## Abstract

This study described differences in vigorous activity participation recalled across the life span, assessed whether reports of past vigorous activity were associated with current participation, and examined factors associated with participation in current vigorous activity among women. After the exclusion of women aged 50–54 years, the study population included 71,837 multiethnic postmenopausal women aged 55–79 years who were participating in the Women’s Health Initiative Observational Cohort Study, 1993–1998. Vigorous activity was assessed retrospectively for ages 18, 35, and 50 years and currently at enrollment into the study (median age, 65 years). Current participation in vigorous activity (>3 days/week) was low and consistent across racial/ethnic groups (13–16%). The prevalence of vigorous activity declined with age, with the largest decrease in vigorous activity occurring after age 50 years for all racial/ethnic groups. Current vigorous activity was generally higher among women with a lower body mass index, not currently smoking, in excellent general health, and of higher socioeconomic status across racial/ethnic groups. These data suggest that a lower prevalence of vigorous activity in the postmenopausal period is part of a complex of health-related attitudes and behaviors that transcends race/ethnicity. The perimenopausal period may be a critical juncture at which targeted and tailored interventions may help to achieve maintenance of physical activity into the postmenopausal period.

Received for publication February 6, 2002; accepted for publication June 12, 2002.

Physical activity is an important health behavior that may change across the life course (1). In general, leisure activity prevalence is negatively associated with age for both women and men (24), with the prevalence of no leisure activity highest in older age groups (5). Cross-sectional data indicate that vigorous and overall physical activity may decline at a faster rate than moderate physical activity (6). The reasons for this age-related decline are not well understood (5).

Few studies have examined participation in vigorous activity across multiple time points in adulthood among women. According to the developmental approach to intervention, life stages and transitions may be effective points to intervene and promote preventive health behavior (7). Determining when lasting habits of physical activity occur in the life course can help to guide public health intervention efforts aimed at increasing population levels of physical activity. Intervention efforts can then be focused more intensively during life periods when favorable changes in physical activity are more likely to occur, thereby potentially increasing success and cost effectiveness.

Therefore, the objectives of this study were the following: 1) to describe differences in vigorous physical activity participation at various times recalled in the life span, 2) to assess whether reports of past participation in vigorous physical activity were associated with self-reported current physical activity participation, and 3) to determine factors related to participation in vigorous physical activity by race/ethnicity among women. Although retrospective reporting has limitations relative to prospective reporting, retrospective reporting has been shown to contribute useful information in this field, particularly when vigorous forms of leisure activity are targeted (810). Moreover, few data are available using either retrospective or prospective methods, particularly for this age group of women and by different racial/ethnic groups. These objectives were examined among a cohort of women participating in the Women’s Health Initiative Observational Cohort Study.

## MATERIALS AND METHODS

The Women’s Health Initiative began in 1992 to investigate the effects of selected interventions and risk factors on the morbidity and mortality of cancer, cardiovascular disease, and osteoporotic fractures among postmenopausal women aged 50–79 years. Women who gave written, informed consent were recruited into the Women’s Health Initiative starting in 1993 at 40 primary clinical centers and nine satellite centers in the United States, mostly through mass mailings to age-eligible women from large mailing lists, such as driver’s license, voter registration, and Health Care Financing Administration or other insurance lists. Women were either specifically recruited for the observational cohort study or entered it because they were ineligible or unwilling to be randomized into the Women’s Health Initiative clinical trial. The results presented here are for the observational cohort only. The details of the Women’s Health Initiative design and recruitment of minorities and older women are reported elsewhere (11).

### Measurement of physical activity

Physical activity was self-reported on a questionnaire that was mailed to the participants prior to the clinic visit and completed at home or during the clinic visit or that was completed at home after the visit and returned by mail. Current vigorous activity was assessed by the question, “Not including walking outside the home, how often each week (7 days) do you usually do the exercises below: strenuous or very hard exercise (you work up a sweat and your heart beats fast); for example, aerobics, aerobic dancing, jogging, tennis, swimming laps?” Consistent with vigorous physical activity recommendations (12), responses were grouped as 3 or more days per week compared with less than 3 days per week. Current moderate activity was assessed by the question, “Not including walking outside the home, how often each week (7 days) do you usually do the exercises below: moderate exercise (not exhausting); for example, biking outdoors, using an exercise machine (like a stationary bike or treadmill), calisthenics, easy swimming, popular or folk dancing?” For consistency, we also defined moderate activity as 3 or more days per week compared with less than 3 days per week. Past vigorous activity was assessed by the question, “For each of the ages below, did you usually do strenuous or very hard exercises at least 3 times a week? This would include exercise that was long enough to work up a sweat and make your heart beat fast.” The ages ascertained were 18, 35, and 50 years and responses of yes or no were collected.

For a subsample of 536 Women’s Health Initiative participants, the measurement of vigorous physical activity at the ages of 18, 35, and 50 years and currently was again ascertained approximately 10 weeks after the first administration of the questionnaire. The reliability of the physical activity questions under study, as measured by the kappa coefficient, was 0.55 for 18 years, 0.55 for 35 years, 0.53 for 50 years, and 0.65 for current activity (J. Emily White, Stratton Veterans Affairs Medical Center, personal communication, December 20, 2000). These kappa coefficients can be interpreted as falling between agreement expected by chance (kappa = 0) and perfect agreement (kappa = 1) (13).

### Measurement of descriptive characteristics

Height and weight were measured during the clinic visit, and body mass index was calculated as weight (kg)/height (m)2. Information on baseline demographic characteristics (age, ethnicity, marital status), socioeconomic status (education, income, occupation), smoking, and current health status was also collected by self-administered questionnaires either mailed to the participants prior to the clinic visit and completed during the clinic visit or completed at home and returned by mail. Self-reported health status was measured with the question, “In general, would you say your health is excellent, very good, good, fair, or poor?” Smoking was categorized as never, former, or current smokers.

