Measuring physical activity with a single question.

Public Health Briefs

Jones RN: Absence of hyperresponsiveness to methacholine in a worker with methylene diphenyl diisocyanate (MDI)induced asthma. Chest 1986; 3:389-393. 23. Baur M, Dewair M, Rommelt H: Acute airway obstruction followed by hypersensitivity pneumonitis in an isocyanate (MDI) exposed worker. JOM 1984; 26:285-287. 24. Peters JM, Murphy RLH, Ferris BG: Ventilatory function in workers exposed to low levels of toluene diisocyanate. Br J Med 1969; 26:115-120. 25. Peters JM: Studies of isocyanate toxicity: Cumulative pulmonary effects in workers exposed to toluene diisocyanate. Proc R

............................

Soc Med 1970; 63:372-375. 26. Peters JM, Murphy LHM, Pagnatto LD, et ab Respiratory impairment in workers exposed to safe levels of toulene diisocyanate. Arch Environ Health 1970; 20:364367. 27. Wegman DH, Peters JM, Pagnatto LD, et at Chronic pulmonary function loss from exposure to toluene diisocyanate. Br J Ind Med 1977; 34:196-200. 28. Diem JE, Jones RN, Hendrick DJ, et at Five-year longitudinal study of workers employed in a new toluene diisocyanate manufacturing plant. Am Rev Respir Dis 1982; 126:420-428.

29. Musk WJ, Peters JM, Di Bernardinis L, Raymond RLH: Absence of respiratory effects in subjects exposed to low concentrations of TDI and MDI. JOM 1982; 24:746750. 30. Gee JB, Morgan WK: A 10-year follow-up study of a group of workers exposed to isocyanates. JOM 1985; 34:263-271. 31. Higgins MW, Keller JB: Seven measures of ventilatory lung function: Population values and a comparison of their ability to discriminate between persons with and without chronic respiratory symptoms and disease. Am Rev Respir Dis 1973; 108:258272.

Measuring Physical Activity With a Single Question

.................................

Kenneth B. Schechtman, PhD, Benico Bazlai, MD, KIthryn Rost, PhD, and Edwin B. Fisher, Jr, PhD

..

..

Inftduton

Methods

Through its association with low levels of HDL (high-density lipoprotein) cholesterol, obesity, and other physiologic parameters, physical inactivity has been linked to increased morbidity and mortality from a variety of diseases. These include cardiovasCUlar diseas,1 diabetes mellitus,2 os_ teoporosis,3 and even breast cancer4 and colon cancer5. In light of these etiologic associations, the measurement of physical activity has become a standard component of epidemiologic research, with at least 30 different measurement techniques having been desCrbed.6 Unfortunately, many of these are complex instruments that are associated with excessive costs and time commitments that make them impractical in large-scale epidemiologic studies.6'7 Thus, there is great value in establishing the validity of brief, easily obtained measures of physical activity. It is the purpose of the present report to discuss the utility of a single exercise question that was generated by the St. Louis Working Hearts Program. That question was "Do you currently participate in any regular activity or program (either on your own or in a formal class) designed to improve or maintain your physical fitness?" We will investigate the validity of this question by evaluating its association with three parameters that are known to be associated with physical activity: body mass index (BMI), HDL cholesterol, and oxygen capacity.

The data discussed herein were generated by Working Hearts, a four-year collaboration between Washington University in St. Louis and a large midwestern communications corporation. All employees at 17 divisions of the corporation were eligible to participate in a multifaceted two-year cardiovascular risk reduction intervention. The intervention included individualized counseling, worksite promotional activities, and 25 awareness seminars and workshops that focused on weight loss, cholesterol reduction, hypertension, exercise, stress reduction, diet, and smoking cessation. Of particular relevance is the fact that the exercise component ofthe intervention included discussions and workshops, but no regular exercise program. Data on three of the four variables pertinent to this report-exercise behavior, BMI, and HDL cholesterol levels-were sought from all study participants. BMI was deAddress reprint requests to Kenneth B. Schechtman, PhD, Division of Biostatistics, Washington University School of Medicine, 660 South Euclid Avenue, Box 8067, St. Louis, MO 63110. Dr. Barzilai is with the Division of Cardiology; Dr. Fisher is with the Center for Health Behavior Research, both at Washington University; Dr. Rost is with the Departnent of Psychiatry, University ofArkansas for Medical Sciences. This paper, submitted to the Journal August 2, 1990, was revised and accepted for publication January 23, 1991.

...............

American Journal of Public Health 771

Public Health Briefs

fined as weight (in kg) divided by height (in meters) squared. Oxygen capacity, in ml/ kg/min, was measured on a self-selected subset of 304 subjects using a sub-maximal bicycle test. Results, expressed as mean ± standard deviation, are analyzed separately for males and females. Analysis of covariance determined whether there were significant age-adjusted differences in the three relevant parameters between exercising and non-exercising subjects. To analyze changes over time, the 591 participants who provided both baseline and final exercise data were divided into four groups as follows: Group NN responded "no" at both times (N = 281, 47.5 percent); Group YY responded

772 American Journal of Public Health

"yes" at both times (N = 130, 22.0 percent); Group NY reported initiating exercise (N = 118, 20.0 percent); and Group YN reported cessation of regular exercise (N = 62, 10.5 percent).

