Contextualisation of findings

In the present study, physical health was ranked as the top priority, with a focus on weight loss. Surprisingly, this theme did not feature in the lists generated from earlier exercises to establish the research priorities in forensic mental health.^{Reference Hodgins9,Reference Crocker, Nicholls, Seto, Roy, Leclair and Brink11,Reference Srivastava, Forrester, Davies and Nadkarni12} The explanation for this notable difference may lie in the recent incentivised drive by commissioners of secure mental healthcare in England for providers to take demonstrable steps to improve physical health, particularly through achieving a healthy weight.^{Reference Johnson, Day, Moholkar, Gilluley and Goyder15,16}

There was overlap with the findings of a previous study showing that service users, as well as professionals, prioritise epidemiological research into factors associated with crime and recidivism, and research oriented toward recovery topics such as the future use of mental health services and employment.^{Reference Crocker, Nicholls, Seto, Roy, Leclair and Brink11} Further consistency was found in the high priority given by both clinicians and service users to research about effective treatments and interventions.^{Reference Duncan, Munro and Nicol10–Reference Srivastava, Forrester, Davies and Nadkarni12}

This finding of common interest is not surprising, given the partnership nature of many treatments involving the professional, who delivers, prescribes or administers the treatment, and the service user, who accepts or refuses the treatment that may cause harm, benefit or no effect. However, although previous studies mentioned treatment as a research priority, service users in the present study clearly linked treatment to cure. Current understanding among mental health clinicians and academics about the nature of disorder, disease, disability and concepts of recovery, diversity, social inclusion, person-centred care and stigma may partly explain why research questions such as ‘could we cure LD [intellectual disability]?’, which are important to some service users, are less likely to be posed by professionals in a research priority-setting exercise.^{Reference Insel and Scolnick17}

The advancement of risk assessment in forensic mental health is consistently reported as a research priority in studies based on literature review, professionals-only groups or mixed professional and service user groups.^{Reference Hodgins9–Reference Crocker, Nicholls, Seto, Roy, Leclair and Brink11} Although it is not clear why risk assessment did not feature as an important research area in the present study, which involved only service users, it is proposed that this topic might be of lesser importance to service users and greater interest to professionals, whose roles involve assessing and managing risk.

It is also notable that in contrast to a larger international study of research priorities for mental health and justice, the present study of service users lawfully detained in either a prison or secure hospital under the Mental Health Act 1983 did not recommend research into legislation and policy.^{Reference Crocker, Nicholls, Seto, Roy, Leclair and Brink11} Length of stay in hospital emerged as one of the new research priority areas in the present study. Although it may appear surprising that this was not a theme, given priority in previous studies, it may be that addressing other themes widely reported, such as treatment, may ultimately have an effect on overall length of stay.^{Reference Davoren, Byrne, O'Connell, O'Neill, O'Reilly and Kennedy18}

Strengths and limitations

It might appear that the responses of service users point to areas that have been extensively researched. For example, much is already known about effective weight loss interventions in the general population. Although such responses may reflect limitations of the phrasing of questions used to elicit the initial views of service users, they may also reveal a lack of evidence of context-specific effectiveness and acceptability of interventions in forensic services.^{Reference Johnson, Day, Moholkar, Gilluley and Goyder15} This is evident by the aforementioned current emphasis on finding effective ways to achieve weight loss among service users in secure mental health settings.

The use of existing consultation structures is recommended for ascertaining the research priorities of service users within forensic mental health services.^{Reference Faulkner and Morris19} The community meeting was, therefore, an appropriate forum in which to obtain the research opinions, because service users were already accustomed to voluntarily making suggestions to improve the service in this group setting. In contrast, a formal panel meeting may have been less accessible to some in-patients and prisoners.^{Reference Faulkner and Morris19} Although the presence of staff and other service users might have influenced responses provided, the wide range of individual answers given, including criticism of staff attitudes, suggests that group bias was not significant.^{Reference Staykova20}

There is little evidence to guide sampling approaches in PPI, although convenience sampling is most commonly used.^{Reference Domecq, Prutsky, Elraiyah, Wang, Nabhan and Shippee21} The modest response during both phases of the study is a significant limitation, and may reveal lack of interest in or apathy toward research among the service user group. It is possible that such explanations may relate to psychopathology experienced by potential participants; for example, anhedonia in a depressive episode, or apathy as a negative symptom of schizophrenia. Alternatively, some service users may have limited understanding of research and the value of service user involvement in research. Unknown factors relating to the level of research interest among staff responsible for presenting the study to service users at community meetings may have contributed to the low response rate. Although the final list of priorities generated from the views of a small proportion of service users is valuable, it may not reflect the views of those service users who did not respond.

