JMIR Publications

ABSTRACT

Background:

Remote monitoring programs have an increasingly larger role in hypertension management. Primary care physicians’ (PCP) referral rates to digital health programs are highly variable. Understanding the factors behind a physicians' likelihood to refer is important in understanding how to increase future adoption of programs through referrals. Prior studies have shown that likelihood to refer is dependent on prior knowledge about clinics and their characteristics, providers' own clinical expertise on the area, communication and relationship with their PCP and patient, and the clinic referral process.

Objective:

The study explores whether knowledge of our remote blood pressure monitoring program program and information on referral patterns influence a PCPs’ intention to refer patients.

Methods:

This is a mixed methods study integrating quantitative analysis of EHR data regarding frequency of PCPs’ referrals of patients with hypertension to a digital health program and quantitative and qualitative analyses of survey data about PCPs’ knowledge of the program (on a scale of 0-10, with 10 being extremely knowledgeable) and their intention to refer patients. PCPs responded to a clinical vignette featuring an eligible patient which approximated their baseline tendencies to refer. They were then randomized to either receive their own referral data or their own plus their peers’ referral data from the EHR. They were assessed their intent to refer eligible future patients. Descriptive and multivariable linear regression analyses examined the characteristics of participants and the factors associated with their intent to refer patients. Narrative reasons for their intention to refer were thematically analyzed.

Results:

Of the 242 eligible PCPs invited to participate, 31% (N=70) responded to the survey. From EHR data, the mean (SD) referral rate of patients per PCP was 11.80% (13.30%). The self-reported knowledge of the digital health program was 6.47 (1.81). The mean likelihood to refer for an eligible patient in a vignette was 8.54 (2.12). The mean likelihood to refer in the group that received their own prior referral data was 8.91 (1.28), while in the group that received their own and peer prior referral data was 8.35 (2.19). Regression analyses suggested the intention to refer vignette patient was significantly associated with their knowledge (coefficient 0.46, 95%CI 0.20 to 0.73, P<.001) whereas the intention to refer future patients was significantly associated with their intent to refer the patient in the vignette (coefficient 0.62, 95% CI: 0.46 to 0.78, P<0.001). No evidence of association was found on receiving own plus peer referral data compared with own referral data and intent to refer future patients (coefficient 0.23, 95% CI: -0.43 to 0.89, P=.48). Respondents called for more communication regarding the program’s superior clinical outcomes, its support of patients in their own care, and the simplicity of the technology to improve PCPs’ awareness and uptake of the program.

Conclusions:

Physicians’ intention to refer patients to a novel digital health program can be extrapolated by examining their intention to refer an eligible patient portrayed in a vignette, and this was found to be significantly influenced by their knowledge of the program. Future efforts should engage PCPs to better inform them so that more patients can benefit from the digital health program.