abstract

This paper analyses official data on cases registered in respect of deaths in police custody in India from 2021-22 to 2025-26. The source is the Ministry of Home Affairs response to Lok Sabha Unstarred Question No. 5115, answered on 24 March 2026. The national totals were 176, 163, 157, 140 and 170 respectively, with the final figure covering the period up to 15 March 2026. Across the five reporting periods, 806 cases were recorded. The paper uses descriptive trend analysis and state-level aggregation to examine temporal change and geographical concentration. Maharashtra recorded the highest five-period total, followed by Bihar, Gujarat, Rajasthan and Uttar Pradesh. The analysis is intentionally cautious: a death in police custody is not automatically proof of police brutality, because deaths may arise from illness, suicide, injury or other causes. Nevertheless, every custodial death engages a heightened state duty of care, documentation, independent inquiry and public explanation. The paper argues that official counts are most useful when integrated with population, arrest, custody-volume, medical, inquiry and accountability data. It recommends standardised definitions, timely publication, independent preservation of evidence, stronger oversight and state-level risk indicators. The findings show that apparent national improvement through 2024-25 was not sustained in the nearly complete 2025-26 period, making continued monitoring essential.