Higher Direct Medical Costs Associated With Physical Inactivity

Michael Pratt, MD, MPH; Caroline A. Macera, PhD; Guijing Wang, PhD

THE PHYSICIAN AND SPORTSMEDICINE - VOL 28 - NO. 10 - OCTOBER 2000

ABSTRACT

BACKGROUND: The benefits of physical activity in reducing morbidity and mortality are well-established, but the effect of physical inactivity on direct medical costs is less clear.

OBJECTIVE: To describe the direct medical expenditures associated with physical inactivity.

DESIGN: Cross-sectional stratified analysis of the 1987 National Medical Expenditures Survey that included US civilian men and nonpregnant women aged 15 and older who were not in institutions in 1987. Main outcome measure was direct medical costs.

RESULTS: For those 15 and older without physical limitations, the average annual direct medical costs were $1,019 for those who were regularly physically active and $1,349 for those who reported being inactive. The costs were lower for active persons among smokers ($1,079 vs $1,448) and nonsmokers ($953 vs $1,234) and were consistent across age-groups and by sex. Medical care use (hospitalizations, physician visits, and medications) was also lower for physically active people than for inactive people.

CONCLUSION: The mean net annual benefit of physical activity was $330 per person in 1987 dollars. Our results suggest that increasing participation in regular moderate physical activity among the more than 88 million inactive Americans over the age of 15 might reduce annual national medical costs by as much as $29.2 billion in 1987 dollars—$76.6 billion in 2000 dollars.

The importance of physical activity in reducing morbidity and mortality from chronic diseases and conditions is well established (1,2). In 1990, 14% of all deaths in the United States were attributed to diet and activity patterns (3), while another study (4) linked sedentary lifestyles to 23% of deaths from the leading chronic diseases.

These data suggest that enormous cost savings to the medical care system could be realized if sedentary persons were to increase their level of physical activity. In fact, it has been suggested that between $4.3 billion and $5.6 billion would be saved annually if 10% of sedentary adults began a walking program (5). Colditz (6) has calculated the direct annual economic costs of physical inactivity in the United States to be $24 billion, or 2.4% of US healthcare expenditures. Another study (7), however, concluded that no net economic benefit results when both direct and indirect medical costs associated with exercise are considered, but that the estimated cost per quality-adjusted life year is favorable compared with other accepted medical and behavioral practices. All of these analyses were based on hypothetical cohorts and decision analysis and used assumptions for values for the costs of services.

A review article (8) published in 1994 on healthcare costs and benefits of physical activity concluded that savings are realized only among adults aged 45 or older, because the costs associated with injuries offset the health costs savings for younger adults. Once again, most of the studies reviewed used cost estimates and assumptions that may not be accurate because of a lack of quantifiable information available in this field.

We addressed these gaps in the literature by examining direct medical expenses of active and inactive men and women using actual medical expenditures of a nationally representative sample.

Methods

Data set. The 1987 National Medical Expenditures Survey (NMES), the most comprehensive healthcare information available, was used for this analysis. The survey, sponsored by the Agency for Health Care Policy and Research, was conducted face-to-face with participants from February 1987 to May 1988 (9,10). It was designed to provide nationally representative estimates of healthcare use and expenditures for US civilians not in institutions. The survey sample consisted of 35,000 persons in 14,000 households. The NMES used a complex, stratified, cluster design that oversampled older adults, low-income families, the disabled, and minority populations (11). All analyses in our study were carried out after weighting the sample to be representative of the US population (table 1).

