports and recreation activities. The index value was calculated using a formula similar to that used by Paffenbarger et al. (17) on the College Alumni Questionnaire: EquationEquation 1BSince the questionnaire used in this study asked for usual frequency and duration of activities over the previous 3 months, we did not adjust the final index value based on the total weeks per year of participation in a specific activity as was done in the College Alumni Study.For the 60 subjects who were administered questions on moderate- and light-intensity activities, a “Total Energy Index” was computed from the first set of questions and the vigorous sports and recreation and labor questions: EquationEquation 1CFor each activity level, energy expenditure was calculated by assigning the appropriate MET value from the Compendium of Physical Activities (light activity, sitting = 1.0 MET; light activity, standing = 2.3 MET; moderate activity = 4.5 MET; vigorous activity, labor = 6 MET; vigorous activity, sports and recreation = total METs expended in up to nine vigorous activities).The validity study used two different types of analysis. The first involved comparisons between several fitness indicators (maximal treadmill time, Quetelet's index of body mass, resting heart rate) and questionnaire variables using first-order partial correlations (controlling for age). The measures of maximal treadmill time, Quetelet's index of body mass, and resting heart rate in this population were approximately normally distributed. Log transformations to further improve the normality of these distributions did not substantially alter the magnitude or significance level of the partial correlations. For each correlation analysis, a plot of standardized residuals was examined. Influential outliers (standardized residual ≥3) on either the fitness indicator or activity measure were excluded (one to two subjects excluded in any particular analysis).The second type of validity analysis involved comparisons between related items on the questionnaire. As most of the questionnaire data were either categorical or indices with nonnormal distributions, Spearman's Rank correlation coefficients were computed.The reliability study required comparisons between identical questionnaire items and computed indices derived from the initial questionnaire and 6-wk follow-up questionnaires. Since these data were also often categorical or not normally distributed, Spearman's Rank correlation coefficients were computed.RESULTSThere were no significant differences between Adventists and non-Adventists of either gender with respect to age, resting heart rate, maximum heart rate, or resting diastolic pressure (Table 2). However, a significantly lower body mass index was found in Adventist males than in non-Adventist males, and resting systolic blood pressure in Adventist females was significantly higher than in non-Adventist females. This latter finding may be related to the greater number of elderly participants among the Adventists (note the larger standard deviation of age). None of the Adventist participants indicated regular smoking within the past year; among the non-Adventist participants, 7% of males and 20% of females were smokers.TABLE 2. Characteristics of the total study population by religious preference. Within each sex strata mean values for Adventist and non-Adventist groups are compared.Treadmill time and sweat frequency were significantly higher in the Adventists than in non-Adventists for both males and females(Table 2). There were no significant gender-specific differences between Adventists and non-Adventists in stair-climbing, Physical Activity Index, Sports and Recreation Index, or Run-Walk-Jog Index.Validity StudyIn the first part of the validity assessment, we determined correlations between maximal treadmill time and selected question items and indices within the four gender-religion groups of participants (Table 3). Thus, comparisons were made between responses to our physical activity questions and an index of physical fitness. In the total study population, there was a significant zero-order correlation between maximal treadmill time and age (R = -0.35, P < 0.00001), Therefore, first-order partial correlations adjusting for age were used in these analyses.TABLE 3. Partial correlation coefficients (age-adjusted) between maximal treadmill time and selected questionnaire items measuring more vigorous activities.There was no question item or index that was significantly correlated with treadmill time across all four groups. Among all the