Summary of main findings
In line with the growing and ageing prison population, our epidemiological analyses found an increasing number of cancers diagnosed in English prisons between 1998 and 2017. The age-standardised incidence rates showed that cancer incidence among people in prison increased from lower levels than among the general population towards similar rates by the end of the study period. Given the higher prevalence of risk factors in the prison population, notably a higher smoking prevalence than in the general population, this increase probably reflects improved awareness and diagnosis.
Although we found a slightly higher stage distribution among the cancers diagnosed in prison, the difference from the general population was not statistically significant. However, there was a clear difference in the route to diagnosis, with cancers diagnosed in prison being less likely to result from an urgent ‘2-week wait’ referral or through screening than those in the general population. We also found significantly lower odds of curative treatment for patients diagnosed in prison than for patients diagnosed with comparable cancers in the general population. Survival from cancers diagnosed in prison was also lower than in the general population. Although some confounding seems to arise from other factors, such as route to diagnosis, half of the survival discrepancy can be explained by fewer among the prison population receiving treatment with curative intent.
Our health economic analyses show that the secondary healthcare cost of cancer care for patients diagnosed in prison is £1284 less per patient on average than that for their peers diagnosed in the general population. This is predominately due to fewer attendances for planned inpatient care and outpatient appointments. The cost of emergency care for patients diagnosed in prison was on average £1018 higher per patient than for the matched general population cohort. This finding fits with the epidemiological findings of poorer access to care and poorer outcomes. This poorer care also comes at the additional cost of escorts and bed-watches for people in prison, a cost that is covered by the NHS.
Our qualitative interviews with patients and health and prison professionals showed that cancer care in prison is complex, not least because people in prison move between a hospital and a prison environment. People in prison follow a similar pathway to diagnosis to those in the community, but with several specific barriers to diagnosis including health literacy, the process of ‘putting in an app’ before seeing a clinician and the logistics and routines of both prisons and hospitals. We found that there were tensions between control and choice in prison health care that impact patients’ experience of cancer care in terms of symptom management, accessing information about their cancer, and family involvement in their care. Communication between professionals within and out of the prison was also disjointed and significantly impacted the care experiences of people with cancer in prison. These interview findings were also reflected in the findings from analyses of previously collected NCPES results. Here patients returning surveys reported worse experiences of cancer care than matched controls from the general population. In interviews we found that oncology professionals took pride in providing ‘the same’ care to people in prison as they did to those in the community. This may, however, unintentionally disadvantage people in prison, as in prison patients have limited access to information and support. The Royal College of General Practitioners warned in 2018 that ‘equivalence of care’ in secure environments should not mean ‘the same’, so more training needs to be provided to healthcare professionals to ensure equity in access to health care.24
Based on all the findings above, four initial policy priorities were identified by stakeholders in the Policy Lab that might be reasonably implemented within the next 2–3 years. Participants’ collective choices were based on the likely feasibility and impact of interventions using their knowledge and experience of working in prison and/or cancer policy. The first was to increase communication to clinical teams in hospitals about how the prison system works. This was seen as being undertaken as part of efforts to join up different parts of ‘the system’ to achieve a more integrated approach for patients. Co-ordinating and promoting an effective approach to cancer screening that significantly increases take-up was the second priority. The third concerned supporting people in prison by developing peer health champions from the prison population who could advise and support others, and the fourth was to devise ways of raising health literacy and increasing awareness of symptoms using different media, especially TV and video, in prison.
Robustness of the results and their limitations
Because in our epidemiological analyses we matched individuals diagnosed in prison on age, sex, diagnosis year, tumour site and disease stage with those in general population settings, we could not investigate the actual influence of these factors, most notably stage of disease, on treatment and survival. There may therefore have been residual confounding that would explain the survival deficit we found. The prison population is also more likely to include individuals from lower socioeconomic backgrounds – a factor that is known to be associated with poorer survival. We were not able to assess the impact of this or of other factors associated with higher levels of deprivation as in England we would normally do so using an individual’s postcode of residence to calculate the Index of Multiple Deprivation. In this study the prison postcode was used to identify the place of diagnosis, which is not reflective of the person’s usual socioeconomic status.
One of the main treatment modalities for curative intent is surgery. We could not identify potentially curative surgical procedures for all cancers in this study. This means that for the 19% of cancers for which relevant OPCS-4 codes have not been identified, potentially curative treatment is missing from the analysis. However, as we matched on cancer site, this affects both the prison and the comparison general population and may attenuate overall findings. We were not able to consider palliative treatments.