“Current resident” was defined according to the US Census classification of clinic site (14) as Northeast (Boston, Massachusetts; Bronx, New York; Buffalo, New York; East Setauket, New York; Newark, New Jersey; Pawtucket, Rhode Island; Pittsburgh, Pennsylvania; Worcester, Massachusetts), South (Birmingham, Alabama; Atlanta, Georgia; Durham, North Carolina; Gainesville, Florida; Houston, Texas; Memphis, Tennessee; Miami, Florida; San Antonio, Texas; Washington, DC (two sites); Winston-Salem, North Carolina), Midwest (Chicago, Illinois (two sites); Cincinnati, Ohio; Columbus, Ohio; Detroit, Michigan; Iowa City, Iowa; Madison, Wisconsin; Milwaukee, Wisconsin; Minneapolis, Minnesota), and West (Honolulu, Hawaii; Orange, California; LaJolla, California; Los Angeles, California; Oakland, California; Portland, Oregon; Reno, Nevada; San Jose, California; Seattle, Washington; Sacramento, California; Torrance, California; Tucson, Arizona).

### Statistical analysis

The Women’s Health Initiative observational cohort comprised 93,725 women. We excluded 12,388 participants aged 50–54 years, in order to allow at least 5 years to recall vigorous physical activity at age 50. Participants were excluded if they were missing information on vigorous activity at present or in the past, height, weight, self-reported health, marital status, or education (n = 8,511). Women who reported their race/ethnicity as “other” (n = 787) or “unspecified” (n = 202) were excluded from analysis because of lack of specific reporting of race/ethnicity.

The prevalence of participation in physical activity (currently and at ages 18, 35, and 50 years) was calculated as the percentage reporting such activity at the baseline cohort visit. Ninety-five percent confidence intervals were computed on the binomial probability distribution. Confidence intervals were not adjusted for multiple comparisons. Logistic regression models were used to examine factors that predicted participation in current vigorous physical activity. Each measure of prevalence or of association was reported by ethnicity and adjusted for age at baseline by the direct method (5-year intervals) using the age distribution of the entire study cohort as the standard population (15). Analyses were conducted using Statistical Analysis System software (version 6.12; SAS Institute, Inc., Cary, North Carolina).

## RESULTS

### Reported vigorous physical activity over the life span

Current participation in vigorous physical activity at least 3 days per week was low and consistent across racial/ethnic groups, ranging from 13.2 percent among Blacks to 16.2 percent among Asians/Pacific Islanders (table 1). The current age averaged 65.3 years across the five racial/ethnic groups, with the median age by race/ethnicity of 63 years for Blacks and Hispanics and 65 years for American Indians, Asians/Pacific Islanders, and Whites. Recalling participation in vigorous physical activity at the ages of 18, 35, and 50 years, Asians/Pacific Islanders reported the lowest prevalence among the five racial/ethnic groups but the highest prevalence at current age. Over half of the Black and American Indian participants and almost half of the Hispanic and White participants reported vigorous physical activity participation at age 18 years. The prevalence of vigorous activity declined monotonically with age (18, 30, and 50 years and current) for Black, Hispanic, White, and American Indian women but was higher than that of Asian/Pacific Islander women at each age category except at current age. For all groups, there was a large decrease in the prevalence of vigorous physical activity from age 50 years to current age.

All possible patterns of vigorous activity (response either yes or no) over the four age periods (18, 35, and 50 years and current) were examined. Women were categorized into 16 different patterns. The most common pattern across racial/ethnic groups was that of no vigorous activity at all time periods. The age-adjusted percentage of women reporting no vigorous physical activity at age 18, 35, and 50 years and currently was highest among Asian/Pacific Islander women (40.5 percent, 95 percent confidence interval (CI): 38.4, 42.7), followed by Whites (33.4 percent, 95 percent CI: 33.0, 33.7), Hispanics (32.3 percent, 95 percent CI: 30.3, 34.4), Blacks (28.4 percent, 95 percent CI: 27.2, 29.7), and American Indians (25.4 percent, 95 percent CI: 19.9, 30.9). Reporting vigorous activity at all four time points was uncommon, with the age-adjusted prevalence in order from highest to lowest as follows: 7.3 percent (95 percent CI: 3.9, 10.7) for American Indians, 4.6 percent (95 percent CI: 4.0, 5.2) for Hispanics, 4.4 percent (95 percent CI: 4.2, 4.6) for Whites, 4.3 percent (95 percent CI: 3.7, 4.9) for Blacks, and 3.3 percent (95 percent CI: 2.5, 4.1) for Asians/Pacific Islanders.

### Association between present and past vigorous leisure activity

Among women who reported vigorous activity at age 18 years (regardless of activity at other ages), 13.0–15.8 percent depending on race/ethnicity reported current vigorous activity (table 2). Vigorous activity at 50 years of age was more predictive of current vigorous activity than participation at age 18 or 35 years. Women who reported vigorous activity at age 50 years were 1.8–2.2 times more likely to report current vigorous activity than were women reporting vigorous activity at age 18 years and 1.5–1.7 times more likely to report current vigorous activity compared with reporting vigorous activity at age 35 years. Women who reported vigorous activity at age 35 years were 1.1–1.4 times more likely to report current vigorous activity compared with women reporting vigorous activity at age 18 years. Among women who reported vigorous physical activity at all past ages (18, 35, and 50 years), reporting current vigorous activity was highest for Asian/Pacific Islander women (27.7 percent) and lowest for Black women (18.1 percent). For women reporting no vigorous activity at age 18, 35, and 50 years, the reporting of current vigorous activity was consistently low across racial/ethnic groups (5.6–7.9 percent).

We hypothesized that past vigorous activity participation might be more predictive of current vigorous plus moderate activity participation, rather than just vigorous activity, which was confirmed (table 2). Vigorous plus moderate activity was defined as 3 or more days per week of activity compared with less than 3 days per week of either intensity of activity. Vigorous activity participation at 50 years of age was most predictive of current moderate plus vigorous activity participation across racial/ethnic groups.

