The effectiveness of the Older prisoner Health and Social Care Assessment and Plan (OHSCAP): a randomised controlled trial

Definition of older prisoners

The definition of an older person in prison is socially constructed. Wahidin et al. 1 noted that minimum cut-off ages as low as 45 years have been used in previous studies of older prisoners to obtain reasonable participant numbers. Both in the USA and the UK, the most frequently used minimum cut-off ages for defining ‘older’ prisoners are 50 or 55 years. 2,3 It is argued that 50 years old is an appropriate age at which to commence preventative health-care measures to reduce the financial burden and improve the health of older prisoners. 4 Furthermore, it has been suggested that prisoners aged ≥ 50 years have the physical appearance and health problems equivalent to those of a person aged 60 years living in the community. 4,5 The usefulness of providing minimum cut-off ages for defining older prisoners has been questioned. Yorston and Taylor6 emphasised the importance of considering levels of service need rather than simply referring to chronological age cut-off points when identifying older prisoners.

Studies exploring older prisoners’ health and social care needs have provided some supporting evidence to suggest that 50 years should be used as a minimum cut-off age to define older prisoners. A cross-sectional survey of older prisoners in the north-west of England found that prisoners aged 50–59 years have similar health and social care needs to those aged ≥ 60 years living in the community, thus suggesting it to be an appropriate cut-off point. 7 Fifty years is, therefore, the minimum cut-off age throughout the current study, unless otherwise indicated.

The increasing number of older prisoners

There has been a recent, considerable increase in the number of older prisoners across many developed countries, including the USA,8 Japan,9 Canada,10 Australia,11 France,12 England and Wales. 13 It is estimated that, by 2030, one-third of all prisoners in the USA will be ≥ 55 years old, equating to > 400,000 prisoners, a 4400% increase from 1980. 8

In England and Wales, people aged ≥ 50 years currently account for 15% of the prison population, with 12,577 in this age group in prison. 14 Those aged ≥ 60 years are the fastest growing age group in the prison estate and, strikingly, the number of prisoners aged ≥ 70 years old is projected to increase by 35% by 2020. 15

The rise in the number of older prisoners is a consequence of a number of factors. The increase is, in part, the result of an ageing population and increases in the number of older people committing crimes. 16 Importantly, it is also a consequence of changes to sentencing practices, with courts sentencing higher numbers of older people to increased periods of imprisonment. 17 The introduction of indeterminate sentences has also contributed to the increase. 18 Furthermore, enhanced forensic evidence and enhanced reporting resulted in greater numbers being convicted for crimes committed in previous decades. 19 In England and Wales, 42% of men in prison aged > 50 years have been convicted of sex offences, the most common offence for this group. 13

The policy context

There is no national strategy for the care of older prisoners, despite repeated recommendations. 19–22 However, prisoners should have access to the same quality and range of health services as they would receive in the community. 22,23 The formal process of passing on the responsibility of employing health-care staff and delivering health care from the prison service to the NHS was completed in 2006 and, since then, all NHS standards and policies apply in prison.

The most relevant NHS policy to older prisoners is the National Service Framework (NSF) for Older People. 24 This aimed to ensure ‘fair, high quality, integrated health- and social-care services for older people’ and outlined eight key standards for the NHS and partners in local authority and community sectors to meet. 24 The NSF incorporated a short paragraph concerning older adults’ care in prison, emphasising the need for collaborative working to support older offenders. Strikingly, Her Majesty’s Chief Inspector of Prisons (HMCIP)18 review of older prisoner care found that that many of the NSF standards were not being met in prison.

A wide range of definitions of social care are employed in the community. 25 Research conducted with staff working in prisons has suggested that there is a lack of agreement concerning this definition. 26 Some staff adopted broad definitions of social care that included assistance with finances, housing and employment, whereas others used narrower definitions, referring only to personal care concerns such as washing and dressing. Without clear agreement of what social care is, it is difficult to determine who is responsible for its provision to older prisoners. This lack of clarity has previously been acknowledged. 27,28 Where social care is provided, it has largely, and inappropriately, been seen as the sole responsibility of prison health-care services, as opposed to a wider multidisciplinary obligation. 18

In July 2012, the government published the white paper, ‘Caring for our future: reforming care and support’, which detailed the reform of adult social care in England and Wales. The white paper described a lack of clarity concerning responsibility for assessing and providing social care support to prisoners. 29 The resultant Care Act,30 introduced in 2014, clearly stipulated that the local authority where the prisoner resides was responsible for providing social care. The extent to which local authorities have become involved in the care of older prisoners and the form this has taken since the introduction of the 2014 Care Act is unknown. Cornish et al. 31 have recommended that a review of the implementation of the 2014 Care Act across the English and Welsh prison estate should be undertaken.

Health and social care needs

Assessment of need and care planning

Professionals conducting health and social care assessments with older people in the community face a number of challenges, namely the under-reporting of need, poor-quality tools, variations in assessments across professions and geographical areas and a lack of agreement among professionals working in different sectors. 50–53 The NSF for older people introduced processes to improve assessment procedures, most notably the introduction of the Single Assessment Process (SAP). 24 The aims of the SAP were to standardise assessments across different organisations and geographical areas, raise the standard of assessment, assist with information sharing, prevent duplication and ensure a comprehensive assessment of need. 24 Studies have suggested some improved identification of older adults’ health and social care needs; however, the extent to which these improvements are a result of the SAP is unknown. 50

Prisons are required to conduct health assessments for all new prisoners on reception into custody. 54 The current standardised reception health assessment tool, introduced in 2004, is designed to identify immediate health concerns, with a recommended second health screen conducted at a later date to allow prisoners to discuss their health needs in more depth. However, research suggests that there are low completion rates for the second, non-mandatory health screen. 55 A further criticism is that it is too often viewed as a one-off process, as opposed to a continuous pathway to care. 56 The tool does not investigate social care need.

The Department of Health57 guidance entitled A Pathway to Care for Older Prisoners: A Toolkit for Good Practice recommends the use of health and social care assessments specifically designed for older prisoners, with reassessments and revised care planning taking place at least every 6 months. There is no standardised older prisoner health and social care assessment in England and Wales; however, some establishments have developed their own. Cooney and Braggins58 reported that 40% of establishments in their survey had no specific assessments in place for older prisoners. Senior et al. 27 found that 81% had not established specialised assessments for older people on prison entry. The specialised assessments introduced in some establishments have not been evaluated, and anecdotal evidence suggests that they are not always delivered systematically to all prisoners. Consequently, in the majority of prisons, the identification of health-care need for older prisoners is largely dependent on information obtained by the generic screening instrument. 59 It is known that, if health issues are not identified at reception, they are unlikely to be detected throughout a person’s time in prison. 60

Health and social care services

There are limited examples of older prisoners being provided with additional specialised services; however, in general, they receive the same treatment as younger prisoners. The implications of being treated the same as younger prisoners have been discussed. 61 Crawley61 conducted interviews with those aged ≥ 65 years in two prisons where there were separate wings for older prisoners, and in two with no such separation. Sustained observations of prisoners’ daily life were also conducted. Crawley61 found that prison officers had to find a balance between consistency and flexibility when supporting older prisoners. She highlighted that ‘treating everyone the same does not always equate to equality’. Crawley61 used the term ‘institutionalised thoughtlessness’ to describe the ways in which ‘prison regimes simply roll on, with little reference to the needs and sensibilities of the old’. Examples included not being provided with sufficient time to move from one location to another or to complete specific activities, being provided with hard chairs and top bunks to sleep on, insufficient warm clothing in cold weather, queuing for long periods of time for medication, unavailability of grab rails in the showers and being unable to go outside for exercise because exercise areas lack seating or a readily accessible toilet. These examples suggest that older prisoners are being ‘doubly punished’ because, in addition to their loss of freedom, they experience inadequate care that is not equivalent to that in the community. 58

Release and resettlement

The post-release period poses particular risks for prisoners in terms of their physical and mental health. 62 Research has identified complex relationships between reoffending and nine key factors, namely education, employment, drug and alcohol misuse, attitudes and self-control, institutionalisation and life skills, housing, financial support and debt, family networks, and mental and physical health. 62 However, 42% of released prisoners are of no fixed abode, 50% have no general practitioner (GP), 50% reoffend within 2 years, debt problems for one-third worsened whilst in custody and 60% are unemployed. 63

Recent research conducted by Wilson64 has explored how prisoners with serious mental health illness seek help after leaving prison. In this study, 63% identified housing and 35% identified financial assistance as one of their two most important service needs; only 12% selected treatment services, thus emphasising the importance of meeting basic needs as well as providing treatment services on release from prison. Released prisoners are also at a greater risk of suicide than the general population, particularly in the first few weeks after their release. 65 Furthermore, discharged prisoners have an increased risk of drug overdose. 66 Consequently, it is essential that contact with care services is maintained on release from prison. 67,68 The National Association for the Care and Resettlement of Offenders recommends that the resettlement needs of every offender should be considered from the start of their prison sentence,69 although little is known regarding the extent to which this is realised because there is a paucity of research concerning discharge planning for prisoners. The UK government has argued that effective release planning is a key priority to reduce barriers to resettlement. 70 However, it has been argued that the current discharge planning process and a lack of effective multiagency working are current barriers to effective resettlement in the UK71 and the USA. 68

The HMCIP review of older prisoner care identified grave concerns that needs were not planned or provided for after release,18 based on identifying only four prisons in England and Wales that provided specific resettlement support for older prisoners. It repeated its previous recommendation72 that the specific resettlement needs of older prisoners should be accurately assessed and provided for on release. The Department of Health’s toolkit for good practice provided recommendations around preparing older prisoners for release and supporting their transition into the community. 73 The Department of Health stipulated that release planning should involve the conduction of a pre-release health and welfare assessment, a face-to-face assessment by a social worker, collaboration with external organisations and the organisation of a care package. This guidance also emphasised the importance of monitoring the progress of released prisoners to ensure that they have access to the appropriate services.

