Spatio-temporal patterns for pneumonia and influenza hospitalization

A.A.Kosheleva, D.A. Castronovo, J.S. Jagai K.C. Chui, Y.N. Naumov, J. Gorski, E.N. Naumova

Background

Hospitalization for pneumonia and influenza (P&I) has a well defined seasonal pattern and is stable across various climatic zones of the USA.
P&I incidence in elderly increases with age and in presence of co-morbid conditions.
Objective
To provide detailed examination for the spatio-temporal distribution of hospitalization rates of P&I in the US elderly (≥65 years old)

Abstracted 6,277,684 individual records of hospitalization (ICD-9M 480-487) from 1998-2002.
Developed a set of curves to reflect age-specific rates of P&I in patients with one of the three preexisting medical conditions: cancer, diabetes, and renal failure.
Examined the annual rates of P&I as a function of age.
Tracked monthly county-specific P&I rates for four influenza seasons, and depicted the spatio-temporal distribution of rates using a series of dynamic maps.
Compiled a county-specific dataset of monthly values of ambient temperature and climate descriptors.

The highest annual rates of P&I hospitalization and the most pronounced dynamic seasonal change were observed in counties with large gradient in ambient temperature, the substantial fraction (≥17%) of older people (RR=1.2; CI95%:1.17, 1.23), and big proportion of rural population (≥75%) (RR=1.27; CI95%:1.25, 1.28).
For both males and females, the P&I rates increase exponentially with age at approximately 72, 83, and 95.
The estimated rates were 14.76, 27.55, and 67.97 cases per year at 65, 75, and 85 respectively per 1000 elderly.
The normalized age-specific curves in patients with cancer and diabetes exhibited steeper increase in age range of 65-90

We suggest that the variations in incidence of P&I in the frail elderly subpopulations stem from differences between individual "immunological" and "chronological" age influenced by the presence of environmental stressors and preexisting medical conditions.