CDSS activation increased appropriate recommendations from 2.35 to 6.20 and reduced inappropriate from 0.29 to 0.07, missing from 3.80 to 0.49 in VHD management.
Does a clinical decision support system increase appropriate clinical recommendations and reduce inappropriate or missing ones in patients with valvular heart disease?
A clinical decision support system significantly improved the appropriateness of clinical recommendations for managing valvular heart disease compared to conventional management.
Absolute Event Rate: 0% vs 0%
Abstract Background The knowledge-based clinical decision support system (CDSS) used in this study provides diagnostic and therapeutic recommendations based on patient’s clinical data. It relies on an expert system of clinical rules that are modelled, validated, and peer-reviewed by healthcare professionals based on reliable knowledge sources. Clinical validation of CDSS is crucial to assess their accuracy, reliability, and potential impact on patient outcomes. Objective To assess the clinical benefit of this CDSS in patients with valvular heart disease (VHD), by evaluating its ability to increase the number of appropriate clinical recommendations and reduce the number of the inappropriate or missing relevant recommendations compared to conventional management. Methods The CDSS product for VHD includes 4122 rules derived from 64 internationally recognised knowledge sources. A clinical validation study was conducted using a pre-test-post-test design with longitudinal repeated measures. A total of 106 clinical cases of aortic stenosis, aortic insufficiency, mitral stenosis, mitral insufficiency and tricuspid insufficiency were selected from a tertiary hospital database through stratified random sampling. Conventional clinical management was retrospectively assessed by a panel of three cardiologist experts using updated clinical guidelines, who labelled clinical decisions as appropriate or inappropriate and identified the missing relevant ones. Assessments were conducted at baseline (conventional management) and after CDSS activation across 4 domains: short-term and long-term diagnostic plans (STD and LTD), cardiological medical treatment (CMT), and indicated therapeutic interventions (ITI). The experts rated their agreement with the CDSS suggestions on a 5-level Likert scale, with higher scores indicating stronger agreement. Results Following CDSS activation, the average number of appropriate recommendations significantly increased (6.20 vs. 2.35, p0.001), and the number of inappropriate and missing relevant ones significantly decreased (0.07 vs 0.29, p0.01 and 0.49 vs 3.80, p0.001, respectively; 0.49 vs 3.80, p0.001 when pooled together) (Fig. 1). Experts’ agreement with the CDSS was very high, with mean Likert scale scores above 4 in all domains (STD: 4.53, LTD: 4.68, CMT: 4.58, ITI: 4.62). The probability of strong agreement (score 5) was significantly higher than for any other rating category (z.ratio (5|1)=30.88, p-value 0.001, z.ratio (5|2)=31.67, p-value 0.001, z.ratio (5|3)=30.21, p-value 0.001, z.ratio (5|4)=22.54, p-value 0.001). Conclusions The CDSS demonstrated a significant clinical benefit by increasing the number of appropriate clinical recommendations while reducing inappropriate and missing relevant ones compared to conventional management. Cardiologist experts showed a high level of agreement with the CDSS, underscoring its potential as a reliable decision-support tool for clinicians in managing VHD.Assessment before and after CDSS support
Peña-Gil et al. (Sat,) reported a other. CDSS activation increased appropriate recommendations from 2.35 to 6.20 and reduced inappropriate from 0.29 to 0.07, missing from 3.80 to 0.49 in VHD management.