TABLE 1. Sample Characteristics of Study Cohort for Measuring the Cost of Inactivity—Men and Nonpregnant Women Ages >15 Years, N=20,041
Variable	N	Unweighted %	Weighted %
Sex
Male	8,975	44.8	47.5
Female	11,066	55.2	52.5
Age (yr)
15-44	10,634	53.1	58.0
45-64	4,940	24.6	26.6
65+	4,467	22.3	15.4
Race
White	15,496	77.3	85.3
African-American	3,514	17.5	9.6
Other	1,031	5.1	5.1
Physically Active*	10,781	53.8	56.8
Male	5,677	63.3	65.4
Female	5,104	46.1	49.0
White	8,583	55.4	58.1
African-American	1,665	47.4	47.3
Physical Limitations+	3,234	16.1	13.0
Cigarette Smoking
Ever smoked >100 cigarettes	10,473	52.3	53.6
Current smoker	5,612	28.0	29.0
Obesity~	2,900	14.5	13.1
Health Status
Excellent	4,984	24.9	28.2
Good	10,276	51.3	51.8
Fair or poor	4,781	23.9	20.0
Income
Low (<200% of poverty line**)	5,982	29.8	26.3
Middle (>200%-400% of poverty line)	6,662	33.2	34.8
High (>400% of poverty line)	7,328	36.6	38.5
Education
Less than high school	5,891	29.4	24.6
Finished high school	7,334	36.6	37.3
More than high school	6,816	34.0	38.1
* Often spent at least half an hour in moderate or strenuous physical activity >3 times/wk. + Health limited the kinds or amount of moderate physical activities (eg, moving a table, carrying groceries, bowling) subject could do. ~ Body mass index >30 kg/m2. ** According to federal poverty levels published by the Bureau of the Census that controlled for family size and age of family head, the poverty line in 1987 ranged from $5,778 for 1 person to $23,105 for a family of 9.

The respondents provided detailed information about sociodemographic factors, health status, use and expense of medical care, and health-risk behaviors (such as regular physical activity and smoking status) for the period from January 1 to December 31, 1987, in three in-person interviews, one telephone interview, and a self-administered health status questionnaire (11). The response rate for completion of all four interviews was an excellent 80.1% (9).

Because survey respondents might have been unable to report accurate costs of medical care because of different payment sources, costs were confirmed through an additional survey of medical providers. Here, we analyze the medical costs for nonpregnant survey participants (N=20,041) aged 15 or older who responded to a minimum set of items in both the health status questionnaire and the access-to-care supplement.

Variables. Respondents who reported spending at least half an hour in moderate or strenuous physical activity three or more times per week were classified as physically active; those who did less physical activity than this were classified as inactive. Because the ability to participate in physical activity may vary by current health and physical status, we created a subgroup that included persons who reported limitations in performing specific kinds or amounts of moderate physical activities, such as moving a table, carrying groceries, or bowling.

Because of the well-documented health effects and costs of smoking, respondents were also stratified by smoking status (12). Those who had ever smoked at least 100 cigarettes were classified as smokers. In addition, we examined current smokers in a subanalysis.

Obesity was defined as a body mass index greater than or equal to 30 kg/m². Low income refers to income at or below 200% of the poverty line, which in 1987 ranged from $5,778 for 1 person to $23,105 for a family of 9.

The following categories of service use were defined. Hospital admissions included admissions for a stay of at least one night. Physician visits included visits to the office of a medical provider or a hospital outpatient department that did not result in an overnight hospital stay and for which the respondent reported seeing a physician. These visits also included emergency room visits that did not result in an overnight stay. Medication use included prescribed medication and refills used during the year. Respondents were asked for detailed information on all of their personal medical care expenditures for 1987, including hospital stays, acute and preventive ambulatory visits, prescription medications, and financing of these expenditures (10,11).

Data analysis. We performed stratified analyses on the relationship between medical expenditures and physical activity. For the primary stratification, we controlled for age, sex, and lifetime smoking status. We started by partitioning the sample into two groups: physically active and physically inactive. Average medical expenditures were calculated for both groups. Means and standard errors were calculated using SUDAAN software to account for the complex sampling frame.

Since both expenditures and participation in physical activity may be strongly influenced by physical limitations due to health status, we analyzed the expenditure comparisons between those who were physically active and inactive after excluding persons whose health limited regular moderate physical activity. Because of the strong influence of smoking on medical costs, we conducted an independent analysis for current smokers only by using similar stratification.

To further elucidate the impact of physical activity on direct medical costs, we investigated the components of the costs and medical care use for physically active and physically inactive persons: hospital stays, physician visits, and medications.

Results

The total sample consisted of 20,041 men and nonpregnant women ages 15 years and older. More than half (52%) were women, 58% were between 15 and 44 years of age, 85% were white, 57% spent at least 30 minutes in moderate or strenuous physical activity three or more times a week, and 13% reported physical limitations (table 1). Although 54% had smoked at some point in their lives, 29% reported current cigarette smoking. Thirteen percent were obese, 20% reported fair or poor health, and 26% reported low income. Three-fourths had at least a high school education.

