To assess the value of promoting health through exercise, by using a cost-benefit analysis of the effect of regular exercise on morbidity and mortality, and the implications for health services resources.

ABI/INFORM was searched from 1987 to 1993, MEDLINE from 1975 to 1993, Social Sciences Citation Index from 1983 to 1993, and SPORTDiscus from 1975 to 1993, using the following search terms: 'exercise', any of the diseases or conditions from a comprehensive bibliography of the benefits of exercise, 'costs of illness', 'cost-benefit analysis', 'cost-effectiveness', 'economic costs', 'catastrophic injuries', 'sudden death', 'sports injuries', 'lifestyle indicators', 'smoking', 'quality of life' and 'well-being'.

Study designs of evaluations included in the review

Studies of any design providng information on the estimates of risk reduction, health benefits, or costs for any of the disease groups identified, were included.

Specific interventions included in the review

Exercise.

Participants included in the review

No restrictions on the participants included were stated.

Outcomes assessed in the review

Risk reduction for coronary heart disease (CHD), stroke and other cerebrovascular diseases (CVD), hip and other fractures related to osteoporosis, mental illness, and diabetes. Risks and costs of exercise-related illness, injury and death.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The authors do not state that they assessed validity.

Relevant estimates of risk-reduction, benefits or costs were taken from studies and combined with national statistics to be used in analysis.

How were the studies combined?

No clear methodology for obtaining estimates of the RR of disease from the studies is reported. The studies were combined differently for each disease:

for CHD, 2 studies of males only gave the same result and this was used as the estimate; for CVD, 2 studies were used to provide an approximate central estimate;

for diabetes, the results from 1 study were taken as the estimate; and for mental health, no numeric data were available and the estimate was made arbitrarily.

Benefits of exercise on health were estimated by combining the estimates of disease incidence in the general population with the estimated RR of the disease in individuals who exercise regularly, compared with those who do not. Healthcare costs were estimated from the in-patient costs associated with each admission that might be avoided by exercising.

How were differences between studies investigated?

The authors do not report any investigation of differences between studies.

A number of references were cited for each disease group studied, but not all of these were included in estimating the relative risk (RR) of disease in people who exercised.

For CHD, 6 studies were cited and 2 were used in the estimate; for CVD 2 studies were cited and used; for diabetes, 4 studies were cited and only 1 was used to estimate risk;

for hip fracture, 2 studies were cited and used; and for mental illness, the estimate of reduction was not based on any study, but 3 were cited to support a relationship.

The RR of disease in individuals who exercise, regularly compared with those who do not, is estimated to be 0.6 for CHD, 0.67 for CVD, 0.61 for diabetes, 0.4 in hip fracture and 0.9 for mental illness. From this, the percentage of incident cases which could be avoided by exercise is estimated at 32% for CHD, 26% for CVD, 22% for diabetes, 51% for hip fracture and 7% for mental disorders.

For diabetes and hip fracture, it is assumed that the RR of disease and benefit of exercise will only vary in those individuals over the age of 44, thus the RR is 1 and no disease incidence would be avoided by exercise.

For all other diseases, the results are the same for all adults (aged 15 and over). Costs avoided are estimated at:

0.31 and 9.75 per person for CHD, aged 15 to 44 years and 45 and over, respectively,

0.35 and 13.67 for CVD,

0.02 and 1.01 for mental disorders; and

for people aged 45 and over only,

1.22 for diabetes and

4.74 for hip fracture.

Overall, these give total costs avoided as 0.68 per person for people aged 15 to 44 and 30.39 for those aged 45 and over. The estimated average annual medical care costs per person incurred due to sport and exercise is 30 for adults aged 15 to 44 and 10 for adults aged 45 and over.

Costs of medical care avoided by exercise and incurred due to sport.

The results show that in adults aged 15 to 44, the average annual medical care costs per person that might be incurred as a result of full participation in sport and exercise (approximately 30) exceed the costs that might be avoided by the disease-prevention effects of exercise (less than 5 per person). However, in adults aged 45 and over, the estimated costs avoided (greater than 30 per person) greatly outweigh the costs that would be incurred (less than 10). With regard to health and medical care costs, there are strong economic arguments in favour of exercise in adults aged 45 years and older, but not in younger adults. These results may not have direct application to health care systems in other countries, but the general results are likely to hold true.

There were a number of quite sweeping assumptions and limitations on the accuracy of the estimates used in the analysis, which are discussed in the paper.

The methodology was lacking in a number of areas:

1. No description of the studies on which the evidence was drawn.

2. Method of estimating benefits of exercise was often ad hoc, and involved simple averages of 2 or more studies.

3. There was little consideration of the validity of different types of study design.

Exercise promotion in the younger adult population (under 45 years of age) may not be efficient. Exercise in the over 45s will save money, in terms of direct health service costs.

Fentem PH, Bassey EJ, Turnbull JB. The new case for exercise. London: Sports Council and Health Education Authority; 1988.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.