中国血液净化

中国血液净化 ›› 2020, Vol. 19 ›› Issue (11): 721-725.doi: 10.3969/j.issn.1671-4091.2020.11.001

• 临床研究 •    下一篇

宁波市鄞州区透析患者30 天再入院影响因素分析及简易评分工具构建

孙小宇1,聂振禹2,孙烨祥3,包蓓艳2,林鸿波3,张路霞1,4   

  1. 1北京大学健康医疗大数据国家研究院学习型智慧医疗体系研究中心
    2宁波市泌尿肾病医院肾内科
    3宁波市鄞州区疾病预防控制中心
    4北京大学第一医院肾内科
  • 收稿日期:2020-07-21 修回日期:2020-09-04 出版日期:2020-11-12 发布日期:2020-11-05
  • 通讯作者: 包蓓艳 baobeiyan2007@sina.com 林鸿波 lin673160@163.com E-mail:lin673160@163.com
  • 基金资助:
    北京市高精尖学科建设项目(BMU2019GJJXK001)

Prediction of readmission within 30 days in maintenance dialysis patients using scoring system developed from electronic health record data in Yinzhou, China

  1. 1National Institute of Health Data Science at Peking University, Beijing 100191, China;  2Division of Nephrology, The Second Ningbo Yinzhou Hospital, Ningbo 315211, China;  3Center for Disease Control and Prevention, Ningbo 315100; 4Renal Division, Peking University First Hospital, Beijing 100034, China.
  • Received:2020-07-21 Revised:2020-09-04 Online:2020-11-12 Published:2020-11-05
  • Contact: Hongbo Lin E-mail:lin673160@163.com

PDF

955

摘要： 【摘要】目的30 天再入院是反映疾病转归和医疗质量的重要指标。本研究通过分析宁波市鄞州区健康医疗大数据平台数据，探讨透析患者发生30 天再入院的影响因素、估算效应大小并建立简易评分工具。方法纳入2010 年1 月～2020 年1 月在鄞州区医疗机构进行过住院治疗，未发生院内死亡且出院诊断ICD 编码符合透析相关诊断的18 岁以上患者。采集患者年龄，性别，住院记录，出院诊断等信息。对于有多次住院记录的患者，随机选择一次作为指示住院记录。统计模型采用COX 比例风险模型。结果最终纳入患者1614 例，发生30 天再入院患者348 例(21.56%)。构建多因素COX 比例风险模型显示，与患者30 天再入院相关的因素为高查尔斯合并症指数(Charles comorbidity index, CCI)评分(对比CCI 2分组, 3～4 分组风险比HR=1.257, 95% CI: 0.917～1.724, ≥5 分组HR= 1.848, 95% CI: 1.374～2.486)、既往半年内住院次数(HR=1.576, 95% CI: 1.471～1.688)、本次住院时间(HR=1.010，95% CI：1.001～1.020)、腹膜透析(对比血液透析，HR=1.505，95% CI：1.207～1.876)。利用以上因素构建简易评分工具，可通过患者评分对应再入院发生概率。结论对宁波市鄞州区住院透析患者的分析显示，高CCI 评分、既往半年内住院次数多、住院时间长、腹膜透析患者发生30 天再入院的风险更高，应用简易评分工具可用于临床中对透析患者再入院风险进行评估。

关键词: 30 天再入院, 透析, CCI 评分, 电子医疗记录, 风险预测

Abstract: 【Abstract】Objective Readmission within 30 days after hospital discharge is an important indicator for disease prognosis and health care quality. This study aims to explore risk factors associated with 30-day readmission in maintenance dialysis patients and to build score sheets for risk prediction. Methods Data were taken from patients receiving maintenance dialysis and discharged from hospitals located in Yinzhou, Ningbo, between January 2010 and January 2020. Age, sex, hospitalization records and diagnoses on discharge were collected. The outcome was the readmission within 30 days after discharge of the index hospitalization. The index hospitalization was arbitrarily selected when there were multiple hospitalizations for a patient. Charles
Comorbidity Index (CCI) was calculated as an indicator of multi- comorbidities status. All patients had 2 scores due to their status of dialysis. COX proportional hazard regression model was utilized to explore the potential risk factors. Score sheets were developed to predict readmission from the β-coefficients of COX proportional hazards model. Results Among the 1614 patients, 348 (21.56%) had a 30-day hospital readmission.Multivariable- adjusted Cox proportional hazard model showed that higher CCI (CCI 3~4 vs. CCI 2, HR=1.257, 95% CI: 0.917~1.724; CCI ≥5 vs. CCI2, HR=1.848, 95%CI:1.374~2.486), more hospitalizations in the prior 6 months (HR=1.576, 95% CI: 1.471~1.688), inpatient days (HR=1.010, 95%CI:1.001~1.020), peritoneal dialysis (HR= 1.505, 95%CI: 1.207~1.876). These variables were used to build score sheets to predict the readmission risk. Conclusions CCI, multiple hospitalizations in the prior 6 months, inpatient days and peritoneal dialysis were associated with 30-day readmission in patients receiving maintenance dialysis, and could be used for risk prediction.

Key words: 30-day readmission, Dialysis, Charles Comorbidity Index, Electronic health record data, Prediction

中图分类号: