Rapid Responses to:

INFORMATION IN PRACTICE: Kensaku Kawamoto, Caitlin A Houlihan, E Andrew Balas, and David F Lobach Improving clinical practice using clinical decision support systems: a systematic review of trials to identify features critical to success BMJ 2005; 330: 765 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Decision support systems are beneficial for interventions on specific populations.

Elizabeth A Garthwaite, Donald Richardson, Cherry Bartlett, Eric J. Wil, Charles G. Newstead   (7 April 2005)

Decision support systems are beneficial for interventions on specific populations. 7 April 2005
Elizabeth A Garthwaite, Specialist Registrar Nephrology Department of Renal Medicine, St James's University Hospital, Beckett Street, Leeds, LS7 9TF, Donald Richardson, Cherry Bartlett, Eric J. Wil, Charles G. Newstead Send response to journal: Re: Decision support systems are beneficial for interventions on specific populations.	Editor In her editorial, Purcell expresses concern at the small number of studies analysing primary outcomes to determine the effectiveness of clinical decision support systems [1]. We evaluated the effect of a decision support algorithm in delivering patient specific prompts to manage cholesterol in renal transplant outpatients. Data was analysed retrospectively for a two-year period with attention to changes in cholesterol levels, prescribing patterns of statins and causes of underperformance. The algorithm was designed to run for each renal transplant outpatient follow up encounter. Using the most recent blood results for individual patients, on-screen treatment recommendations were given for those with a serum cholesterol of greater than or equal to 5.0mmol/l with instructions regarding starting treatment, dose adjustment, subsequent monitoring, dose titration and follow-up. The system incorporated the features identified in Kawamoto’s paper suggested to improve patient care [2]. At baseline, 36.7% of patients achieved a total serum cholesterol level <5.0mmol/l, compared to 67.2% at two years, with mean values of 5.6±0.1 and 4.8±0.1mmol/l,(p<0.0001). At baseline 24% of the patients were on statin therapy; rising to 61% at two years, with trends towards increasing doses of atorvastatin, representing changes suggested by the algorithm at subsequent out patient attendances. There were no significant adverse biochemical changes or drug interactions. For patients followed concurrently in two units without the algorithm, serum cholesterol measurements fell from 5.57 and 5.34mmol/l to 5.31 and 5.27mmol/l respectively, both higher than that achieved contemporaneously at St James’s,(p=0.05). Underperformance depended less on medical non-compliance than with systematic features of the methodology and patient preference/collaboration with treatment. The introduction of the system coincided with significant reductions in cholesterol levels in a population at specific risk, with increasing numbers of patients on appropriate therapy and no serious adverse effects. Our results illustrate the positive effect of decision support software on populations. References 1 Purcell GP. What makes a god clinical decision support system. BMJ 2005; 330:740-1 2 Kawamoto K, Houlihan CA, Balas EA, Lobach DF. Improving clinical practice using clinical decision support systems: a systematic review of trials to identify features critical to success. BMJ 2005; 330:765-8 3 Garthwaite EA, Will EJ, Bartlett C, Richardson D, Newstead CG. Patient specific prompts in the cholesterol management of renal transplant outpatients: results and analysis of underperformance. Transplantation 2004; 15;78(7):1042-7 Competing interests: None declared