Marta Mikolajewska, University of Sheffield
Introduction
This paper is based on an ESRC-funded doctoral study aimed at examining the ways in which vulnerable adult detainees are defined, identified and safeguarded in police custody suites in England and Poland. Given the scarcity of comparative research on police detention – particularly in non-English-speaking, post-communist countries – the overarching aim of this study was to critically explore and compare the police practices surrounding vulnerability on an international level, with a view to improving them. Drawing on preliminary qualitative data collected in 2025, this paper specifically explores the complexities of detainee vulnerability identification and police decision-making in the English context, offering the first empirical findings on vulnerability identification post the 2018 change of the vulnerability definition within the Code C of the Police and Criminal Evidence Act (PACE) 1984.
Identification of vulnerable adult suspects – the challenges
The issue of successful identification of vulnerable suspects by custody sergeants (CSs), which often takes place in cooperation with other professionals such as Liaison and Diversion (L&D) services and Health Care Professionals (HCPs), has been explored by scholars and practitioners over the past decades. This research has shown multiple problems with identifying vulnerability, owing to the self-reporting nature of the risk assessment carried out upon booking in detainees, lack of effective training and guidance for staff on vulnerability identification, CSs feeling unprepared to deal with vulnerable suspects, no standard risk assessment model for the English police forces, and subsequent reliance on instincts, experiences, and judgement (e.g. Bath and Dehaghani, 2020; College of Policing, 2024; Cummins, 2011). These difficulties are worsened by the definitional and conceptual ambiguities of the term, vulnerability, and its broad, socially constructed nature (Dehaghani, 2017). Though this paper reports on the problems with vulnerability identification, such problems are inherently connected to how vulnerability is understood prior to identifying it, a point that merits brief attention. Indeed, there are multiple conceptualisations of vulnerability present in the literature, which can be clustered into five distinct, yet overlapping themes that formed the theoretical basis of this study. These are the notion of i) universal vulnerability (e.g.. the idea that we all, as humans, are inherently vulnerable due to our corporeality), ii) risk-based vulnerability (e.g. as in risk being closely connected to the concept of vulnerability), iii) structural vulnerability (e.g. connected to race, gender or social class), iv) situational vulnerability (e.g. linked to immediate circumstances) and lastly v) innate vulnerability (e.g. connected to childhood) (e.g. Brown, 2015). The juxtaposition of the aforementioned theoretical constructions of vulnerability with its legal definition contained in paragraph 1.13(d) of Code C, reveals the latter’s unidimensional nature (e.g. Dehaghani, 2021), given Code C’s heavy focus on psychological, innate vulnerabilities and reliability of evidence. Interestingly though, despite the legal definition of vulnerability opening with a clause ‘because of a mental health condition or mental disorder’, psychological, innate vulnerabilities were not central either to how vulnerability was understood or to how it was identified by the participants in this study, who generally understood vulnerability as a multidimensional concept.
Methodology
In order to explore the police decision-making around vulnerability identification, two qualitative research methods were employed, namely participant observation and semi-structured interviews. Participants were recruited using a combination of opportunity and purposive sampling, ensuring access to a range of relevant roles within the suite. The observation took place in a large custody suite in the north of England over a one-month period and included both structured activities, such as observing the booking-in of detainees, and informal interactions with detention officers, L&D services, HCPs, Drug Workers, CSs, and custody inspectors working in the suite on a given day. Following the observation, interviews with CSs, custody inspectors, Appropriate Adults and detention officers were conducted online and transcribed. The transcripts and observation notes were coded thematically using NVivo through an inductive approach, allowing for triangulation across data sources.
