A Systematic Review of Risk Scoring Methods and Clinical Decision Aids Used in the Primary Prevention of Coronary Heart Disease (Supplement) [Internet]

Excerpt

Background: Guidelines recommend that cardiovascular disease risk scoring methods are used to identify patients who are asymptomatic, but at high risk of future cardiovascular disease.

Methods: Using standard Cochrane methods we performed a series of systematic reviews to assess the accuracy and effectiveness of risk-scoring methods in different populations.

Results: A comprehensive search strategy identified 3,439 articles of which 996 related to the use of risk scoring methods. One hundred and ten risk scoring methods were identified of which 70 were specifically aimed at guiding primary prevention of cardiovascular disease. Most of the risk scores included a core set of variables reflecting the contemporary established independent risk factors, although several widely recommended methods do not include family history of cardiovascular diseases. Some methods incorporated an adjustment for use in geographic regions with differing levels of cardiovascular disease risk.

Studies reporting the generalisability of Framingham-derived scores in 112 different population groups revealed a tendency to over-estimate the risk of coronary heart disease in groups with low observed risk and to under-estimate risk in high-risk groups. There were no consistent sex differences in the calibration of risk score, but discrimination was better in women than in men.

Only four randomised controlled trials examined the effectiveness of risk scoring methods in improving health outcomes in hypertensives (3 studies) and diabetics (1 study). Effectiveness in moderate risk individuals was not examined. None showed any improvement in absolute risk or their primary outcomes compared to controls.

Conclusions: The generalisability of risk scores in different populations cannot be assumed and there is currently little evidence that they are effective in improving health outcomes. Future research needs to be directed at refining the accuracy of prediction models and, most importantly, examining ways of turning them into effective clinical tools.