A limited number of studies have explored older prisoners’ concerns and issues about release prior to discharge. 74,75 Findings suggest that older prisoners struggle disproportionately with resettlement as a result of reduced support networks and their increased likelihood of health and mobility problems. In addition, older prisoners experience intense anxieties about their release, and inadequately understand the resettlement process. Key concerns include where they are going to live and how they will get there; their physical safety (for convicted sex offenders in particular); loss of personal possessions and support networks; and access to health-care support for chronic illness. Concerns prior to discharge are so intense that many feel that it would be better to stay in prison. In spite of these increased needs, older prisoners’ resettlement needs are often ignored; it has been suggested that this is because they are generally considered less of a risk75 and are less assertive than younger peers. 76

One study explored whether or not older prisoners’ fears about release became reality. Forsyth et al. 77 conducted interviews with older prisoners prior to and after release, and reported that older prisoners perceived release planning to be effectively non-existent. Those due to reside in Probation Approved Premises were very anxious about the prospect and were concerned about sharing accommodation with younger people who may abuse substances and be physically violent towards them. However, interviews with older prisoners after release revealed that the immediate health and social care needs of those housed in Probation Approved Premises were generally fairly well met. Many with complex social care needs were inappropriately housed; for example, some wheelchair users were housed in accommodation accessed by steps, meaning that they could not enter or exit the premises independently. Similarly, a lack of suitably adapted bathroom facilities compromised safety and independence in this aspect of self-care. It is possible that the Offender Rehabilitation Bill (2013–14)78 will improve this situation. The Bill proposed the introduction of 70 resettlement prisons, with offenders being sent to a prison close to their release destination at least 3 months prior to discharge. The extent to which this will improve release planning for all prisoners, including those in older age, is as yet unknown.

Primary hypothesis

The OHSCAP will significantly increase the proportion of met health and social care needs 3 months after prison entry, compared with TAU controls.

Secondary hypothesis

Compared with TAU controls, the OHSCAP will significantly improve the following at 3 months after prison entry:

1. HRQoL

2. depression

3. functional health and well-being

4. quality of health and social care planning.

Ethics approval

Ethics approval for the study was granted by the Research Ethics Committee for Wales in May 2013 (reference number 13/WA/0108). National Offender Management Service (NOMS) research approval was provided in July 2013 (reference number 2013-115). The trial was registered with the International Standard RCT Number ISRCTN11841493. Additionally, all required site-specific permissions and research governance approvals (research and development) were obtained from the relevant NHS trusts.

Changes to protocol

1. Increase in number of sites.

   We closely monitored data collection to ensure that we would meet our targets. It became apparent that we needed to add further sites in order to meet the follow-up target of 392 participants. Additional sites were selected for pragmatic reasons while still ensuring that we had included a range of prison types. Consequently, the number of sites increased from 4 to 10. We are confident that the increase in number of sites had no impact on the overall study design.

2. Increase in baseline target.

   As we progressed into the trial, it became apparent that our attrition rate was much higher than the expected 10%, at almost 20%. This was mainly because of retention issues in the local (remand) prison sites, where it proved to be harder than expected to identify individuals who would remain in custody for the 3-month follow-up period. As a result of this, we extended our recruitment period and increased our baseline recruitment target to a maximum of 502 participants at baseline, from the original target of 462.

3. Changes to assessment tools.

   A number of changes were made to the assessment tools before data collection commenced. The SF-3679 was replaced by the Bristol Activities of Daily Living Scale (BADLS)80 because it was considered more appropriate for use in prison. The Client Service Receipt Inventory81 was replaced by the Secure Facilities Service Us Schedule (SF-SUS)82 for the same reason. We also added the following tools in order to describe the sample: the Operational Criteria Checklist for Psychotic and Affective Illness (OPCRIT),83 PriSnQuest84 and the Burvill grid. 36

Inclusion criteria

To be eligible for inclusion, participants had to:

1. be aged ≥ 50 years

2. have a known release date (convicted) or likely release date (unconvicted) of at least 3 months after their prison entry date.

Exclusion criteria

The following individuals were excluded:

1. those who did not have the capacity to consent

2. those considered by prison or health-care staff not safe to interview alone as a result of their current risk assessment

3. those previously included in the study.

Recruitment procedure

An administrator within each of the prisons identified potential participants who fulfilled the inclusion criteria. They did this by running a search on all prisoners newly received from court into their establishment on the prison computer system [Computer – National Offender Management Information System (C-NOMIS)]. An administrator was selected for this role because they did not need to access health information, but they did need to access sentence information. It would not have been appropriate for the members of staff delivering the OHSCAP to conduct the initial approach because this would have affected anonymity, and may have then had an impact on the support the TAU group received. The administrators were also required to inform potential participants of the proposed study. If the service user expressed an interest in learning more, the administrator requested their permission to pass their name on to a member of the research team. A researcher then arranged a time to talk to the potential participant to discuss the study further and ask them to consider participating in it.

Release dates for unconvicted prisoners were predicted using an adapted version of an algorithm developed for a previous study. 85 The algorithm and accompanying offence list is appended (see Appendix 2).

Consent

Informed consent was sought from all potential participants prior to taking part. Researchers explained the project, provided an information sheet and described the relevant ethical rights as part of the consent process. Sensitivity was shown to the high levels of learning difficulties and vulnerability in this population, with researchers reading and explaining the information sheet, when required, and remaining aware of the potential for any coercion.

All participants were informed that participation was voluntary and that they were free to withdraw at any point, with their decision to participate or otherwise having no bearing on the future care they received or their other legal rights.

Confidentiality

Participants were all informed of the arrangements to ensure confidentiality, including the limits of this, and data protection.

During the consent process, the limits of confidentiality were clearly outlined with participants. Participants were informed that all information disclosed during the research process was confidential unless (1) the information imparted revealed real risks of harm (e.g. self-harm, suicide, violence towards others) that needed to be acted on to safeguard the participant or others; and/or (2) the participant revealed criminal activity previously unknown to a relevant authority. This covered the potential for reporting previously undisclosed offending outside custody, or criminal offences committed while in prison including, but not limited to, illicit drug importation/use, importation or possession of other prohibited items (e.g. mobile phones), assaults on other prisoners and/or other criminal activity (e.g. continuing involvement with crime outside prison). How a required breach of confidentially would be dealt with depended on the circumstances. Risks of self-harm or suicide would involve the researcher either starting self-harm management processes [Assessment, Care in Custody and Teamwork (ACCT)] if the risk had not been previously identified, or liaising with staff directly to contribute to a person’s ongoing care under ACCT if the risk was already known. Reporting of previous/current criminal activity would be reported to the prison’s security department using routine procedures.

No circumstances arose that required a breach of confidentiality.

Each participant was allocated a unique participant identification number. This identifier was used to link participants’ study data with identifiable data, which were stored securely and separately.

Individuals lacking capacity

Researchers received training in assessing capacity using the two-stage process outlined in the Mental Capacity Act (2005). 86 If there was any indication that an individual lacked the capacity to consent, that individual was excluded from participation.

Randomisation

An individual-level randomised design was selected for two key reasons. First, it was anticipated that there would be minimal contamination because older prisoners are not usually systematically identified on entry into prison and, therefore, the older prisoner lead does not usually come into contact with the older prisoners unless specific issues arise. Second, a clustered or stepped-wedge design was considered; however, such designs would not have been feasible to implement because many more institutions would be required to participate, thus having an impact on cost and time.

Randomisation was undertaken by the Manchester Academic Health Science Centre Clinical Trials Unit (MAHSC-CTU). Participants were randomised to receive the OHSCAP or TAU. The MAHSC-CTU provided a telephone-based central randomisation service for the trial. The allocation method was minimisation with a random element using imbalance scores over the margins of two factors: institution and baseline number of unmet needs (0, 1, 2, 3, ≥ 4). With minimisation, the group allocated to the next participant is dependent on the characteristics of existing participants. 87 The aim is that the allocation of each participant should minimise the imbalance across groups. In order to achieve this, provisional imbalance scores were calculated (one for each trial arm assuming allocation of the case to that arm). The imbalance score was Sum (|n1 – n2|), where the sum is taken over the observed levels of each factor of the ‘case at hand’ and n1 and n2 are the accrued cases to date in the two trial arms for the given levels, including the provisional allocation. If the imbalance scores were tied, we considered imbalance in the totals in each arm without reference to the factors. Allocation was made to the arm that would yield lower imbalance with probability 0.75 or with probability 0.5 if scores were tied. This random allocation sequence was generated by David Ryder [statistician, Clinical Trials Unit (CTU)].

