Can doctors and patients correctly estimate cardiovascular risk? A cross-sectional study in primary care

Abstract

Accurate cardiovascular risk estimations by patients and doctors are important as these affect health behaviour and medical decision making. We aimed to determine if doctors and patients were accurately estimating the absolute cardiovascular risk of patients in primary care. Methods A cross-sectional study was carried out in primary care clinics in Malaysia in 2014. Patients aged 35 years and above without known cardiovascular disease (CVDs) were included. Face-to-face interviews with a structured questionnaire were used to collect sociodemographic and clinical data as well as patients' perception and doctors' estimate of the patients' CVD risk. Associations were tested using ‡ 2, correlation and independent t-tests. Results We recruited 1094 patients and 57 doctors. Using the Framingham Risk Score (FRS) alone, 508 patients (46.4%) were in the high-risk group. When diabetes was included as high risk, the number increased to 776 (70.9%). Only 34.4% of patients and 55.7% of doctors correctly estimated the patient's CVD risk in comparison with the reference FRS. Of the high-risk patients, 664 (85.6%) underestimated their CV risk. Factors associated with underestimation by patients included not having family history of CVD (adjusted OR (AOR): 2.705, 95% CI 1.538 to 4.757), smaller waist circumference (AOR: 0.979,95% CI 0.960 to 0.999) and ethnicity in comparison with the Malay as reference group (indigenous/others: AOR: 0.129, 95% CI 0.071 to 0.235). Doctors underestimated risk in 59.8% of the high-risk group. Factors associated with underestimation by doctors were patients factors such as being female (AOR: 2.232, 95% CI 1.460 to 3.410), younger age (AOR: 0.908, 95% CI 0.886 to 0.930), non-hypertensive (AOR: 1.731, 95% CI 1.067 to 2.808), non-diabetic (AOR: 1.931, 95% CI 1.114 to 3.348), higher high-density lipoprotein levels (AOR: 3.546, 95% CI 2.025 to 6.209), lower systolic blood pressure (AOR: 0.970, 95% CI 0.957 to 0.982), non-smoker (AOR: 2.246, 95% CI 1.354 to 3.726) and ethnicity in comparison with the Malay as reference group (Indian: AOR: 0.430, 95% CI 0.257 to 0.720; indigenous/others: AOR: 2.498, 95% CI 1.346 to 4.636). Conclusions The majority of consultations occurring between doctors and patients are being informed by inaccurate cardiovascular risk estimation.