Background: The atherosclerotic cardiovascular disease (ASCVD) 10-year risk estimate is recommended by cardiologists for determining risk of a cardiac event. However, the majority of patients presenting to primary care with chest pain have noncardiac etiologies. Therefore, we determined if high versus low ASCVD risk was associated with primary care physicians’ referral to cardiology in patients with and without chest pain. Methods: Deidentified electronic health record (EHR) data was obtained from 5795 patients treated in academic primary care clinics from 2008 to 2015. Referral to cardiology was defined by an EHR code, chest pain was defined by ICD-9-CM code (786.5) and ASCVD was modeled as high versus low risk. Separate logistic regression models were computed to estimate the association between chest pain and referral to cardiology, ASCVD risk and referral, and both chest pain and ASCVD risk and referral with adjustment for potential confounding factors. Results: More patients with (n = 95, 7.8%) versus without (n = 75, 2.0%) chest pain were referred to cardiology (P < .0001). Separate unadjusted models revealed chest pain and high versus low ASCVD risk were significantly associated with referral (odds ratio [OR] = 4.20; 95% confidence interval [CI] 3.07-5.73 and OR = 1.41; 95% CI 1.04-1.91, respectively). After adjusting for ASCVD risk and confounders, chest pain but not high ASCVD risk remained significantly associated with referral (OR = 1.75; 95% CI 1.24-2.47 and OR = 1.15; 95% CI 0.72-1.82, respectively). Conclusions: In primary care patients presenting with chest pain, ASCVD risk scores are not associated with referral to cardiology. Overall, less than 8% of patients with chest pain were referred. While there is no evidence to indicate excessive referral to cardiology, we posit that implementing ASCVD risk tools in decision aids could contribute to referring those most in need of cardiology care.
Read full abstract