Evaluation of two screening methods for undiagnosed diabetes in China: an cost-effectiveness study.

AIMS To evaluate the performance and cost-effectiveness of two screening methods to identify undiagnosed diabetes at primary care settings among a Chinese population. METHODS Two screening methods using a fasting capillary glucose (FCG) test or a Chinese diabetes risk score (DRS) at primary care settings followed by diagnostic tests were compared. The performance of FCG and DRS was evaluated by using receiver operating characteristic (ROC) curve analysis. The main economic outcome measures were the total cost of screening per 1000 persons, proportion of undiagnosed diabetes detected, and cost per undiagnosed diabetes identified from the societal perspective. RESULTS Among all participants, 14.6% (1349/9232) had undiagnosed diabetes defined by fasting plasma glucose ≥ 7.0mmol/l and/or 2h plasma glucose ≥ 11.1mmol/l and/or hemoglobin A1c ≥ 6.5%. At the optimal cutoff point of 6.1mmol/l for FCG and 14 for DRS, the sensitivity was 65.1% and 65.8%, and specificity was 72.4% and 55.2%, respectively. The area under the ROC curve was 75.3% for FCG and 63.7% for DRS (P<0.001). Based on the input costs, the total cost of screening 1000 persons was ¥64,000 ($9143) for FCG and ¥81,000 ($11,571) for DRS. The average cost per case identified was ¥674 ($96) for FCG at cutoff point of 6.1mmol/l and ¥844 ($121) for DRS at score of 14. The incremental cost per case identified was ¥17,000 ($2429) for DRS compared to FCG. The dominance relations between strategies remained with the changed in sensitivity analysis. CONCLUSIONS As a first-line screening tool for undiagnosed diabetes, the FCG test performed better than the DRS in primary care settings in China. The non-invasive and layperson-oriented DRS was feasible and detected more cases but more expensive. No strategy has strong dominance that was both more effective and less costly. The favorable strategy will depend on if the purpose of the screening program is to identify more cases or to have lower cost per case.