Physical activity can reduce illnesses and deaths linked to chronic diseases (1,2). The health benefits of physical activity increase with increasing frequency, duration, and intensity of exercise (2-4). Data from longitudinal cohort studies suggest that physical inactivity is associated with at least a 1.5-fold to 2.0-fold higher risk of most chronic diseases of lifestyle, such as coronary heart disease, type 2 diabetes, and hypertension (1,5), and accounts for an estimated 1.3% of lost disability-adjusted life-years worldwide. Furthermore, studies corroborate the public health recommendation that 30 minutes of accumulated, moderate-to-vigorous intensity physical activity on most days is protective for these chronic diseases (3). The associated risk of inactivity is similar in magnitude to many other well-known risk factors, such as overweight, smoking, hyperlipidemia, and low fruit and vegetable intake (1,6).

Cross-sectional studies have estimated the economic costs associated with inactivity, or the cost savings associated with regular physical activity, at a national level in many industrialized countries (7-9). In Canada, where more than two-thirds of the population is considered to be insufficiently active, physical inactivity is estimated to be responsible for 2.5% of the total direct health care costs or the equivalent of 21,000 lives lost prematurely each year (7). Efforts to model the cost of inactivity to health care plans have typically yielded similar or higher costs compared with national estimates. Using a cost-of-illness approach, another study examined medical claims among approximately 1.5 million health plan members aged 18 years or older in Minnesota (10). In this model, more than 30% of cases of stroke, cancer of the colon, cardiovascular disease, and osteoporosis were attributable to inactivity. Health care providers are recognizing the role of physical activity in reducing risk for noncommunicable diseases (11).

A small number of studies have demonstrated actual reduction in health care costs and cost savings in physically active members of health plans (12-14). For example, claims data from approximately 23,000 health plan members showed that average annual health care claims were approximately $250 lower for those who were either moderately active (1-2 times per week) or very active (3-4 times per week), compared with their sedentary counterparts, on the basis of self-report (12). After 2 years, Medicare members who received a health club benefit as part of their health plan had significantly fewer inpatient admissions and lower total health care costs than did matched controls (13).

Further evaluation of physical activity programs offered by health plans is needed to establish the cost savings of such strategies. We designed our study to examine the association between levels of participation in fitness-related activities, as part of the incentive-based wellness program Vitality on medical claims and hospital admissions among members of the largest national private health insurer in South Africa, Discovery Health.

In 2006, approximately 60% of all members were registered for Vitality. Of these, 71% were inactive and 12% were high active (Table 1).

Men represented 59% of the high active group and 46% to 50% of all other groups. The proportion of people registered for chronic illness benefits was higher among people not registered for Vitality than people in any of the other groups.

Our study found an unequivocal and inverse relationship between fitness-related activities among insured persons and hospital claims and admissions. The cost savings were similar to those reported in previous research (12,13); average annual health care costs were approximately $250 lower among active compared with inactive members, even considering those who exercised only 1 to 2 times per week. A study of Medicare members receiving a health club membership as part of their health plan had fewer inpatient admissions and lower total health care costs than matched controls not receiving the benefit (13). The actual uptake of this benefit remained low; less than 7% of the total plan membership participated.

Differences in savings between studies may be explained, in part, by the fact that we analyzed only medical claims associated with hospitalization, whereas the comparable studies typically report total health care expenditures. Furthermore, study populations differed in terms of age and demographics. Among health plan members aged 50 years or older, changing physical activity status from inactive to active was associated with approximately $2,000 in savings in health care claims during 2 years compared with remaining inactive during the same time (14).

The strength of our study is that gym visits and sports participation were documented and not based on self-report. However, it may be argued that the definition of engagement in fitness-related activities (eg, >48 gym visits per year) lacked sufficient sensitivity to accurately reflect dose-response exposure to physical activity. These criteria are not compatible with physical activity and public health recommendations of 30 minutes of moderate-to-vigorous intensity activity on most, preferably all, days of the week (17). Despite this, the apparent dose-response effect suggests that the definitions may be sufficiently discriminating. Previous cohort studies have found that even 1 to 2 bouts of physical activity per week showed significant risk reduction for diabetes mellitus (18,19) and cardiovascular mortality (20).

The facts that the activity participation was documented and that most of the high active people were highly engaged in Vitality suggest that the potential health benefits that accrued were in part related to participation, even if there was potential for selection bias. Furthermore, the significant association between participation in fitness-related activities and reduced medical claims or admissions was also demonstrated in the disease-related subgroups. This is in line with cohort studies in which relative risk for cardiovascular death, for example, in women diagnosed with diabetes decreased by 7% with as little as 1 to 2 hours of moderate physical activity per week (21). Similarly, as much as a 40% savings in health care expenditures was demonstrated in members of a managed care cohort with diabetes who attended a community-based fitness program at least once per week (22).

We cannot rule out the fact that physical activity clusters with other positive health behaviors. This underlying association may explain, at least in part, some of the relationship between gym visits and medical claims. For example, in the Aerobics Center Longitudinal Study cohort, cardiorespiratory fitness was inversely associated with consumption of dietary fat, saturated fat, and cholesterol in approximately 10,000 people measured during an 8-year period (23).

The challenge remains that while those members who are highly engaged have significantly lower health care claims and hospitalization, they are underrepresented in the larger plan membership. Various health care providers have developed strategies to increase adoption of physically active lifestyles, including full or partial subsidy of a health club or fitness center memberships. Third-party monitoring of fitness center visits was associated with increased use (24). We cannot say whether the incentives and rewards associated with Vitality influenced participation in fitness-related activities.

The data concerning fitness-related activities were limited in that not all members would have submitted information or recorded gym visits. Alternatively, members may have participated in physical activities that the rewards program did not capture. This is the case particularly for members not registered in the Vitality program.

Participation in fitness-related activities within an incentive-based health insurance wellness program was associated with significantly lower health care costs. However, as in other studies, the involvement in fitness-related activities was generally low, and further research is required to identify and address the barriers to participation in such programs.