measures taken from the questionnaire, the Run-Walk-Jog Index and the response to the question on sweat frequency showed the strongest age-adjusted associations with maximal treadmill time. In the males, significant positive correlations with treadmill time were found for the Run-Walk-Jog Index and the sweat frequency item. Similar, although weaker, correlations for the Run-Walk-Jog Index and sweat frequency item were found in non-Adventist females. These measures were not significantly correlated with treadmill time in the Adventist females.It was noted that the proportion of elderly subjects among Adventist females (20% over the age of 75) was much higher than in the other three groups (<2% over the age of 75). When subjects over the age of 75 were excluded from the analysis of Adventist females, the magnitude of the age-adjusted correlation with treadmill time increased modestly for both sweat frequency (R = 0.16, P = 0.27) and the Run-Walk-Jog Index(R = 0.27, P < 0.07).The Run-Walk-Jog Index was further examined by stratifying the subjects within each group on speed (<3 mph, ≥3 mph). In all four groups, the correlation between treadmill time and the index was substantially higher for subjects reporting run-walk-jog workouts of 3 mph or higher. For Adventists and non-Adventists, these correlations were significant in the males and close to significance (P < 0.06) in the females. In both genders, restricted to those with workout speeds exceeding 3 mph, the Run-Walk-Jog Index showed correlations of higher magnitude in the non-Adventists.No significant correlations between treadmill time and question items measuring pace of walking, stair-climbing, and blocks walked were found in any group. Energy expenditure from up to nine sports and recreation activities was highly correlated with treadmill time in Adventist males only. The Physical Activity Index measuring energy expenditure computed from stair-climbing, blocks walked, and sports and recreation activities was significantly correlated with treadmill time in both Adventist and non-Adventist males.Associations between selected vigorous activity measures and nontreadmill fitness indicators were also investigated. There were generally nonsignificant, but negative, age-adjusted correlations between resting heart rate and both sweat frequency and the Physical Activity Index. Negative correlations (age-adjusted) between resting heart rate and the Run-Walk-Jog Index were somewhat stronger (Adventist males, R = -0.032, P = 0.82; non-Adventist males, R = -0.32, P = 0.02; Adventist females, R =-0.28, P = 0.05; non-Adventist females, R = -0.41, P = 0.02). Quetelet's index of body mass showed nonsignificant negative or positive age-adjusted correlations with sweat frequency, the Run-Walk-Jog Index, and the Physical Activity Index in all four gender-religion groups.There were no significant positive correlations between treadmill time and responses to the moderate- and light-intensity activity questions administered only to a subset of the Adventist population (Table 4). However, the “sleeping/reclining” question showed a weak negative correlation with treadmill time in Adventist males and females. Correlations between Quetelet's index of body mass and the moderate- and light-intensity activity questions were of low magnitude and nonsignificant in the gender-specific Adventist groups. Among Adventist females, resting heart rate was significantly correlated (age-adjusted) with the response to the question of moderate-intensity activities in set 2 (R = -0.43, P = 0.05). No other important correlations between resting heart rate and the moderate- and light-intensity activities in the Adventists were found.TABLE 4. Partial correlation coefficients (age-adjusted) between maximal treadmill time and questionnaire items measuring moderate- and light-intensity activities and an index of total energy expenditure.