The procedures for randomisation were as follows. Once a participant had consented to participate and had been confirmed as eligible for the trial, and the baseline assessments were completed, contact was made with the MAHSC-CTU to be allocated a participant identification number and allocated to either the intervention or the TAU group. The following information was provided: a trial password (allocated by the project manager), the centre name, the participant’s initials and date of birth, and the caller’s name. A participant identification number was allocated immediately, followed by, via e-mail, the result of allocation to intervention or TAU.

Blinding

When possible, RCTs should be double blinded, that is, the participants and researchers should not be aware of which group they have been allocated to. 88 Participants unavoidably became aware of which group they had been allocated to when they received the intervention. Furthermore, the researchers did know which group some of the participants belonged to because 14 of the participants in the intervention group were invited to take part in qualitative semistructured interviews. Within the current study, however, quantitative data analysis was conducted blind. Identifying variables were removed by a statistician at the CTU before the data were provided to the researcher conducting the analysis.

Intervention

The OHSCAP was developed and implemented as part of a previous study funded by the National Institute for Health Research (NIHR) Service Delivery and Organisation programme. An action learning group (including prisoners, NHS staff and prison staff) at one prison in England developed the OHSCAP. 27

The OHSCAP is a structured approach for better identifying and managing the health and social care needs of older prisoners. The previous study showed that the OHSCAP was acceptable to prisoners and staff, could be integrated into current prison/health-care processes, assisted effective multiagency working, provided an opportunity for prisoners to raise concerns that would have otherwise gone unreported and could be successfully conducted by a prison officer.

The OHSCAP is paper based and information collected is uploaded onto existing prison, health and offender management computer programs. A copy of the OHSCAP is appended (within the OHSCAP manual; see Appendix 3). The first page of the OHSCAP includes instructions for completion and background information. A table for collecting basic demographic information including name, age, date of birth and NOMS number is also included. The OHSCAP consists of an assessment, a care plan and reviews of these.

The assessment includes a series of open questions to facilitate discussion, and is divided into three key parts, namely social, well-being and discharge planning. The social assessment includes open questions around relationships, activities and mobility. The well-being assessment includes exploratory questions around emotional well-being, physical well-being, and medications and treatment. A section for other concerns is also incorporated. The final section of the assessment includes open questions around discharge planning. A series of ‘trigger’ open questions are included for each of these sections. A place for the signatures of those conducting the assessment and the prisoner is also incorporated.

The care plan consists of a matrix with five columns. These are (1) issue raised from assessment, (2) aim of the proposed action, (3) action (including by whom and when), (4) date to be reviewed and rationale and (5) status of action.

The review section includes space for a date and details of the reviewer. It also takes the form of a matrix and includes the following columns: (1) progress since last review, (2) action planned and (3) next review with rationale.

The assessment is conducted approximately 1–2 weeks after an older prisoner enters prison. This was based on discussion in the action learning group around the wealth of information that is both asked of and provided to other prisoners, immediately after they arrive in prison, and around how prisoners are suffering from ‘entry shock’ and would find it difficult to cope with a further assessment during the initial entry period. In addition, it was felt that older prisoners require a period to settle into the prison in order to be able to identify their needs effectively. The older prisoner lead accesses the prison’s computer system, C-NOMIS, on a daily basis to identify any prisoners aged ≥ 50 years newly received from court into the prison, whose known release date (convicted prisoners) or likely release date (unconvicted), is at least 3 months after prison entry. The older prisoner lead conducts the assessment on a one-to-one basis with the older prisoner. The care plan is completed in conjunction with the older prisoner, who is provided with a copy of the OHSCAP. In addition, a summary of the OHSCAP is entered onto the prison computerised information system (C-NOMIS) and a copy of the OHSCAP is scanned onto the prison computerised clinical records (SystmOne; The Phoenix Partnership, Leeds, UK) and probation computer records (Offender Assessment System). The older prisoner lead conducts reviews as considered necessary and develops further action plans. Reviews of care plans involve ensuring that actions have been completed and pursuing these as necessary.

It was initially intended that the older prisoner leads, who are usually prison officers, would deliver the intervention. This followed the recommendation of an earlier action learning group comprising prisoners, NHS staff and prison staff at one particular prison who decided that they would be the most appropriate individuals to conduct the OHSCAP in their establishment. The previous study found that prisoners were happy to discuss their health and social care issues with prison officers. However, in 6 of the 10 sites in this study, health-care workers delivered the OHSCAP, as this was deemed by senior managers at the sites to be more appropriate or more achievable within their prison at the time the project was being set up. This was largely because of the benchmarking process that was taking place at the time, which was resulting in a reduction in prison officers and the loss of some roles, including the disability liaison officer (DLO) in some establishments. The DLO is responsible for supporting prisoners with disabilities. They assess the needs of disabled prisoners and try to ensure that adaptations are made when necessary. The DLO is often given the responsibility of supporting older prisoners as well. How they do this varies from prison to prison, but may include facilitating older prisoner groups. The OHSCAP facilitator received training before commencing this work.

Training

All of the OHSCAP facilitators were trained to deliver the OHSCAP, in line with the OHSCAP manual (see Appendix 3). Throughout the study, two training sessions were held at the University of Manchester, which were attended by facilitators from all study sites. Some of the facilitators attended both sessions and were able to share experiences and good practice, having already completed the OHSCAP process with some prisoner participants. Ongoing support was also offered by Dr Elizabeth Walsh, who has vast experience as a clinician within prison settings. She acted as a mentor to facilitators and was contactable by telephone and e-mail, should they have any questions or need any reassurance. Additional site-specific training sessions were provided at prisons that joined part-way through the study in an attempt to bolster recruitment.

Treatment as usual

Treatment as usual included the standard, non-age-specific health assessment carried out at prison entry. 59 Support provided as TAU varied from prison to prison, but included interventions such as older prisoner social groups, peer carers and healthy man checks. Ongoing assessments and interventions followed local procedures at each establishment. Previous research has indicated that the identification of health and social needs and subsequent care planning is generally ad hoc and inadequate. 55

Primary outcome measure

The primary outcome measure was the mean number of unmet health and social care needs at 3 months, as measured by the Camberwell Assessment of Need – Short Forensic Version (CANFOR). 89 CANFOR includes 25 domains, namely accommodation, food, looking after the living environment, self-care, daytime activities, physical health, psychotic symptoms, information about condition and treatment, psychological distress, safety to self, safety to others, alcohol, drugs, company, intimate relationships, sexual expression, childcare, basic education, telephone, transport, money, benefits, treatment and access to anti-recidivist interventions for sexual offending and arson. Participants are asked to identify if they have no need, a met need or an unmet need in each of these domains. The extended version of this tool incorporates the perspectives of carers or members of staff working with participants. The short version of the tool was selected because of previous difficulties experienced during the pilot study27 in identifying a member of staff who knew the older prisoner well enough to answer these questions. This was a particular challenge immediately after entry into the prison. A limitation of the tool is that it provides respondents only with an opportunity to state whether their needs in each domain are met or unmet, rather than any indication of the extent to which this is the case. Furthermore, it does not provide specific details about which areas of the domains are met/unmet. However, the CANFOR is the only validated tool available for assessing need among forensic populations. In addition, it has been successfully used in previous studies with older prisoners. 7,20,27

Secondary outcome measures

1. Functional health and well-being and ADL as measured by BADLS. 80 The BADLS measure incorporates the 20 domains of preparing food, eating, preparing drink, drinking, dressing, hygiene, teeth, bath/shower, toilet/commode, transfers, mobility, orientation (time), orientation (space), communications, telephone, housework/gardening, shopping, finances, games/hobbies and transport. The scale was originally designed for patients with dementia, but has recently been used with an older prisoner population. 7,20 Hayes et al. 7,20 omitted items 18 (money) and 20 (transport) because they are not relevant to prisoners, and more relevant versions of these items are covered in the CANFOR. 89 This approach was also adopted for the current study. Total scores were not used because this would affect the internal consistency of the measure. Items were, therefore, to be examined separately.

2. Depression as measured by the Geriatric Depression Scale – short form (GDS-15). 90 The GDS-15 contains 15 questions, which may be answered with ‘yes’ or ‘no’. Each item to which a response of ‘yes’ indicates depression generates a score of 1. A total scale score of ≥ 5 suggests mild depression, and guidelines suggest that further investigation is required, and a total score of ≥ 10 almost always indicates severe depression. The scale has been validated for use with older people and has been used with older prisoners in a previous study. 41 Murdoch et al. 41 adapted question 12 of the assessment from ‘Do you prefer to stay at home rather than go out and do new things?’ to ‘Do you go on association?’. Association refers to prisoners leaving their cells to mix with other prisoners and participate in games and/or other social events. The current study also made this alteration to the question to make it appropriate for prison use.

