A protocol in action: Recovery approach for patients within high secure care: A 20+ year follow‐up

Abstract Objectives A person‐centred approach to recovery is increasingly represented within mainstream mental health literature. Little examination of recovery among forensic mental health patients is evidenced. This study plans to address that insufficiency. Methods This protocol paper details a novel approach to exploring recovery among a cohort of 241 patients detained under conditions of high secure care in Scotland during August 1992 to August 1993. Under discussion is the repurposing of previous research to circumnavigate length of inpatient stay commonly associated with forensic mental health care. The methodology adopted, while considering data leakage given the vulnerable participant group, will be discussed. Results Repurposing and extending previous research attempts to address the file cabinet effect with 85% of health care research being wasted and future uncertainty regarding research funding in a post‐Brexit era. This is an ongoing study. Ethical, confidentiality, privacy issues, and permissions are considered within the methodology. Conclusions Ethical arguments can be made for tracing and attempting contact with vulnerable groups under‐represented in the literature. A well‐considered methodology putting the focus on participant welfare and confidentiality at every step is essential. The reported methodology provides an opportunity to expand and re‐examine previously collected data through a contemporary lens.

long stay psychiatric patient in Scotland is approximately £2000 (2511 USD 4 ) whereas for a long stay forensic patient, the cost is approximately £5500 (6905 USD 4 ). Patients within forensic mental health services have not only to engage in recovery tasks related to their mental health but also to address the behaviours or offending that has led to their detention within forensic services. As such, the obstacles to be negotiated on the path to recovery will be significantly different for patients within forensic mental health services as opposed to those within mainstream mental health inpatient services.
When considering how to explore recovery within the context of patients passing through high secure care in Scotland, the length of stay associated with forensic inpatient events was a factor in deciding how to supplement existing studies assessing recovery. All patients resident within high secure forensic settings are detained under mental health legislation. Within 2016/2017, patients detained at the State Hospital, Scotland, were spending on average 6 years in high secure care with 65.9% of discharges for that period representing transfers to other National Health Service (NHS) hospitals. 5 The literature relating to recovery within mental health services is continuing to grow, but there is little representation of forensic patients. The need for research within this area has been previously highlighted. 6,7 More recently, Clarke et al 8

and Shepherd et al 9 have
answered this call by publishing meta synthesis of 11 and 5 qualitative studies, respectively, exploring the subjective recovery of forensic mental health patients. Recovery has been most commonly explored qualitatively using a cross-sectional design. Examination of mortality among forensic patients is one area where longitudinal studies can be found, [10][11][12] as obtaining information relating to deceased individuals is less bureaucratically intensive 13 and avoids many thorny ethical issues. We, however, are not aware of any longitudinal studies that have adopted a mixed methods approach, marrying quantitative measures with interviews to explore the subjective experience of recovery, following decade long breaks in contact.
Given the current economic climate of budget constraints and with 85% of health care research reportedly being wasted, 14 the decision was made to pull some previously collected data out of the filing drawer.
Examining it through a contemporary lens cuts costs, breathes new life into previously collected data, and facilitates exploration of issues reliant upon time passed, such as recovery within a forensic mental health context. Uncertainty surrounding Brexit and European funding has also made repurposing and enhancing existing data a more attractive pursuit.
This protocol paper therefore aims to demonstrate how a cohort of patients, first recruited as a descriptive whole population survey 15 in 1992/1993 then revisited during 2000/2001 to examine risk, offending, and violence 16 can be repurposed, together with previously collected data, to create a descriptive longitudinal design. This new recovery focused study intends to build on the previous work conducted by the chief investigator to describe in detail the recovery of patients since their detention in high secure care 20+ years ago.
As previously mentioned, we have adopted the stance that recovery within the forensic environment is different to that navigated by patients within mainstream mental health services. To that end, we are conceptualising recovery in broad terms to ensure that we examine all facets of the progression from high secure care towards the community. For the purposes of this study, we are examining clinical recovery, functional recovery, social recovery, personal recovery, and offender recovery as described by Drennan and Alred. 17 Repurposing previous work with this cohort provides a unique opportunity to explore obstructive and supportive factors towards recovery within forensic mental health services.
The overall aims and specific research questions of the recovery approach for patients within a high secure setting: a 20+ year followup are therefore:  15 Only individuals included within the original survey and currently deemed to have the capacity and wellness to consent to participate in the study by their current mental health care team or general practitioner (GP) will be eligible for inclusion.
Exclusion: Nonparticipation within the original study sample in 1992/1993 is the main exclusion criteria. This extends to participants from the original cohort who are deemed by their current care team to lack the capacity to consent and/or are not well enough (physically and/or mentally). Those for whom a suitable gatekeeper (mental health care team member/GP) cannot be located are also excluded.

