A Model to Estimate the Potential for a Physical Activity-Induced Reduction in Healthcare Costs for the Elderly, Based on Pedometer/Accelerometer Data from the Nakanojo Study

The rising healthcare costs associated with an aging population have become an urgent fiscal problem. However, evidence of the efficacy of preventive programmes is limited, since almost all studies have involved only small numbers of highly selected participants. This article examines potential physical activity-induced decreases in healthcare expenses, applying a theoretical model to the Nakanojo Study of habitual physical activity and health in an entire elderly community. The Nakanojo Study has shown substantial associations of health with both step count and the duration of moderate effort (intensity >3 metabolic equivalents [METs]). Participants are classed as 'dependent' (n = 800) or 'independent' (n = 4400); the latter category is divided arbitrarily into quartiles, based on physical activity patterns (Q1-Q4; n = 1100 for each quartile). The five groups show a graded prevalence of various morbidities, including dependency, depression, osteoporosis, fractures, hypertension, diabetes mellitus, hyperlipidaemia, ischaemic heart diseases, cerebrovascular diseases, cancer and dementia. Consequently, annual healthcare expenditures (based on 2009 published Japanese costs associated with each of these conditions) differ by about yen (Y)197 900 ($US1979) per person between dependent individuals and those in group Q1, Y20 700 ($US207) between Q1 and Q2, Y14 600 ($US146) between Q2 and Q3, and Y5300 ($US53) between Q3 and Q4. Accepting a causal relationship between physical activity and health, and assuming that an increase in physical activity induces a benefit that is uniform across conditions and diseases, respective morbidity prevalences and associated healthcare costs seem likely to decrease as physical activity increases. Thus, if the physical activity of only 5% of each group could be increased by a single ranking (pedometer/accelerometer scores of 2000 steps/day and 5-10 min/day at >3 METs and/or an adjusted questionnaire score of 10 MET hours/week), one might predict average savings across this population of about Y12600 (SUS126) per person, or 3.7%, of total medical expenses, including Y9800 ($US98) of public nursing care insurance costs and an additional Y2800 ($US28) of national health insurance expenditures. The impact of various changes in the prevalence of physical activity can be simulated using our model. In principle, savings should increase if more people increase their physical activity, and/or the magnitude of individual increases in physical activity is greater. Nevertheless, our analysis suggests that if even a small fraction of individuals in the three least active groups were to make a single-rank increase in their habitual physical activity as a result of focused health support and the promotion of physical activity, a significant reduction in medical expenses might be anticipated, justifying investment in preventive programmes. We now propose to test the validity of the present simulations on a national basis, obtaining accurate and objective evidence of change in individual physical activity patterns using an advanced design of pedometer/accelerometer.