Correlates of current vigorous physical activity participation are reported by race/ethnicity in table 3. The prevalence of vigorous activity generally declined monotonically with body mass index categories. The prevalence of vigorous activity generally declined monotonically with self-reported health categories, indicating that those reporting fair or poor health were much less active. Current vigorous activity was reported more often among women who did not presently smoke (never or former smokers) compared with those that were current smokers. Participation in current vigorous physical activity did not consistently vary with either current marital status or region of the country. For the socioeconomic indicators, participation in current vigorous physical activity was generally higher for women with at least a college education, family income of at least $75,000, and managerial or professional occupations across racial/ethnic groups. However, physical activity patterns within categories of education, income, and occupation were not always consistent. Age-adjusted and fully adjusted logistic models to predict current vigorous activity participation were also calculated, with current body mass index, self-reported health, smoking, marital status, geographic residence, and education in the model (data not shown). In most cases, further adjustment attenuated the odds ratios only slightly. However, education was meaningfully attenuated when adjusting for other co-variates in the model. ## DISCUSSION A recent review of the correlates of physical activity in ethnically diverse women concluded that very few data exist regarding physical activity in Hispanics, Asians/Pacific Islanders, and American Indians, and the review called for more research with these ethnic groups (16). The present study addresses this call and adds to the small body of literature currently available on the patterns of physical activity observed across the life span for women. Most of the previous studies of activity and race/ethnicity have, like the present study, been based on retrospective reports. Although this method is subject to recall bias, prospective collection of data across the four to six decades captured here is difficult to obtain prospectively because of logistic and cost constraints. Although stronger prospective designs are desirable, retrospective reports can nonetheless contribute useful information about physical activity patterns, particularly in understudied populations such as the different racial/ethnic groups of women focused upon in this report. ### Vigorous leisure activity over the life span The prevalence of current vigorous leisure activity ranged from 13 percent to 16 percent for women aged 55–79 years across race/ethnicity. This finding is consistent with the prevalence of regular vigorous leisure activity for women aged 45–64 years (17.7 percent) and aged 65–74 years (16.5 percent) reported by the 1992 Behavioral Risk Factor Surveillance System (5). This finding is also consistent with data for women 40 years or older participating in the US Women’s Determinants Study, in which the prevalence of vigorous activity ranged from 8 percent to 14 percent across race/ethnicity (17). Over half of Black and American Indian participants and almost half of Hispanic and White participants reported vigorous physical activity participation at age 18 years. This is in contrast to the prevalence of current regular vigorous leisure activity data for women aged 18–29 years (11.4 percent) from the 1992 Behavioral Risk Factor Surveillance System data (5). However, these women should not be compared directly, because they come from a different birth cohort, and different methods of data collection were applied (e.g., retrospective vs. cross-sectional). The current results reflect the patterns of reduced vigorous activity over the course of the life span that have been reported in cross-sectional surveillance studies (5). Of note, Asian/Pacific Islander women reported the lowest percent vigorously active at ages 18, 35, and 50 years compared with women of a different race/ethnicity. Such results underscore the importance of continued epidemiologic study of racial/ethnic groups that comprise a growing segment of the US population, yet often have not been included in epidemiologic research of physical activity. For all racial/ethnic groups, the most precipitous drop in the percentage of those vigorously active occurred between the ages of 50 years and current age. For Asian/Pacific Islander women, the current prevalence was half of what was reported at age 50 years, and for Black, Hispanic, White, and American Indian women the prevalence was approximately one third of what is reported at age 50 years. Moreover, a large drop in the percentage of those vigorously active occurred between the ages of 35 and 50 years among the non-Asian women in this sample. Why these declines occurred is not explained by these data, but the fourth and fifth decades may represent a particularly important period for further study and intervention. ### Association between present and past vigorous leisure activity Tracking is defined as the ability to predict a subsequent observation from an earlier observation or the ability to maintain over time one’s relative rank of an observation among a group of peers (18). One of the most consistent predictors of current behavior is past behavior. A study of both men and women has concluded that competitive sport participation as early as 10–19 years of age is a significant predictor of physical activity at the current age of 65–84 years (19). In the Black Women’s Health Study, a cohort of women aged 20–69 years, participation in vigorous activity (>2 hours/week) in high school was significantly related to current leisure walking (above