Woldhek Annemarie L, Rijkenberg Saskia, Bosman Rob J, van der Voort Peter H J
Dept of intensive care, OLVG hospital, Amsterdam, The Netherlands.
Dept of intensive care, OLVG hospital, Amsterdam, The Netherlands; TIAS School for Business and Society, Tilburg University, Tilburg, The Netherlands.
J Crit Care. 2017 Apr;38:328-334. doi: 10.1016/j.jcrc.2016.12.001. Epub 2016 Dec 6.
Readmission rate is frequently proposed as a quality indicator because it is related to both patient outcome and organizational efficiency. Currently available studies are not clear about modifiable factors as tools to reduce readmission rate.
In a 14year retrospective cohort study of 19,750 ICU admissions we identified 1378 readmissions (7%). A multivariate logistic regression analysis for determinants of readmission within 24h, 48h, 72h and any time during hospital admission was performed with adjustment for patients' characteristics and initial admission severity scores.
In all models with different time points, patients with older age, a medical and emergency surgery initial admission and patients with higher SOFA score have a higher risk of readmission. Immunodeficiency was a predictor only in the at any time model. Confirmed infection was predicted in all models except the 24h model. Last day noradrenaline treatment was predicted in the 24 and 48h model. Mechanical ventilation on admission independently protected for readmission, which can be explained by the large number of cardiac surgery patients. All multivariate models had a moderate performance with the highest AUC of 0.70.
Readmission can be predicted with moderate precision and independent variables associated with readmission are age, severity of disease, type of admission, infection, immunodeficiency and last day noradrenaline use. The latter factor is the only one that can be modified and therefore readmission rate does not meet the criteria to be used as a useful quality indicator.
再入院率常被用作质量指标，因为它与患者预后和组织效率都相关。目前的研究对于可作为降低再入院率手段的可改变因素并不明确。
在一项对19750例重症监护病房（ICU）入院病例进行的14年回顾性队列研究中，我们确定了1378例再入院病例（7%）。对患者特征和初始入院严重程度评分进行调整后，对24小时、48小时、72小时以及住院期间任何时间的再入院决定因素进行多因素逻辑回归分析。
在所有不同时间点的模型中，年龄较大、初始入院为内科和急诊手术以及序贯器官衰竭评估（SOFA）评分较高的患者再入院风险较高。免疫缺陷仅在任何时间模型中是一个预测因素。除24小时模型外，在所有模型中均可预测确诊感染。在24小时和48小时模型中可预测最后一天使用去甲肾上腺素治疗。入院时机械通气可独立预防再入院，这可以用大量心脏手术患者来解释。所有多因素模型的表现中等，最高曲线下面积（AUC）为0.70。
再入院可以得到中等精度的预测，与再入院相关的独立变量包括年龄、疾病严重程度、入院类型、感染、免疫缺陷和最后一天使用去甲肾上腺素。后一个因素是唯一可以改变的因素，因此再入院率不符合用作有用质量指标的标准。
