Modelling physiological deterioration in post-operative patient vital-sign data

Abstract

Patients who undergo upper-gastrointestinal surgery have a high incidence of post-operative complications, often requiring admission to the intensive care unit several days after surgery. A dataset comprising observational vital-sign data from 171 post-operative patients taking part in a two-phase clinical trial at the Oxford Cancer Centre, was used to explore the trajectory of patients’ vital-sign changes during their stay in the post-operative ward using both univariate and multivariate analyses. A model of normality based vital-sign data from patients who had a “normal” recovery was constructed using a kernel density estimate, and tested with “abnormal” data from patients who deteriorated sufficiently to be re-admitted to the intensive care unit. The vital-sign distributions from “normal” patients were found to vary over time from admission to the post-operative ward to their discharge home, but no significant changes in their distributions were observed from halfway through their stay on the ward to the time of discharge. The model of normality identified patient deterioration when tested with unseen “abnormal” data, suggesting that such techniques may be used to provide early warning of adverse physiological events.

A prospective multicentre observational study of adverse iatrogenic events and substandard care preceding intensive care unit admission (PREVENT)

“…The majority of the medical events were delays in management or seeking advice, or inadequate resuscitation. Ten of the nursing events were misapplications of the early warning score, hence perhaps a new approach is required, for instance, the continuous analysis of telemetrically monitored vital signs, which automatically identifies patients at high risk of an acute deterioration .…”

Physiological trajectory of patients pre and post ICU discharge¹

“…This is a non-parametric technique, and thus no a priori assumptions about the underlying distribution were made. Our notation follows that of [9]. The data distribution p(x) was modeled using the N = 336 set of observations, each with D = 5 dimensions, as shown in (1).…”

“…Secondly, as the model is multivariate and multimodal, the interactions between various vital signs can be modeled and assessed for normality. Data-driven models of normality have been successfully applied to track the trajectory of recovery in cancer patients following upper gastro-intestinal surgery [9].…”

Personalized Risk Scoring for Critical Care Prognosis Using Mixtures of Gaussian Processes

“…C RITICALLY ill patients who are hospitalized in regular wards with solid tumors, hematological malignancies, neutropenia, or those who are recipients of stem cell (or bone marrow) transplants, or upper-gastrointestinal surgeries, are vulnerable to a wide range of adverse outcomes, including post-operative complications [1]- [8], cardiopulmonary arrest [9], [10], and acute respiratory failure [11]. All these adverse events can lead to an unplanned ICU transfer [2], the timing of which plays a major role in determining clinical outcomes, since the efficacy of acute care interventions (including thrombolytic agents, aspirin and β-blockers, mechanical ventilation, etc) depends substantially on the timeliness of their application.…”