Temperature-Related Hospitalization Burden under Climate Change (Formal)

This section outlines the data and computational implementations associated with the  main results in the research 'Temperature-Related Hospitalization Burden under Climate Change'. The analyses were primarily conducted using R version 4.3.2, with the following package versions: DLNM (2.4.7), dplyr (1.1.4), splines (4.3.2), tsModel (0.6-2), mvmeta (1.0.3), lubridate (1.9.3), stringr (1.5.1), readxl (1.4.3), forecast (8.24.0). 

The key scripts used in the analysis include:

1. Data Description

(1).ChinaCity_DailyHospitalizationCases_2021-2023.xlsx This file contains daily hospital admission counts for five climate-sensitive diseases across 301 cities in China from 2021 to 2023, used in this study.

(2)ChinaCity_FiveDisease_HospitalizationCost_2021-2023.xlsx This file provides the average hospitalization costs for the  five diseases in 301 Chinese cities during the 2021–2023 period.

2. Code Description

(1)Historical Temperature-Related Hospitalization Risk Analysis.R – Conducts a historical analysis of the relationship between temperature and hospitalizations using the NLNM (Nonlinear Lagged Model) approach.

(2)Temperature_thresholds-days_calculations.R – Implements extreme weather projections based on three different temperature threshold calculation methods.

(3)Project_Hospitalization_risk.R – Builds on Historical Temperature-Related Hospitalization Risk Analysis.R to project health risks under extreme high and low temperatures for each city from 2030 to 2100. (correct:  pred <- crosspred(bvar_new, coef = blup_coef, vcov = blup_vcov,
                    model.link = "log", at = temp_for_analysis,...).

(4)attr.R-from the dlnm ecosystem/extension functions.

The computational updates of this series of functions will be released in the near future, enabling researchers to extend the application of the method.

Article Title : Temperature-related hospitalization burden under climate change
DOI : 10.1038/s41586-025-09352-w