Comment on: "impact and cost-effectiveness of a universal strategy to promote physical activity in primary care".

Abstract

The paper by Gulliford et al. [1] analyzes the cost-effectiveness of promoting physical activity in healthy adults in a UK primary care setting by a Markov state transition model. The intervention consists of annual screening for physical activity plus brief consultation through a family physician. The model assumes low intervention costs (£35 per person year) and significant added health benefits in terms of sedentary adults becoming active after 1 year. The number needed to treat for one additional sedentary adult to meet internationally recommended levels of activity is 12 (95 % confidence interval 7–33) [5]. Yet, the intervention was not found to be clearly cost-effective at a willingness to pay of £30,000 per quality-adjusted life year (QALY) gained. That is, the probability of being cost-effective was only 72 % over 10 years (Table 4). The result is surprising as effective low-cost interventions such as generic drugs or vaccines are usually highly cost-effective. On closer scrutiny, the underlying reason for the lack of cost-effectiveness is not the base-case costeffectiveness ratio (which is £13,685/QALY = 80.744/ 0.0059, based on data in Table 4) but an astonishing 29 % probability that the intervention leads to lower (!) life expectancy (Table 3). This finding, however, is at odds with the prevailing viewpoint in the literature that the relationship between the level of exercise and health/survival is linear or curvilinear [2, 3]. For example, a recently conducted large pooled cohort analysis [4] comprising 654,827 individuals, 21–90 years of age, showed that even a low level of leisure time physical activity of moderate to vigorous intensity was associated with reduced risk of mortality relative to no such activity. The overall dose– response relationship was curvilinear, with diminishing returns at high activity levels (see Fig. 1). Yet, for none of risk groups considered the returns to exercise were negative, even at the limits of the 95 % confidence interval. Therefore, the result of the analysis by Gulliford et al. [1] is not externally valid. Fig. 1 Leisure time physical activity level and hazard ratios for mortality and gains in life expectancy after age 40 [4]