| Procedures
The cohort of 241 patients identified within the original State Hospital survey will be traced and their current location found through a combination of means. Although no consent to follow up was obtained at baseline, it was considered by both the South East Scotland Regional Ethics Committee 01 and Public Benefit and Privacy Panel (PBPP) to be in the public interest to examine the pathway through services and personal recovery outcomes for this cohort. We also considered that even if consent to follow up had been requested at baseline, the length of time that has passed since then without contact could invalidate such consent due to changes in the previous participants personal circumstances.
The original paper-based participant key/cipher that listed patient identifiers, including their unique Scottish Community Health Index (CHI) number (CHI is a population register, which is used in Scotland for health care purposes) against their study index, was still in existence and held securely in a filing cabinet drawer. This key/cipher had been created at baseline (1992/1993) to ensure no duplication of participants and to address ward movement within the hospital.
The key/cipher was subsequently used for a first follow-up (2000/ 2001) and to inform analysis of the Forensic Network Census data (first collected in 2013, then annually thereafter 19 ). The key/cipher was stored as data belonging to the State Hospital, and as such, retention did not breach local retention guidelines. Regarding this current study, ethical approval was sought from the local Caldicott Guardian and research committee who provided ethical approval and use of the key/cipher was brought into line with modern data protection practice.

| Locating previous participants
Due to the length of time elapsed since initial interview, it is essential that robust recent information be obtained as to the location and status (alive or dead) of participants. Our primary approach is detailed in this section with supplemental strategies examined in 'Ethical Considerations' which also outlines how we contemplated the ethical concerns raised by this study. Support to locate participants was obtained from eDRIS (electronic Data Research and Innovation Service). "The eDRIS service is designed to provide a single point of contact and to assist ous participant remained within their service, held the capacity to consent to the study, and was physically and/or mentally well enough to be approached for participation. Only following approval from the RMO (or proxy) will an introduction to the participant be arranged.
For individuals no longer in contact with mental health services, their GP will be approached for a gatekeeper decision. If approved, letters of study introduction will be forwarded via their GP practice.
If any of the original cohort are traced to the prison service, they will initially be treated as any other previous participant and we will attempt to establish an RMO to seek a gatekeeper decision. If approved for approach, then we will liaise with the prison service to access the individual through the prison health service.
National Health Service Central Registry (NHSCR) will provide notification of those individuals who have died since baseline and continuing notifications of any deaths that occur within the cohort across the United Kingdom for the duration of the study. Obtaining cohort tracing data through eDRIS has to occur without participant consent as we need to find the previous participants to ask them for informed consent to study interview and/or case note review and/or summary of criminal history again. Structured interviews were completed to assess psychopathology and neurological side effects of medication. Tables 2 and 3 detail the   tools applied and data collected by participant interview and case note review, respectively, for each stage of the study.

| Data to be repurposed: first follow-up
One hundred and sixty-nine individuals identified from the original cohort who attracted a primary diagnosis of schizophrenia 16    For those within the cohort who remain alive, ethical approval has also been granted for a member of the patient's multidisciplinary team to be interviewed with reference to the patient regardless of the ability of the individual to take part, their refusal or consent to study. This interview was conducted at baseline, 15 and the tools used are detailed in Table 4.
We are seeking to repeat this interview to asses any objective development in functional recovery. We specifically sought ethical approval for this interview in an attempt to obtain data regarding functional recovery for every individual who is alive. We acknowledge that this approval does however raise some concerns where an individual refuses to participate. It is possible that patients could construe the staff interview as being against their wishes and subsequently impact upon the therapeutic relationship. The patient is made aware during study introduction of the involvement of their care team in terms of locating them, the gatekeeper decision and knowledge of their involvement (or refusal) from the perspective of post interview support (if required) and helping them advance their recovery. The staff interview is not explicitly discussed; however, any strong views expressed about their care team involvement would be acknowledged, and in such cases, the staff interview would not be completed. We also acknowledge the expert role of the Regional Ethical Committee who have provided approval for this aspect of the study.