3. Health-related quality of life as measured by the EuroQoL-5 Dimensions, five-level version (EQ-5D-5L). 91 The EQ-5D-5L is a standardised assessment that has been widely adopted and validated in a variety of different populations. 92 The EQ-5D-5L encompasses the following five domains of health: mobility; ability to self-care; ability to undertake usual activities; pain/discomfort; and anxiety/depression. For each of these, five options are provided (e.g. no problem, slight problem, moderate problem, severe problem, unable). The EQ-5D-5L will be used because the current study is part of a broader study measuring the cost-effectiveness of the OHSCAP. The EQ-5D-5L is recommended by the National Institute for Health and Care Excellence (NICE) to enhance the comparability of different studies. 92 The visual analogue scale aspect of this tool was excluded to reduce participant burden, as this aspect of the tool is not widely used. This was administered face to face. The face-to-face version of this tool was not available when this research commenced.

4. Bespoke OHSCAP research tool. The extent to which specific health and social care needs had been addressed was measured by the bespoke OHSCAP tool specifically designed for the study. It has been necessary to design a specific tool because no standardised assessments will measure improvements in specific issues for older prisoners, such as hearing instructions, receiving information about their release or delays in receiving medication. The bespoke OHSCAP research tool is based around the three sections of the OHSCAP assessment, that is, physical health, well-being and discharge from prison. A number of specific topics are addressed in each of these sections. For each topic, participants are asked if they are experiencing difficulties, if they are receiving help and whether or not they have an unmet need in this area.

The following tools were also used at baseline to describe the sample.

1. PriSnQuest. 84 PriSnQuest consists of seven questions, each with yes or no responses. A score of three or more indicates that a further assessment for mental health is required. PriSnQuest has been selected because it has been developed from other standardised assessments, is widely used in prison and is short, reducing participant burden.

2. Burvill grid to measure physical health. 85 The Burvill grid was used to obtain data on the physical health of participants. The Burvill grid categorises physical disorders into different body systems. Each system is rated according to severity of the disorder (coded 0 = absent, 1 = mild, 2 = moderate, 3 = severe) and disability as a consequence of the disorder (coded 0 = none, 1 = little, 2 = some, 3 = great deal). Physical problems are also defined as acute or chronic. A disorder is considered chronic if it has been present for at least the previous 3 months. The Burvill grid has been used in previous studies of older prisoners. 20,35

All data were collected between 28 January 2014 and 6 April 2016.

Fidelity

The fidelity, and quality, of implementation of the OHSCAP were assessed using an audit tool specifically designed for this study (see Appendix 4 and Chapters 4 and 5).

Sample size

From our previous work (a cross-sectional study assessing the unmet needs of 100 older prisoners at baseline) the mean number of unmet needs was assumed to be 2.71 [standard deviation (SD) 2.65 unmet needs]. The distribution of unmet needs, ranging from 0 to 25, was unsurprisingly positively skewed, with a median number of unmet needs of 2. Although we did not have supporting data, we assumed that this distribution would be broadly similar at 3 months’ follow-up in the TAU group. The purpose of the current study was to see if the average number of unmet needs can be reduced with the OHSCAP intervention. For the study to be practice-changing, we believed that at least a 30% reduction, to a mean of 1.90 unmet needs, was required, and thus powered the study proposal accordingly [mean 1 = 2.71, mean 2 = 1.90 (= 0.7 × mean 1), common SD 2.65 unmet needs, implies n = 169 participants per group for 80% power in a two-tailed t-test at a 5% level of significance]. A 1 : 1 randomised trial employing a two-tailed t-test at the 5% level of significance would require 169 per group for 80% power if the true means are 2.71 and 1.90 unmet needs and the common SD is 2.65 unmet needs. As the distributions would be quite skewed, it was considered preferable at the planning stages to use a non-parametric Mann–Whitney U-test. The asymptotic relative efficiency of this test is at worst 0.864 compared with the t-test, and so a conservative approach is to inflate the proposed sample size accordingly, that is, 169/0.864 = 196 per group. Linear regression was used to analyse the primary outcome measure with bootstrapping to account for skewness, and adjust for minimisation factors (institution and baseline measures of unmet need). This allowed for a more sophisticated approach than would have been adopted if we had used the Mann–Whitney U-test as originally planned and, consequently, we will have > 80% power. The trial stopped when we achieved sufficient numbers at baseline to estimate that we would reach our follow-up target.

Primary outcomes

The primary hypothesis for the change in the mean number of unmet needs, as measured by the CANFOR, was analysed using appropriate regression models. We adjusted for baseline characteristics used in the minimisation process, for example site and number of unmet needs at baseline. We used bootstrapping to account for the skewness in the outcome of the data. The 95% CIs were calculated around all key effect size measures and two-sided p-values were reported.

In addition to analysing the mean number of unmet needs, we used a Poisson model to analyse the data as counts.

The CANFOR was used because it was the most appropriate available tool for assessing unmet health and social care needs within the prison population. The research team were, however, aware that there were certain domains of the CANFOR that the OHSCAP specifically aimed to address and some domains of the CANFOR that the OHSCAP did not aim to address. The research team therefore felt that it would be useful to analyse the data separately for the specific domains of the CANFOR that were considered most relevant to the OHSCAP. The aim of this analysis was to gain a more detailed understanding of the specific domains of the CANFOR that the OHSCAP appeared to assist more with and of which domains the OHSCAP was less able to address.

Logistic regression was used to conduct this analysis, with adjustment for site and number of unmet needs at baseline. Table 2 shows the ratings the research team applied to each of the domains for relevance to the OHSCAP. The highest value is 3 and the lowest is 1.

Secondary outcomes

Similar approaches were adopted for the secondary outcomes, with the form of regression depending on the distribution of the particular outcome. Linear models were used for continuous outcomes, and logistic regression for binary outcomes. Bootstrapping was used only with linear regression. For each of the secondary outcome measures we adjusted for establishment at baseline and for that specific secondary outcome measure at baseline.

Missing data

Data completeness and accuracy were confirmed by the MAHSC-CTU during the data entry process. If, during the data collection and inputting processes, a field was found to have been left blank, the Data Manager at the CTU raised a query and the research team clarified whether the missing information could be obtained or confirmed that it was not available. This assisted in preventing unexplained missing data. The research team conducted further checks to ensure that the data were complete and accurate. Missing data were minimal and, therefore, it was not necessary to compute any missing data.

Database and data entry checks

All data entry checks were conducted by the CTU throughout the duration of the trial. Any missing or inconsistent data were clarified with sites through the data query/correction process. In line with the CTU’s policies, 100% of critical fields and 2% of non-critical fields were quality checked.

Definitions

Reporting serious adverse events

In this study, SAEs were reported to the Chief Investigator (JS), regardless of relatedness, within 24 hours of the principal investigator (or authorised delegate) becoming aware of the event. All SAEs deemed to have a causal relationship were reported to the Trial Steering Committee. Any non-serious AEs, regardless of relatedness, were not reported in this study.

Sample

Overall, 220 OHSCAPs were completed, of which 150 (68.2%) were available for potential audit. Reasons for loss can be seen in Table 14.

Only one site was able to provide 100% of completed OHSCAP assessments to the research team for audit. In most sites, between 1 and 10 OHSCAP assessments were lost, with the exception of site 2, where an entire folder of filed assessments went missing during office moves. The prison health-care provider changed at this site and large volumes of paperwork were securely destroyed as part of the transition.

Assessments were audited from 9 of the 10 study sites. One site, from which only four participants were ever recruited, was unable to provide copies of the two OHSCAP assessments that it completed.

Relationships

In total, 40 problems were identified in this area, but only 21 (52.5%) were directly addressed in subsequent care plans.

Within the category of ‘other relationship needs’, issues identified included concerns for personal safety or family members, problems using prison telephones, financial difficulties, lost property and wanting transfers between wings or prisons. Only 16 out of 26 (61.5%) of these problems were addressed in care plans. Financial problems were the least likely to be addressed, with four out of five not translating into care plan actions. Two specific problems that were identified and not addressed are cause for particular concern. One assessment stated that the individual ‘doesn’t feel safe’ and another stated that the participant was ‘worried about children in general’, but neither of these problems led to actions in the subsequent care plan.

Activities

Across all subcategories, 21 problems relating to activities were identified. Of these, 15 (68.2%) resulted in a care plan action.

Mobility

Of the 31 mobility problems identified during the OHSCAP assessments, only 16 (51.6%) were actioned in the care plans that followed. Difficulties with getting in or out of bed were more often not addressed (63.6%) than any other type of problem within this category.

With regard to ‘other mobility needs’, a small majority (57.1%) were addressed in care plans, but difficulties that remained unaddressed included complaints of leg and back pain and swollen feet.

Emotional well-being

Only 7 of the 17 (41.2%) emotional well-being problems raised during assessment resulted in care plan actions. Difficulties sleeping were the most common type of problem reported (n = 7), and also the least often addressed (28.6%).

Half of the problems within the ‘other emotional well-being’ category were not addressed in care plans. These included reports of low mood, ‘concern regarding child contact’ and ‘lost mum and family doesn’t speak’.

Physical well-being

Within this category were five reports of difficulties with accessing health care, eight problems relating to glasses or contact lenses and 17 ‘other physical well-being’ concerns. With 70% of problems identified leading to a subsequent care plan action, physical well-being problems were addressed more often than any other type of problem. However, some concerning issues remained unaddressed despite the OHSCAP process, including five within the ‘other’ category relating to pain, a chest infection and a prisoner’s ventilator.

Medication

Over one-third (35.7%) of identified outstanding medication needs were not addressed through the care planning process. All of the issues that remained outstanding related to the appropriateness and/or timing of medication.