ResuIts The Working Hearts data set contained 1,004 subjects whose age was 36.6 ± 9.5 years; 73.1 percent were female. Among the 986 subjects who provided baseline exercise information, 32 percent (N = 316) reported participation in regular exercise. Subjects who provided oxygen capacity data (N = 304) were younger (p = 0.001 in women, p = 0.012 in men), had

a lower baseline BMI (p = 0.036 in women, p = 0.049 in men), and were more likely to exercise at baseline among women (p = 0.008) but not among men (p = 0.265), than those not providing oxygen capacity data. However, after age adjustment, logistic regression found that BMI had no independent association with the availability of bicycle test data in either sex, but that baseline exercise was still significant in women (p = 0.030). The availability of bicycle test data was not associated with baseline HDL cholesterol in either sex. Table 1 provides the key data summarizing the validity of the simplified exercise measure we have utilized. It indicates that at baseline, and after age adjustment, subjects who said they were exercisers had a lower BMI (p < 0.0001 in women and p = 0.001 in men), greater oxygen capacity (p = 0.0007 in women and p = 0.002 in men), and had a higher HDL cholesterol (p < 0.0001 in women) than those who did not report participation in regular exercise. Using the four groups defined in the Methods section, Table 2 contains the mean change from baseline to final measurement in both BMI and HDL cholesterol. Oxygen capacity was not evaluated in this context because it was measured at only one time point. Following age adjustment, mean HDL cholesterol was not the same in all four groups of men (p = 0.007). Statistical contrasts show that this reflected an increase in HDL in group YY (2.2 ± 11.5, N = 58) that was significantly

June

1991, Vol. 81, No. 6

Public Health Briefs

different from the decrease in HDL cholesterol in group NN (-3.0 ± 6.4, p = 0.0006, N = 70). Although the overall ageadjusted ANOVA showed no difference in BMI among the four groups of women (p = 0.100), it is worth noting that the statistical contrast comparing the increase in BMI in group NN (0.59 + 1.9) with the small change in group NY (0.05 + 2.2) was significant (p = 0.028).

Discussion These results indicate that among Working Hearts subjects, a single selfreported measure of participation in regular exercise at baseline was significantly associated, after age adjustment, with body mass index and oxygen capacity in both sexes and with HDL cholesterol in women. Thus, our data support several other studies8-10 which have suggested that simple self-reported instruments can yield valid information about participation in regular exercise. Based on only two exercise questions, for example, the Lipid Research Clinics Study8 found that subjects who reported regular vigorous exercise had higher HDL cholesterol, lower resting heart rates, and longer exercise test durations than those not reporting reg-

June 1991, Vol. 81, No. 6

ular exercise. Siconolfi et al9 found that a single question concerning the self-reported frequency of exercise that induced sweatingwas associated with maximal oxygen uptake. If a prime focus of a study is to gain detailed and precise information about levels of physical activity, extensive exercise questionnaires are required. Our results suggest, despite some positive associations in Table 2, that simplified exercise measures may also be inadequate when changes over time are of primary interest. However, as with the reports noted above,"10 our data argue strongly that one or two self-administered questions can provide a valid measure ofwho is and who is not currently involved in regular exercise. Thus, when extensive questionnaires place great demands on participants in a multifaceted epidemiologic study, and when detailed information about exercise behavior is not essential, a few simple exercise questions may be an appropriate alternative to the complex instruments that are frequently utilized. []

References 1. Powell KE, Thompson PD, Caspersen CJ, et al: Physical activity and the incidence of coronary heart disease. Annu Rev Public Health 1987; 8:253-287.

2. Vranic M, Berger M: Exercise and diabetes mellitus. Diabetes 1979; 28:147-163. 3. Montoye HJ, McCabe JF, Metzner HL, et al: Physical activity and bone density. Human Biol 1976; 48:599-610. 4. Frisch RE, Wyshak G, Albright NL, et al: Lower prevalence of breast cancer and cancer of the reproductive system among former college athletes compared to nonathletes. Br J Cancer 1985; 52:885-891. 5. Gehardsson M, Norell SE, Kiviranta H, et al: Sedentary jobs and colon cancer. Am J Epidemiol 1986; 123:775-780. 6. Laporte RE, Montoye HJ, Caspersen CJ: Assessment of physical activity in epidemiologic research: Problems and prospects. Public Health Rep 1985; 100:131146. 7. Montoye RI, Taylor HL: Measurement of physical activity in population studies: A review. Human Biol 1984; 56:195-216. 8. Haskell WL, Taylor HL, Wood PD, Schrott H, Heiss G: Strenuous physical acitvity, treadmill exercise test performance and plasma high-density lipoprotein cholesterol: The Lipid Research Clinics Program Prevalence Study. Circulation (suppl IV) 1980; 62:IV53-IV61. 9. Siconolfi SF, Lasater TM, Snow RC, Carleton RA: Self-reported physical activity compared with maximal oxygen uptake. Am J Epidemiol 1985; 122:101-105. 10. Godin G, Jobin J, Bouillon J: Assessment of leisure time exercise behavior by selfreport: A concurrent validity study. Can J Public Health 1986; 77:359-362.

American Journal of Public Health 773