An online tool for investigating clinical decision making

Abstract

Background: Induction of labour is a common clinical intervention. There has been a recent rise in rates of induction of labour and wide variation between published hospital rates without obvious explanation. Clinician variation has been suggested as a reason.

Objective: The study described aimed to examine clinical decision making, whilst removing individual patient bias. To achieve this clinical behaviour was studied by the use of imaginary clinical scenarios presented to clinicians by computer. Unlike retrospective audit, the rates thus generated are unaffected by differences in casemix, pressure of time, work or other factors and allow direct comparison between clinicians and comparison with clinical guidelines.

Methods: Data about 15 imaginary pregnant women are presented to the clinician, each may have symptoms or signs of hypertensive disorders, intrauterine growth restriction (IUGR) and/or postdates. From the decision made in each scenario, and the information revealed about each scenario, a set of ‘decision rules’ is created for each clinician, describing in what circumstances they would induce labour. Data from the National Women's Hospital (Auckland, New Zealand) is then examined using these rules and the induction of labour rate thus generated presented to the clinician.

Results: Sixteen clinicians were interviewed. Their induction of labour rate ranged from 10 – 31%.

Conclusions: Clinician variation in decision making is evident about the intervention when to induce labour. The system is available on the WWW at http://csrs2.aut.ac.nz/scenario

• Knowledge acquisition computer
• Decision making
• Labour induced