得克萨斯州农村医院关闭与时间敏感性疾病的死亡率
该县-年队列研究纳入得州175个县、2006—2019年数据,采用相邻县暴露与医院服务区(HSA)暴露两种双重差分设计,比较医院关闭前后3—5年时间敏感性疾病(急性心肌梗死、卒中、脓毒症、哮喘或COPD)的县级年死亡率,校正县固定效应、年固定效应及随时间变化的县级社会人口学特征。研究期内14所农村医院关闭。相邻县设计ATT为每10万居民每年增加11.60例死亡(95%CI 4.62-18.58;P=.001),HSA设计ATT为11.52(95%CI 0.46-22.58;P=.04),相当于基线死亡率升高5.4%。补充分析提示死亡关联集中于急性急症,采用替代时间敏感性疾病定义时仍持续(相邻县ATT 8.21/10万,95%CI 2.08-14.34),但在更宽泛定义下不显著(ATT 6.85/10万,95%CI -2.31至16.01)。
要点
- 在得州175个县中,农村医院关闭与受影响县每年每10万居民增加约12例时间敏感性疾病死亡相关(相邻县ATT 11.60,95%CI 4.62-18.58;HSA ATT 11.52,95%CI 0.46-22.58)。
- 该增幅相当于受影响县基线死亡率的5.4%,研究期间共14所农村医院关闭(约占得州农村医院的8%)。
- 死亡关联主要集中在急性急症(急性心肌梗死、卒中、脓毒症、哮喘或COPD),采用替代时间敏感性疾病定义时仍持续(相邻县ATT 8.21/10万,95%CI 2.08-14.34)。
- 更宽泛定义下关联不显著(相邻县ATT 6.85/10万,95%CI -2.31至16.01),提示关联对结局定义敏感。
- 结果在多种敏感性分析中稳健,采用相邻县暴露与HSA暴露两种设计方向一致。
适用人群与场景
- 得克萨斯州175个县(相邻县设计:71个处理县、104个对照县;HSA设计:53个处理县、122个对照县),研究时段2006年1月1日至2019年12月31日。
- 暴露定义为居住在发生农村医院关闭的县或邻近受影响区域(相邻县暴露或HSA暴露),非个体患者层面。
诊断标准与关键阈值
- 原文未提及
处置流程与药物剂量
- 原文未提及
红旗征与不能漏的情况
- 原文未提及
研究设计与关键结果
- 研究类型:县-年队列研究,采用两种互补的双重差分设计(相邻县暴露和HSA暴露),数据覆盖2006—2019年得州175个县,共14所农村医院关闭。
- 主要终点:每10万居民年时间敏感性疾病(急性心肌梗死、卒中、脓毒症、哮喘或COPD)死亡数;相邻县设计ATT 11.60(95%CI 4.62-18.58;P=.001),HSA设计ATT 11.52(95%CI 0.46-22.58;P=.04),相当于基线死亡率升高5.4%。
- 补充分析:替代时间敏感性疾病定义下相邻县ATT 8.21/10万(95%CI 2.08-14.34),更宽泛定义下ATT 6.85/10万(95%CI -2.31至16.01),关联集中于急性急症。
- 结果对多种敏感性分析稳健。
证据类型、等级与局限
- 证据类型:回顾性县-年队列研究(双重差分设计),非随机对照试验。
- 局限:观察性设计不能排除未测量混杂;暴露为县级生态学指标而非个体水平;研究仅限得州,外推需谨慎;部分设计置信区间较宽(HSA设计下限0.46)。
对床旁实践的改变
- 提示农村医院关闭后,急诊与ICU团队应关注时间敏感性疾病(急性心肌梗死、卒中、脓毒症、哮喘或COPD)的院前延误与转运风险,但原文未提供具体处置方案。
- 适用人群为受影响县居民;不适用于个体化临床决策,不能据此改变具体药物或干预措施。
- 常见陷阱:将县级生态学关联误读为个体因果效应;将补充分析中不显著的宽泛定义结果等同于无风险。
指南或共识出处
- JAMA Netw Open
以上提炼依据论文摘要(全文未开放获取,或许可不允许翻译转载);全文请点「阅读原文」。
英文原文(摘要)· 点击展开
IMPORTANCE: Rural hospital closures in the US have accelerated since 2010, with Texas experiencing more rural closures than any other state. Whether closures are associated with population-level mortality remains unclear, and rural hospital closures in Texas have not been examined.
OBJECTIVE: To examine the association between rural hospital closures in Texas and county-level mortality from 4 time-sensitive conditions: acute myocardial infarction, stroke, sepsis, and asthma or chronic obstructive pulmonary disease.
DESIGN, SETTING, AND PARTICIPANTS: This county-year cohort study of Texas counties used 2 complementary difference-in-difference study designs with data from January 1, 2006, to December 31, 2019. Statistical analyses were performed from November 2025 to June 2026.
EXPOSURE: Residence in a county where a rural hospital closed or in the nearby affected area, assessed using 2 designs: adjacent-county exposure (a closure county or a county bordering a closure county) and Hospital Service Area (HSA)-based exposure (a county being within an HSA where a closure occurred).
MAIN OUTCOMES AND MEASURES: The main outcome was annual deaths per 100 000 residents from time-sensitive conditions (acute myocardial infarction, stroke, sepsis, and asthma or chronic obstructive pulmonary disease). County-level mortality rates were compared for treated and control counties for 3 to 5 years before and after closure. Regressions were adjusted for county fixed effects, year fixed effects, and time-varying sociodemographic county characteristics.
RESULTS: In the sample of 175 counties (71 treated and 104 control counties in the adjacent-county design and 53 treated and 122 control counties in the HSA treatment design), 14 rural hospitals (approximately 8% of rural hospitals in Texas) closed during the study period. Closures were associated with 12 additional annual time-sensitive deaths per 100 000 residents in an affected county (adjacent-county treatment: average treatment effect on the treated [ATT], 11.60 [95% CI, 4.62-18.58]; P = .001; HSA treatment: ATT, 11.52 [95% CI, 0.46-22.58]; P = .04). This represented a 5.4% increase compared with the baseline mortality rate in affected counties. Estimates were robust to several sensitivity analyses. Supplementary analyses suggested that mortality associations were concentrated in acute emergencies, persisting under an alternate definition of time-sensitive conditions (adjacent-county ATT, 8.21 [95% CI, 2.08-14.34] per 100 000 residents) but not under a broader definition (adjacent-county ATT, 6.85 [95% CI, -2.31 to 16.01] per 100 000 residents).
CONCLUSIONS: In this cohort study, rural hospital closures in Texas were associated with increased population-level mortality from time-sensitive conditions. The findings suggest that there were adverse patient outcomes in the aftermath of rural hospital closures, emphasizing a need to manage the emergency response and the progressive deterioration of clinical infrastructure and patient health.
参考来源
- Poisler MA, Gujral K. Rural Hospital Closures and Mortality From Time-Sensitive Conditions in Texas. JAMA Netw Open. 2026. PMID: 42804204. doi:10.1001/jamanetworkopen.2026.36325
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。