6: No. 4, October 2009
Fitness-Related Activities and Medical Claims Related to Hospital Admissions — South Africa, 2006
Estelle V. Lambert, PhD; Rosanne da Silva; Deepak Patel, MD, MSc; Libero Fatti, PhD; Tracy Kolbe-Alexander, PhD; Adam Noach; Craig Nossel, MBChB, MBA; Wayne Derman, MBChB, PhD; Thomas Gaziano, MD,
Suggested citation for this article: Lambert EV, da Silva R, Patel D, Fatti L, Kolbe-Alexander T, Noach A, et al. Fitness-related activities and medical claims related to hospital admissions — South Africa, 2006. Prev Chronic Dis 2009;6(4):A120.
oct/08_0226.htm. Accessed [date].
We report on the effect of an incentive-based wellness program on medical claims and hospital admissions among members of a major health insurer. The focus of this investigation was specifically on fitness-related activities in this insured population.
Adult members of South Africa’s largest private health insurer (n = 948,974) were grouped, a priori,
on the basis of documented participation in fitness-related activities, including gym visits, into inactive (80%, equivalent to
≤3 gym visits/y), low active (7.0%, 4-23 gym visits/y), moderate active (5.2%, 24-48 gym visits/y), and high active (7.4%, >48 gym visits/y) groups. We compared medical claims data related to hospital admissions between groups after
adjustment for age, sex, medical plan, and chronic illness benefits.
Hospitalization costs per member were lower in each activity group compared with
the inactive group. This same pattern was demonstrated for admissions rates. There was good agreement between level of participation in fitness-related activities and in other wellness program offerings; 90% of people only nominally engaged in the wellness program also were low active or inactive, whereas 84% of those in the high
active group also had the highest overall participation in the wellness program.
Participation in fitness-related activities within an incentive-based health insurance wellness program was associated with lower health care costs. However, involvement in fitness-related activities was generally low, and further research is required to identify and address barriers to participation in such programs.
Back to top
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.
Back to top
In South Africa, participation in private medical plans is inversely associated with income, despite the copayment by employers. Only 34% of people earning above R5,000 (US $600) per month are members of private medical
plans, and this proportion more than doubles at incomes of R10,000 (US $1,200) or more. Discovery Health is more than 3 times the size of its nearest competitor and alone accounts for approximately 35% of the open
plan market and 25% of all
medical plan beneficiaries in South Africa.
Discovery Health has offered Vitality, an incentive and reward-based health promotion program, to its members since 1998. Membership is voluntary and offered separately from the health plan because legislation in South Africa precludes differential insurance premiums based on health status or engagement with health promotion programs. The program is offered to plan members for a nominal monthly fee of approximately R100 (US $12) per family. The sample included both principal and
spouse members whose benefits had been effective for a full 12 months during 2006 and who, during that time, were either registered for Vitality or were not registered for at least a full 12 months.
The final sample was 948,974 members. All data were analyzed unlinked to any personal identifiers. The Research Ethics Committee of the Faculty of Health Sciences, University of Cape Town, approved the study protocol.
Levels of engagement in Vitality
The activities of the program are fitness-related activity, assessment and screening, healthy choices, and health knowledge. Specific activities include subsidized gym memberships, visits to dietitians and exercise specialists, smoking cessation and weight reduction programs,
and access to online or in-person risk assessments and online and print media material for health and wellness. Participation in the various wellness services and programs earns the participants points, which we used as a
proxy measure for the level of participation in the health promotion program. Points are redeemable as discounts (ranging from 15% to 45%) on various goods and services. The level of engagement in Vitality was classified as 1) not registered for Vitality; 2) registered for Vitality but with no points in any of the 4 categories, defined as registered but not engaged; 3) registered for Vitality and accumulating up to a threshold level of points, defined as low engagement; and 4) registered for
Vitality and accumulating more than the threshold level of points, defined as high engagement.
Categories of fitness-related activity
Vitality program participants were awarded points specifically for fitness-related activities according to the total number of recorded gym visits to participating commercial fitness center partners. Members
could also accumulate points for participation in major sporting or fitness events such as road running or cycle races (members register to participate through a commercial partner organization, SA Active). The fitness status was defined as 1) high active — points equivalent
to more than 48 gym visits per year; 2) moderate active — points equivalent to 24 to 48 gym visits per year; 3) low active — points equivalent to 4 to 23 gym visits per year; and 4) inactive — points equivalent to 3 or fewer gym visits per year.
Claims data categories
Members of the Discovery Health insurance plan can subscribe to 2 different plan types: comprehensive and core. Plans differ in the degree of coverage for ambulatory care.
