PS-061 Developing and optimising a clinical decision support system

Background The development of clinical decision support systems (CDSS) has become an ongoing process of increasing the sophistication, generating systems that link patient characteristics with computerised knowledge bases by using algorithms (clinical rules) and generating patient-specific assessments or treatment recommendations. Purpose To develop a more efficient CDSS by tackling algorithmic differentiation, CDSS possibilities and data delivery. Materials and methods In early 2011 a multidisciplinary team started developing a CDSS, the CRR (Clinical Rule Reporter). The CRR possibilities were expanded making it possible to integrate the electronic medical record system (medical history and laboratory data) and the computerised physician order entry system (drug record and contraindications). The data delivery was also optimised by standardising the format so that the CRR could interpret more data such as starting/stopping dates, the number of administrations, and the dose per administration. As for the algorithmic differentiation, evidence and/or literature-based clinical rules were developed making them as sensitive and specific as possible. Results In a previous CDSS around 90 alerts were generated per day, of which only 3.6% were relevant (an intervention had to be performed). In the first CRR version around 55 alerts were generated per day, of which approximately 10% were relevant. After revising and further optimising the CRR, around 35 alerts are now generated per day, of which 25% are relevant. Conclusions Optimising the CRR implies a decrease in the number of alerts and an increase in relevant alerts. Constant development and updates are of great importance to further optimise the CRR making it more efficient. Conflict of interest: All authors are Pfizer employees.