Discharge planning

The majority8,13 of problems in this area were not addressed during the OHSCAP care planning process. None of the identified financial concerns was actioned and only half of those relating to accommodation had corresponding actions in the subsequent care plan. ‘Other’ needs, including employment problems and concerns about relationships outside prison, were also apparently overlooked by facilitators.

Other needs

In total, 27 ‘other’ needs were highlighted in the final section of the OHSCAP assessment, and over 80% of these appeared in subsequent care plans. Problems relating to privileges, transfers, cell sharing and a request for suitable winter clothing were not explicitly addressed.

External

Only five external referrals were made across all of the 150 OHSCAPs and 65 care plans reviewed. One individual was put in touch with a solicitor, one with the probation service and three with housing support. No individual received multiple external referrals.

Internal

In total, 43 internal referrals were made. Table 17 shows them broken down by type.

Within the ‘Other’ category, the most common referral was to a prison officer/wing staff (n = 6), followed by the equalities department, offender management unit and carers (presumably prisoner carers; n = 2). Other referrals documented included to reception, education, an anxiety group and remedial gym.

Referrals by facilitator role

Prison officer facilitators referred 23 individuals to 26 professionals. Only nine (34.6%) of these referrals were to health-care professionals. Health-care referrals were made by prison officers to a psychiatrist, GPs, opticians and the mental health in-reach team.

In contrast, health-care staff facilitators made 21 referrals in total for 13 individuals, with 15 (71.4%) of them being to health-care professionals. The ‘non-health-care’ referrals they made included education, the kitchen (participant’s workplace), the offender management unit and the gym.

Staff shortages

Prison officers described a staffing crisis across the prison estate. This was considered to have had an impact on day-to-day prison practices, as illustrated by this prison officer’s quotation:

The staffing levels are just ridiculous, to be honest, every day there’s a wing shut down because there’s no staff, prisoners are noticing it, prison officers are stressed, it’s dangerous at the minute, to be honest, the staffing levels.

Prison officer 3

All of the prison officers delivering the OHSCAP were based within wider safety custody and diversity teams. These teams conducted a variety of tasks to support prisoners. This included monitoring prisoners who were at risk of suicide and self-harm, resolving bullying issues and assessing the needs of prisoners with disabilities. The prison officers described how they were frequently redeployed to general wing security duties, resulting in them being unable to appropriately support vulnerable prisoners:

I’ve had probably 10 hours in my job this month in 23 days. Ten hours and you’re just playing catch-up all the time and that’s when you start missing things. I like. I love this job, or like/love. If I could do it properly but I don’t feel I can do my job properly. It’s redeployment, it just, it messes you up.

So how much notice would you get if you’re going to be redeployed?

About a minute.

Prisoners also discussed how a shortage of prison officers had an impact on the care they received and how implementing initiatives such as the OHSCAP were consequently given a low priority:

If it’s [implementing OHSCAP] relying on prison staff, they’re so short of staff, I think things like that would be bottom-drawer stuff.

Prisoner 4

There were also staff concerns across health-care departments within a number of the prison establishments involved in this study:

Well, when I first went one of the sisters was leaving, the other sister was on long-term sick. There was a couple of bank staff. We were working on, sort of, sometimes just on one nurse a shift with one HCA [health-care assistant], sometimes two nurses a shift and one HCA. And we had clinics all day and went to the emergencies, and it’s very poorly staffed.

Health-care worker 2

The staff shortages across health-care departments were considered by health-care staff to be a result of high sickness levels and also difficulties and delays in recruiting new staff:

They’ve shortlisted about three or four [health-care worker candidates] each time, there have been three sessions and not one’s turned up. Another one last week, last Monday, there were supposed to be four coming for band 5 nurses, no one showed up.

Health-care worker 10

Job satisfaction

Strikingly, 6 out of the 11 staff participants we interviewed were in the process of leaving their job role, three of whom were leaving the prison service altogether. Many referred to the current staffing crisis as an explanation for their departure. Staff members stated that there was low morale among the workforce as a result of the recent drastic reductions in staff numbers and the added pressures this placed on them. As one health-care assistant surmised:

I think we’ve got quite low morale haven’t we at the minute, in the health-care staff.

I’m totally disillusioned by it all.

And the prison staff aren’t happy, I think as well, ‘cause they’ve had a lot of changes so they’re feeling very negative about things, and it has a knock-on effect really doesn’t it?

Staff members stipulated that staff shortages were having a negative impact on their health. Prison officers and health-care staff described how staff shortages were causing high levels of stress among their colleagues, resulting in high levels of staff absence:

Staff are stressed, sickness has gone up massively. Plus you’re doing three people’s jobs instead of one person’s job and you can’t, it’s a lot to deal with, it’s a lot to deal with.

Prison officer 3

Prisoners had also noticed the high levels of stress among prison officers, and this led them to be less likely to request support or approach prison officers for advice. As one prisoner described:

On [A] wing I had a very hard bed, good for my back. It was a brand new mattress, downstairs but this one is very old. They don’t come in to ask how I am, they don’t care. If you ask too much, the officers go angry because they had too many people come asking, all complaining.

Prisoner 7

Some prison officers expressed concerns that the current prison environment was a dangerous place to work. Consequently, staff would prioritise remaining safe over meeting prisoners’ health and social care needs:

I think some people [prison officers] kind of come in and just think, as long as I go home at the end of the day safe then it’s alright.

Prison officer 8

Basic needs unmet

Both prisoners and staff described how current staff shortages were resulting in prisoners’ basic needs (such as access to showers, socialisation, etc.) frequently not being met:

Well, because we’ve put lots of security reports in, we’ve heard the prisoners saying that’s it going to go off soon because they’re always locked behind their doors because they’re getting back from work, they’re getting locked up, they’re not even able to have a shower, make a phone call, it’s all those little things that mean a lot to people. And if those things are taken away from them, I mean things like showers it’s a basic human right being able to wash especially if you’ve been in, say for example, the kitchen where it’s hot and sweaty all day and stuff.

Prison officer 3

One prisoner, who suffered from mobility difficulties, endorsed the prisoner officers’ perceptions. In this excerpt, he describes how a lack of available prison officers to take prisoners to a suitable, safe shower facility meant that they were unable to have their basic hygiene needs met:

I’m using wet wipes to shower still. There is a wet room on A wing and they [prison officers] say they’ll give me access but it’s when they’ve got time on their hands.

Prisoner 8

Process focused

Throughout the interviews, both prison and health-care staff focused on describing the processes they adhered to rather than addressing individuals’ needs. Frequently, staff would continue to refer back to the recognised process even if this was failing to meet prisoners’ needs. This appeared to be a survival tactic for dealing with the challenging prison environment. This focus on process is captured in the transcription of an interview with two health-care workers described below:

Yeah, some of the lads [prisoners] obviously didn’t have their medication when they first came in, but that’s because of processes that we’ve got, it can take up to 2 weeks.

They’re quite upset about it. But obviously it was sorted in the end. But I agree with them, why should they have to go without their medication basically. It’s been an ongoing issue for a long time.

Right, and is there anything you can do to speed things up?

It’s just the process.

I mean the process is the process.

One prisoner also described how prison staff focused on processes, stipulating:

They come out with pathways this and pathways that but I don’t think any of them could find a path up their own garden.

Prisoner 4

Applications, frequently referred to as ‘apps’, were described as the written process by which prisoners could access help or support while incarcerated. Prisoners detailed how officers often simply told them to complete an application for every aspect of support they required, without listening to their needs:

Oh yeah, ‘put an app in. See you in a week’. No one’s really interested.

Prisoner 9

Reflecting the quotation above, prisoners stated that they would frequently not get a response from an ‘app’ or would experience long delays:

I’ve gone into the office sometimes and said ‘how do I get . . . ?’ – ‘put in an app’. I put one in and I get no response.

Prisoner 4

Although there were, reportedly, ‘processes’ to manage and address prisoners’ health and social care needs, participants stipulated that the documented processes were often not adhered to in practice, as one health-care worker explained:

On the [staff] induction they actually told us about . . . well it sounded all wonderful, about how they were helped and how they were helped with housing and work, but what most of them say, the older ones [prisoners] that I’ve spoken to, it’s not true, it doesn’t happen. You’re just told where you’re going to go, you’re allocated somewhere far away from where you come from, away from any sort of family network.

Health-care worker 2

Prison and health-care ‘silos’

Staff members and prisoners considered prison issues and health-care issues to be very separate, and prison staff and health-care staff did not generally appear to work collaboratively to support older prisoners:

And one [prisoner] . . . he’d lived in children’s homes all his life and he’d had a girlfriend that had died and he had nothing to support him on the outside, and he really didn’t want to come out of prison. So I really didn’t know what to do about that. I didn’t have enough experience with the prison to know about that, and unless it’s a safety issue we’re not allowed to share that sort of information with the prison side of it.

Health-care worker 4

However, there was some evidence that the introduction of the OHSCAP had begun to support the development of positive working relationships between prison and health-care staff:

And it [attending the OHSCAP training] made it easier as well, to bridge the gap, when I’m going to get it on SystmOne; I can say, [health-care worker’s name], can you put this on SystmOne for me please; because I don’t really know many of the nurses very well, because I don’t work with them. I know their names, but I don’t really know them very well to talk to; so many people that work in there.