The average annual direct medical costs for each nonpregnant respondent (active and inactive combined) were $1,690. The costs were considerably less for those who reported regular physical activity than for those who did not ($1,242 vs $2,277, P<0.01) (figure 1). However, much of the medical costs may be incurred by persons who, because of chronic health problems, have difficulty being physically active. When we took this factor into account by stratifying on the presence of physical limitations, there was still a cost advantage for people who were active (see figure 1). Those 15 and older without physical limitations and regularly engaged in physical activity had average annual direct medical costs of $1,019, vs $1,349 for those who reported being inactive (P<0.01).

Among persons without physical limitations in each age-group, those who were physically active had lower medical costs than those who were physically inactive (figure 2A). When stratified by sex, this effect held true except for men in the youngest (15 to 24 years) and oldest (75 years and older) age-groups (figure 2B). For women, after age 54, the medical costs for those who were physically inactive increased substantially relative to those who were physically active (figure 2C).

Among study participants who had ever smoked, the medical costs were higher than they were among those who had never smoked (see figure 1). Even among "ever smokers," physically active men and women had lower direct medical costs than inactive men and women did, with the exception of men 15 to 44 years old who had never smoked. The greatest cost savings were for physically active men 45 and older who had never smoked ($949) and for physically active women 45 and older who had ever smoked ($643).

Current smokers who were physically active had lower medical costs than those who were physically inactive ($1,081 vs $1,987, P<0.01) (see figure 1). Among those with physical limitations, current smokers who were physically active had lower direct medical costs than did their inactive counterparts ($3,290 vs $4,845, P=0.02). Among those without limitations, the direct medical costs were also lower for physically active persons than for inactive persons ($947 vs $1,268, P<0.01). This finding applied to men and women in younger (15 to 44) and older (45 and older) age-groups. The greatest cost savings among current smokers without physical limitations were for physically active men 45 and older ($424).

Resource use varied by physical activity status. When we eliminated survey participants who reported physical limitations, physically active persons used fewer resources than physically inactive persons did: They had fewer hospital stays and physician visits and used less medication (table 2). The major savings derived from lower hospitalization costs for physically active persons compared with inactive persons ($391 vs $613, P<0.01).

TABLE 2. Annual Use of Medical Services and Average Direct Medical Costs—Men and Nonpregnant Women Ages >15 Years Without Physical Limitations, N=16,807
Medical Service	Physically Active* (N=9,976)	Physically Inactive (N=6,831)	Cost Difference+ (95% CI)
Percent hospitalized	6.4%	9.7%
Length of hospital stay (nights)	6.2	6.5
Medical cost per person	$391	$613	$222 ($135-$309)
Physician visits
None	17.8%	15.8%
1-5	47.2%	44.8%
>5	35.0%	39.3%
Medical cost per person	$526	$599	$73 ($19-$127)
Medication~
None	37.3%	32.1%
1-5	41.0%	38.6%
>5	21.6%	29.3%
Medical cost per person	$102	$136	$34 ($24-$44)
Total direct medical cost per person	$1,019	$1,349	$330 ($214-$446)
* Often spent at least half an hour in moderate or strenuous physical activity >3 times/wk. + Dollar amount that inactive persons paid over what active persons paid. ~ Number of prescriptions and refills purchased. CI = confidence interval

Discussion

We have identified real and substantial medical cost differences between persons who are physically active and those who are not. Consistent cost savings for regularly active persons were observed for men and women, smokers and nonsmokers, and those with and without physical limitations. In contrast to previous theoretically modeled estimates of the direct medical costs associated with physical activity, we observed actual cost savings associated with regular physical activity for both younger and older study participants (8).

Our finding that the largest difference in direct medical costs was among women 55 and older is consistent with the widely held belief that the potential health gain associated with physical activity is especially high for older women (13-15). Older women are clearly a good target for physical activity interventions, but our data suggest that a population-wide strategy might produce cost savings among most adult age-groups.

Patterns of medical care use were consistent with the cost findings. Regularly active persons were hospitalized less often than inactive persons, made fewer physician visits, and filled fewer prescriptions.

