Modelling Clinician Interaction With a Clinical Decision Support System for Wound Management and Early Recognition of Deteriorating Patients

Principal Investigator (PI): Assoc Prof Christopher Khoo

Co-Principal Investigator (Co-PI): Asst Prof Chang Yun-Ke

Collaborators: Nil

Start Date: Nov 2013

End Date: On-going

Abstract: Clinical decision support systems (CDSS) can potentially improve clinical decision making, work efficiency and patient outcomes, and reduce hospital length-of-stay. However, little is known about how to design CDSS that effectively supports clinical decision making. Researchers have only identified broad factors associated with successful CDSS implementations, without detailed experimental studies.

The focus of this study is on the CDSS interface design, and how information provided by the interface influences decision making. The study will be carried out in the domain of wound management and early recognition of deteriorating patients. A basic CDSS has been implemented at TTSH to support the management of pressure ulcers. In this study, alternative interfaces with different kinds of help information will be implemented and tested. The CDSS will also be extended to cover other types of wounds, and patient monitoring for early recognition of clinical deterioration.

The objectives of the study are:

  1. to find out what kinds of information and information presentation on CDSS interfaces will support diagnosis, treatment selection and patient monitoring;
  2. to develop a model of clinician-CDSS interaction to predict effective decision making in different situations.

The purpose is to develop more effective CDSS interfaces.

Diagnosis, treatment selection and patient monitoring involve different mental processes that need to be modelled differently:

  • Diagnosis will be modelled as a categorization behaviour using categorization theories from cognitive psychology;
  • Treatment selection will be modelled as selecting from a set of decision options by weighing potential benefits, risks and probabilities of success, using theories from decison sciences;
  • Patient monitoring involves monitoring of a patients’ physiological parameters that are charted graphically, and will be modelled using theories of information visualization.

The data collection will be carried out using semi-structured interviews, observation of how the current CDSS is used, and controlled experiments with alternative CDSS interface designs.