Editors Selection IGR 20-3

Prevention and Screening: Health Economics of Screening in China

Steve Mansberger
Seth Bricel

Comment by Steve Mansberger & Seth Bricel on:

81123 Cost-effectiveness and cost-utility of population-based glaucoma screening in China: a decision-analytic Markov model, Tang J; Liang Y; O'Neill C et al., The Lancet. Global health, 2019; 7: e968-e978

Find related abstracts

Glaucoma seems a disease well-suited for screening: it is relatively common, often asymptomatic, easily screened-for, and treatable. However, several review articles suggest that population-wide screening is not cost-effective when studied in high-income countries. Tang et al. examined whether this holds true in China, where the costs of screening are lower and prevalence of PACG in particular is far higher than in European-derived populations such as the UK and Finland where prior, similar studies had focused. Overall, screening is more effective when it occurs in locations and groups with high prevalence of disease with high risk of morbidity.

The authors used Markov modeling simulations using the following data sources: 1) a meta-analysis of glaucoma prevalence studies in China; 2) inferred probability of transition to blindness from studies on closely-related populations such as in Mongolia; and 3) cost data from the Wenzhou Medical University Hospital. These data were used to model a proposed screening examination of Van Herrick testing for PACG and fundus photos to screen for POAG. The authors used a cost utility analysis using the World Health Organization established thresholds for cost effectiveness of an intervention based on the incidental cost-utility ratio (ICUR) compared to per-capita GDP. ICUR was defined as the cost difference between screened and unscreened populations divided by the difference in number of quality-adjusted life-years attributable to the intervention.

We applaud the authors for their important study. Several caveats include that the Van Herick method, while non-invasive and inexpensive, lacks sensitivity. It is unclear how the results would be altered with newer methods such as anterior segment OCT or digital gonioscopy. Several previous studies suggest that the follow-up proportion after a glaucoma screening is poor (from 40-60% with follow-up), when conducted in urban areas. It is unclear whether the Markov model includes lack of follow-up after screening and how this would impact the ICUR results especially in rural areas. The large confidence intervals of ICUR estimates may suggest variable results. The current manuscript found screening for PACG and POAG to meet cost effectiveness criteria in China. Their findings could be used to support public health policy changes.

Issue 20-3

Change Issue