Lauren Morgan, University of Oxford
Deterioration in hospital inpatients frequently goes unrecognised despite widespread introduction of vital sign-based “early warning scores”. Patient data including demographics, laboratory results and vital-sign observations may all help identify patients who will deteriorate. However, the multiple data types are not integrated to provide clinically useful information and support decision-making based on an overall risk index. Consequently, patients deteriorate because they are not reliably identified to clinical teams equipped to deliver timely, often life-saving treatments. Severe deterioration requiring intensive care unit (ICU) admission occurs in ~40,000 UK inpatients annually. If early recognition of deterioration reduced ICU admissions and subsequent mortality by only 10%, each year 4000 unnecessary deaths could be avoided, and nearly £4 million of critical care bed capacity would be made available.
Our team at the University of Oxford are developing a hospital-wide patient surveillance system fusing electronically-available patient information to provide a continuously-updated risk index to clinicians (HAVEN). This will allow them to identify, review and treat patients who without acute medical intervention will deteriorate to the point of requiring ICU admission.
Evidence of Human Factors Integration (HFI) in medical IT projects is rare. We have previously developed in-house a bedside data entry and analysis system for a subset of clinical data, where we employed a variety of HF tools and techniques. This system is live, in use and across multiple hospital sites and has proven instrumental in getting HFI recognised as a necessary project requirement.
We propose to present the HF methods used in the development of the new HAVEN system. We would specifically benefit from the discussion of the approach with the wider HF in complex systems community. We will present the preliminary results of an applied cognitive work analysis, and initial resultant screen designs and proposed workflows. We will be exploring novel data visualisation approaches to presenting the clinical data in a way that supports the clinical decision-making and ease of interpretation.