Cost-Effectiveness of Population Screening Strategies for Gastric Cancer in India: A Markov Model Calibrated to High- and Low-Incidence Districts

Abstract Background. Gastric cancer imposes a substantial but geographically uneven burden in India, with age-adjusted incidence differing roughly seven-fold between registries. India has no organised gastric cancer screening programme, and no economic evaluation has assessed whether screening would represent value for money at Indian incidence levels and Indian prices. Methods. We constructed an eight-state Markov cohort model (well; precancerous lesion; localised and advanced cancer, each undiagnosed and diagnosed; cancer death; other-cause death) with annual cycles, following a cohort of 100,000 adults from age 40 over a lifetime horizon from a health-system perspective, discounting costs and outcomes at 3% per annum. Four strategies were compared: no screening, Helicobacter pylori test-and-treat (5-yearly), serum pepsinogen (3-yearly), and universal upper gastrointestinal endoscopy (5-yearly). Natural-history transition probabilities were calibrated by least-squares to two empirical targets: an observed stage distribution at diagnosis of 39.6% localised, and registry age-adjusted incidence. Two district scenarios were calibrated separately, representing the extremes of the published Indian range. Costs were anchored to Central Government Health Scheme (CGHS) tariffs and Indian episode-of-care estimates. Incremental cost-effectiveness ratios (ICERs) were compared against an empirically derived Indian threshold of 1–1.52 times GDP per capita. Results. In the high-incidence district scenario, all screening strategies improved outcomes but at high cost. Universal endoscopy averted the most deaths (228 per 100,000; number needed to screen 439 per death averted) and produced the greatest stage shift (71.1% versus 39.6% localised at diagnosis), at an ICER of ₹456,927 per QALY versus no screening. H. pylori test-and-treat was the least costly effective option (ICER ₹396,832 per QALY). On full incremental analysis, serum pepsinogen was dominated in both scenarios. In the low-incidence district scenario, ICERs rose approximately seven-fold (H. pylori ₹2,922,758; endoscopy ₹3,249,010 per QALY). No strategy fell below 1.52 times GDP per capita in either scenario. Threshold analysis indicated that incidence would need to reach approximately 20 and 23 per 100,000 per year for H. pylori test-and-treat and endoscopy respectively to become cost-effective at 1 times GDP per capita — above every incidence rate reported by Indian population-based registries. In probabilistic sensitivity analysis (10,000 iterations), no screening was the optimal strategy in 93.9% of iterations at 1 times GDP per capita and 63.8% at 1.52 times in the high-incidence scenario, and in 100.0% of iterations at every threshold in the low-incidence scenario. Conclusions. Under the assumptions modelled, no population-wide gastric cancer screening strategy is cost-effective at contemporary Indian incidence levels and thresholds, even in the highest-incidence districts. The determining variable is not test performance but background incidence. These findings argue against universal screening and in favour of restricting any future programme to genuinely high-risk subpopulations, while prioritising the primary-data gaps — India-specific test accuracy, stage-stratified survival, and reimbursement-based episode costs — that currently limit confidence in any such evaluation.