COMPARATIVE AND COST-EFFECTIVENESS ANALYSES OF LABORATORY-BASED AND NON-LABORATORY-BASED CARDIOVASCULAR RISK SCREENING IN UZBEKISTAN

Purpose: Cardiovascular disease (CVD) risk scores requiring laboratory testing  (e.g., cholesterol testing) are often used to identify high- and low-risk individuals for primary CVD prevention interventions, such as statin treatment. However, laboratory testing can be expensive and difficult to conduct in resource-limited settings. We sought to evaluate and compare the performance and cost-effectiveness of a non-laboratory-based and six laboratory-based CVD risk characterization approaches and statin initiation policies among adults in Tashkent City, Uzbekistan.

Methods: CVD risk for 853 adults (376 men, 477 women) was assessed using 2002 Uzbekistan Health Examination Survey (UHES) data. Non-laboratory-based CVD risk prediction was compared to six laboratory-based scores (SCORE for high-risk and low-risk countries, three versions of Framingham risk scores, and CUORE) using Spearman rank correlation. UHES data were also used to generate hypothetical populations of 10,000 men and 10,000 women without CVD to evaluate the cost-effectiveness of various statin initiation policies (proportion of population treated ranged from 0-20%) that used either non-laboratory or laboratory-based (SCORE) risk screening approaches. We developed a micro-simulation model, calibrated to national and regional CVD incidence data, to project the lifetime discounted CVD-related costs and quality-adjusted life years (QALYs) accrued under each strategy. We defined an intervention as "cost-effective" if the incremental cost-effectiveness ratio (ICER) was less than three times the per capita GDP ($5,150).

Results: The Spearman rank correlation coefficients for the laboratory-based and non-laboratory-based CVD risk scores ranged from 0.872-0.984 for men and 0.937-0.980 for women. The ICERs for the non-laboratory-based strategies on the efficient frontier ranged from $840 to $6,600 for men and $6,200 to $16,000 for women (Table 1 and 2). Almost all SCORE strategies for both men and women were dominated; the only exception was an ICER of $92,000 in men. Cost-effectiveness results were most sensitive to variations in all types of costs (laboratory, statin, and CVD event), and statin effectiveness and compliance.

Conclusions: There was a high correlation between the laboratory-based and non-laboratory-based risk scores in both men and women. It is cost-effective to use the non-laboratory based risk score to assess risk and treat the top 1.6% of men (in terms of CVD risk) in Uzbekistan. Statin treatment does not appear to be cost-effective for primary CVD prevention in women in Uzbekistan.