the median), moderate activity, and vigorous activity (20). Furthermore, in this study, vigorous activity (or vigorous plus moderate activity) at 50 years of age was more predictive of current vigorous activity than was participation at age 18 or 35 years. Although the potentially greater accuracy of reporting activity in more recent years cannot be ruled out, it is also possible that participation in vigorous activity in the middle years sets the stage for continued participation heading into the older years of life. The reporting of current vigorous activity according to past participation in activity was clearest among Asian/Pacific Islander women, indicating that physical activity tracked more strongly in this group. However, our data indicate that few women (3–5 percent across race/ethnicity) sustained regular vigorous activity across the life span, as indicated by the questions at age 18, 35, and 50 years and currently. Moreover, of any of the 16 potential patterns of activity, the most common pattern reported was no regular vigorous activity across these same time periods. The prevalence of this pattern ranged from 28 percent for Black women to 41 percent for Asian/Pacific Islander women, indicating strong tracking of not being vigorously active. Since past vigorous activity predicted a present lack of vigorous activity, research is needed on understanding and developing interventions to encourage young adults to become more active and stay active throughout the life course. ### Correlates to current vigorous leisure activity participation This study also examined correlates of current vigorous leisure activity. Better self-reported current health was positively associated with reporting current vigorous activity. A monotonic relation for Asian/Pacific Islander, Black, Hispanic, and White women indicated that approximately 4–6 percent of the women reporting poor health currently engaged in vigorous activity compared with about 24–27 percent of participants reporting excellent health. Body mass index also had a clear inverse relation with vigorous activity, declining monotonically with increasing body mass index categories, except for Hispanics. Overall, about 16–26 percent of nonobese or nonoverweight subjects (body mass index, <25 kg/m2) reported current vigorous activity as compared with 2–11 percent of obese participants (body mass index, >35 kg/m2). Across all racial/ethnic groups, women who reported current smoking were less likely to report current vigorous activity than were women who were not current smokers. Status indicators, such as marital, regional, and socioeconomic variables, had more complex relations with racial/ethnic group and vigorous activity. For example, the lowest proportion of vigorous activity was reported in “never married” Asian/Pacific Islander and White women, but the prevalence of vigorous activity was highest among “never married” Black women. Analogous relations were observed for the region of the country in which the participants resided. The highest prevalence for vigorous activity was reported in western clinics for Blacks and Whites, midwestern clinics for Asians/Pacific Islanders and Hispanics, and northeastern clinics for American Indians. Differences in residence may reflect differences in the proportion of different subgroups within a particular racial/ethnic group (e.g., more Mexican Americans in the West, Puerto Rican Americans in the Northeast, and Cuban Americans in the Southeast). The association of socioeconomic level and vigorous activity was more consistent, with a positive relation noted with both educational level and family income across racial/ethnic groups. In addition, more women in higher income occupations (e.g., managerial or professional) reported engaging in vigorous activity. However, in Asian/Pacific Islander women, the second highest income category had the lowest proportion of women reporting vigorous activity. These results support those of other studies (16), showing positive associations of health, leanness, and higher socioeconomic status with vigorous activity. They do not, however, reveal causal relations. Inactivity may produce and/or result in poorer health and obesity. Higher socioeconomic status may afford both the opportunity for exercise and access to better health care, both of which could affect health status. Long-term prospective studies with frequent measurements are the best vehicles for unraveling and defining these effects. ### Strengths and limitations The Women’s Health Initiative provides an opportunity to examine adult life course physical activity in a large, multiethnic cohort of women. This is one of the first studies that reports physical activity levels among Asian/Pacific Islander women living in the United States, and this study also contributes to a small but growing literature on physical activity among Hispanic and American Indian women. However, this study has several limitations that should be recognized. Although we were able to report all associations by race/ethnicity, the sample size for American Indian women was low. The Women’s Health Initiative is composed of volunteers, so the generalizability of these data is limited because of self-selection. In addition, some women were included because they did not meet the health requirements for the Women’s Health Initiative clinical trials. The measurement of vigorous leisure activity consisted of only one question for each of the time periods. The question did not ascertain the duration of activity or the specific types of leisure activities. The women were also not queried about other types of vigorous activity, such as occupational, transportation, or household activities. Women recalled vigorous physical activity participation at several ages (18, 35, and 50 