Prisoner officer 6

Both staff and prisoners stipulated that there was a lack of accountability for resolving prisoners’ social care needs, resulting in needs remaining unmet. For example, one health-care worker described the difficulties they experienced when trying to meet the prisoners’ needs identified during the OHSCAP, as a result of a lack of accountability:

There is a wheelchair based in the centre but the brake broke so we could not use it. You ask people to look into it and no one gets back to you.

Health-care worker 5

Discontinuity of care

Prisoners described their frustration at not being able to resolve their concerns in relation to both prison and health-care issues, as a result, at least partly, of inconsistent staffing:

You don’t see the same person [GP] more than once. It’s a lucky dip. You don’t know if you’re coming or going.

Prisoner 8

My problem is I have a bad leg, had an operation, pain all down the side, leg and back. I have been on the top bunk for one week. There is another cell empty, single one, so people say why don’t I ask for it but I don’t get the chance. I showed the officer here my letter [from the GP], but nothing happens. What happens here is you show one officer one day, then next day he’s not here, so you start again, explain again to another officer.

Prisoner 7

Perceptions of service provision

There were many examples of perceived poor service provision, including delays in seeing a GP and receiving medication; limited appropriate education opportunities; a failure to organise home leave (where applicable); being located far from family, impeding visiting opportunities; and a lack of input from offender managers and psychologists. Prison induction procedures were also deemed to be suboptimal by both prison staff and prisoners. An illustrative example of a quotation regarding the limited induction procedures is provided below:

It’s like being transported back to the dark ages sometimes. It’s helplessness. If you’re not depressed when you come in, you are when you get out. I feel like I’ve been fighting all the time. [During induction] they should tell you who you can see for help and how to access services. What they tell you [during induction] is very limited – ‘go to the kiosk . . . ask one of the cons [prisoners]’.

Prisoner 8

In addition, all prisoners should reportedly be allocated a personal officer. However, many had never met theirs:

I haven’t had a chance to speak to my case officer since I’ve been here. I saw him once after I searched him out but he’s been on another wing. His job as a personal officer is redundant, made redundant by their absence.

Prisoner 1

There were also some examples of what was perceived to be a good standard of health-care service provision, particularly in relation to prisoners with complex or serious health concerns:

I’ve never had so much health-care as I have in here. It’s been very good. They do look after you.

Prisoner 1

One prisoner felt it was unlikely that the OHSCAP facilitator would identify any health concerns that the health-care department had not previously detected:

The health-care department in here is very good. The health care up the top there is very good. They’re on top of everything. I don’t think [OHSCAP facilitator] would pick up much that they haven’t found before him.

Prisoner 9

Uninterested in social care needs

Prison officers inadvertently described how they did not believe that social care needs to be particularly important, as illustrated in the quotation below. Consequently, they were less interested in supporting prisoners with social care issues than they were with concerns that had more traditionally been considered to be the responsibility of prison officers:

Well, I just consider working on the wings like the actual bread and butter of the job, that’s why you join the job, to be on the wings and dealing with issues. But it’s not issues like ‘can I have a new flask because I can’t open mine properly’, or . . . you know, because we do get that; or, ‘can I have a sock aid to help me get my socks on because I can’t bend over properly’; it’s proper issues, like ‘I need to ring my mum because she’s not well’ and ‘I’ve got no money on my PIN [a person’s individual payphone accounts], can you sort this for me’; ‘I can’t get this person cleared for a visit, can you help me’; or just the general day-to-day of supervising the prisoners, making sure they’re not up to no good really.

Prison officer 6

Some prisoners described how they felt that prison officers lacked insight into the social care issues faced by older prisoners. For example, one prisoner described his experience of mandatory drug testing:

I have to stand up for an hour or so because the seat isn’t anywhere near the wall to lean back on. I’ve told them before that I need a seat to sit on but they won’t give me one.

Prisoner 5

One prisoner stipulated that officers failed to consider the varying nature of his illness but highlighted that, when officers were responsible for caring for fewer prisoners, they were able to better meet his needs:

I have ups and downs. They don’t understand that I can have good days and bad. I finally got to the library once and it’s caused me grief ever since. I’m being penalised for going to the library ‘cause they now expect me to be able to walk everywhere. On [other wing] they had more understanding, a smaller wing – they know you better.

Prisoner 8

Support for ‘genuine’ issues

There were many occasions when staff were reportedly supportive of prisoners’ needs. This appeared to happen only if staff perceived the need to be genuine and usually of a serious nature. For example, when prisoners were considered to be at risk of suicide or self-harm, there were examples of staff developing positive, supportive relationships with prisoners:

The officers are good. They’ve been coming to see if I’m ok because I’m on an ACCT, this is the system used to provide monitoring and support to those considered to be at risk of self-harm.

Prisoner 5

A further example was provided by one prisoner, who had reportedly been given Spice (a synthetic cannabinoid substance) unwittingly. He described how he had attempted suicide after this event but, after receiving support from officers, was no longer experiencing suicidal thoughts:

‘Cause some people were giving me Spice and I didn’t know what it was. People just gave me, like, a rollie [rolled-up cigarette] and I smoked them and I went under. I nearly died. And then tried to kill myself . . . Well . . . you’ve got to get on with them [prison officers], haven’t you. It’s where you live ‘til you get out. ‘Til you get out those gates you say to yourself, they’ve looked after me, do you know what I mean? Now if those officers weren’t here, you know, they’d turned their backs, I wouldn’t be sitting here now . . . Like, that SO’s [senior officer’s] done a lot for me.

Prisoner 14

‘Walking on egg shells’

Within open establishments, prisoners can walk around relatively freely within the confines of the prison, without being reliant on prison officers unlocking doors. Prisoners residing in open prisons may also be eligible for home leave. Both prisoners and prison officers described how prisoners housed in open prisons were ‘walking on egg shells’ as a result of not wanting to lose their right to home leave or be returned to closed conditions. This inability to raise concerns freely within open establishments was the most striking difference between sites in relation to how the OHSCAP works in practice. Within closed conditions, there are far more security restrictions and prisoners are locked in their cells or on a prison wing without freedom of movement for longer periods of time. These concerns prevented prisoners from raising any issues with staff:

And you’re then in a situation where you can’t create a fuss over it . . . Because anything that they give you can always be taken off you . . . I’ve done it [raise concerns] in the past . . . Doesn’t get you nowhere. Just another rod for your own back. That’s all’s it is.

Prisoner 9

One prison officer stated that this reluctance to raise concerns in open establishments was particularly pertinent to discussions around safety:

And the stupidest question of the lot [within the OHSCAP], is, do they feel safe? Because they’re always going to say, yes, I feel safe . . . Because if they don’t feel safe, they wouldn’t be in open conditions, they’d have to go back to closed conditions. Where they get protection. We don’t offer any protection.

Prison officer 1

Facilitators

A recurring theme was that it was perceived to be inappropriate that prison officers were OHSCAP facilitators. Over half of the participants expressed concerns about prison officers leading the process. Concerns were expressed about knowledge, trust and confidentiality:

At the end of the day, they’re still in uniform aren’t they? And she [prison officer] could go and say to another officer, ‘oh, he doesn’t like this place’, and all that, like, they do.

Prisoner 2

However, some prisoners did consider it acceptable for officers to deliver the OHSCAP:

It doesn’t matter who it is as long as I can speak to someone when I need to. I’ve nothing to hide about my health have I?

Prisoner 5

No real pattern could be identified regarding the characteristics of prisoners who considered it acceptable for prison officers to deliver the OHSCAP. Their perspective was influenced by a complex set of factors, including their previous experience of prison officers, the personalities of the prison officers in question and the nature of the prisoners’ health and social care needs. Some prisoner participants stipulated that they were able to build trust over time with prison officers:

I get on OK with the officers, I know them well.

Prisoner 5

Other prisoners felt that their experience of prison had taught them not to divulge information to prison officers:

In the days of old, I would have been very disquieted but after a while, you learn that some things are best kept under your hat. It’s not wise to blab everything the first time you see someone. I wouldn’t want to talk about personal things . . . not to an officer because I’ve seen officers elsewhere use knowledge and position to dehumanise or belittle prisoners in front of others.

Prisoner 4

Prison officers did not generally perceive there to be any concerns with them delivering the OHSCAP:

I’ve had no negative reactions to me being an officer.

Prison officer 1

Some prison officers lacked skills in conducting assessments. One prison officer described how she shadowed a colleague to support them conducting the OHSCAP (both members of staff had previously attended the training):

The actual assessment is really self-explanatory, the fact that you’ve done all the example questions; I mean I fly through them now. When I went with [prison officer colleague] and she said, ‘can I shadow you doing one’ . . . ‘No’, I was shadowing her, she said, ‘well, you can watch me because I’m nervous’, I don’t know how; I says, ‘it is easy’, you just ask all the questions that are on there; ‘oh, please, come with me’. And I could see that she was struggling, so every now and again I were jumping in; and afterwards she went, ‘oh, it’s just so easy for you, isn’t it, you could just do it straightaway’; I went, it’s not that, I says, I’ve done more than you, that’s all . . . But all I do is ask all the questions that are on there.

