Relation between temperature and heatstroke ambulance transport in Japan

Abstract

Listen

Translate article

PDS 66: Climate change, Exhibition Hall (PDS), Ground floor, August 28, 2019, 1:30 PM - 3:00 PM Background/Aim: The temperature-mortality relationship has been well described in multi-city studies. However, temperature effect on heatstroke incidence has not been well analyzed in the multi-city setting. We aim to investigate the temperature-heatstroke morbidity using Japanese ambulance transport data. Methods: We collected data from 47 prefectures in Japan between 2008 and 2018; ambulance transport data from the Fire and Disaster Management Agency and weather data from the Meteorological Agency. We restricted the season for summer (July, August and September), because Heatstroke data were not available outside of this season to analyze through the observation period. We applied a distributed lag non-linear model with 7 days of lag using quasi-Poisson regression, expecting a relatively short lag for heatstroke. Results: In most of the prefectures, the relation was inverse L-shaped; the risk is almost null up to a certain temperature and it increased monotonously as the temperature becomes higher. Some prefectures, mostly but not restricted to southern ones, had obvious “maximum morbidity temperature.” So far we found no clue to explain this phenomenon. Unlike temperature-mortality relation, the lag effect existed only on lag 0 day. In some prefectures, there were small amount of residual risk for lag 1+ days, but the amount was negligible. This lag phenomenon was virtually unchanged when we controlled for year-trend and relative humidity. Conclusion: The relation between heatstroke ambulance transport and daily maximum temperature showed inversed L-shape. In some prefectures, there was a maximum morbidity temperature. The lag was mainly on lag 0 day, partly because we used ambulance transport data, in which instance only very acute cases may have been brought to hospitals via ambulance cars.