*aValidity was also assessed by correlating responses to related items and indices in the questionnaire. The questions were analyzed in two groups: 1) correlations between vigorous activity questions (stair-climbing, blocks walked, Sports and Recreation Index, Physical Activity Index, pace of walking, sweat frequency, and the Run-Walk-Jog Index) (Tables 5 and 6), and 2) correlations between two sets of moderate- and light-intensity activity questions (Table 7).TABLE 5. Spearman's rank correlations: Adventist subjects. Relate variables along lower edge to those on the left side for females or the right side for males.TABLE 6. Spearman's rank correlations: non-Adventist subjects. Relate variables along lower edge to those on the left side for females or the right side for males.TABLE 7. Internal validity. Spearman's rank correlations between two sets of moderate- and light-activity questions in 60 Adventist subjects. Adventist females (N = 27)Among the vigorous activity questions, significant positive correlations, often 0.5-0.8 in magnitude, were found between the Physical Activity Index, Run-Walk-Jog Index, sweat frequency item, and city blocks walked item across all four groups. Stair-climbing showed no significant positive correlations with the vigorous activity questions in any group. Pace of walking showed no consistent association with the vigorous activity questions.Among the 60 Adventist subjects who were administered moderate- and light-intensity questions, correlations between the moderate and light(standing, sitting) activity questions in set 1 were compared to similar questions in set 2 (Table 7). Generally the diagonal elements of the table, comparing similar activities, are statistically significant with correlations in the 0.3-0.6 range.Reliability StudyCorrelations between identical variables on the initial questionnaire and the 6-wk follow-up questionnaire were determined (among those who indicated no change in exercise habits at follow-up) in order to evaluate the reliability of selected question items and indices (Table 8). Across all four groups, there were significant correlations, generally 0.5-0.8 in magnitude, between measures on the initial questionnaire and the 6-wk follow-up questionnaire for the stairs climbed item, city blocks walked item, Sports and Recreation Index, the Physical Activity Index, the Run-Walk-Jog Index, the sweat frequency item, and the sweat duration item.TABLE 8. Test-retest reliability. Spearman's rank correlations between variables on the questionnaire and the 6-wk follow-up questionnaire.There were 30 Adventists (among Adventists who indicated no change in exercise habits at follow-up) who responded to the moderate- and light-intensity questions on both the initial questionnaire and the follow-up questionnaire. Reliability of these items was evaluated in the whole group rather than within gender strata due to the small size of the sample. For set 1, there were significant correlations in the range of 0.5-0.75 between responses to the two questionnaires for moderate- and light-intensity activity items. For set 2 (weekend and weekday activities separated), significant correlations were found only for the light-intensity questions.DISCUSSIONWe administered a physical activity questionnaire to study subjects to measure levels of vigorous and moderate- and light-intensity activity and to assess the validity and reliability of these items in an Adventist and comparison non-Adventist population. Seventh-day Adventist populations have contributed much to the epidemiology of chronic diseases as investigations have taken advantage of certain dietary and socioreligious factors associated with church membership that may decrease risk of chronic disease. Such factors include almost no current use of tobacco or alcohol, adherence to a lactoovo vegetarian diet by a large proportion of church members, an above average educational level, and a 50% greater frequency of vigorous exercise(9). Of particular note, there is the possibility of important differences in the distribution of usual activities due to Sabbath observance among Adventists. Saturday is usually a day devoid of usual work and recreational activities, although activities such as hiking or walking are not uncommon in the afternoon. This raises the possibility of systematic differences in responses to physical activity items on a questionnaire. Therefore, in this report we have investigated whether questions that have been successfully used to classify habitual vigorous physical activity in studies of general population samples yield similar results in an Adventist population.One way of evaluating the validity of reported physical activity(particularly vigorous activities) is by comparison with a fitness indicator or other variables expected to be correlated with physical activity. The three questions taken directly from the College Alumni Questionnaire(stair-climbing, city blocks walked, and pace of walking) were not significantly correlated with treadmill time or body mass index in Adventists of either gender. Similarly, in the non-Adventists, stair-climbing and city blocks walked showed no significant associations with treadmill time or body mass. These findings confirm