A strength of the study is effectively reaching a wide range of service users, with differing risk and health needs. The methodology of involvement was successful in increasing research involvement access to service users who, (because of reasons relating to health or risk) may not have had the opportunity to leave the ward or wing to attend a formal group meeting of service user volunteers, without limiting involvement to a select sample of existing service user representatives.^{Reference Domecq, Prutsky, Elraiyah, Wang, Nabhan and Shippee21}

Consideration was given from the outset to the evidence-based approaches to avoid tokenism and collaborate with user-led groups in a way that clearly explained the philosophy of the project. However, a formal reference research group of service users did not exist at the time of the study, although the generic (non-research) service user group was already established. Consultation with this service user group helped to shape the study design; ongoing consultation with the group may have been valuable in finding ways to increase the response,^{Reference Ennis and Wykes5} and to reduce sample bias.^{Reference Staykova20} Although the involvement of professionals from in-patient and prison settings, as well as nursing and medical disciplines, ensured a healthy range of perspectives during the thematic analysis, service user involvement at this stage for collaborative data analysis would have further strengthened the study.^{Reference Jennings, Slade, Bates, Munday and Toney22}

Implications and future directions

To our knowledge, this is the first study to use PPI principles and the Delphi technique to establish research priorities from the exclusive perspective of service users within a forensic service for mental health, intellectual disabilities and autism. The methods showed how a mental health service can overcome barriers^{Reference Telford and Faulkner3} and involve its users to identify priority areas of research. By contributing to research priority-setting exercises, service users demonstrated that they wish to express their views on the greatest research needs for forensic mental health services. It also revealed a willingness to be involved at the earliest stage of the research process, and a desire to influence the work of researchers in the field. Comparisons with similar studies highlighted the importance of understanding the service user perspective separately from that of professionals.

The eight research priorities were adopted immediately by the forensic service, serving as a checklist against which all proposed research is considered before service-level approval. Additional weight is given to proposed research in an area falling within one of the top three themes. Rarely should research be undertaken within the service that does not link directly or indirectly to this list. Embedding the service user perspective to this degree avoids tokenistic involvement,^{Reference Paul and Holt6} and allows service users, as experts by experience, to directly shape research strategy and influence future research. Findings have been shared with service users, senior managers and staff working across the forensic service. Although the scope of this study was limited to the service user perspective, further study is required to explore the extent to which the research priorities of clinicians working in this forensic service are aligned with those proposed by service users, and to understand any differences.

There is a lack of evidence to guide the best method of achieving engagement.^{Reference Domecq, Prutsky, Elraiyah, Wang, Nabhan and Shippee21} There is benefit in exploring the enablers and barriers to PPI in research that involves service users in a forensic service. Early conversations should commence with representatives from community service users and carers/friends/families of service users, to develop appropriate methodologies for obtaining the views on research priorities from these two groups. The development of a carer research reference group may be an effective approach to carer involvement with this process.^{Reference Lammers and Happell8}

Further consultation with service users is required to develop a strategy to support ongoing involvement, ensuring that future research questions, methods and outcomes are acceptable and relevant to service users. The exercise of setting research priorities from the service user perspective could be repeated at 5-year intervals, to ensure that the service continues to prioritise research that is relevant to those who would benefit from the service. Given the expected benefits of PPI, the challenge is to evaluate the impact of this early service user involvement in setting research priorities, on subsequent phases of the research process and future service user involvement in, and engagement with, research.^{Reference Tritter and McCallum23,Reference Collins, Long, Page, Popay and Lobban24}