1Department of Medical Informatics, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands. 2Student Research Committee, Mashhad University of Medical Sciences, Mashhad, Iran. 3Cancer Informatics Department, Breast Cancer Research Center, ACECR, Tehran, Iran. 4Pharmaceutical Research Center, School of Pharmacy, Mashhad University of Medical Sciences, Mashhad, Iran. 5Department of Medical Informatics, Faculty of Medicine, Mashhad University of Medical Sciences, Mashhad, Iran. 6Clinical Research Unit, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands. 7Department of Intensive Care, Leiden University Medical Center, Leiden, The Netherlands. 8Health e-Research Centre, Division of Imaging, Informatics and Data Sciences, School of Health Sciences, The University of Manchester, Manchester, United Kingdom.
Crit Care Med. 2017 Feb;45(2):e222-e231. doi: 10.1097/CCM.0000000000002054.
We systematically reviewed models to predict adult ICU length of stay.
We searched the Ovid EMBASE and MEDLINE databases for studies on the development or validation of ICU length of stay prediction models.
We identified 11 studies describing the development of 31 prediction models and three describing external validation of one of these models.
Clinicians use ICU length of stay predictions for planning ICU capacity, identifying unexpectedly long ICU length of stay, and benchmarking ICUs. We required the model variables to have been published and for the models to be free of organizational characteristics and to produce accurate predictions, as assessed by R across patients for planning and identifying unexpectedly long ICU length of stay and across ICUs for benchmarking, with low calibration bias. We assessed the reporting quality using the Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies.
The number of admissions ranged from 253 to 178,503. Median ICU length of stay was between 2 and 6.9 days. Two studies had not published model variables and three included organizational characteristics. None of the models produced predictions with low bias. The R was 0.05-0.28 across patients and 0.01-0.64 across ICUs. The reporting scores ranged from 49 of 78 to 60 of 78 and the methodologic scores from 12 of 22 to 16 of 22.
No models completely satisfy our requirements for planning, identifying unexpectedly long ICU length of stay, or for benchmarking purposes. Physicians using these models to predict ICU length of stay should interpret them with reservation.
我们系统地回顾了预测成人 ICU 住院时间的模型。
我们在 Ovid EMBASE 和 MEDLINE 数据库中搜索了关于 ICU 住院时间预测模型的开发或验证的研究。
我们确定了 11 项研究，描述了 31 个预测模型的开发和 3 项研究，描述了其中一个模型的外部验证。
临床医生使用 ICU 住院时间预测来规划 ICU 容量，识别意外的 ICU 住院时间延长，并对 ICU 进行基准测试。我们要求模型变量已发布，并且模型不受组织特征的影响，并根据 R 对患者进行规划和识别意外的 ICU 住院时间延长以及对 ICU 进行基准测试，以评估准确性，并且校准偏差较低。我们使用预测模型研究的关键评估和数据提取清单来评估报告质量。
入院人数从 253 人到 178503 人不等。ICU 住院时间中位数为 2 至 6.9 天。有两项研究未公布模型变量，有三项研究包括组织特征。没有一个模型产生了低偏差的预测结果。R 在患者中的范围为 0.05-0.28，在 ICU 中的范围为 0.01-0.64。报告评分范围为 78 分中的 49 分至 78 分中的 60 分，方法评分范围为 22 分中的 12 分至 22 分中的 16 分。
没有任何模型完全满足我们在规划、识别意外的 ICU 住院时间延长或进行基准测试方面的要求。医生使用这些模型来预测 ICU 住院时间时，应持保留态度。
