Electronic Prescribing Systems as Tools to Improve Patient Care: A Learning Health Systems Approach to Increase Guideline Concordant Prescribing for Venous Thromboembolism Prevention

Background: Venous thromboembolism (VTE) causes significant mortality and morbidity in hospitalised patients. In England, reporting the percentage of patients with a VTE risk assessment is mandated, but this does not include whether that risk assessment resulted in appropriate prescribing. Full guideline compliance (defined as an assessment and an appropriate action) is rarely reported. Education, audit and feedback enhance guideline compliance but electronic prescribing systems (EPS) can mandate guideline-compliant actions. We hypothesised that EPS-based interventions would increase full VTE guideline compliance more than other interventions. Methods: All admitted patients within University Hospitals Birmingham NHS Foundation Trust were included for analysis between 2011-2020. The proportion of patients who received a fully compliant risk assessment with the recommended action (prophylaxis or not, depending on the risk score) was assessed over time. Interventions included face-to-face feedback based on measured performance (an individual approach) and mandatory risk assessment and prescribing rules into an EPS (a systems approach). Findings: Data from all 235,005 admissions and all 5503 prescribers were included in the analysis. Face-to-face feedback gradually improved full VTE guideline concordance from 70% to 77% (p=<0.001). Changes to the EPS to mandate assessment with prescribing rules increased full VTE compliance to 95% (p=<0.001). Further amendments to the EPS system to reduce erroneous VTE assessments slightly reduced full compliance to 92% (p<0.001), but this was then maintained during two changes to the low molecular weight heparin drug used for VTE prophylaxis. Interpretation: While both feedback and a rules-based EPS intervention improved guideline-compliant VTE management, a systems approach was more effective and benefits were sustained even during changes to formulary drugs. Non-compliance (an assessment without appropriate action or no assessment) was not eradicated despite this mandated system. Further work is needed to understand this and to ensure software changes enhance full guideline compliance. Funding Statement: This work was supported by PIONEER, the Health Data Research Hub in acute care and the HDR-UK Better Care programme. Declaration of Interests: S Gallier, P Nightingale, T Pankhurst, I Woolhouse, MA Berry, M Garrick report no conflicts of interest. S Ball reports funding support from the HDRUK. E Sapey reports funding support from HDRUK, MRC, Wellcome Trust, NIHR, Alpha 1 Foundation, EPSRC and British Lung Foundation. All other authors have nothing to declare. Ethics Approval Statement: This research was conducted under the ethical approvals of PIONEER, a Health data research hub in acute care (East Midlands – Derby Research ethics committee, reference 20/EM/0158).