Coverage does not substantially differ for conditions requiring hospital admission. Members with specified chronic conditions, such as hypertension, diabetes, and hypercholesterolemia, were expected to register for chronic illness benefits paid from the insurer’s risk pool (as opposed to the member’s personal
medical savings account).
For this analysis hospital claims data included the admission rate, cost per member for the entire population, cost per patient admitted to a hospital, number of days hospitalized per patient, number of hospitalizations per patient, length of stay per patient, and cost per hospitalization. Because the insurance pool does not cover claims for acute ambulatory care, we considered those data as incomplete; therefore, they
were not analyzed. A subsample analysis included only those members who
had been hospitalized at least once.
A further diagnosis-related subgroup analysis was conducted for high active status compared with all other fitness-related activity groups for hospital admission rates. Preselected subgroups included cancers, cardiovascular diseases, musculoskeletal conditions, and endocrine and metabolic conditions such as diabetes, which are complications of conditions responsive to interventions for health risk behaviors.
The adjusted means were first calculated for those engaged in the Vitality
program, taking into account the
effect of the weighted covariates. Factors that were likely to independently influence medical claims data, irrespective of participation in the wellness program, were preselected as covariates for the multivariate analysis of covariance (ANCOVA),
with the Tukey-Kramer t test (for multiple
comparisons) to determine significance. This approach combines regression with experimental design into a single model (15,16).
The covariates selected included age (in 5-year bands), sex, chronic illness status (single or multiple risk factors or comorbid conditions), and health plan options. We used a tree analysis implemented in SAS Enterprise Miner (SAS Institute, Inc, Cary, North Carolina) to assess the
effect of these covariates under each claim cost category.
Back to top
In 2006, approximately 60% of all members were registered for Vitality. Of these, 71% were inactive and 12% were high active
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.
Fitness-related activity group and Vitality engagement
A strong relationship existed between engagement in fitness activities and engagement in Vitality
(Table 2). Approximately 84% of the high active group had a high level of engagement with the Vitality program. Approximately 27% of those highly engaged in Vitality
went to gym more than 96 times in 2006 (average of 1.9 times per week), and approximately 62% reported
going to the gym more than 48 times. We did not determine the level of participation in other wellness activities. Those who were
highly engaged in the Vitality program also had proportionally more participation in health knowledge activities
(eg, online health risk assessment and feedback, online nutrition assessment and
feedback) (56% vs 17%) and assessments and screening (48% vs 8%) compared with those who had low engagement.
Fitness-related activity group and hospitalization
For members who were admitted to a hospital in 2006, the adjusted means for cost per patient,
total number of days hospitalized per patient, number of admissions per patient, length of stay per patient, and cost per hospitalization were significantly lower in the high active group compared with all other groups (P < .001) (Table 3). In addition, number of days hospitalized per patient and length of stay were significantly lower among moderate active patients compared with those not
registered or who were low active (P < .001). Among those patients with at least 1 hospital admission, both costs per patient and number of days of hospitalization per patient
decreased in relation to increased levels of participation in fitness-related activities. The high active members who were hospitalized in 2006 experienced a mean annual savings in associated medical claims of R5,025 (US $603) compared with inactive members.
The cost per member, number of admissions per member, and length of stay per member were all significantly lower in the high active group (Table 4). Furthermore, hospitalization costs per member
decreased in each group from the inactive to the high active group (P < .001).
This same pattern was demonstrated for admissions rate (P < .001).
Participants in the high active group saved an average of R1,535 (US $184) in
health care costs compared with the inactive group.
Diagnosis-related subgroup analysis
The admission rates per member for the high active group were significantly lower when compared
with all other groups for diagnosis-related subgroups (Table 5). Admissions associated with cancer and mental illness were approximately 35% lower, and admissions associated with endocrine,
nutritional, and metabolic disorders and kidney and urinary tract disorders were 20% lower.
Back to top
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
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
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.
Back to top
The source of funding for this research study was from Discovery Health, a South African
health insurance company. Discovery Health also provided the researchers with unlinked data for analysis. The University of Cape Town is not constrained in any way by the funder and is free to publish all results of the research study.
Back to top
Corresponding Author: Estelle V. Lambert, PhD, UCT/MRC Research Unit for Exercise Science and Sports Medicine, University of Cape Town, PO Box 115 Newlands, Cape Town, Western Cape 7725, South Africa. Telephone: 27-21-6504571. E-mail:
Author Affiliations: Rosanne da Silva, Libero Fatti, University of the Witwatersrand, Johannesburg, South Africa; Deepak Patel, University of Cape Town, Cape Town, South Africa, and Discovery Health, Johannesburg, South Africa; Tracy Kolbe-Alexander,
Wayne Derman, University of Cape Town, Cape Town, South Africa; Adam Noach, Craig
Nossel, Discovery Health, Johannesburg, South Africa; Thomas Gaziano, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts.