| DATA MANAGEMENT AND ANALYSIS
The data collected through the use of the psychiatric rating tools, criminal records, and case note review (quantitative data) will be coded and entered into Statistical Package for Social Sciences (SPSS 44 v23), which will allow descriptive data in the form of tables and graphs to be created and the relationships between factors (characteristics) to be explored by regression analysis. Data will be fully anonymised and care taken to ensure that combinations of variables do not lead to the identification of individuals. Data will be processed into individual datasets, eg, patient interview, case note data, and staff interview; however, variables will be merged between datasets to facilitate a deeper exploration of the various aspects of recovery we are examining.
When all data have been processed and checked, it will be assessed using SPSS v23 to ensure that data are missing completely at random. The intention is then to use a process embedded in the software that multiply imputes missing values, each missing data point will then be assigned a value. Scale variables are modelled with a linear regression and categorical variables with logistic regression. Each model uses all other variables as main effects.

| Statistical analysis
Statistical analyses will be conducted using SPSS v23. Descriptive statistics will be used (numbers and proportions/percentages) for categorical variables; and means, medians, and ranges for continuous variables.
Bivariate analyses including the chi-square test will be used when comparing 2 groups on categorical variables (eg, gender) and independent samples t test when comparing 2 groups on continuous variables (eg, age). Paired samples t test will be used when comparing continuous variables within patients at 2 points in time. Where association between 2 continuous variables is examined, Pearson correlation coefficient will be used. If data are not normally distributed, appropriate nonparametric tests will be used.
With time to event data (eg, time to leave high security, time to reach community, and time to remission of psychosis), survival analysis will be used and Kaplan Meier survival curves will be plotted.

| Baseline/follow-up variable regression model
Logistic regression will be used to examine the independent association between a number of variables (eg, age, previous convictions, Variables will be selected based on the literature, bivariate analyses (see above), and clinical relevance. Backward conditional withdrawal of variances will be used identify the variables that best predict the particular outcomes.

| Clinical correlates regression model
A regression model will be developed using only variables relating to course of primary diagnosis and co-morbid conditions as independent variables to look more specifically at the clinical correlates of outcomes.

| Qualitative analysis
The data collected to explore personal and social recovery will be the words, phrases, and themes that participants use in response to the questions they will be asked about elements of recovery. An inductive interpretive thematic approach will be taken to the qualitative data captured. Analysis will be supported with NVivo 11 45

| Monitoring
Two members of the research team will undertake reviews of medical records in 5% of cases to explore interrater reliability. For 5% of interviews, the researcher will be joined by an external rater who will score the questionnaires independently. Due to the nature of the participant group who are very familiar with having medical sessions recorded, it is considered effective for data checking to record the full interview rather than just the qualitative sections.

| ETHICAL CONSIDERATIONS
The study was reviewed and approved by the South East Scotland Regional Ethical Committee 01, reference 15/SS/0015. The methodological approach approved is reflected in Figure 1. Where a participant is no longer in contact with mental health services, their GP will be approached for a gatekeeper decision. Recovery is the focus of this study, and it is deemed appropriate to attempt contact with former participants who are no longer engaged with community mental health teams. We appreciate the pressures on GP's and that they may not have as defined a relationship with the previous participant; however, the expectation is that very few individuals will require a GP gatekeeper decision and it is anticipated that GP's may actually be more cautious than RMO's when considering capacity to consent and general wellness. Only previous participants for whom we can obtain an appropriate gatekeeper decisions regarding capacity/wellness will be included in the study. Due to sensitivities surrounding the high secure care State Hospital and the vulnerability of the patient group, every participant information sheet, consent, letter, and reply slip has been carefully written and approved by the Regional Ethical Committee, National Caldicott Guardian, and PBPP. This is an ongoing study, and we are actively contacting care teams. For those living within the community, another line of protection has been noted in the form of the individuals' community psychiatric nurse/multidisciplinary care team. A researcher introduction to community-based individuals is sought from their community team following an appropriate gatekeeper decision. If the community team feel that an approach would be unsettling for their patient despite the RMO giving approval, then this is discussed with the RMO and the decision to approach reversed.
The methodology flowchart, Figure 1, outlines the 3 pronged approach originally mooted and ethically approved to locate the current whereabouts of the previous participants and their current mental health care team. Associated ethical issues for each approach and confidentiality/privacy concerns later realised are discussed.
1. Tracking discharge/transfer notes through the 241 previous participant's paper and electronic records Despite the shortcomings of the methodology adopted for the first follow-up study, 16 it seemed reasonable to include this method as a supplementing strategy for tracing the whereabouts of the previous participants current care team. Despite securing ethical approval, accessing case note data prior to consent is a privacy and confidentiality minefield. National, regional, and local permissions would need to be secured to view records for this purpose, and the local network of confirmations required to source and view case notes makes this nonfeasible. The risk of accidental data disclosure becomes too high, particularly when viewing paper records and previous experience has demonstrated that these records are not always accurate. 16  Engaging with this process ensured that every piece of correspondence and participant information sheet was not only ethical in approach but was carefully worded to ensure confidentiality and privacy particularly among previous participants who were now living independent lives within the community. Amendments to correspondence were requested by both PAC and NC post ethical approval.
Our research group was aware of the Privacy Advisory Group (PAC), National Caldicott Guardian (NC), and CHIAG requirements from previous work; however, when developing a methodology, the focus is very much on obtaining ethical approval as the first and generally essential requirement. We would like to suggest that privacy and confidentiality considerations be more closely integrated with the ethical approval process at all levels of research from student projects to applications through the online Integrated Research Application System (IRAS). The IRAS could be easily amended to highlight when further approvals or privacy and confidentiality issues require deeper consideration. In this age of "Big Data," we would welcome a more cohesive approach to ethical, confidentiality, and privacy concerns regarding the use of patient data.