While we were able to identify cancers diagnosed in prison, the total burden of cancers in prison is likely to be higher. To estimate the prison cancer prevalence, work will need to be undertaken to match clinical data from SystmOne and MOJ to the cancer registry to determine those diagnosed with cancer before arriving in prison and who therefore may be living with active or recurrent cancer that could require treatment.
Our health economic study is limited to including only secondary care contacts and does not include primary care or prison healthcare costs. Further work is required to understand the quantity and cost of prison health care for patients diagnosed with cancer in prison. The equivalent would also need to be undertaken for a matched cohort in the community. We were also unable to include accident and emergency data and hence additional costs of these attendances could have been missed, particularly for patients in the community. The cost of appointments that patients did not attend was also excluded from the analysis. This is due to the issues associated with estimating such costs, particularly the cost implications for prison compared with the community.
The use of routine data is a strength of this study as research in prisons is notoriously difficult to conduct, and loss to follow-up is common.46 Using routine data reduces the bias inherent in prison studies, but people in prison can still be commonly missing from routine healthcare data.47 Routine data are also known to contain some errors, particularly when calculating costs.48 These errors should have been equal across both groups, so although the total costs might not be correct, the difference in costs should be close to the true value.
While we sought to calculate the costs for people diagnosed in prison, some of the patients diagnosed in prison may have been released into the community during the 6 months after diagnosis and their contact with care may have changed. People who are released on temporary licence may also not have escort and bed-watches costs, although they make up only 5% of the prison population. As a result, the cost of escorts and bed-watches might have been overestimated and should be interpreted with caution. When looking at costs it is also important to look at outcomes and potentially conduct a full economic evaluation to fully quantify the inefficiency of providing cancer care in prison. This was not possible given that, other than on mortality, we had limited outcome data. The number of different cancer sites and potential outcomes would have made any decision modelling or estimation of outcomes prohibitive.
Owing to the COVID-19 pandemic we had to stop our qualitative interviews early; nevertheless, the 55 interviews conducted provided detailed information from a broad range of participants on the main themes. One-off interviews offer a snapshot of people’s lives, and follow-up interviews and a longitudinal approach could have offered even greater insight into the lived experiences of people with cancer in prison and those who care for them. However, as this is the first study to explore this issue, our interviews already offer ample evidence of how cancer care is experienced and potentially could be improved. This is the first time that NCPES data have been analysed for patients in prison, and we do not know how representative the patients responding are of all those diagnosed in prison. We also know that patients from lower socioeconomic groups are less likely to respond to surveys and that the lower literacy levels among the prison population would likely decrease the response rates further. We were not able to match patients’ CPES response on tumour site.
Finally, it is possible that a different group of practitioner stakeholders attending a second Policy Lab might have viewed the study findings in different ways and chose different policy priorities as a result.
Review of evidence
This study is the first to use comprehensive NDRS data to reliably establish cancers diagnosed in English prisons, based on postcode of residence for the two decades 1998–2017. As far as we can determine it is also the first study worldwide to consider an entire national population of patients diagnosed with cancer in prison, quantify the cost of cancer care for a national population of people in prison, particularly compared with those in the community, offer an in-depth multiperspective account, including the prisoner-patient voice, and report stakeholders’ priorities based on findings that demonstrate the complexity of cancer care in custodial settings.
We identified several barriers to and enablers of getting a diagnosis and treatment. Not all the barriers we found are specific to patients in prison, with some (i.e. late diagnosis) experienced by people residing in the general population, but these are exacerbated by the prison environment. Prisons are designed to take away elements of control and choice,18 yet hospitals aim to be inclusive and empowering. The overlap and tensions between the social roles of ‘prisoner’ and ‘patient’ have been documented17 and were also present in the study as the professional groups used a range of terms, including ‘prisoner’, ‘patient’, ‘prisoner/patient’ and ‘the men and women in our care’.
While surgeons, radiotherapists and oncologists might treat people in prison ‘the same’, this sameness can unintentionally lead to further disadvantage. We also found that each prison adopts its own way of caring for its prison population. National guidance and the sharing of best practice could therefore improve cancer care. Better methods for communicating within and between institutions could also improve cancer care and specifically the patient experience. Our results support findings from other studies on health care in prison showing that there is awareness of the disadvantages and structural barriers to care for those residing in prison.17,18,23 With a growing and ageing prison population these barriers need to be addressed in both policy and practice to ensure good patient care.