Back to top
- Rodgers A, Ezzati M, Vander Hoorn S, Lopez AD, Lin RB, Murray CJ, et al.
Distribution of major health risks: findings from the Global Burden of Disease Study. PLoS Med 2004;1(1):e27.
- Blair SN, Cheng Y, Holder JS.
Is physical activity or physical fitness more important in defining health benefits? Med Sci Sports Exerc 2001;33(6 Suppl):S379-99.
- Bouchard C.
Physical activity and health: introduction to the dose-response symposium. Med Sci Sports Exerc 2001;33(6
- Haskell WL.
JB Wolffe Memorial Lecture.
Health consequences of physical activity:
understanding and challenges regarding dose-response. Med Sci Sports Exerc 1994;26(6):649-60.
- Brown WJ, Burton NW, Rowan PJ.
Updating the evidence on physical activity and health in women. Am J Prev Med 2007;33(5):404-11.
- Farrell SW, Kampert JB, Kohl HW 3rd, Barlow CE, Macera CA, Paffenbarger RS
Jr, et al. Influences of cardiorespiratory fitness levels and other predictors on cardiovascular disease mortality in men. Med Sci Sports Exerc 1998;30(6):899-905.
- Katzmarzyk PT, Gledhill N, Shephard RJ.
The economic burden of physical inactivity in Canada. CMAJ 2000;163(11):1435-40.
- Wang G, Pratt M, Macera CA, Zheng ZJ, Heath G.
Physical activity, cardiovascular disease, and medical expenditures in U.S. adults. Ann Behav Med 2004;28(2):88-94.
- Allender S, Foster C, Scarborough P, Rayner M.
The burden of physical activity-related ill health in the UK. J Epidemiol Community Health 2007;61(4):344-8.
- Garrett NA, Brasure M, Schmitz KH, Schultz MM, Huber MR.
Physical inactivity: direct cost to a health plan. Am J Prev Med 2004;27(4):304-9.
- Fitzner K, Madison M, Caputo N, Brown E, French M, Bondi M, et al.
Promoting physical activity: a profile of health plan programs and initiatives. Manag Care Interface 2002;15(12):29-41.
- Wang F, McDonald T, Champagne LJ, Edington DW.
Relationship of body mass index and physical activity to health care costs among employees. J Occup Environ Med 2004;46(5):428-36.
- Nguyen HQ, Ackermann RT, Maciejewski M, Berke E, Patrick M, Williams B, et al. Managed-Medicare health club benefit and reduced health care costs among older adults. Prev Chronic Dis 2008;5(1).
- Martinson BC, Crain AL, Pronk NP, O’Connor PJ, Maciosek MV.
Changes in physical activity and short-term changes in health care charges: a prospective cohort study of older adults. Prev Med 2003;37(4):319-26.
- Kim K, Timm N. Univariate and multivariate general linear models: theory and
application with SAS, second edition. Boca Raton (FL): Chapman and Hall/CRC; 2006.
- Breiman L, Friedman J, Olshen RA, Stone CJ. Classification and regression trees. Boca Raton (FL):
Chapman and Hall/CRC; 1984.
- Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, et al.
Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation 2007;116(9):1081-93.
- Manson JE, Rimm EB, Stampfer MJ, Colditz GA, Willett WC, Krolewski AS, et al.
Physical activity and incidence of non-insulin-dependent diabetes mellitus in women. Lancet 1991;338(8770):774-8.
- Lee IM, Rexrode KM, Cook NR, Manson JE, Buring JE.
Physical activity and coronary heart disease in women: is “no pain no gain” passé? JAMA 2001;285(11):1447-54.
- Lee IM, Sesso HD, Oguma Y, Paffenbarger RS Jr.
The “weekend warrior” and risk of mortality. Am J Epidemiol 2004;160(7):636-41.
- Hu FB, Stampfer MJ, Solomon C, Liu S, Colditz GA, Speizer FE, et al.
Physical activity and risk for cardiovascular events in diabetic women. Ann Intern Med 2001;134(2):96-105.
- Nguyen HQ, Ackermann RT, Berke EM, Cheadle A, Williams B, Lin E, et al.
Impact of a managed-Medicare physical activity benefit on health care utilization and costs in older adults with diabetes. Diabetes Care 2007;30(1):43-8.
- Brodney S, Mcpherson RS, Carpenter RS, Welten D, Blair SN.
Nutrient intake of physically fit and unfit men and women. Med Sci Sports Exerc 2001;33(3):459-67.
- Shepich J, Slowiak JM, Keniston A.
Do subsidization and monitoring enhance adherence to prescribed exercise? Am J Health Promot 2007;22(1):2-5.
Back to top