Annals of epidemiology · 2025
Algorithm development for the automation of death certificate analysis and coding.
СтатьяПоддержка NIH: внешняя программаНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
McCord Kaylee, Gause Emma L, McLone Suzanne, Bard Hannah, Gilson Thomas, Jay Jonathan
Аннотация
PURPOSE: We developed an open-source, rule-based algorithm to automate cause of death coding for analyzing mortality in understudied populations, such as people experiencing homelessness, and dynamic public health crises including overdoses and climate-related deaths.
METHODS: Death categories of immediate public health concern were selected and keyword lists representing each category were developed in consultation with a domain expert. A rule-based keyword matching algorithm was built to assign death records into the selected death categories. The algorithm was trained on death certificate data from five counties across the United States. A case study applying the algorithm to deaths among people experiencing homelessness in Clark County, NV from 2015 to 2018 (N = 646) tested the accuracy of the program against a manual coder.
RESULTS: There was strong agreement between the algorithm and the manual coder in the all-cause identification (κ 0.905) and mutually exclusive sorting (κ 0.853) methods. Our findings illustrate the algorithm's ability to accurately classify death certificates into useful categories.
CONCLUSION: This open-source, customizable algorithm may be utilized by researchers, journalists, and others to conduct mortality analyses with publicly available death certificate data, bridging gaps in existing mortality tracking efforts.
METHODS: Death categories of immediate public health concern were selected and keyword lists representing each category were developed in consultation with a domain expert. A rule-based keyword matching algorithm was built to assign death records into the selected death categories. The algorithm was trained on death certificate data from five counties across the United States. A case study applying the algorithm to deaths among people experiencing homelessness in Clark County, NV from 2015 to 2018 (N = 646) tested the accuracy of the program against a manual coder.
RESULTS: There was strong agreement between the algorithm and the manual coder in the all-cause identification (κ 0.905) and mutually exclusive sorting (κ 0.853) methods. Our findings illustrate the algorithm's ability to accurately classify death certificates into useful categories.
CONCLUSION: This open-source, customizable algorithm may be utilized by researchers, journalists, and others to conduct mortality analyses with publicly available death certificate data, bridging gaps in existing mortality tracking efforts.