years) and, because this information was not gathered prospectively, it is, as noted earlier, subject to recall bias. Differences in activity levels may reflect cultural or cohort effects. The accuracy of self-reported physical activity might also vary by characteristics, such as age, race/ethnicity, education, body mass index, and health status, although a recent article indicates that this may not be the case (21). It is also possible that reporting current activity or reporting moderate and mild activity may have affected the reporting of past activity. Finally, the kappa coefficients for the measurement of vigorous physical activity ranged from 0.53 to 0.55 for past recalled activity, but they were higher for current activity (kappa = 0.65), as expected. This reliability is consistent with evidence from other studies on the recall of physical activity for earlier ages. In a study conducted in Sydney, Australia, participants 65–95 years of age recalled their physical activity at the ages of 20 and 50 years (22). The reliability of recalling the main sport or exercise participated in at these ages (categorized as none, light, moderate, or vigorous activity) was 0.65 at age 20 years and 0.61 at age 50 years, as assessed by weighted kappa statistics taken 1–3 months apart. The literature also indicates that vigorous leisure activity has generally been easier to recall relative to other forms of leisure time activity (e.g., moderate, light, or intermittent) (810, 2325). Of note, in a prospective study, the measurement of intraclass correlation coefficients ranged from 0.36 to 0.38 for recalled hours per week of vigorous activity reported 32–36 years earlier (24). ### Conclusions Physical activity studied within the same cohort of individuals can provide clues as to why people change their physical activity patterns and help to discern how sustained physical activity behaviors are developed (26). These patterns can also help to target population groups more likely to benefit and adhere to an interventional program (6). These data suggest that a lower prevalence of vigorous activity in the postmenopausal period is part of a complex of health-related attitudes and behaviors that transcend race and ethnicity. These findings suggest that the perimenopausal period may be a critical time at which targeted and tailored interventions may help to achieve maintenance of physical activity patterns from earlier stages of the life course, in order to achieve the benefits of physical activity in the postmenopausal period. ## ACKNOWLEDGMENTS This study was supported by grant N01-WH-4-2119 from the National Institutes of Health, Department of Health and Human Services. The authors wish to acknowledge all the Women’s Health Initiative centers and their principal investigators for their participation in this research. The following is a short list of Women’s Health Initiative investigators: Program Office:National Heart, Lung, and Blood Institute, Bethesda, Maryland: Jacques E. Rossouw, Linda Pottern, Shari Ludlam, Joan McGowan, and Nancy Morris; Clinical Coordinating Centers:Fred Hutchinson Cancer Research Center, Seattle, Washington: Ross Prentice, Garnet Anderson, Andrea LaCroix, Ruth Patterson, and Anne McTiernan; Bowman Gray School of Medicine, Winston-Salem, North Carolina: Sally Shumaker and Pentti Rautaharju; Medical Research Laboratories, Highland Heights, Kentucky: Evan Stein; University of California at San Francisco, San Francisco, California: Steven Cummings; University of Minnesota, Minneapolis, Minnesota: John Himes; University of Washington, Seattle, Washington: Susan Heckbert; Clinical Centers:Albert Einstein College of Medicine, Bronx, New York: Sylvia Wassertheil-Smoller; Baylor College of Medicine, Houston, Texas: Jennifer Hays; Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts: JoAnn Manson; Brown University, Providence, Rhode Island: Annlouise R. Assaf; Emory University, Atlanta, Georgia: Lawrence Phillips; Fred Hutchinson Cancer Research Center, Seattle, Washington: Shirley Beresford; George Washington University Medical Center, Washington, DC: Judith Hsia; Harbor-UCLA Research and Education Institute, Torrance, California: Rowan Chlebowski; Kaiser Permanente Center for Health Research, Portland, Oregon: Cheryl Ritenbaugh; Kaiser Permanente Division of Research, Oakland, California: Bette Caan; Medical College of Wisconsin, Milwaukee, Wisconsin: Jane Morley Kotchen; Medstar Research Institute, Washington, DC: Barbara V. Howard; Northwestern University, Chicago/Evanston, Illinois: Linda Van Horn; Rush-Presbyterian St. Luke’s Medical Center, Chicago, Illinois: Henry Black; Stanford Center for Research in Disease Prevention, Stanford University, Stanford, California: Marcia L. Stefanick; State University of New York at Stony Brook, Stony Brook, New York: Dorothy Lane; The Ohio State University, Columbus, Ohio: Rebecca Jackson; University of Alabama at Birmingham, Birmingham, Alabama: Cora Beth Lewis; University of Arizona, Tucson/Phoenix, Arizona: Tamsen Bassford; University at Buffalo, Buffalo, New York: Maurizio Trevisan; University of California at Davis, Sacramento, California: John Robbins; University of California at Irvine, Orange, California: Allan Hubbell; University of California at Los Angeles, Los Angeles, California: Howard Judd; University of California at San Diego, LaJolla/Chula Vista, California: Robert D. Langer; University of Cincinnati, Cincinnati, Ohio: Margery Gass; University of Florida, Gainesville/Jacksonville, Florida: Marian Limacher; University of Hawaii, Honolulu, Hawaii: David Curb; University of Iowa, Iowa City/Davenport, Iowa: Robert Wallace; University of Massachusetts, Worcester, Massachusetts: Judith Ockene; University of Medicine and Dentistry of New Jersey, Newark, New Jersey: Norman Lasser; University