Correction: Temperature extremes and infant mortality in Bangladesh: Hotter months, lower mortality (PLoS One (2018) 13:1 (e0189252) DOI: 10.1371/journal.pone.0189252)

Abstract

In Table 1, values for the mean “Monthly under 5 mortality” variable and mean “Monthly death count(age30days)” variable are incorrect. Please see the complete, correct Table 1 here. (Table Presented) There are errors in the S1 Table and its caption. The correct caption is: “S1 Table. Models of mean temperature effects on child mortality and gender mortality. Monthly under 5 mortality ratio ((U5MR); Deaths before 60 months of age per 1000 live births) or gender mortality (count of monthly deaths before 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC.” Please view the complete, correct S1 Table below. There are errors in the S2 Table and its caption. The correct caption is: “S2 Table. Models of maximum temperature effects on child mortality and gender mortality. Monthly under 5 mortality ratio (Deaths before 60 months of age per 1000 live births) or gender mortality (count of monthly deaths before 153 days) regressed on MAXIMUM monthly temp and MAXIMUM temp in the prior month. All models use first differences of all variables to correct for non-stationarity.” Please view the complete, correct S2 Table below. There are errors in the S3 Table caption. The correct caption is: “S3 Table. Models of mean temperature effects on neonatal and post neonatal. Monthly neonatal mortality (Death count before 1 month) and monthly post neonatal mortality (Death count between 30 and 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC. Both sexes analysed together.” There are errors in the S4 Table caption. The correct caption is: “S4 Table. Models of maximum temperature effects on neonatal and post neonatal. Monthly neonatal mortality (Deaths before 1 month) and monthly post neonatal mortality (Death count between 30 and 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC. Both sexes analysed together.” There are errors in the S5 Table and its caption. The correct caption is: “S5 Table. ARIMA AIC rankings for model residuals at lag 0. Table S5 below shows the AIC of several ARMA models fitted to the residuals (e1) of the various regressions at time lag = 0” Please view the complete, correct S5 Table below. There are errors in the S6 Table and its caption. The correct caption is: “S6 Table. ARIMA AIC rankings for model residuals at lag 1. Table S6 below shows AIC of ARMA models fitted to the residuals (e2) of the various regressions at time lag = 1” Please view the complete, correct S6 Table below. There are errors in S1 File. Please view the complete, correct S1 File below. Supporting information S1 Table. Models of mean temperature effects on child mortality and gender mortality. Monthly under 5 mortality ratio ((U5MR); Deaths before 60 months of age per 1000 live births) or gender mortality (count of monthly deaths before 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC. (DOCX) S2 Table. Models of maximum temperature effects on child mortality and gender mortality. Monthly under 5 mortality ratio (Deaths before 60 months of age per 1000 live births) or gender mortality (count of monthly deaths before 153 days) regressed on MAXIMUM monthly temp and MAXIMUM temp in the prior month. All models use first differences of all variables to correct for non-stationarity. (DOCX) S3 Table. Models of mean temperature effects on neonatal and post neonatal. Monthly neonatal mortality (Death count before 1 month) and monthly post neonatal mortality (Death count between 30 and 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC. Both sexes analysed together. (DOCX) S4 Table. Models of maximum temperature effects on neonatal and post neonatal. Monthly neonatal mortality (Deaths before 1 month) and monthly post neonatal mortality (Death count between 30 and 153 days) regressed on MEAN monthly temp and MEAN temp in the prior month. All models use first differences of all variables to correct for non- stationarity. ARIMA terms included to minimize AIC. Both sexes analysed together. (DOCX) S5 Table. ARIMA AIC rankings for model residuals at lag 0. Table S5 below shows the AIC of several ARMA models fitted to the residuals (e1) of the various regressions at time lag = 0 (DOCX) S6 Table. ARIMA AIC rankings for model residuals at lag 1. Table S6 below shows AIC of ARMA models fitted to the residuals (e2) of the various regressions at time lag = 1 (DOCX) S1 File. Data used in the analysis of mortality and temperature from Matlab, Bangladesh. Monthly reports of temperature, maximum temperature, minumum temperature, infant mortality, female infant mortality, male infant mortality, male mortality less than 30 days, male mortality greater than 30 days, female mortality less than 30 days, female mortality greater than 30 days, both sex mortality less than 30 days, and both sex mortality greater than 30 days. (XLS).