| CONCLUSION
Against the backdrop of budget constraints and the funding uncertainty being ushered in by "Brexit," this protocol paper seeks to address these issues along with the apparent "waste" of health care General questions: incidents, medication,drug/alcohol, staff contact.
In the last month have you been involved in any capacity, in any incidents that have led to police involvement.
In the last month have you been involved in any capacity, in any other incidents involving verbal or physical aggression or violence?
Nomove onto next question.

Medication
Are you currently prescribed any regular medications?

Yes/No (if no move onto next question
What medications are you prescribed? (Complete table below) Are there any medications that you do not take as prescribed? (Complete table below).
What are the reasons for not taking those medicines?
Alcohol and substance use.
Have you been drinking alcohol over the last month? Yes/No (if no move onto next question).
How do you feel when you stop drinking? (ask questions exploring withdrawal/ dependence symptoms).
Have you taken any non-prescribed drugs in the last month? Yes/ No (if no move onto next question).

Contact with services
How often do you see your keyworker?
How often do you see a psychiatrist? Section 3.
Questionnaire about Process of Recovery (QPR) 33 Semi structured interview based on 7 elements of personal recovery (additional probing as required).
Recovering from an illness or finding ways to change the way you behave is a very personal experience.
Although there are many similarities, no two people recover in the same way.
The journey towards recovery can mean different things to different people. If Yes: When did you first really believe you could recover?
What helps you hold onto that belief when you are finding things difficult?
If No: Have you ever believed you could recover?
Yes What happened to make you feel that you couldn't get better?
What do you think would help you to find that belief again?

OR
No What do you think would need to happen to make you believe you could get better?

Secure Base
What do you think of the place where you live?
If Positive: do you think it has been helpful in your recovery? Explore.
If Negative: do you think it been unhelpful in your recovery journey? Explore.
(If applicable) What do you think about the staff? Explore.

Sense of self
It has been 20 or so years since you were first interviewed in the State Hospital… Has the way you see yourself changed over that time? Explore.
What has been the impact of that (change/lack of change) on your recovery journey? Explore.

Supportive relationships (should know a bit about relationships from the CPRS etc)
Have you been able to develop friendships where you can be yourself (not hide your past illness/difficulties)?
Can you tell me a bit about them?
Do you have any other people (family or professionals) who give you support?
Do you find yourself having to tell lies about your past? (Exploreto avoid stigma, fill in blanks in history…)

Empowerment and Inclusion
Can you make decisions about your recovery?
Does someone else make any decisions?
How does this make you feel?
Do you feel part of a community?
Do you feel part of the wider world?
Coping strategies Have found the right skills to help you manage your illness?
What works the best for you?
Do you still find yourself using strategies that you know aren't good for you in order to cope? Explore (self-harm, destructive, aggressive etc.) Meaning and purpose.
What sorts of things make you feel happy with life?
What sorts of things do you do that make you feel like you have achieved something? (work/placement activities/sports other activities) If Negative: What do you think would allow you to feel happy?
Post interview the researcher rates observed items of CPRS and completes the Standardised Psychiatric Assessment for Chronic Psychotic Disorders (The Manchester) 21