of Miami, Miami, Florida: Mary Jo O’Sullivan; University of Minnesota, Minneapolis, Minnesota: Karen Margolis; University of Nevada, Reno, Nevada: Robert Brunner; University of North Carolina, Chapel Hill, North Carolina: Gerardo Heiss; University of Pittsburgh, Pittsburgh, Pennsylvania: Lewis Kuller; University of Tennessee, Memphis, Tennessee: Karen C. Johnson; University of Texas Health Science Center, San Antonio, Texas: Robert Schenken; University of Wisconsin, Madison, Wisconsin: Catherine Allen; Wake Forest University School of Medicine, Winston-Salem, North Carolina: Gregory Burke; and Wayne State University School of Medicine/Hutzel Hospital, Detroit, Michigan: Susan Hendrix. Reprint requests to Dr. Kelly R. Evenson, Department of Epidemiology, University of North Carolina-Chapel Hill, 137 East Franklin Street, Suite 306, Chapel Hill, NC 27514 (e-mail: kelly_evenson@unc.edu). TABLE 1. Percentage of women who reported participation in vigorous activity at ages 18, 35, and 50 years and current age by race/ethnicity, adjusted for current age, Women’s Health Initiative Observational Cohort Study, 1993–1998  Age of vigorous activity participation* Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI† % 95% CI % 95% CI % 95% CI % 95% CI 18 years 35.4 33.3, 37.5 53.4 52.0, 54.8 48.0 45.8, 50.2 44.1 43.7, 44.5 56.3 50.1, 62.6 35 years 31.1 29.1, 33.1 50.2 48.8, 51.6 47.4 45.3, 49.6 42.3 41.9, 42.7 54.4 48.1, 60.7 50 years 32.0 30.0, 34.0 39.0 37.7, 40.4 37.8 35.7, 39.9 39.6 39.2, 40.0 45.5 39.2, 51.7 Current age 16.2 14.6, 17.8 13.2 12.2, 14.1 13.4 11.9, 14.9 14.8 14.5, 15.1 15.1 10.5, 19.7  Age of vigorous activity participation* Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI† % 95% CI % 95% CI % 95% CI % 95% CI 18 years 35.4 33.3, 37.5 53.4 52.0, 54.8 48.0 45.8, 50.2 44.1 43.7, 44.5 56.3 50.1, 62.6 35 years 31.1 29.1, 33.1 50.2 48.8, 51.6 47.4 45.3, 49.6 42.3 41.9, 42.7 54.4 48.1, 60.7 50 years 32.0 30.0, 34.0 39.0 37.7, 40.4 37.8 35.7, 39.9 39.6 39.2, 40.0 45.5 39.2, 51.7 Current age 16.2 14.6, 17.8 13.2 12.2, 14.1 13.4 11.9, 14.9 14.8 14.5, 15.1 15.1 10.5, 19.7 * These categories are not mutually exclusive. † CI, confidence interval. TABLE 2. Percentage of women who reported current participation in vigorous or vigorous plus moderate activity stratified by their past vigorous activity participation by race/ethnicity, adjusted for current age, Women’s Health Initiative Observational Cohort Study, 1993–1998  Age of past vigorous activity* and current activity Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI† % 95% CI % 95% CI % 95% CI % 95% CI 18 years Vigorous 15.8 13.1, 18.5 13.0 11.7, 14.2 14.6 12.3, 16.8 15.0 14.6, 15.4 14.8 8.7, 20.9 Vigorous + moderate 33.8 30.3, 37.3 26.5 24.9, 28.2 30.7 27.7, 33.6 35.0 34.5, 35.6 23.0 15.8, 30.2 35 years Vigorous 21.6 18.4, 24.9 14.8 13.4, 16.2 15.9 13.6, 18.3 17.2 16.7, 17.6 16.2 9.7, 22.6 Vigorous + moderate 39.2 35.3, 43.0 28.7 27.0, 30.5 32.4 29.5, 35.4 37.6 37.0, 38.2 23.7 16.3, 31.1 50 years Vigorous 534.2 30.5, 37.9 22.8 20.9, 24.7 25.7 22.5, 28.9 28.7 28.2, 29.3 27.5 19.0, 36.0 Vigorous + moderate 51.1 47.3, 55.0 36.7 34.5, 38.9 41.5 37.9, 45.0 48.8 48.1, 49.4 35.1 26.0, 44.2 18, 35, and 50 years Vigorous 27.7 21.9, 33.5 18.1 15.9, 20.3 21.4 17.4, 25.4 23.1 22.4, 23.9 25.1 14.7, 35.4 Vigorous + moderate 44.5 38.1, 50.9 31.5 28.9, 34.2 37.6 32.9, 42.3 43.0 42.1, 43.9 32.9 21.8, 44.0 None at 18, 35, and 50 years Vigorous 7.0 5.3, 8.7 6.2 5.0, 7.5 5.6 3.8, 7.4 5.8 5.4, 6.1 7.9 1.0, 14.9 Vigorous + moderate 22.8 20.1, 25.6 18.7 16.7, 20.7 15.0 12.3, 17.8 27.6 27.0, 28.2 26.1 15.1, 37.0  Age of past vigorous activity* and current activity Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI† % 95% CI % 95% CI % 95% CI % 95% CI 18 years Vigorous 15.8 13.1, 18.5 13.0 11.7, 14.2 14.6 12.3, 16.8 15.0 14.6, 15.4 14.8 8.7, 20.9 Vigorous + moderate 33.8 30.3, 37.3 26.5 24.9, 28.2 30.7 27.7, 33.6 35.0 34.5, 35.6 23.0 15.8, 30.2 35 years Vigorous 21.6 18.4, 24.9 14.8 13.4, 16.2 15.9 13.6, 18.3 17.2 16.7, 17.6 16.2 9.7, 22.6 Vigorous + moderate 39.2 35.3, 43.0 28.7 27.0, 30.5 32.4 29.5, 35.4 37.6 37.0, 38.2 23.7 16.3, 31.1 50 years Vigorous 534.2 30.5, 37.9 22.8 20.9, 24.7 25.7 22.5, 28.9 28.7 28.2, 29.3 27.5 19.0, 36.0 Vigorous + moderate 51.1 47.3, 55.0 36.7 34.5, 38.9 41.5 37.9, 45.0 48.8 48.1, 49.4 35.1 26.0, 44.2 18, 35, and 50 years Vigorous 27.7 21.9, 33.5 18.1 15.9, 20.3 21.4 17.4, 25.4 23.1 22.4, 23.9 25.1 14.7, 35.4 Vigorous + moderate 44.5 38.1, 50.9 31.5 28.9, 34.2 37.6 32.9, 42.3 43.0 42.1, 43.9 32.9 21.8, 44.0 None at 18, 35, and 50 years Vigorous 7.0 5.3, 8.7 6.2 5.0, 7.5 5.6 3.8, 7.4 5.8 5.4, 6.1 7.9 1.0, 14.9 Vigorous + moderate 22.8 20.1, 25.6 18.7 16.7, 20.7 15.0 12.3, 17.8 27.6 27.0, 28.2 26.1 15.1, 37.0 * These categories are not mutually exclusive. † CI, confidence interval. TABLE 3. Percentage of women who reported that they currently engage in vigorous activity by sociodemographic and health characteristics and race/ethnicity, adjusted for current age, Women’s Health Initiative Observational Cohort Study, 1993–1998  Sociodemographic and health characteristics Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI* % 95% CI % 95% CI % 95% CI % 95% CI Current BMI* (kg/m2) <25.0 18.1 16.0, 20.2 16.4 14.1, 18.7 19.2 16.0, 22.5 18.3 17.8, 18.7 26.3 15.2, 37.3 25.0–<30.0 13.9 11.1, 16.7 14.2 12.6, 15.9 12.1 9.8, 14.5 14.1 13.7, 14.6 14.4 6.5, 22.3 30.0–<35.0 12.2 5.8, 18.5 11.1 9.3, 12.9 9.3 6.5, 12.2 10.4 9.8, 11.0 9.3 1.6, 17.1 ≥35.0 5.0 0, 12.5 10.9 9.0, 12.9 10.1 6.0, 14.2 7.8 7.1, 8.6 2.1 0, 8.5 Current self-reported health Excellent 26.9 21.2, 32.6 26.4 21.7, 31.2 23.9 18.1, 29.6 24.1 23.3, 24.9 24.1 9.8, 38.4 Very good 18.7 16.0, 21.4 16.1 14.2, 18.1 19.2 16.0, 22.3 15.3 14.9, 15.8 17.9 7.7, 28.1 Good 10.9 8.6, 13.1 11.9 10.5, 13.2 9.6 7.4, 11.7 