the work of others who have used these items from the College Alumni Questionnaire (1,11,25). Ainsworth et al. (1) found a low, nonsignificant correlation between the stairs and blocks items and physical fitness measures(peak V2, percent body fat) in a population of 78 men and women (mean age, 38 ± 9 yr) with a broad range of physical activities. It was suggested that subjects in this population may be underestimating daily walking and stair-climbing, particularly in household and transportation activities. Also, these activities are usually not particularly vigorous and may not greatly affect fitness.A Physical Activity Index was computed as the sum of energy expended on stair-climbing, city blocks walked, and up to nine sports and recreation activities. In our study population, a significant correlation between the Physical Activity Index and maximal treadmill time was found only in Adventist and non-Adventist males, perhaps partly reflecting the higher values of this index in the males. This index is somewhat similar to the “Physical Activity Index-College Alumni Questionnaire” (PAI-CAQ) measure used by Paffenbarger et al. (17) in the population of male alumni from Pennsylvania and Harvard University. In previous studies, significant associations between the PAI-CAQ and cardiorespiratory fitness, recorded body motion, body mass, and highdensity lipoprotein cholesterol have also been found in males (1).The higher correlation between fitness measures and vigorous activity questions (sweat frequency, Physical Activity Index, Run-Walk-Jog Index) seen in males as compared to females in our data set is similar to findings by Siconolfi et al. (23). Such differences may be expected considering data from large community surveys indicating a higher level of participation in running/jogging and vigorous sports and recreational activities in males 4,16. Using a modification of the Minnesota Leisure Time Physical Activity (LPTA) questionnaire, Ainsworth et al. (2) specifically showed a significantly lower participation in heavy-intensity LPTA and a larger contribution of household activities to overall LPTA among women as compared to men. Taken together, these findings suggest that some of the commonly used questions on physical activity surveys are not gender specific and therefore may measure a smaller proportion of the important contributors to energy expenditure and fitness in female subjects.The Run-Walk-Jog Index and sweat frequency item showed the strongest associations with maximal treadmill time. As would be expected, subjects in all groups whose run-walk-jog workouts were more vigorous (≥3 mph) showed an even stronger association between the Run-Walk-Jog Index and treadmill time. The lower magnitude correlations for sweat frequency and the Run-Walk-Jog Index in Adventist females have been shown to be partly attributable to the older age distribution of this group. These results suggest that it is mainly running, walking, or jogging greater than 3 mph that influences fitness, particularly among males and younger females. This speed is probably not often exceeded by elderly subjects in their usual exercise programs.The associations between resting heart rate and Run-Walk-Jog Index and sweat frequency were less clear but suggestive of effect. Resting heart rate was negatively correlated with these two measures in all groups. Again, the Run-Walk-Jog Index showed the most consistent associations, providing further evidence for the usefulness of this index. It should be noted that while a lower resting heart rate may reflect increased fitness levels, the role of other important determinants of resting heart rate (genetic, dietary, psychosocial) limit the use of this measure as a fitness indicator.Other studies have found similar correlations between measures of physical fitness and a Run-Walk-Jog Index or items measuring the frequency of exercise-induced sweating. In a population of Cooper Clinic men, Kohl et al.