Out-of-hours discharge from intensive care, in-hospital mortality and intensive care readmission rates: a systematic review and meta-analysis

Abstract

PurposeDischarge from an intensive care unit (ICU) out of hours is common. We undertook a systematic review and meta-analysis to explore the association between time of discharge and mortality/ICU readmission.MethodsWe searched Medline, Embase, Web of Knowledge, CINAHL, the Cochrane Library and OpenGrey to June 2017. We included studies reporting in-hospital mortality and/or ICU readmission rates by ICU discharge “out-of-hours” and “in-hours”. Inclusion was limited to patients aged ≥ 16 years discharged alive from a non-specialist ICU to a lower level of hospital care. Studies restricted to specific diseases were excluded. We assessed study quality using the Newcastle Ottowa Scale. We extracted published data, summarising using a random-effects meta-analysis.ResultsOur searches identified 1961 studies. We included unadjusted data from 1,191,178 patients from 18 cohort studies (presenting data from 1994 to 2014). “Out of hours” had multiple definitions, beginning between 16:00 and 22:00 and ending between 05:59 and 09:00. Patients discharged out of hours had higher in-hospital mortality [relative risk (95% CI) 1.39 (1.24, 1.57) p < 0.0001] and readmission rates [1·30 (1.19, 1.42), p < 0.001] than patients discharged in hours. Heterogeneity was high (I2 90.1% for mortality and 90.2% for readmission), resulting from differences in effect size rather than the presence of an effect.ConclusionsOut-of-hours discharge from an ICU is strongly associated with both in-hospital death and ICU readmission. These effects persisted across all definitions of “out of hours” and across healthcare systems in different geographical locations. Whether these increases in mortality and readmission result from patient differences, differences in care, or a combination remains unclear.Electronic supplementary materialThe online version of this article (10.1007/s00134-018-5245-2) contains supplementary material, which is available to authorized users.

Evaluation of Readmission to Intensive Care Unit Prediction Score in a Training Hospital

“…Yapılan çalışmalarda YBÜ'ye başvuru sırasındaki hastalığın ciddiyeti yüksek olmasının ve gece veya mesai dışı taburculuğun YBÜ'ye yeniden başvuruda bağımsız bir risk faktörü olduğunu göstermiştir (9,10). Çalışmamızda aynı ekip tarafından yalnızca mesai içinde yapılan taburculukları değerlendirerek yeniden başvuru için ayrıca risk olan bu Yoğun bakımda yatışları sırasında mekanik ventilasyon desteği alan hastaların almayanlara göre taburculukları sonrasında daha sık yoğun bakıma yeniden başvurduğunu belirledik.…”

Towards a decision support tool for intensive care discharge: machine learning algorithm development using electronic health care data from MIMIC-III and Bristol, UK.

“…We note that it is more conventional to use readmission (or mortality) within 48 hours to define a negative outcome related to ICU care [12,26]. However, this practice is not universal [27] and in our case it was not possible because of limitations in the data available locally. Given the low rates of negative outcome following callout in both MIMIC and GICU (see table 2), we generated further instances of the negative class, in order to balance the class sizes.…”

Is research from databases reliable? Not sure

“…The observational study by Connors et al [5], questioning the use of pulmonary artery catheterization in the management of critically ill patients, is a classic example of a study that revealed unexpected results using existing data, and consequently led to critical reappraisal of a common practice (even though suggested harm was not confirmed by experimental studies) [6]. Observational studies using administrative data can also offer insights in situations where an intervention simply cannot be randomly allocated in any practical way, e.g., in studying the (potentially harmful) effects of premature discharge at night [7].…”