Prison officer 6

‘Above and beyond’

There were occasions when some OHSCAP facilitators clearly went ‘above and beyond’ what was normally expected of them in their role. For example, some officers created ‘distraction packs’ (paper packs including crosswords, word searches, etc.) to prevent boredom and potentially depressive symptoms. There were examples of health-care workers who were OHSCAP facilitators making GP appointments for prisoners on release and organising repeat prescriptions for older prisoners who were confused by the system. There was one example of a health-care worker telephoning the health-care department of the establishment where prisoners were previously held, in order to locate their glasses and have them sent to them to prevent them having to put in a new application with months of delays. One health-care worker described how she introduced an older prisoner to other prisoners in an attempt to make him less isolated:

I would introduce the chap so he could make friends. Especially on A wing, they all sit around in a group, you know, and I took him over and introduced him to people . . . I think it’s just . . . putting that foot forward, isn’t it, sometimes?

Health-care worker 3

These examples of going ‘above and beyond’ were reportedly relatively rare and the majority of identified needs would simply be fed back into the prison systems through the applications procedure, as previously discussed:

I didn’t find it [the OHSCAP] helpful because they didn’t go over anything more than you [the researcher]. It felt like a double-up. No useful information apart from saying I’d have to put ‘apps’ in. They [prison officers delivering the OHSCAP] didn’t say they could help with anything.

Prisoner 4

Frequently, prisoners reported being told that there was nothing that could be done to help them and nowhere to refer them to. This was particularly evident with financial and housing concerns:

I ended up passing him on to his solicitor, because there was nobody in the prison that can help him with that [housing]. Yeah, it’s a separate issue to the fact, completely separate to prison service, so there were nobody that could help him; I said, you’re going to have to get a solicitor to help you with it.

Prison officer 6

Prisoner involvement

There was an apparent lack of genuine prisoner involvement in developing the care plans. One prison officer described his attempt at involving prisoners in the development of the care plans. However, he appeared to have a lack of understanding of what genuine prisoner involvement would involve, and discussed how he simply informed prisoners what would happen instead of meaningfully collaborating with them to involve them in decisions:

I always tell him [prisoner] what we’re going to do, if he’s happy with what we’re going to do, and would he mind if we went ahead and did it. And that’s it [how they involved the prisoner in the development of care plans].

Prison officer 1

However, prisoners greatly appreciated the rare opportunity to talk to someone about their needs, even when little was done to try and address these needs. As one prisoner explained:

Blimey, after 15 years, someone is finally taking notice that I’m here.

Prisoner 4

The OHSCAP facilitators also recognised that prisoners appreciated the opportunity to be listened to. Prison officers described how the OHSCAP process became an isolation prevention tool:

So generally even if people didn’t have issues they’d still say it was nice to chat to somebody. I don’t know if it’s made that much difference to their needs, to what I could help them with, but in general as an isolation tool, I suppose, I think that’s been the positive things and being able to chat to them because it’s nice to chat to someone and then I’ve linked to others who maybe feel a bit isolated, can you ask one of the cleaners or a carer just regularly check up on him and that kind of stuff.

Prison officer 1

External services used in secure facility

This comprised information on both contacts the individual had with external professionals and other community-based/visiting services received while inside the secure facility. The number of personal contacts with each service and, when relevant, the number of telephone calls and letters received were recorded. The services recorded were the following: Samaritans or Listeners, Citizens Advice Bureau, solicitor, barrister, legal advocate, social worker, care worker, organised prison visitors (National Association of Prison Visitors), older prisoner organisation, police officer and other.

Daily activities inside secure facility

This refers to group and other activities inside the secure facility. The name of the group or activity, the number of contacts over the study period and the average duration of each activity were recorded. The daily activities recorded were the following: therapeutic group, creative activity, work, education course, sports activity, older prisoner group and other.

Assessment, Care in Custody and Teamwork

Whether or not an individual was on an ACCT, and, if so, the start and end dates of this, were recorded.

Professional contacts inside secure facility

Information on one-to-one contacts with professionals inside the secure facility was recorded. The number of contacts with each professional and the average duration of these contacts was recorded. The professional contacts recorded were the following: GP/prison doctor, older prisoner lead nurse, practice/prison nurse, psychiatric nurse, psychiatrist, psychologist, counsellor/therapist, drug and alcohol treatment staff/Counselling, Assessment, Referral, Advice and Throughcare (CARAT) staff, dentist, optician, chiropodist, physiotherapist, chaplain and other.

Hospital contacts inside secure facility

Contacts with hospital services and visiting hospital specialists running outpatient clinics inside the secure facility were recorded. The number of contacts and the duration were recorded. The hospital contacts recorded were the following: inpatient stays and contacts with visiting specialists.

Medication

Information was collected covering all medication prescribed. The name of the medication, the daily dose and the number of days that the medication was prescribed during the study period were recorded.

Hospital contacts outside secure facility

Contact with hospital services outside the secure facility were recorded, for example inpatient stays, outpatient attendances, accident and emergency (A&E) attendances. The number of contacts and the length of any inpatient stays were collected.

The OHSCAP intervention

The resources required to deliver the OHSCAP comprised training sessions for staff on how to deliver the intervention, and an average of 40 minutes’ delivery per individual receiving the OHSCAP. The three main training sessions were delivered by a combination of project research staff and one prison officer, lasting 3 hours each. These sessions also involved some explanation of the research process, and, thus, if the OHSCAP were to be implemented, these would, in practice, be shorter. As the proportion of the session spent on training versus explanation of the research process was not recorded, the cost of training was based on the full 3-hour session. As additional sites were recruited to the trial, four informal training sessions were run by research staff to train additional staff members at these new sites.

The resources required for training and delivery of the OHSCAP are presented in Table 19. It was estimated that each OHSCAP delivered cost an average of £25.52 per prisoner, consisting of £16.42 for the costs of delivery and the remaining £9.10 for training costs.

Additional resource utilisation

In addition to the direct cost of providing the OHSCAP intervention, the economic analysis considered the wider resource utilisation of both the TAU and intervention arms. This information was available only for the complete-case sample. Of the 497 individuals in the trial at baseline, 404 were retained at follow-up and thus provided resource utilisation data. Table 20 provides the average resource utilisation for the complete-case sample over the trial period. Resource utilisation is presented separately by trial arm and summarised as the mean number of contacts with each service per individual.

The number of missing data for each resource-use category was zero. This reflects the method of data collection, with resource use data being collected from a combination of health-care record examination and participant recall. Resource use was recorded as zero if there was no mention of the service being accessed in any prisoner records or if the participant did not report having accessed the service for categories obtained via participant recall. It is not possible to tell whether all of these values are true zeros or, in reality, they reflect missing information in the individuals’ records. It is possible that a prisoner accessed a service and this was not recorded, but there is no way to verify this information. There is, however, no reason to suggest that the level of missingness should vary by trial arm, and any missing data should, therefore, not bias the comparative results.

The data presented in Table 20 indicate that, although the prisoners in our sample are accessing a wide range of services and there is some variation in the use of each service category between the two trial arms, there is no systematic pattern of greater resource utilisation in one treatment group than in the other.

Of the external services utilised inside the secure facility, solicitors were the most commonly accessed. No prisoners in either treatment group accessed a care worker. The daily activity most frequently reported was work and the least frequently reported activity was participation in older prisoner group meetings. During the study period 7% of the TAU group and 10% of the OHSCAP group were placed on an ACCT. Practice and prison nurses were the most frequently accessed professionals inside the secure facility, followed by GPs. The professionals inside the secure facility who were least contacted by individuals in our sample were psychologists and counsellors/therapists. Three per cent of the control group were admitted to the health-care wing for an inpatient stay during the study, and 2% saw a visiting specialist within the secure facility. In the OHSCAP group, 1% of the sample were admitted to the health-care wing and 2% saw a visiting specialist. Although, wherever possible, health care is provided within the prisons, some individuals in our sample did leave the secure facility to receive health care. Among the TAU group, 3% were admitted to hospital as inpatients, 19% left the secure facility to attend outpatient appointments and 6% attended A&E departments. Within the OHSCAP arm, 1% were admitted to hospital outside the secure facility as inpatients, 16% attended outpatient appointments and 2% were taken to A&E.

To assign a monetary value to the resource utilisation consequences in both arms of the trial, unit costs were applied to each individual’s resource utilisation. Table 21 presents the unit cost figures used in the analysis.

Table 22 multiplies the resource utilisation of the complete-case sample by the unit costs presented in Table 21 to obtain the average costs associated with the resource utilisation of individuals by trial arm. Only costs falling under the public sector perspective are included, with the services supplied by charitable organisations excluded from these calculations. The mean costs per individual are presented separately for each type of service used, in addition to the mean total costs for each resource use category, and the mean total costs of all services used.

The most costly category of resource use was daily activities inside the secure facility, with these costs driven by the costs of education. These education costs were taken from the PSSRU UCCJ,107 based on information provided by The Skills Funding Agency, and include Offenders’ Learning and Skills Service overheads.

For individuals receiving TAU, mean costs were £54.22 for external services used inside the secure facility, £2281.85 for daily activities inside the secure facility, £81.40 for time spent on ACCT, £284.55 for professional contacts inside the secure facility, £67.75 for hospital contacts inside the secure facility, £282.49 for hospital contacts outside the secure facility and £93.68 for medications. The mean total cost for individuals in the TAU group was £3145.94 over the study period.