(12) found that age, a Run-Walk-Jog Index, and frequency of sweating were significant predictors of treadmill time in multiple-regression analyses. The magnitude of their age-adjusted correlations between treadmill time and both sweat frequency and the Run-Walk-Jog Index were about 0.5. In our population of Adventist and non-Adventist males, the magnitude of the correlations with sweat frequency and the Run-Walk-Jog Index were somewhat lower, but in a similar range when a run-walk-jog speed of >3 mph was required. It should be noted that the Cooper Clinic men were younger(mean age, 46.1 yr) than the males in our population and, in addition, the Run-Walk-Jog Index in the Cooper Clinic study was calculated by a different method based on a regression-derived measure using speed, distance, and frequency of workouts as predictors of treadmill time.In a volunteer sample of men and women, (mean age, 42 ± 15 yr), Siconolfi et al. (23) found a significant association between sweat episodes per week and maximal oxygen uptake on a cycle ergometer among males (R = 0.54, P < 0.01) but no significant association among females (R = 0.24). Washburn et al. (25) reported a weak age-gender adjusted correlation (R = 0.11, P < 0.05) between sweat hours per week and high-density lipoprotein cholesterol in a large random sample of Boston area residents (mean age, 40 ± 11 yr). In an earlier study of a sample of college students, Washburn et al.(26) found a negative association (R = -0.29,P < 0.01) between resting heart rate and sweat episodes per week. We found an age-adjusted negative correlation in non-Adventist females (R =-0.47, P < 0.01), but this was not clearly seen in the other groups.Our results are generally in agreement with those of others in finding rather weak, though often statistically significant, correlations between simple questions of vigorous physical activity and measures of physical fitness. Constitutional factors aside from exercise habits are important determinants of physical fitness (18,21). Hence, the modest associations found for the activities measured are to be expected. It is generally considered that it is primarily vigorous activity that increases physical fitness. This last fact may explain the absence of any significant positive correlations between maximal treadmill time and the moderate- and light-intensity items, exercise programs where running, walking, or jogging was [lte]3 mph, casual walking (city blocks), or stair-climbing(often an activity of brief duration). It is also not surprising that our findings indicate that fitness measures such as maximal treadmill time are not appropriate for the validation of reported moderate- or light-intensity activity levels. Alternate validation methods (correlations with pedometer readings, continuous heart rate monitoring, physical activity diaries) should be explored.Validation of questionnaire data by internal comparison of related items/indices produced high correlations between the vigorous activity measures. The Physical Activity Index, Run-Walk-Jog Index, and the sweat frequency item were highly correlated with each other in all groups (R = 0.50-0.78, P < 0.0001). These results are similar to those of Siconolfi et a. (23), Laporte et al.(14), and Washburn et al. (26), who also examined correlations between sweat frequency and a Physical Activity Index based on the PAI-CAQ in their study populations. This presumably indicates that these questions are each estimating some underlying physical activity factor, as required. Furthermore, the high correlation among these vigorous activity items would indicate that surveys may need to include only one or a few of these items as an estimator of physical fitness.The correlations between the two sets of moderate- and light-intensity questions among Adventist subjects were of lower magnitude (R = 0.14-0.60) than would be expected given the similarity in question content between them. The variability in subject responses to these items was not substantially different, relative to mean levels, from those found in vigorous activity items. These findings may indicate that differences in question format, particularly with respect to time interval and discrimination between weekday vs weekend activities, contributed to within-subject variation in responses to questions of moderate- and light-intensity activities. It is probably most important that these lower-intensity activities are continual, hence less memorable, so they are likely to be recalled and reported with less accuracy.Test-retest reliability coefficients were high (generally 0.6-0.8) for the vigorous activity questions and the simple city blocks walked and stairs climbed questions. The magnitude of these estimates is partly attributable to the exclusion of subjects indicating a change in physical activity patterns during the test-retest interval. Our objective in reporting test-retest correlations was to establish whether the questionnaire items could perform similarly in two independent tests under the same circumstances. Hence, by definition, subjects reporting changes in activity level during the test-retest interval had to be excluded. Reliability