Findings
Preliminary findings regarding vulnerability identification indicate that CSs’ decision-making is shaped less by the legal definition of vulnerability, and much more by their personal judgement, accumulated professional experience, and how a detainee is ‘presenting’ at the time of booking in (a finding consistent with earlier research). Although, naturally, the CSs were very well aware of Code C’s definition of vulnerability, none of the participants could cite or paraphrase it nor did they say that they actively applied it when determining who should be deemed vulnerable. Instead, decisions tended to rest on instinctive appraisals of who ‘appeared’ vulnerable, which were often informed by behavioural cues such as visible distress, disorientation, lack of eye contact, or erratic conduct as well as officers’ own, at times very broad and vague, conceptualisations of vulnerability (e.g. detainee seen as vulnerable if “they are not right” – CS6).
The apparent irrelevance of the legal definition and the central role of personal judgement led to CSs’ ambiguous descriptions of how one actually identifies someone as vulnerable. CSs reflected that it is something “you just know, it’s a feeling” (CS8) or something “you just pick up” (CS6) based on exposure to custody work. While the risk assessment and the Vulnerability Assessment Framework (VAF) used by the CSs were seen as the core tools offering a structure and useful prompts – ‘nudges’ to (re)consider certain risk factors – they were not seen as definitive guides, nor were they seen as capable of providing an objective indication whether or not a detainee should be considered vulnerable. It is worth mentioning, though, that despite the subjectivity in the decision-making, there was a strong and shared risk-averse orientation, with each participant stating that they always “err on the side of caution” if in doubt.
Similarly to previous studies, CSs consistently emphasised the importance of HCPs and L&D services in helping them to reach a decision, especially in cases involving mental health, substance misuse or neurodivergence. Their advice was seen to offer both reassurance and expertise, given that they were perceived as more knowledgeable with regard to mental health etc. However, input from HCPs and L&D services was not always readily available. HCPs were at times overstretched, leading to delays, while L&D practitioners expressed that they often lacked access to relevant records due to a recent change of service provider and varying computer systems and access levels. L&D practitioners ultimately felt like they were “(…) going in blind, making decisions about someone not knowing the risks fully” (L&D2).
Conclusion
The preliminary findings underscore a clear gap between the legal definition of vulnerability and the discretionary nature of policing work. CSs make vulnerability-related decisions based less on legal frameworks and more on accumulated experience, instinct and detainee’s immediate presentation. While tools like the VAF and input from the specialists are useful, ultimately the decision remains subjective – a “judgement call” (CS11) based upon contextual factors, behavioural cues, and strong risk-averse tendencies. This raises important implications for both academic debate and policy design. This suggests that academic debate should reconsider the emphasis placed on critiquing the legal definition in isolation given that frontline officers rarely use it as a reference point. Instead, the focus should be placed, such as during police training, more on how operational realities shape understanding and application. Furthermore, improving timely access to HCPs and L&D services, addressing barriers with accessing records, and investing in practical, experience-based training could help standardise vulnerability identification and provide clearer guidance. Facilitating support for frontline practitioners at the point of decision-making may yield more tangible benefits than focusing on theoretical refinements of definitions.
References
Bath, C. and Dehaghani, R. (2020) There to Help 3: The identification of vulnerable adult
suspects and application of the appropriate adult safeguard in police investigations in 2018/19, National Appropriate Adult Network.
Brown, K. (2015) Vulnerability and young people: care and social control in policy and practice, Bristol: Policy.
Cummins, I. (2011) ‘The other side of silence: The role of the appropriate adult: Post Bradley’,
The Journal of Ethics and Social Welfare, 5(3), pp. 306-312.
College of Policing (2024) Detention and custody risk assessment, College of Policing, available at:https://www.college.police.uk/app/detention-and-custody/detention-and-custody-risk-assessment (accessed: 25th of July 2025).
Dehaghani, R. (2017) Vulnerability in Police Custody: Definition, Identification and Implementation in the context of the Appropriate Adult Safeguard, Thesis submitted for the degree of Doctor of Philosophy at the University of Leicester, Leicester Law School, University of Leicester.
Dehaghani, R. (2021) ‘Interrogating vulnerability: reframing the vulnerable suspect in police custody’ in Social and Legal Studies, 30(2), pp. 251-271.
Table of legislation
Police and Criminal Evidence Act (PACE) 1984
Defending Vulnerability 