The measure of physical activity used in the NMES is a single simple question assessing participation in half an hour of moderate or vigorous physical activity at least three days per week. This is a more easily reached standard than the current public health recommendation of 30 minutes of moderate physical activity 5 or more days per week (1). The prevalence estimate for participation in regular physical activity derived from this question (56.8%) is higher than published estimates from the National Health Interview Survey (NHIS) from that time period (men 44.3%, women 38.3%) (16). However, the NMES includes all physical activity and persons as young as 15, while the NHIS includes only leisure time physical activity and adults 18 and older. Both differences would tend to increase the prevalence estimate for physical activity from the NMES. Gender and racial differences in participation in regular physical activity are similar between the two surveys. We believe that the NMES categorizes persons as active or inactive with reasonable accuracy.

A key issue in interpreting our data is the relationship between physical inactivity and health conditions that might limit participation in physical activity. Physical inactivity may be both an effect and a cause of disabling health conditions (17,18). Through a stratified analysis, we attempted to control for the potential negative influence of physical limitations on participation in physical activity. Since we could not control for all aspects of poor health, it is possible that unreported health problems might have limited participation in physical activity among some persons. However, we were also unable to take into account the potential causal effect of inactive lifestyles on the development of physical limitations, and we therefore may have underestimated the potential benefit associated with regular physical activity.

It is of interest that even among participants reporting limitations in carrying out moderate physical activities, the medical costs were $1,053 lower per person among those who were regularly active. Some of this difference might be accounted for by less severe health limitations, but it may also reflect a benefit of physical activity.

Coronary heart disease, diabetes, and obesity are estimated to account for $51.1, $44.1, and $51.6 billion respectively in direct medical costs and $44.5, $54.1, and $47.6 billion respectively in indirect medical costs each year in the United States (19-21). With the well-documented relationship between physical activity and reduced risk for these and other chronic diseases, it seems reasonable to expect that physically active persons incur lower medical costs. Our data quantify the differences in direct medical costs between those who are active and those who are not.

Although causality cannot be determined from a cross-sectional study such as ours, the data suggest that increasing participation in moderate physical activity among sedentary adults may reduce direct medical expenditures. If the 43.2% of the Medical Expenditures Survey participants who reported no regular physical activity were to become regularly active and acquired the healthcare expenditure profile of active persons, between $330 and $1,053 per person would be saved in direct medical costs. If a conservative estimate of $330 is used, and this difference is applied to the US population over the age of 15 (N=205,030,000) (22), a savings of $29.2 billion in 1987 dollars ($76.6 billion in 2000 dollars) per year in direct medical costs could be realized (.432 * 205,030,000 * $330 [$866 for 2000]) (23,24). This estimate of potential savings of at least $29.2 billion is remarkably close to Colditz's (6) admittedly conservative $24 billion calculation of annual direct medical costs due to physical inactivity.

Our data do not allow us to address the potential costs associated with increasing participation in regular physical activity, or the indirect costs and benefits of physical activity that would be required to calculate an overall estimate of the cost effectiveness of physical activity in reducing medical costs. However, it is likely that the pattern of direct and indirect costs attributable to physical inactivity is similar to those for coronary heart disease, diabetes, and obesity. For these three conditions, the ratio of indirect to direct costs ranges from 87% to 123% (19-21). Thus, the total of direct and indirect costs associated with physical inactivity may well be in excess of $150 billion in 2000 dollars.

The medical costs of physical inactivity appear to be similar to those due to smoking. Two analyses of medical costs based on the NMES produced estimates of $53.4 and $50.0 billion in 1993 for smoking-attributable direct medical care (11,12). In comparison, we estimate that 1993 medical costs attributable to physical inactivity would have been $45.6 billion (23,24).

A Clear Public Health Concern

Our study suggests that the economic burden of physical inactivity is large. Additional economic analyses and longitudinal studies of physical activity and medical costs are needed to better define both direct and indirect costs associated with physical inactivity. However, the health and economic consequences of physical inactivity are sufficiently great to merit public health action now. Coupled with the clear causal relationship between physical inactivity and chronic disease and the availability of effective public health intervention strategies, our findings of substantial medical cost differences between active and inactive people reinforce the need to make reducing physical inactivity a public health priority.

