Nationwide analysis of medicine search behaviour and system performance: a one-year evaluation of MediVerify in Sri Lanka (July 2024-July 2025) based on 1.49 million search queries.

BACKGROUND

National-scale digital health platforms generate large volumes of usage data that provide insights into population-level healthcare needs. MediVerify, Sri Lanka's national online medicines information platform, providing public access to the registry of all National Medicines Regulatory Authority (NMRA)-approved products, recorded over 1.49 million queries in its first year of operation. However, systematic analysis of such large-scale medicine search behaviour in low- and middle-income countries (LMICs) remains limited.

OBJECTIVES

(1) Characterise national medicine information-seeking behaviour using large-scale search logs; (2) identify mismatches between public demand and essential medicines policy; and (3) evaluate the technical performance and computational efficiency of a national digital health system.

METHODS

We conducted a retrospective observational analysis of 1,497,304 anonymised search queries on MediVerify between July 2024 and July 2025. Queries were normalised and mapped to a registry of 11,933 approved medicines using a multi-step pipeline including fuzzy matching (Levenshtein distance, threshold ≤5). Therapeutic categories were assigned using an Anatomical Therapeutic Chemical (ATC)-aligned classification. Query distributions, zero-query and zero-result patterns, and system performance metrics were analysed. Energy consumption was estimated from observed processing times.

RESULTS

Vitamins/minerals (11.74%), antibiotics (10.57%), diabetes medications (7.02%), and antihypertensives (6.84%) dominated searches. The top 20 medicines comprised 37.3% of all queries. Approximately 40%-50% of registered medicines were never queried. Zero-result queries (∼1%) revealed unmet information needs. A significant divergence was observed between frequently searched medicines and essential medicines list (EML) designations. System performance was robust, with a median latency of 8 ms and >99% query resolution. Estimated energy consumption was extremely low (∼0.12 kWh per million queries).

CONCLUSIONS

Large-scale medicine search data provide actionable insights into national healthcare information demand and policy alignment. MediVerify demonstrates that a national digital health platform can achieve high utilisation with minimal computational and environmental cost. Findings support integration of real-time usage analytics into pharmaceutical policy and highlight the potential of low-resource digital infrastructure in strengthening health systems in LMICs.