Correction to: Modification of the PM2.5- and extreme heat-mortality relationships by historical redlining: a case-crossover study in thirteen U.S. states (Environmental Health, (2024), 23, 1, (16), 10.1186/s12940-024-01055-5)

Edgar Castro, Abbie Liu, Yaguang Wei, Anna Kosheleva, Joel Schwartz

Research output: Contribution to journalComment/debate

Abstract

Following publication of [1], errors were found in the code used to prepare the cohort for a case-crossover analysis and the resulting data that was used for the analysis. Despite these errors, results were only marginally effected and all conclusions remain the same. A few typos were also found in the manuscript. A table of all affected texts is shown below. (Table presented.) Section Lines Text Abstract 48-51 Individuals who lived in redlined areas had an interaction odds ratio for mortality of 1.00931.0104 (95% confidence interval [CI]: 1.00841.0095, 1.01011.0114) for each 10 µg m-3 increase in same-day ambient PM2.5 compared to individuals who did not live in redlined areas. For extreme heat, the interaction odds ratio was 1.02181.0146 (95% CI 1.00311.0039, 1.04081.0457). Methods 159-161 To derive measures of extreme heat, we first calculated various percentiles of minimum temperature in each block group in each year. For our main analysis, we considered the 95th90th percentile. Methods 163-165 In other words, if the minimum temperature on a certain day met or exceeded the 95th90th percentile of minimum temperature in that block group in that year, then that day was marked as an extreme heat day. Results 229-237 We obtained 11,115,38011,076,020 mortality records from the twelvethirteen state departments of public health. From these records, we sequentially excluded 466,874453,754 deaths involving external causes; 139,908133,348 deaths involving individuals younger than 18 years old; 196,558 deaths with geocodes that were missing or coarser than block group-level; 331 deaths involving individuals whose home locations were outside of the state that reported their death; 1,392,4231,372,743 deaths before January 5th, 2001 or after December 31st, 2016 and 537 deaths whose home block groups had a population of zero according to the preceding Decennial Census (for which 4-day moving averages of population-weighted PM2.5 could not be calculated); and 34,016 deaths with lag days from 0 to 4 that included December 31st on leap years (for which Daymet predictions are not available; Figure 3) Results 272-278 We found a significant interaction with exposure to any extreme heat (interaction odds ratio 1.02181.0246; 95% CI 1.00311.0039, 1.04081.0457) while we did not observe significant interactions for singleton heat events or when looking at length-specific exposures. In absolute terms, this amounts to a 2.157%2.434% (95% CI 0.307%0.386%, 4.036%4.521%) increase in the daily risk of death death from non-external causes by exposure to any extreme heat in historically-redlined neighborhoods compared to other neighborhoods. The highest overall effects were observed for exposure to any extreme heat, followed by 3, 12, and 21 consecutive days of extreme heat, respectively. Results 283-287 We found a significant interaction with same-day ambient PM2.5 (interaction odds ratio for each 10 µg/m-3 increase: 1.00931.0104; 95% CI 1.00841.0095, 1.01011.0114) while we did not observe interactions for different moving averages of ambient PM2.5. In absolute terms, this amounts to a 0.930%1.029% (95% CI 0.831%0.940%, 1.000%1.128%) increase in the daily risk of death from non-external causes for each 10 µg/m-3 increase in ambient PM2.5 in historically-redlined neighborhoods compared to other neighborhoods. Results 295-296 However, for PM2.5, we did observe that the interaction with same-day ambient PM2.5 was not significant for a population cutoffs of 50% and 99%. Results 302-305 We also observed that the 85th and 95th percentile cutoffs of minimum temperature had higher interactions than the 90th percentile cutoff on earlier days, with the 85th percentile being the highest for any exposure or the 1st day of extreme heat and the 99th percentile being the highest for any exposure or the 1st or 2nd days of extreme heat. Firstly, 39,360 deaths from the raw data were duplicated (0.36% of total deaths), resulting in an erroneous Fig. 1 and numbers in the accompanying paragraph. However, the total number of deaths used for the analysis remained the same, though this was misreported in the original Fig. 1 (884,733 deaths rather than 8,884,733). A few percentages were also incorrect. The corrected Fig. 1 is shown below next to the original Fig. 1, followed by a table of changes in the text. Original and corrected Fig. 1 with changed bolded. Additionally, code that was meant to restrict control days only to those occurring in the same month of the case was not run, resulting in a mean of 7.91 controls for each case (SD: 0.57). After corrections, this was reduced to a mean of 3.38 controls per case (SD: 0.49). Unlike the previous error, this slightly altered the results, though only marginally. The corrected text, figures, and coefficient tables are presented below alongside what was originally reported.

Original languageEnglish
Article number34
JournalEnvironmental Health: A Global Access Science Source
Volume23
Issue number1
DOIs
StatePublished - Dec 2024
Externally publishedYes

Cite this