10.0 9.6, 10.5 9.6 3.2, 16.1 Fair or poor 9.8 5.5, 14.1 7.9 6.2, 9.5 6.4 4.0, 8.8 7.2 6.4, 7.9 11.0 1.9, 20.1 Current smoking† Never 15.9 14.1, 17.8 12.6 11.3, 14.0 12.0 10.2, 13.8 13.8 13.4, 14.2 19.3 12.1, 26.5 Former 18.0 14.5, 21.5 14.5 13.0, 16.1 17.0 14.0, 20.1 16.5 16.1, 16.9 12.4 5.3, 19.6 Current 7.5 0.5, 14.4 8.6 6.0, 11.1 10.1 4.6, 15.6 9.9 8.9, 10.9 10.7 0, 24.3 Current marital status Never married 14.1 7.2, 21.0 15.1 10.8, 19.4 9.2 3.3, 15.1 12.2 11.0, 13.5 17.0 0, 43.7 Divorced/separated 17.4 12.2, 22.7 13.1 11.3, 14.8 13.8 10.5, 17.1 14.2 13.4, 14.9 17.5 6.4, 28.5 Widowed 19.5 15.3, 23.7 13.8 11.9, 15.7 14.2 10.5, 17.9 12.8 12.2, 13.4 9.3 1.5, 17.0 Married 15.7 13.8, 17.7 12.5 11.0, 14.0 14.1 12.0, 16.2 15.4 15.1, 15.8 17.0 9.8, 24.2 Marriage-like relationship 31.2 3.4, 59.1 13.2 2.9, 23.4 6.4 0, 15.1 18.4 15.8, 21.1 5.4 0, 27.3 Current geographic residence Northeast 18.1 10.0, 26.2 12.4 10.3, 14.5 13.1 9.3, 16.9 13.5 13.0, 14.0 25.8 11.2, 40.4 South 20.0 11.8, 28.2 12.0 10.6, 13.4 11.5 9.0, 13.9 15.2 14.6, 15.8 14.0 4.1, 23.9 Midwest 21.1 11.6, 30.6 13.2 11.3, 15.1 18.6 9.8, 27.4 12.9 12.4, 13.4 13.6 0, 28.3 West 15.5 13.9, 17.2 18.2 15.1, 21.3 14.6 12.3, 16.8 17.2 16.7, 17.8 11.4 5.6, 17.2 Education Less than high school 12.7 6.9, 18.6 9.7 7.4, 12.1 7.7 5.3, 10.1 10.4 9.0, 11.7 8.1 0.2, 16.0 High school/GED* 12.8 9.2, 16.3 10.5 8.1, 12.8 13.4 9.7, 17.2 9.7 9.2, 10.3 8.6 0, 18.0 Some after high school 16.1 13.3, 18.9 14.7 13.0, 16.3 15.0 12.4, 17.7 14.0 13.5, 14.4 19.8 11.6, 28.0 College degree or higher 17.5 15.0, 20.1 13.9 12.3, 15.5 17.2 13.8, 20.7 17.8 17.4, 18.3 20.1 8.4, 31.8 Current family income ($)† ≤19,999 13.6 9.3, 17.8 11.6 9.9, 13.2 10.3 8.0, 12.7 10.5 9.8, 11.1 11.1 4.2, 17.9 20,000–34,999 16.1 12.4, 19.7 12.0 10.1, 13.9 13.4 10.3, 16.6 11.9 11.4, 12.4 17.2 7.4, 27.1 35,000–49,999 16.8 12.9, 20.7 15.5 13.0, 18.1 16.1 12.0, 20.3 14.1 13.5, 14.7 18.6 3.8, 33.4 50,000-74,999 12.7 9.4, 15.9 13.5 11.0, 16.0 14.9 9.8, 19.9 16.4 15.7, 17.0 18.0 2.6, 33.4 ≥75,000 19.5 15.7, 23.4 16.8 13.2, 20.4 18.4 12.1, 24.7 20.3 19.6, 21.1 23.1 0.1, 46.0 Current occupation† Managerial, professional 16.9 14.3, 19.5 14.2 12.6, 15.7 18.3 14.9, 21.7 16.9 16.4, 17.3 17.8 8.2, 27.3 Technical, sales, administrative 14.9 12.2, 17.7 13.5 11.5, 15.6 12.6 9.6, 15.5 12.0 11.5, 12.5 11.8 3.1, 20.4 Service, labor 14.6 11.0, 18.2 11.2 9.3, 13.0 12.9 9.9, 15.8 13.7 13.0, 14.4 17.2 7.5, 26.9 Homemaker 16.1 10.0, 22.2 9.4 6.3, 12.4 9.7 6.6, 12.8 15.5 14.6, 16.4 13.5 1.8, 25.1
 Sociodemographic and health characteristics Asian/Pacific Islander(n = 2,098) Black(n = 5,056) Hispanic(n = 2,066) White(n = 62,360) American Indian(n = 257) % 95% CI* % 95% CI % 95% CI % 95% CI % 95% CI Current BMI* (kg/m2) <25.0 18.1 16.0, 20.2 16.4 14.1, 18.7 19.2 16.0, 22.5 18.3 17.8, 18.7 26.3 15.2, 37.3 25.0–<30.0 13.9 11.1, 16.7 14.2 12.6, 15.9 12.1 9.8, 14.5 14.1 13.7, 14.6 14.4 6.5, 22.3 30.0–<35.0 12.2 5.8, 18.5 11.1 9.3, 12.9 9.3 6.5, 12.2 10.4 9.8, 11.0 9.3 1.6, 17.1 ≥35.0 5.0 0, 12.5 10.9 9.0, 12.9 10.1 6.0, 14.2 7.8 7.1, 8.6 2.1 0, 8.5 Current self-reported health Excellent 26.9 21.2, 32.6 26.4 21.7, 31.2 23.9 18.1, 29.6 24.1 23.3, 24.9 24.1 9.8, 38.4 Very good 18.7 16.0, 21.4 16.1 14.2, 18.1 19.2 16.0, 22.3 15.3 14.9, 15.8 17.9 7.7, 28.1 Good 10.9 8.6, 13.1 11.9 10.5, 13.2 9.6 7.4, 11.7 10.0 9.6, 10.5 9.6 3.2, 16.1 Fair or poor 9.8 5.5, 14.1 7.9 6.2, 9.5 6.4 4.0, 8.8 7.2 6.4, 7.9 11.0 1.9, 20.1 Current smoking† Never 15.9 14.1, 17.8 12.6 11.3, 14.0 12.0 10.2, 13.8 13.8 13.4, 14.2 19.3 12.1, 26.5 Former 18.0 14.5, 21.5 14.5 13.0, 16.1 17.0 14.0, 20.1 16.5 16.1, 16.9 12.4 5.3, 19.6 Current 7.5 0.5, 14.4 8.6 6.0, 11.1 10.1 4.6, 15.6 9.9 8.9, 10.9 10.7 0, 24.3 Current marital status Never married 14.1 7.2, 21.0 15.1 10.8, 19.4 9.2 3.3, 15.1 12.2 11.0, 13.5 17.0 0, 43.7 Divorced/separated 17.4 12.2, 22.7 13.1 11.3, 14.8 13.8 10.5, 17.1 14.2 13.4, 14.9 17.5 6.4, 28.5 Widowed 19.5 15.3, 23.7 13.8 11.9, 15.7 14.2 10.5, 17.9 12.8 12.2, 13.4 9.3 1.5, 17.0 Married 15.7 13.8, 17.7 12.5 11.0, 14.0 14.1 12.0, 16.2 15.4 15.1, 15.8 17.0 9.8, 24.2 Marriage-like relationship 31.2 3.4, 59.1 13.2 2.9, 23.4 6.4 0, 15.1 18.4 15.8, 21.1 5.4 0, 27.3 Current geographic residence Northeast 18.1 10.0, 26.2 12.4 10.3, 14.5 13.1 9.3, 16.9 13.5 13.0, 14.0 25.8 11.2, 40.4 South 20.0 11.8, 28.2 12.0 10.6, 13.4 11.5 9.0, 13.9 15.2 14.6, 15.8 14.0 4.1, 23.9 Midwest 21.1 11.6, 30.6 13.2 11.3, 15.1 18.6 9.8, 27.4 12.9 12.4, 13.4 13.6 0, 28.3 West 15.5 13.9, 17.2 18.2 15.1, 21.3 14.6 12.3, 16.8 17.2 16.7, 17.8 11.4 5.6, 17.2 Education Less than high school 12.7 6.9, 18.6 9.7 7.4, 12.1 7.7 5.3, 10.1 10.4 9.0, 11.7 8.1 0.2, 16.0 High school/GED* 12.8 9.2, 16.3 10.5 8.1, 12.8 13.4 9.7, 17.2 9.7 9.2, 10.3 8.6 0, 18.0 Some after high school 16.1 13.3, 18.9 14.7 13.0, 16.3 15.0 12.4, 17.7 14.0 13.5, 14.4 19.8 11.6, 28.0 College degree or higher 17.5 15.0, 20.1 13.9 12.3, 15.5 17.2 13.8, 20.7 17.8 17.4, 18.3 20.1 8.4, 31.8 Current family income (\$)† ≤19,999 13.6 9.3, 17.8 11.6 9.9, 13.2 10.3 8.0, 12.7 10.5 9.8, 11.1 11.1 4.2, 17.9 20,000–34,999 16.1 12.4, 19.7 12.0 10.1, 13.9 13.4 10.3, 16.6 11.9 11.4, 12.4 17.2 7.4, 27.1 35,000–49,999 16.8 12.9, 20.7 15.5 13.0, 18.1 16.1 12.0, 20.3 14.1 13.5, 14.7 18.6 3.8, 33.4 50,000-74,999 12.7 9.4, 15.9 13.5 11.0, 16.0 14.9 9.8, 19.9 16.4 15.7, 17.0 18.0 2.6, 33.4 ≥75,000 19.5 15.7, 23.4 16.8 13.2, 20.4 18.4 12.1, 24.7 20.3 19.6, 21.1 23.1 0.1, 46.0 Current occupation† Managerial, professional 16.9 14.3, 19.5 14.2 12.6, 15.7 18.3 14.9, 21.7 16.9 16.4, 17.3 17.8 8.2, 27.3 Technical, sales, administrative 14.9 12.2, 17.7 13.5 11.5, 15.6 12.6 9.6, 15.5 12.0 11.5, 12.5 11.8 3.1, 20.4 Service, labor 14.6 11.0, 18.2 11.2 9.3, 13.0 12.9 9.9, 15.8 13.7 13.0, 14.4 17.2 7.5, 26.9 Homemaker 16.1 10.0, 22.2 9.4 6.3, 12.4 9.7 6.6, 12.8 15.5 14.6, 16.4 13.5 1.8, 25.1

