["Criminal Behaviour and Mental Health, Volume 31, Issue 4, Page 262-274, August 2021. ", "\nAbstract\n\nBackground\nThe level of mental state incidents dealt with by police and the police resource involved is under‐researched, often giving rise to un‐evidenced claims around demand, response and resources. The 2019 National Police Chiefs' Council and College of Policing definition of such incidents provides a useful basis for more accurate calculation: ‘Any police incident thought to relate to someone's mental health where their vulnerability is at the centre of the incident or where the police have had to do something additionally or differently because of it’.\n\n\nAims\nTo establish the nature and frequency of incidents involving the police when mental state is a primary reason for the involvement.\n\n\nMethods\nIn this mixed methods study, we first analysed data from records in two mixed inner city/urban/rural forces and one large multi‐local authority metropolitan force. Secondly, we made an in‐depth analysis of a sample of mental state‐related incidents (n = 320) in two of these forces. Thirdly, we took a 24‐hour snapshot of all such incidents in England and Wales.\n\n\nResults\nMental state‐related incidents accounted for 5.1% of recorded police contacts from the public, rising to just 7.8% when confining attention only to contacts that generated a police response beyond taking the call and recording it. Length of time between an incident being reported and first response was similar between mental state‐related and non‐mental state‐related incidents, but response to closure time was shorter for the former.\n\n\nConclusions\nWhile incidents relating to mental state problems do consume police resources, they do not represent disproportionate demand in terms of numbers or time spent. That said, only about a quarter of the police work recorded was related to possible crimes, and the possibility of conflating perception of wider social need with mental state problems may further account for an apparent mismatch between the perceived and actual proportion of the workload spent on these incidents.\n\n"]