BACKGROUND: The mortality rates of older people changes with the seasons. However, it has not been properly investigated whether the seasons affect medical care expenditure (MCE) and institutionalization. Seasonal variation in MCE is plausible, as MCE rises exponentially before death. It is therefore important to investigate the impact of the seasons on MCE both mediated and unmediated by mortality.

METHODS: Data on mortality, MCE and institutionalization from people aged 65 and older in a region in the Netherlands from July 2007 through 2010 were retrieved from a regional health care insurer and were linked with data from the Netherlands Institute for Social Research, and Statistics Netherlands (n = 61,495). The Seasonal and Trend decomposition using Loess (STL) method was used to divide mortality rates, MCE, and institutionalization rates into a long-term trend, seasonal variation, and remaining variation. For every season we calculated the 95% confidence interval compared to the long-term trend using Welch's t-test.

RESULTS: The mortality rates of older people differ significantly between the seasons, and are 21% higher in the winter compared to the summer. MCE rises with 13% from the summer to the winter; this seasonal difference is higher for the non-deceased than for the deceased group (14% vs. 6%). Seasonal variation in mortality is more pronounced in men and people in residential care. Seasonal variation in MCE is more pronounced in women. Institutionalization rates are significantly higher in the winter, but the other seasons show no significant impact.

CONCLUSIONS: Seasonal changes affect mortality and the level of MCE of older people; institutionalization rates peak in the winter. Seasonal variation in MCE exists independently from patterns in mortality. Seasonal variation in mortality is similar for both institutionalized and community-dwelling elderly. Policy-makers, epidemiologists and health economists are urged to acknowledge and include the impact of the seasons in future policy and research.