胰腺手术死亡率的医院层面决定因素:全球基础设施分析
该研究将2021年PancreasGroup.org全球前瞻性队列的患者结局数据与全球调查的医院基础设施数据链接,纳入60个国家214个中心的2928例患者,构建医院资源指数(HRI)。结果显示6项医院因素与较低死亡率显著相关,HRI每增加1分与死亡 odds 降低12%相关(OR 0.88,95%CI 0.81-0.95,P=0.001),且独立于病例量。
要点
- 6项可干预医院因素与胰腺手术后较低死亡率独立相关。
- HRI每增加1分,90天死亡 odds 降低12%(OR 0.88,95%CI 0.81-0.95)。
- 具备全部6项资源的医院较无任何资源者死亡 odds 降低54%(OR 0.46,95%CI 0.30-0.70)。
核心结论
- HRI对胰腺手术死亡率显示累积保护效应,病例量在调整模型中仍独立保护。
- 结果为超越集中化、通过资源分配改善胰腺手术结局提供可操作目标。
研究设计与样本
- 将2021年PancreasGroup.org全球前瞻性队列的患者结局数据与全球调查的医院基础设施数据链接。
- 评估20项医院层面因素的二元可及性,将单变量多水平回归中与死亡率显著相关的因素合并为加性HRI。
- 采用多变量logistic回归评估HRI与90天死亡率的关联。
- 分析纳入60个国家214个中心的2928例患者。
主要发现
- 6项医院因素与较低死亡率显著相关:介入放射/内镜可及性、结构化营养支持、循证实践采用、员工培训、设备维护、有利的护患比。
- HRI每增加1分与死亡 odds 降低12%相关(OR 0.88,95%CI 0.81-0.95,P=0.001),独立于病例量、国家发展指数及患者和手术特征。
- 病例量在调整模型中独立保护(OR 0.73,P=0.014,最佳截断值每年55例)。
- 具备全部6项HRI资源的医院患者90天死亡调整 odds 较无任何资源者低54%(OR 0.46,95%CI 0.30-0.70)。
关键数值与推荐
- HRI由6项医院因素组成:介入放射/内镜可及性、结构化营养支持、循证实践采用、员工培训、设备维护、有利的护患比。
- HRI每增加1分,90天死亡 odds 降低12%(OR 0.88,95%CI 0.81-0.95,P=0.001)。
- 全部6项资源 vs 无资源:90天死亡调整 OR 0.46(95%CI 0.30-0.70)。
- 病例量最佳截断值为每年55例(OR 0.73,P=0.014)。
临床怎么用
- 除病例量集中化外,应优先改善介入放射/内镜可及性、结构化营养支持、循证实践、员工培训、设备维护和护患比。
- 可采用HRI评估医院资源短板,指导资源分配以降低胰腺手术死亡率。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVE: To identify modifiable hospital-level determinants of mortality after pancreas surgery and assess their cumulative protective effects through the development of a Hospital Resource Index (HRI) on a global scale. BACKGROUND: Ninety-day mortality rates after pancreas surgery vary globally (2%-10%); however, institutional factors affecting mortality remain poorly understood. Although caseload volume has driven centralization policies in many health care systems, it is not fully explanatory, and the impact of specific modifiable structural and process-related hospital factors remains unclear. METHODS: We linked patient outcome data from the 2021 PancreasGroup.org global prospective cohort to hospital infrastructure data from the PancreasGroup.org global survey. Binary availability of 20 hospital-level factors was assessed. Factors significantly associated with mortality on univariable multilevel regression were combined into an additive HRI. Multivariable logistic regression evaluated associations between HRI and 90-day mortality. RESULTS: Analyses included 2928 patients from 214 centers across 60 countries. Six hospital factors were significantly associated with lower mortality, including interventional radiology/endoscopy availability, structured nutrition support, evidence-based practice adoption, staff training, equipment maintenance, and favorable nurse-patient ratios. Each additional HRI point was associated with 12% lower mortality odds (OR: 0.88, 95% CI: 0.81-0.95, P =0.001), independent of caseload volume (OR: 0.73, P =0.014, optimal cutoff 55 cases per year), country's development index, and patient and procedural characteristics. Patients treated at hospitals with all 6 HRI resources had 54% lower adjusted odds of 90-day mortality than those at hospitals with none (OR 0.46, 95% CI 0.30-0.70). CONCLUSIONS: The developed HRI showed cumulative protective effects on mortality after pancreas surgery, with volume remaining independently protective in the adjusted model, providing actionable targets for resource allocation to enhance pancreas surgery outcomes beyond centralization globally.
引用
- Hidalgo Salinas C, Tinguely P, Raptis DA, et al. Hospital-Level Determinants of Mortality After Pancreas Surgery: A Global Infrastructure Analysis. Ann Surg. 2026. PMID: 42711758. doi:10.1097/SLA.0000000000007196
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。