* CI, confidence interval; BMI, body mass index; GED, general equivalency diploma.

† Sample sizes are reduced for these variables.

## References

1.
Dishman R, Dunn A. Exercise adherence in children and youth: implications for adulthood. Champaign, IL: Human Kinetics Books, 1988.
2.
Crespo C, Keteyian S, Heath G, et al. Leisure-time physical activity among US adults.
Arch Intern Med

1996
;
156
:
93
–8.
3.
Stephens T, Jacobs D, White C. A descriptive epidemiology of leisure-time physical activity.
Public Health Rep

1985
;
100
:
147
–58.
4.
Stephens T, Caspersen C. The demography of physical activity. In: Bouchard C, Shephard R, Stephens T, eds. Physical activity, fitness, and health. Champaign, IL: Human Kinetics Publishers, 1994:204–13.
5.
US Department of Health and Human Services. Physical activity and health: a report of the Surgeon General. Atlanta, GA: National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, US Department of Health and Human Services, 1996.
6.
Dishman R, Sallis J. Determinants and interventions for physical activity and health. In: Bouchard C, Shephard R, Stephens T, eds. Physical activity, fitness, and health. Champaign, IL: Human Kinetics Publishers, 1994.
214
–38.
7.
Felner R, Farber S, Primavera J. Transitions and stressful life events: a model for primary prevention. In: Felner R, Jason L, Moritsugu J, et al, eds. Preventive psychology: theory, research and practice. New York, NY: Pergamon Press, 1983:199–215.
8.
Albanes D, Conway H, Taylor P, et al. Validation and comparison of eight physical activity questionnaires.
Epidemiology

1990
;
1
:
65
–71.
9.
Blair S, Dowda M, Pate R, et al. Reliability of long-term recall of participation in physical activity by middle-aged men and women.
Am J Epidemiol

1991
;
133
:
266
–75.
10.
Slattery M, Jacobs D. Assessment of ability to recall physical activity of several years ago.
Ann Epidemiol

1995
;
5
:
292
–6.
11.
Design of the Women’s Health Initiative clinical trial and observational study. The Women’s Health Initiative Study Group.
Control Clin Trials

1998
;
19
:
61
–109.
12
American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. Philadelphia, PA: Lippincott, Williams, and Wilkins, 2000.
13.
Armstrong B, White E, Saracci R. Principles of exposure measurement in epidemiology. Oxford, United Kingdom: Oxford University Press, 1994.
14.
US Bureau of the Census. Statistical abstract of the United States. 117th ed. Washington, DC: Economics and Statistics Administration, US Department of Commerce, 1997.
15.
Selvin S. Statistical analysis of epidemiologic data. New York, NY: Oxford University Press, 1996.
16.
Eyler A, Wilcox S, Matson-Koffman D, et al. Correlates of physical activity among women from diverse racial/ethnic groups.
J Womens Health Gend Based Med

2002
;
11
:
239
–53.
17.
Brownson R, Eyler A, King A, et al. Patterns and correlates of physical activity among women aged 40 years and older, United States.
Am J Public Health

2000
;
90
:
264
–70.
18.
Foulkes M, Davis C. An index of tracking for longitudinal data.
Biometrics

1981
;
37
:
439
–46.
19.
Hirvensalo M, Lintunen T, Rantanen T. The continuity of physical activity—a retrospective and prospective study among older people.
Scand J Med Sci Sports

2000
;
10
:
37
–41.
20.
Adams-Campbell L, Rosenberg L, Washburn R, et al. Descriptive epidemiology of physical activity in African-American women.
Prev Med

2000
;
30
:
43
–50.
21.
Falkner K, McCann S, Trevisan M. Participant characteristics and quality of recall of physical activity in the distant past.
Am J Epidemiol

2001
;
154
:
865
–72.
22.
Cumming R, Klineberg R. A study of the reproducibility of long-term recall in the elderly.
Epidemiology

1994
;
5
:
116
–19.
23.
Lee M, Whittemore A, Lung D. Reliability of recalled physical activity, cigarette smoking, and alcohol consumption.
Ann Epidemiol

1992
;
2
:
705
–14.
24.
Falkner K, Trevisan M, McCann S. Reliability of recall of physical activity in the distant past.
Am J Epidemiol

1999
;
150
:
195
–205.
25.
Chasan-Taber L, Erickson J, McBride J, et al. Reproducibility of a self-administered lifetime physical activity questionnaire among female college alumnae.
Am J Epidemiol

2002
;
155
:
282
–9.
26.
Anderssen N, Jacobs D 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
:
351
–62.