•  
  •  
 

Sustaining Use of a Clinical Decision Support Tool for Primary Care Providers

Publication Date

4-30-2015

Keywords

clinical decision support, cardiovascular risk

Abstract

Background/Aims: Achieving and maintaining high rates of use of clinical decision support in primary care settings has been challenging. Our goal was to reach and maintain high use rates throughout the study period through ongoing feedback and incentives to clinics and consented providers.

Methods: We conducted a clinic randomized trial of an electronic health record (EHR)-based point-of-care clinical decision support (CDS) tool (called CV Wizard) that provides prioritized treatment recommendations to optimize management of six reversible cardiovascular (CV) risk factors: lipids, blood pressure, glucose, tobacco use, aspirin use and weight. We assessed use (the number of times the tool was opened and printed) at targeted office visits for two groups of primary care providers (PCPs) at 11 intervention clinics: (a) those who provided informed consent to use and evaluate the tool (n=54), and (b) those who did not provide consent but still had access to the CDS (n=69). CDS use rate per provider was calculated in three postintervention months as the number of eligible visits at which the tool was used relative to the number of targeted outpatient visits that month. The use goal was 80% of targeted visits, and we reported monthly use rates to clinic leaders for all PCPs, with clinic compensation totaling $2,000 over the intervention period to achieve and maintain the goal. Generalized linear models tested whether PCP consent predicted use of the CDS system.

Results: Among consented PCPs, average CDS use rates at 4, 8 and 12 months after full intervention implementation were 57.0%, 73.9% and 75%. Among PCPs at the same intervention clinics who did not provide consent, average use rates were 57.3%, 70.7%, and 58.9% (significant difference only at 12 months, P<0.05).

Discussion: We observed robust use of the CDS tool by PCPs and rooming nurses at targeted primary care visits, in the context of targeted use to high CV risk patients only, leadership support and PCP design input, implementation process measurement and feedback, and small financial incentives to clinics that achieved high use rates. Additional evaluation to explain why use rates declined at 12 months in the nonconsented PCPs is of interest.

Share

COinS
 

Submitted

April 2nd, 2015

Accepted

April 28th, 2015