results for the moderate- and light-intensity activities were less consistent with coefficients of lower magnitude, which is not surprising. Similalry, lower test-retest correlations for moderate and light activities relative to higher-intensity leisure activities have also been found in other studies(6,11). Despite this, the simpler set of questions(set 1) had acceptable reliability, with coefficients in the range of 0.50-0.75. The small sample used in the test-retest analyses of the moderate- and light-intensity questions must also be acknowledged.There are several limitations of this study that must be considered. The subjects were a nonrandom sample of healthy volunteers recruited from health testing programs and therefore may not be representative of the population subgroups (gender-religion) described in this report. However, the Adventists are probably representative of the type of subject who volunteers for epidemiologic and other health-related studies, where recruiting is also through church advertisements.The use of treadmill performance as an estimate of physical fitness has some limitations as body size and familiarity with the test may influence outcome. This may have been especially true of the elderly females in the Adventist group. The objective was to investigate these questions as measures of both physical fitness and exercise habits. The treadmill as a criterion is a reasonable choice for the first but a very indirect measure of the second.The questionnaire used in this study asks for a report of usual physical activity over the past 3 months. While this shorter time period facilitates greater recall of physical activity patterns, it does not allow for seasonal variation in activities. It should be noted that activity levels more than 3 months in the past should not be expected to directly impact treadmill performance, which was our major validity criterion.Also, we were unable to evaluate the validity of vigorous occupational labor, as only 17 subjects indicated that their work involved such activities. This may indicate that, in this population, these activities contribute little to the total energy expenditure from vigorous activities.Finally, we have reported validity and reliability of physical activity questions in a population where subjects with positive treadmill tests(ischemia, > 1.25 mm ST depression, sustained arrhythmia) or bundle branch block on the ECG were excluded. This may have resulted in higher correlations between activity questions and fitness measures in this presumably healthier population than would be found in the total study population of a large cohort study. However, in using treadmill performance to validate physical activity, such exclusions are appropriate since physically active heart patients may have impaired maximal exercise capacity. Also, prospective analyses frequently exclude subjects with known or suspect prevalent disease.In summary, our results indicate that certain items from questionnaires used in studies of other populations can probably be used to classify habitual vigorous physical activity patterns of white Seventh-day Adventists. These include a simple question measuring frequency of exercise-induced sweating, a short series of questions about run-walk-jog workouts (especially at an intensity of least 3 mph), and set 1 of the light- and moderate-intensity activity questions. Indices of total energy expenditure correlated less well with physical fitness, presumably because of the substantial component of nonvigorous activities. As expected, our light- and moderate-intensity activity questions also correlated poorly with physical fitness. Overall, there were no clear differences in validity between Adventists and non-Adventists. In both groups, the vigorous activity items provided a more valid estimate of fitness among males as compared to females. Test-retest reliability in all groups was also good for most questions of vigorous activities and for set 1 of the moderate- and light-intensity activity questions.REFERENCES1. Ainsworth, B. E., A. S. Leon, M. T. Richardson, D. R. Jacobs, and R. S. Paffenbarger. Accuracy of the College Alumnus Physical Activity Questionnaire. J. Clin. Epidemiol. 46:1403-1411, 1993. [Context Link]2. Ainsworth, B. E., M. Richardson, D. R. Jacobs, and A. S. Leon. Gender differences in physical activity. Women Sport Phys. Act. J. 2:1-16, 1993. [Context Link]3. Ainsworth, B. E., W. L. Haskell, A. S. Leon, et al. Compendium of Physical Activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc. 25:60-70, 1993. [Context Link]4. Blair, S. N., W. L. Haskell, R. D. Paffenbarger, K. M. Vranizan, J. W. Farquhar, and P. D. Wood. Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments. Am. J. Epidemiol. 112:794-804, 1985. [Context Link]5. Blair, S. N., H. W. Kohl, C. E. Barlow, R. S. Paffenbarger, L. W. Gibbons,[and] and C. A. Macera. Changes in physical fitness and all-cause mortality. J.A.M.A. 273:1093-1098, 1995. [Context Link]6. Blair, S. N., M. Dowda, R. R. Pate, et al. Reliability of long term recall of participation in physical activity by middle aged men and women. Am. J. Epidemiol. 133:266-275, 1991. [Medline Link] [Context Link]7. Blair, S. N., H. W. Kohl, R. S. Paffenbarger, D. G. Clark, K. H. Cooper, and L. W. Gibbons. Physical fitness and all-cause mortality. J.A.M.A. 262:2396-2401, 1989. [Context Link]8. Ekelund, L. G., W. L. Haskell, J. L. Johnson, F. S. Whaley, M. H. Criqui, and D. S. Sheps. Physical fitness as a predictor of cardiovascular mortality in asymptomatic North American men. The Lipid Research Clinics Mortality Follow-up Study. N. Engl. J. Med. 319:1379-1384, 1988. [Context Link]9. Fraser, G. E., P. W. Dysinger, C. Best, and R. Chan. IHD risk factors in middle aged Seventh-day Adventist men and their neighbors.Am. J. Epidemiol. 126:638-646, 1987. [Medline Link] [Context Link]10. Fraser, G. E., M. Strahan, J. Sabate, W. L. Beeson, and D. Kissinger. Effects of traditional coronary risk factors on rates of incident coronary events in a low-risk population. The Adventist Health Study.Circulation 86:406-413, 1992. [Context Link]11. Jacobs, D. R., B. E. Ainsworth, T. J. Hartman, and A. S. Leon. A simultaneous evaluation of 10 commonly used physical activity questionnaires. Med. Sci. Sports Exerc. 25:81-91, 1993. [CrossRef] [Full Text] [Medline Link] [Context Link]12. Kohl, H. W., S. N. Blair, R. S. Paffenbarger, C. A. Macera, and J. J. Kronenfeld. A mail survey of physical activity habits as related to measured physical fitness. Am. J. Epidemiol. 127:1228-1239, 1988. [Medline Link] [Context Link]13. Laporte, R. E., H. J. Montoye, and C. J. Caspersen. Assessment of physical activity in epidemiologic research: problems and prospects. Public Health Rep. 100:132-146, 1985. [Context Link]14. Laporte, R. D., R. Black-Sandler, J. A. Cauley, et al. The assessment of physical activity in older women: analysis of the interrelationship and reliability of activity monitoring, activity surveys, and caloric intake. J. Gerontol. 38:394-397, 1983. [CrossRef] [Medline Link] [Context Link]15. Morris, J. N., M. G. Everitt, R. Pollard, and S. P. Chave. Vigorous exercise in leisure time: protection against coronary heart disease. Lancet 2:1207-1210, 1980. [CrossRef] [Medline Link] [Context Link]16. National Center for Health Statistics. Caloric and selected nutrient values for persons 1-74 years of age: First Health and Nutrition Examination Survey United States 1971-1974. Vital Health and Statistics, Series 11. Data from the National Health Survey, no. 209. DHNEW Population no. (PHS) 79-1657. Rockville, MD: National Center for Health Statistics. 1979, p. 37. [Context Link]17. Paffenbarger, R. S., A. L. Wing, and R. T. Hyde. Physical activity as an index of heart attack risk in college alumni.Am. J. Epidemiol. 108:161-175, 1978. [Medline Link] [Context Link]18. Paffenbarger, R. S., S. N. Blair, I. Lee, and R. T. Hyde. Measurement of physical activity to assess health effects in a free living population. Med. Sci. Sports Exerc. 25:60-70, 1993. [CrossRef] [Full Text] [Medline Link] [Context Link]19. Paffenbarger, R. S., R. T. Hyde, A. L. Wing, and C. Hsieh. Physical activity, all cause mortality, and longevity of college alumni. N. Engl. J. Med. 314:605-613, 1986. [CrossRef] [Medline Link] [Context Link]20. Pekkanen, J., B. Marti, A. Nissinen, J. Tuomilehot, S. Punsar, and M. J. Karvonen. Reduction of premature mortality by high physical activity: a 20-year follow-up of middle aged Finnish men. Lancet 1:1473-1477, 1987. [CrossRef] [Medline Link] [Context Link]21. Perusse, A., C. Tremblay, C. Leblanc, and C. Bouchard. Genetic and environmental influences on level of habitual physical activity and exercise participation. Am. J. Epidemiol. 129:1012-1022, 1989. [Medline Link] [Context Link]22. Powell, K. E., P. D. Thompson, C. J. Caspersen, and J. S. Kendrick. Physical activity and the incidence of coronary heart disease.Annu. Rev. Public Health 8:253-287, 1987. [CrossRef] [Medline Link] [Context Link]23. Siconolfi, S. F., T. M. Lasater, R. C. K. Snow, et al. Self-reported physical activity compared with maximal oxygen uptake.Am. J. Epidemiol. 