Effects of computerized clinical decision support systems on practitioner performance and patient outcomes
Effects of Computerized Clinical Decision Support Systems (CDSS) for Disease Management Programs on Practitioner Performance and Patient Outcomes
Dr Garg et al published a systematic review of controlled trials to determine the effectiveness of CDSS on practitioner performance and patient outcomes. They searched MEDLINE, EMBASE, Cochrane Library, Inspec, and ISI databases and identified 100 studies that trialled the effect of diagnostic systems, reminder systems, disease management systems and drug-dosing (prescribing) systems. In this review, I would only focus on their findings about disease management systems. Improvement in practitioner performance in disease management was defined as adherence to recommended practice guidelines and improvement in patient outcomes was defined as statistically significant positive effect in at least 50% of the outcomes measured.
They identified 40 disease management studies. They found that in 62% of the studies that evaluated practitioner performance, there was an improvement. On the other hand, only 27 studies evaluated patient outcomes, of which only 5 (18%) demonstrated improvements. After that, they provided an account of outcomes measures per specific area of disease management (i.e. diabetes care, cardiovascular prevention, thrombolytic prescribing, urinary incontinence, HIV management etc.) and provided the number and percentage of studies in each disease category that demonstrated either improvement in performance or patient outcomes.
The authors provide some useful comments about the success and failure of CDSSs. They found that in studies in which users were automatically prompted to use the system, demonstrated better performance compared with studies in which users were required to actively initiate the system. They also found that systems developed internally were significantly more successful than systems developed externally.
Comments: A very good review with rigorous methodology, broad scope and relatively simple outcome measures to synthesise. It cannot be regarded as a meta-analysis in which a pooled effect size could be calculated, because the CDSSs and outcome measures were quite different and heterogeneous in each primary study that they identified and therefore, each study had to be evaluated on its own. There are two other reasons that can justify the poor results of the studies about disease management CDSS: the first is that there is a dearth of evidence about the effectiveness of standard (not supported with a CDSS) disease management programmes, which might undermine the effectiveness of CDSSs developed for them; and second, most seemingly successful disease management programmes have not been evaluated under a controlled trialled condition. Non-inclusion of pre-post studies in this systematic review could result in under-representation of the literature in this area.
Finally, the authors of the systematic review did not provide a summary account of the main functionalities or technical specification of the CDSSs in each study. For a systematic review on the effectiveness of CDSSs, this data seems to be a key to better understand the reasons for success and failure of each system.
