Background: Many studies have investigated the association between schizophrenia and daily mean temperature, but the conclusions were not consistent. Apparent temperature (AT), a comprehensive indicator of temperature and humidity, reflected human thermal sensation more objectively than temperature. We aimed to investigate the relationship between high/low AT and schizophrenia hospitalizations in Hefei, China, and further to identify vulnerable subpopulations.
Methods: Poisson generalized linear regression model combined with distributed lag non-linear model (DLNM) was applied to quantify the relationship between AT and schizophrenia hospitalizations, compared with the AT of minimum admissions (3.3 °C). Stratified analysis was conducted by sex, marital status and age. Separate models for AT and temperature were compared as well.
Results: The high AT (90th, 28.3 °C) exerted the greatest effect at lag0, with RR = 1.062 (95%CI, 1.019-1.106). The effect of the low AT (10th, -4.7 °C) was first observed at lag3 (RR = 1.007, 95%CI: 1.000-1.015), and increased to the maximum at lag7 (RR = 1.018, 95%CI: 1.009-1.027). The male patients, the married patients and patients aged 41-60 years appeared to be more vulnerable to both high and low ATs. The values of Akaike Information Criterion and mean squared error (MSE) in model with AT were smaller than that with temperature.
Conclusions: Both high and low ATs were associated with increase in schizophrenia hospitalizations. More protective measures should be taken to avoid adverse AT exposure in susceptible groups.
Keywords: Apparent temperature; Schizophrenia; Time-series analysis.
Copyright © 2019. Published by Elsevier B.V.