Abstract
OBJECTIVE AND RATIONALE: Hospital-acquired falls are common and serious adverse events in medical institutions, with high incidence and injury rates. Studying the occurrence patterns of hospital-acquired falls is important for preventing falls in hospitalized patients. However, the effect of meteorological factors on hospital-acquired falls has not been elucidated. Therefore, this study explored the impact of meteorological parameters on hospital-acquired falls in Chongqing, China, and provided new ideas for the clinical prevention of falls in patients. METHODS: Correlation analysis and distributed lag nonlinear models were employed to analyze the relationship between 3890 cases of hospital-acquired falls and meteorological data in 13 hospitals in 11 districts and counties in Chongqing from January 2013 to April 2023. RESULTS: The number of hospital-acquired falls demonstrated a nonlinear correlation with the daily average relative humidity and negatively correlated with sunshine duration; however, temperature, air pressure, and wind speed were not correlated. Compared to the reference humidity (87 %), the immediate effects of daily average relative humidity (65-68 % and 90-97 %) increased the risk of hospital-acquired falls on the same day (relative risk [RR]:1.027-1.243). When the daily average relative humidity was 95-97 %, lags of 0-1 d and 8-12 d had greater effects on falls (RR:1.073-1.243). The daily average relative humidities of 62-74 % and 91-97 % were statistically significant at cumulative relative risk (CRR)of 4, 7, 10, and 14 d with a cumulative lag (CRR: 1.111-4.277). On sex and age stratification, the lag and cumulative effects of relative humidity more significantly impacted falls in women and patients aged ≥65 years. CONCLUSION: Daily average relative humidity had a nonlinear correlation and lag effect on hospital-acquired falls; therefore, medical institutions should pay attention to the effect of relative humidity on hospital-acquired falls in patients, especially old and female patients.