Selective screening for high cholesterol in Australian general practice: the Newcastle Cholesterol Prediction Study

Abstract

Objective: To test whether information simply obtained from adults can identify those who are likely to have blood cholesterol levels higher than 6.5 mmol/L (250 mg/dl).

Design: Cross-sectional study.

Setting: Seven general practices in the lower Hunter Region of Australia.

Participants: Of 693 men and women 25 to 65 years old attending the general practices, 616 (89%) participated. Twelve pregnant women and four without blood samples were excluded, leaving 600 subjects (208 men, 392 women).

Measurements and main results: Twenty-two percent of subjects had blood cholesterol higher than 6.5 mmol/L (250 mg/dl). In women, high cholesterol was mainly found among those over 40 years of age, but in men high cholesterol was more evenly spread across the different age groups. Stepwise logistic regression identified age, history of hypertension, and a past history of heart attack as significant independent predictors of high cholesterol. A simple model developed from these variables identified 81% of men and women (95% CI = 72-90%) with high cholesterol while testing only 49% of the population (95% CI = 44-54%). This model was developed in a random subset of 331 of the 600 subjects, and when applied to the remaining 269 subjects, it identified 77% of those with high cholesterol (95% CI = 67-87%) after testing 48% (95% CI = 42-54%).

Conclusions: Selective screening using this simple model can identify adults who are likely to have high blood cholesterol and could complement case-finding or provide an alternative high-risk strategy for communities that cannot afford to screen all individuals.