Associations Between Natural Language Processing-Enriched Social Determinants of Health and Suicide Death Among US Veterans.
Associations Between Natural Language Processing-Enriched Social Determinants of Health and Suicide Death Among US Veterans.
复制标题
DOI：
10.1001/jamanetworkopen.2023.3079
复制
发表时间：
2023-03-01
影响因子：
13.8
通讯作者：
中科院分区：
文献类型：
--
作者：
从结构化和非结构化临床数据中提取的健康社会决定因素(SDOH)是否与美国退伍军人自杀死亡风险的增加有关？在这项病例对照研究中，来自结构化数据和非结构化数据(使用自然语言处理系统提取)的SDOH都与自杀死亡风险增加相关，该研究涉及8821例病例和35名 匹配的对照组。这项研究的结果表明，SDOH是美国退伍军人自杀的危险因素，可以利用自然语言处理从非结构化数据中提取SDOH信息。这项病例对照研究通过使用自然语言处理提取的电子健康记录中的结构化数据和非结构化数据确定的社会健康决定因素(SDOH)来评估退伍军人的自杀风险。众所周知，健康的社会决定因素(SDOH)与自杀行为风险的增加有关，但很少有研究使用非结构化电子健康记录笔记中的SDOH。调查退伍军人自杀死亡与最近使用结构化和非结构化数据确定的SDOH之间的关联。这项嵌套病例对照研究包括2010年10月1日至2015年9月30日期间接受美国退伍军人健康管理局护理的退伍军人。为了从非结构化的临床病历中提取SDOHs，开发了一个自然语言处理(NLP)系统。结构化数据产生了6个SDOH(即，社会或家庭问题、就业或财务问题、住房不稳定、法律问题、暴力和非特定的心理社会需求)，非结构化数据的NLP产生了8个SDOH(社会隔离、工作或经济不安全、住房不稳定、法律问题、护理障碍、暴力、护理过渡和食品不安全)，将它们结合起来产生了9个SDOH。对2022年5月的数据进行了分析。与未发生SDOHs相比，SDOHs出现的最长跨度为2年。自杀死亡病例在出生年份、队列进入日期、性别和随访持续时间等方面与4个对照组相匹配。自杀被国家死亡指数确定，患者在进入队列后进行了长达2年的随访，研究结束日期为2015年9月30日。使用条件Logistic回归估计调整后的优势比(AORS)和95%的CI。在6名 122 785退伍军人中，8 821人在23 725 382人年的随访中自杀(发病率为37.18/10万 000人年)。这8,821名退伍军人与35名 -284对照组参与者进行了匹配。队列主要是男性(42 540(96.45%))和白人(34 930(79.20%))，黑人退伍军人6 227人(14.12%)。平均年龄(SD)为58.64(17.41)岁。在5种常见的SDOH中，NLP提取的SDOH平均保留了49.92%的结构化SDOHs，覆盖了所有SDOHs的80.03%。通过结构化数据和/或NLP获得的SDOH与自杀风险的增加显著相关。当将结构化数据和自然语言处理相结合时，影响程度最大的3个SDOH分别是法律问题(AOR，2.66；95%CI，2.46-2.89)、暴力问题(AOR，2.12；95%CI，1.98-2.27)和非特定心理社会需求(AOR，2.07；95%CI，1.92-2.23)。在这项研究中，NLP提取的SDOH，无论有没有结构化SDOH，都与退伍军人中自杀风险的增加有关，这表明NLP在公共卫生研究中的潜在用途。
Are social determinants of health (SDOHs), extracted from both structured and unstructured clinical data, associated with an increased risk of suicide death among US veterans? In this case-control study of 8821 cases and 35 284 matched controls, SDOHs from both structured data and unstructured data (extracted using a natural language processing system) were associated with an increased risk of suicide death. The findings of this study suggest that SDOHs are risk factors for suicide among the US veterans and that natural language processing can be leveraged to extract SDOH information from unstructured data. This case-control study assesses suicide risk among veterans by social determinants of health (SDOHs) identified using both structured data and unstructured data in the electronic health record extracted using natural language processing. Social determinants of health (SDOHs) are known to be associated with increased risk of suicidal behaviors, but few studies use SDOHs from unstructured electronic health record notes. To investigate associations between veterans’ death by suicide and recent SDOHs, identified using structured and unstructured data. This nested case-control study included veterans who received care under the US Veterans Health Administration from October 1, 2010, to September 30, 2015. A natural language processing (NLP) system was developed to extract SDOHs from unstructured clinical notes. Structured data yielded 6 SDOHs (ie, social or familial problems, employment or financial problems, housing instability, legal problems, violence, and nonspecific psychosocial needs), NLP on unstructured data yielded 8 SDOHs (social isolation, job or financial insecurity, housing instability, legal problems, barriers to care, violence, transition of care, and food insecurity), and combining them yielded 9 SDOHs. Data were analyzed in May 2022. Occurrence of SDOHs over a maximum span of 2 years compared with no occurrence of SDOH. Cases of suicide death were matched with 4 controls on birth year, cohort entry date, sex, and duration of follow-up. Suicide was ascertained by National Death Index, and patients were followed up for up to 2 years after cohort entry with a study end date of September 30, 2015. Adjusted odds ratios (aORs) and 95% CIs were estimated using conditional logistic regression. Of 6 122 785 veterans, 8821 committed suicide during 23 725 382 person-years of follow-up (incidence rate 37.18 per 100 000 person-years). These 8821 veterans were matched with 35 284 control participants. The cohort was mostly male (42 540 [96.45%]) and White (34 930 [79.20%]), with 6227 (14.12%) Black veterans. The mean (SD) age was 58.64 (17.41) years. Across the 5 common SDOHs, NLP-extracted SDOH, on average, retained 49.92% of structured SDOHs and covered 80.03% of all SDOH occurrences. SDOHs, obtained by structured data and/or NLP, were significantly associated with increased risk of suicide. The 3 SDOHs with the largest effect sizes were legal problems (aOR, 2.66; 95% CI, 2.46-2.89), violence (aOR, 2.12; 95% CI, 1.98-2.27), and nonspecific psychosocial needs (aOR, 2.07; 95% CI, 1.92-2.23), when obtained by combining structured data and NLP. In this study, NLP-extracted SDOHs, with and without structured SDOHs, were associated with increased risk of suicide among veterans, suggesting the potential utility of NLP in public health studies.