122:101-105, 1985. [Medline Link] [Context Link]24. Slattery, M. L. and D. R. Jacobs. Physical fitness and cardiovascular disease mortality. The US Railroad Study. Am. J. Epidemiol. 127:571-580, 1988. [Context Link]25. Washburn, R. A., S. R. W. Goldfield, K. W. Smith, and J. B. McKinlay. The validity of self-reported exercise-induced sweating as a measure of physical activity. Am. J. Epidemiol. 132:107-113, 1990. [Medline Link] [Context Link]26. Washburn, R. A., L. L. Adams, and G. T. Haile. Physical activity assessment for epidemiologic research: the utility of two simplified approaches. Prev. Med. 16:636-646, 1987. [CrossRef] [Medline Link] [Context Link]APPENDIXTableTable [Context Link]RELIABILITY; GENDER DIFFERENCES; PHYSICAL ACTIVITY QUESTIONNAIRE; FITNESS;|00005768-199608000-00013#xpointer(id(R6-13))|11065405||ovftdb|SL00000429199113326611065405P84[Medline Link]|00005768-199608000-00013#xpointer(id(R9-13))|11065405||ovftdb|SL00000429198712663811065405P87[Medline Link]|00005768-199608000-00013#xpointer(id(R11-13))|11065213||ovftdb|00005768-199301000-00012SL000057681993258111065213P89[CrossRef]|00005768-199608000-00013#xpointer(id(R11-13))|11065404||ovftdb|00005768-199301000-00012SL000057681993258111065404P89[Full Text]|00005768-199608000-00013#xpointer(id(R11-13))|11065405||ovftdb|00005768-199301000-00012SL000057681993258111065405P89[Medline Link]|00005768-199608000-00013#xpointer(id(R12-13))|11065405||ovftdb|SL000004291988127122811065405P90[Medline Link]|00005768-199608000-00013#xpointer(id(R14-13))|11065213||ovftdb|SL0000482519833839411065213P92[CrossRef]|00005768-199608000-00013#xpointer(id(R14-13))|11065405||ovftdb|SL0000482519833839411065405P92[Medline Link]|00005768-199608000-00013#xpointer(id(R15-13))|11065213||ovftdb|SL0000553119802120711065213P93[CrossRef]|00005768-199608000-00013#xpointer(id(R15-13))|11065405||ovftdb|SL0000553119802120711065405P93[Medline Link]|00005768-199608000-00013#xpointer(id(R17-13))|11065405||ovftdb|SL00000429197810816111065405P95[Medline Link]|00005768-199608000-00013#xpointer(id(R18-13))|11065213||ovftdb|00005768-199301000-00010SL000057681993256011065213P96[CrossRef]|00005768-199608000-00013#xpointer(id(R18-13))|11065404||ovftdb|00005768-199301000-00010SL000057681993256011065404P96[Full Text]|00005768-199608000-00013#xpointer(id(R18-13))|11065405||ovftdb|00005768-199301000-00010SL000057681993256011065405P96[Medline Link]|00005768-199608000-00013#xpointer(id(R19-13))|11065213||ovftdb|SL00006024198631460511065213P97[CrossRef]|00005768-199608000-00013#xpointer(id(R19-13))|11065405||ovftdb|SL00006024198631460511065405P97[Medline Link]|00005768-199608000-00013#xpointer(id(R20-13))|11065213||ovftdb|SL0000553119871147311065213P98[CrossRef]|00005768-199608000-00013#xpointer(id(R20-13))|11065405||ovftdb|SL0000553119871147311065405P98[Medline Link]|00005768-199608000-00013#xpointer(id(R21-13))|11065405||ovftdb|SL000004291989129101211065405P99[Medline Link]|00005768-199608000-00013#xpointer(id(R22-13))|11065213||ovftdb|SL000018541987825311065213P100[CrossRef]|00005768-199608000-00013#xpointer(id(R22-13))|11065405||ovftdb|SL000018541987825311065405P100[Medline Link]|00005768-199608000-00013#xpointer(id(R23-13))|11065405||ovftdb|SL00000429198512210111065405P101[Medline Link]|00005768-199608000-00013#xpointer(id(R25-13))|11065405||ovftdb|SL00000429199013210711065405P103[Medline Link]|00005768-199608000-00013#xpointer(id(R26-13))|11065213||ovftdb|SL0000664519871663611065213P104[CrossRef]|00005768-199608000-00013#xpointer(id(R26-13))|11065405||ovftdb|SL0000664519871663611065405P104[Medline Link]3684976Validity of selected physical activity questions in white Seventh-day Adventists and non-AdventistsSINGH, PRAMIL N.; TONSTAD, SERENA; ABBEY, DAVID E.; FRASER, GARY E.Epidemiology828InternalMedicine & Science in Sports & Exercise20033591537-1545SEP 2003Validity of a Modified CHAMPS Physical Activity Questionnaire among African-AmericansRESNICOW, K; MCCARTY, F; BLISSETT, D; WANG, T; HEITZLER, C; LEE, RE & Science in Sports & Exercise10.1249/01.mss.0000183851.94261.d22006382208-216FEB 2006The Physical Activity Recall Assessment for People with Spinal Cord Injury: ValidityLATIMER, AE; MARTIN GINIS, KA; CRAVEN, BC; HICKS, AL & Science in Sports & Exercise2001333468-475MAR 2001Validity of a physical activity questionnaire among African-American Seventh-day AdventistsSINGH, PN; FRASER, GE; KNUTSEN, SF; LINDSTED, KD; BENNETT, HW 2001The effect of menopause on the relation between weight gain and mortality among womenSingh, PN; Haddad, E; Knutsen, SF; Fraser, GE