Electronic health records are being adopted due to numerous potential benefits. This requires the development of objective metrics to characterize morbidity, comparable to studies performed in centers without an electronic health record. We outline the development of an electronic version of the postoperative morbidity score for integration into our electronic health record.
Twohundred and three frail patients who underwent elective surgery were reviewed. We retrospectively defined postoperative morbidity score on postoperative day 3. We also recorded potential electronic surrogates for morbidities that could not be easily extracted in an objective format. We compared discriminative capability (area under the receiver operator curve) for patients having prolonged length of stay or complex discharge requirements.
One hundred thirty-nine patients (68%) had morbidity in ≥1 postoperative morbidity score domain. Initial electronic surrogates were overly sensitive, identifying 173 patients (84%) as having morbidity. We refined our definitions using backward logistic regression against “gold-standard” postoperative morbidity score. The final electronic postoperative morbidity score differed from the initial version in its definition of cardiac and neurological morbidity. There was no significant difference in the discriminative capability between electronic postoperative morbidity score and postoperative morbidity score for either outcome (area under the receiver operator curve: 0.66 vs 0.66 for complex discharge requirement, area under the receiver operator curve: 0.66 vs 0.67 for a prolonged length of stay; P> .05 for both). Patients with postoperative morbidity score or electronic postoperative morbidity score–defined morbidity on day 3 had increased risk of prolonged length of stay (P < .001 for both).
We present a variant of postoperative morbidity score based on objective electronic metrics. Discriminative performance appeared comparable to gold-standard definitions for discharge outcomes. Electronic postoperative morbidity score may allow characterization of morbidity within our electronic health record, but further study is required to assess external validity.
From the *University Division of Anaesthesia
†Department of Medicine for the Elderly, Addenbrooke’s Hospital, Cambridge, United Kingdom.
Published ahead of print 17 October 2018.
Accepted for publication October 17, 2018.
R. Romero-Ortuno is currently affiliated with the Discipline of Medical Gerontology, Trinity College Dublin, Mercer’s Institute for Successful Ageing, St Jame’s Hospital, Dublin, Ireland.
Funding: D.J.S. is supported by a Wellcome Trust Clinician PhD Fellowship. D.K.M. is supported by the National Institute for Health Research ([NIHR] UK) through funding for the Cambridge NIHR Biomedical Research Centre and an NIHR Senior Investigator Award.
The authors declare no conflicts of interest.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website.
Reprints will not be available from the authors.
Address correspondence to Daniel J. Stubbs, BMBCh, University of Cambridge, University Division of Anaesthesia, Addenbrooke’s Hospital, Hills Rd, Cambridge CB2 0QQ, United Kingdom. Address e-mail to firstname.lastname@example.org.