For individuals receiving the OHSCAP, in addition to the mean cost of £25.52 associated with the intervention, the mean costs were £67.04 for external services used inside the secure facility, £2145.28 for daily activities inside the secure facility, £136.39 for time spent on ACCT, £289.62 for professional contacts inside the secure facility, £19.21 for hospital contacts inside the facility, £160.37 for hospital contacts outside the secure facility and £102.23 for medications. The mean total cost for individuals receiving the OHSCAP was £2945.65 over the study period.

Randomised controlled trial

Limitations in relation to the trial include the possible impact of increasing the number of sites involved in the study and limitations of the outcome measures used. On balance, the impact of adding further sites on the trial processes was insignificant but the qualitative study revealed some potential issues with outcome measures.

The CANFOR was the tool used to measure the primary outcome. We found that, on average, prisoners had 2.05 unmet health and social care needs, as measured by the CANFOR. This accords with the findings of a previous study, using similar methodology, that reported an average of 2 unmet needs per older prisoner. 8 Senior et al. 28 reported a slightly higher figure of 2.74 unmet needs. However, participants in the Senior et al. study were newly received from court and it is therefore, perhaps, unsurprising that they had needs that the prison system had not yet had time to address, in contrast to those interviewed in later stages of their sentence (as in the current study and Hayes et al. 7

Some participants indicated that many of the discrete domain items were not applicable to their current situation in prison, or at all, given their age (e.g. childcare responsibilities). In addition, the CANFOR considered needs to be either met or unmet, but it is unlikely that some health and social care needs are ever fully met as they are ongoing and changeable in nature/severity. Findings from our previous research suggest that older prisoners are less likely than their younger counterparts to raise concerns27 and, consequently, participants in this study may not have always disclosed if they were experiencing unmet needs. In spite of the limitations of this tool, it was considered by the authors to be the best available tool for measuring health and social care needs within the prison setting and has been successfully used with this population in previous studies. 21,27

The GDS-15 was not designed for use in prison. However, the scale has been used with older prisoners in a previous study. 41 In that study, one question was adapted from ‘Do you prefer to stay at home rather than go out and do new things?’ to ‘Do you go on association?’ to better reflect prison life. We adopted the same approach. The current study found that 31% of participants had GDS-15 scores indicating symptoms of clinical depression. Previous studies have reported that the percentage of older prisoners experiencing depressive symptoms ranges from 12% to 56%. 7,27,32,35,41 However, the Murdoch et al. 41 study, reporting a rate of 51%, focused on life-sentenced prisoners, and the Senior et al. 27 study’s reported rate of 54% included only prisoners newly received from court. It is therefore, perhaps, to be expected that both of these figures are higher than those found in the current study because this study included participants with a wider range of prison sentence lengths and participants were at various stages of their sentence.

Very few ADL needs were identified using the BADLS. This tool has been used in previous older prisoner studies;7 however, it is designed for use with dementia patients and is perhaps not sensitive enough to identify ADL needs among either older adults not experiencing dementia or those living in a limiting institutional setting.

It was decided that a 3-month follow-up period should allow sufficient reasonable time for initial needs to be met. The CANFOR measures whether or not prisoners are receiving some assistance that is helping them. The research team perceived 3 months to be sufficient time for the prisoners to begin to get suitable help. We also wanted to minimise attrition.

It is possible that the OHSCAP may have initially helped to speed up the meeting of health and social care needs, but that, by 3 months, this had evened out between the two groups. Previous research has illustrated that early identification of mental health needs is important as issues are unlikely to be picked up later in custody. In order to address this early potential to meet needs and to establish whether or not more needs are eventually met with the OSCHAP, multiple follow-ups at 1, 3 and 6 months would be useful.

Audit

Unfortunately, not all of the completed OHSCAPs were available for audit. All of one site’s assessments were lost entirely, reportedly because documents were destroyed as part of the process of changing health-care provider. In another site, only 2% of documents were recoverable for audit. Unfortunately, the latter was the site where the greatest number of OHSCAP assessments had been completed by a health-care member of staff. The availability of these records would have enabled us to draw clear inferences about the impact of the facilitator’s core role on the OHSCAP process.

The original protocol specified that 10% of the OHSCAPs should be reviewed. The research team decided to review 100% of the audits. This decision was made because the number of sites involved in the RCT increased from 4 to 10 and we wanted to ensure that we had a clear understanding of how well OHSCAPs were completed at each site. The auditor was trained by the research team, independent of the trial. Some sections of the bespoke audit tool required a subjective judgement to be made; therefore, the judgement of one individual may not have been entirely reliable. However, 10% of the OHSCAPs were also reviewed by one of the researchers employed on the RCT, with good agreement between raters.

Regardless of which or how many individuals were involved in the audit process, the audits were based on the actual documented comments within the OHSCAP plans. Without being present for the assessment interview, it is not possible to know exactly what happened in practice and whether or not this is fully reflected in the documentation. For example, where ‘no issues’ was noted in relation to an entire area of need, the auditor could not possibly know whether or not a full and meaningful discussion had taken place in order to establish this. The same applies in relation to whether or not prisoners were actively involved in the care planning process. The auditor was able to establish only whether or not there was documented evidence of involvement, but this may not accurately reflect what happened in practice. Examination of the OHSCAPs and rating them in this way was a useful process that, with adaptation, is fit for purpose for other studies of this type to assess one aspect of fidelity.

Qualitative interviews

As a result of security restrictions, we were unable to audio record 6 out of the 14 prisoner participants’ qualitative interviews. In these circumstances, detailed notes were produced by an additional researcher in order to allow the first researcher to focus on the interview itself without distraction. These notes were typed up and analysed alongside the transcriptions. We were able to obtain very detailed notes, but obtaining audio recordings and transcribing them verbatim would likely have added greater richness to the data generated for these participants.

The research team decided to conduct interviews with prisoner participants on several occasions in order to capture rich data on progress with health and social care needs over time. One of the findings was that there was little further input from OHSCAP facilitators after initial assessment. Furthermore, as very few OHSCAP reviews took place, the additional interviews did not yield very rich data. Little had changed for the majority of participants between the initial and later interviews. Thus, through conducting a series of interviews with each individual, we were able to demonstrate that health and social care needs often remained unmet while new problems emerged.

The researcher who generated and analysed the majority of the qualitative data has undertaken research concerning the health and social care needs of older prisoners for the last 7 years. Her previous experience includes the study within which the OHSCAP was created and piloted, although she did not have direct involvement with the action learning group processes that created the OHSCAP itself. Her previous experience will, however, have had an impact on the way in which the interview content guides were designed, the issues she chose to focus on within interviews and the way in which she analysed the data. We aimed to reduce the potential impact that the researcher’s previous experience may have had on the findings by ensuring that a second researcher analysed a selection of transcriptions and assisted with the development of themes.

It is also important to consider the potential impact of the data being collected by three female researchers within a male-dominated environment, particularly as a large proportion of the prisoner participants were convicted of sexual offences. This could have led prisoner participants to be more or less open or truthful when discussing issues with female researchers, depending on broader influences on their personalities and their own views on their offending behaviour. It is possible that the research participants may have responded differently to a male researcher. Despite a shared view among researchers that prisoner participants were not always truthful about the type of offence they had committed, it was felt that prisoners did respond openly and honestly about their experience of health and social care within the prison.

Conducting randomised controlled trials in prison

Research in prison and, perhaps most acutely, the conduct of RCTs in prison, is challenging. Sometimes changes to the landscape may occur as a result of positive policy initiatives. For example, a RCT conducted by Abel et al. 119 among women who repeatedly self-harm was affected by the publication of the Corston Report on women’s imprisonment,118 which made numerous recommendations for a holistic person-centred approach to supporting female prisoners. Consequently, general improvements were made to support women in prison at risk of suicide and self-harm across England and Wales, and the impact of the Women Offenders’ Repeated Self-Harm Intervention Pilot could not easily be established. 119 However, other RCTs conducted by our research group have been affected in more negative ways. For example, the current study was affected by the loss of the dedicated time given to prison officers and others as older prisoner leads and/or DLOs, meaning that the individual role identified in the pilot study as core to the OHSCAP process simply no longer exists in many establishments. Where it does still exist, individuals within the role are often redeployed to other duties. Other trials into through-the-gate support to promote engagement with mental health services for people with serious and/or common mental health problems have been similarly negatively affected by reduced staff time and availability (Professor Pamela Taylor, Cardiff University, 2016, personal communication).

The quite dramatic shift in the prison landscape throughout this study illustrates the challenges of conducting a RCT in a real-life setting. It may have been possible to fund a designated role to deliver the OHSCAP for the duration of the trial, which may have reduced some of the problems with staffing shortages, but wider staff shortages would still have had an impact on the ability of the OHSCAP facilitator to make referrals to other services, etc. Furthermore, it is likely that such funding would not have been made available after the study ended, so the effectiveness of the OHSCAP would not have been evaluated within a real-life setting.

The effectiveness of the OHSCAP: a randomised controlled trial – semistructured qualitative initial patient interview guide

Semistructured qualitative follow-up patient interview guide