Content area
Full Text
Introduction
The development of even minor postoperative complications is a major determinant of hospital readmission, long-term adverse outcomes and death (Khuri et al., 2005; Moonesinghe et al., 2014). Postoperative morbidity can be recorded using a number of tools, but POMS (Post Operative Morbidity Survey) has emerged as a useful survey for assessing short-term morbidity following moderate-major surgery in clinical (Bennett-Guerrero et al., 1999; Grocott et al., 2007; Ackland et al., 2010; Ackland et al., 2015; Ackland et al., 2011; Ackland et al., 2007; Ausania et al., 2012; Davies et al., 2013; Jones et al., 2013; Pearse et al., 2014; Sanders et al., 2012; Snowden et al., 2010; Wakeling et al., 2005) and translational perioperative studies (Edwards et al., in press). However, POMS has thus far only utilized prospective data collection, requiring research staff to record morbid events. The potential for electronic capture of these data is under-explored. However, determining whether retrospective, rather than prospective, data collection can capture POMS-defined morbidity is a first step that may help exploit these data for enhanced, large scale bioinformatic interrogation. We hypothesized that retrospective data collection from charts, medical and nursing notes was non-inferior to prospective data collection for capturing POMS-defined morbidity. We tested this by three different approaches. First, we established whether POMS- defined morbidity captured retrospectively was statistically non-inferior (Walker & Nowacki, 2011) to prospective, real time data collection, by calculating differences based on Nam’s RMLE method (Nam, 1997). Second, we assessed whether POMS-defined morbidity captured retrospectively or prospectively altered the trajectory of patients becoming free of postoperative morbidity. Third, we assessed whether POMS-defined morbidity on postoperative day 3 captured retrospectively or prospectively was linked with delayed hospital discharge, the predictive value for which has been established in previous studies (Ackland et al., 2011).
Materials and Methods
We analysed POMS in 85 patients undergoing major elective surgery at University College London Hospital, having obtained written informed consent (institutional board review- Medical Research Ethics Committee: 10/WNo03/25). POMS domains are detailed in Table 1. Morbidity was recorded as defined by POMS prospectively on postoperative days 3 and 7. Both investigators were trained in prospectively collecting POMS data at the bedside. Thereafter, one was assigned to prospective data collection at the bedside, while the other assessed POMS from the same...