References

1. US Department of Health and Human Services: Physical Activity and Health: A Report of the Surgeon General. Atlanta, DHHS, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996
2. Centers for Disease Control and Prevention: Public health focus: physical activity and the prevention of coronary heart disease. MMWR Morb Mortal Wkly Rep 1993;42(35):669-672
3. McGinnis JM, Foege WH: Actual causes of death in the United States. JAMA 1993;270(18):2207-2212
4. Hahn RA, Teutsch SM, Rothenberg RB, et al: Excess deaths from nine chronic diseases in the United States, 1986. JAMA 1990;264(20):2654-2659
5. Jones TF, Eaton CB: Cost-benefit analysis of walking to prevent coronary heart disease. Arch Fam Med 1994;3(8):703-710
6. Colditz GA: Economic costs of obesity and inactivity. Med Sci Sports Exerc 1999;31(11 suppl):S663-S667
7. Hatziandreu EI, Koplan JP, Weinstein MC, et al: A cost-effectiveness analysis of exercise as a health promotion activity. Am J Public Health 1988;78(11):1417-1421
8. Nicholl JP, Coleman P, Brazier JE: Health and healthcare costs and benefits of exercise. Pharmoeconomics 1994;5(2):109-122
9. Cohen SB, Carlson BL: Characteristics of reluctant respondents in the National Medical Expenditure Survey. J Econ Soc Meas 1995;21:269-296
10. Cohen SB, DiGaetano R, Waksberg J: Sample design of the 1987 household survey (AHCPR Pub. No. 91-0037). Rockville, MD, Agency for Health Care Policy and Research, Public Health Service, 1991
11. Miller VP, Ernst C, Collin F: Smoking-attributable medical care costs in the USA. Soc Sci Med 1999;48(3):375-391
12. Centers for Disease Control and Prevention: Medical-care expenditures attributable to cigarette smoking: United States, 1993. MMWR Morb Mortal Wkly Rep 1994;43(26):469-472
13. Wagner EH, LaCroix AZ, Buchner DM, et al: Effects of physical activity on health status in older adults. I: observational studies. Annu Rev Public Health 1992;13:451-468
14. Stewart AL, King AC: Evaluating the efficacy of physical activity for influencing quality-of-life outcomes in older adults. Ann Behav Med 1991;13(3):108-116
15. Nelson ME, Fiatarone MA, Morganti CM, et al: Effects of high-intensity strength training on multiple risk factors for osteoporotic fractures: a randomized controlled trial. JAMA 1994;272(24):1909-1914
16. Caspersen CJ, Christenson GM, Pollard RA: Status of the 1990 physical fitness and exercise objectives: evidence from NHIS 1985. Public Health Rep 1986;101(6):587-592
17. Vita AJ, Terry RB, Hubert HB, et al: Aging, health risks, and cumulative disability. N Engl J Med 1998;338(15):1035-1041
18. Clark DO: The effect of walking on lower body disability among older blacks and whites. Am J Public Health 1996;86(1):57-61
19. American Heart Association: 1998 Heart and Stroke Statistical Update. Dallas, American Heart Association,1997
20. American Diabetes Association: Economic consequences of diabetes mellitus in the US in 1997. Diabetes Care 1998;21(2):296-309
21. Wolf AM, Colditz GA: Current estimates of the economic cost of obesity in the United States. Obes Res 1998;6(2):97-106
22. National Center for Health Statistics: Health, United States 1996-97 and Injury Chartbook. Hyattsville, MD, National Center for Health Statistics, 1997
23. Agency for Health Care Policy and Research: Trends in personal health care expenditure, health insurance, and payment sources, community-based population, 1987-1995. Rockville, MD, Agency for Health Care Policy and Research, Public Health Service, March 1997
24. Agency for Health Care Policy and Research: Trends in personal health care expenditure, health insurance, and payment sources, community-based population, 1996-2005. Rockville, MD, Agency for Health Care Policy and Research, Public Health Service, August 1998

Dr Pratt is a medical epidemiologist, Dr Macera is a senior epidemiologist, and Dr Wang is a health economist in the Division of Nutrition and Physical Activity, Physical Activity and Health Branch, at the Centers for Disease Control and Prevention in Atlanta. Address correspondence to Michael Pratt, MD, MPH, Centers for Disease Control and Prevention, Division of Nutrition and Physical Activity, Physical Activity and Health Branch, 4770 Buford Hwy NE (K-46), Atlanta, GA 30341; e-mail to mxp4@cdc.gov.