Recommendations for future research
At a health system level there is a need to map and understand how NHS governance structures in England, including ICS and Cancer Alliances, currently operate for cancer patients in prison. In addition, at this level an assessment is needed of how previously unanalysed data on people in prison, currently held by NHS and used in this study, could be used to routinely report cancer outcomes for patients in prison and explore variations in referral and outcomes. Because of the complexity of obtaining prison healthcare medical records, patient data for people in prison have not been used to monitor quality of care. Further work with MOJ, prison health providers and commissioners is required to explore how these data could be accessed with the aim of conducting data linkage studies to give a complete picture of the burden of cancer in this population. As there may be time delays in reporting these data, more real-time measures that might be used to track the impact of any planned improvements on outcomes should also be explored. The cost implications to the NHS of escorts and bed-watches could also be better quantified using linked data.
To design new interventions more could be learnt from health and cancer care systems in Scotland, Wales and Northern Ireland, which were not included in this study. Some models of care already applied in other parts of health and social care, and the voluntary sector around health inclusion and decreasing health inequalities, may be relevant. For example, there may be lessons learnt for cancer care for people who are sex workers, homeless or refugees that could be applied to prison care. As far as the development of specific interventions in cancer care are concerned, more work is needed to understand barriers to achieving equivalence in treatment and care. These may be related to the organisation of clinical care, restrictions in the prison system due to security measures, the availability of personnel, communication between health professionals within the prison system and hospitals, communication between health professionals and prison staff, or patient choice affected by availability of information. It will also be important to determine as far as possible the reasons for the lower rates of curative treatment and survival and to determine where these are amenable to change. Technological solutions including remote consultations and access to online information are potentially attractive for improving co-ordination of care between the NHS and prisons, but they do potentially risk creating or embedding further inequities relating to technological literacy, and a multiplicity of approaches may be needed. Finally, there is a need to know what happens to cancer care for patients once they leave prison and the kind of ongoing support that these patients may need. This would include data from primary care and over a longer time horizon to obtain a better understanding of long-term community care.
Implications for decision-makers
This study provides new knowledge about the cancers that most commonly affect people in English prison and the areas on which prevention and national cancer screening programmes should focus. Prisons have been smoke-free since 2006, but the effect of previous exposure will endure for many years, and lung and bladder cancer will remain a significant issue for this population. National screening programmes currently exist for three cancers – breast, cervical and bowel cancer – and need to be both effective and monitored. The use of cancer data already within the system needs to be extended for reporting and used much more actively to monitor incidence, prevalence, and patients’ experiences and their outcomes routinely.
Although the secondary care component of cancer care costs less for patients diagnosed in prison than for their peers diagnosed in the community, this comes at a significant cost in terms of outcomes, such as increased emergency attendances and reduced survival. It also comes at the additional cost associated with escorts and bed-watches. Further work is needed to improve access to planned care following a cancer diagnosis in prison. It is important that any future evaluations of interventions to improve access to cancer care in prisons take account of the impact of costs as well as attempting to quantify the health benefits of the intervention. Equity considerations also need to be explicitly considered as investments should be made such that outcomes for people in prison begin to better reflect those for people in the community, which may require significant additional investment.
Several barriers to and enablers of getting a diagnosis and treatment exist within the prison setting and between prisons and NHS services. Limited guidance is available on care provision for this group and so national guidance and the sharing of best practice could help drive improvements in cancer care. As communication processes are disjointed, better methods for more seamless communication within and between institutions could improve cancer care, the patient-prisoner experience and understanding among clinical teams in hospital of how the prison system works. Co-ordinating and promoting an effective approach to cancer screening that significantly increases take-up and effective follow-up and treatment would assist with earlier diagnosis of cancer in the prison population, particularly in relation to cervical screening in women’s prisons. Psychological support of people with cancer in prison could be improved by developing peer health champions from the prison population who could advise and support others. Finally, raising health literacy levels and increasing awareness of cancer symptoms using different media, especially TV and video, in prison, could help detect cancer at an earlier stage.
Equality, diversity and inclusion
The composition of the research team remained almost entirely female throughout the study, with one male research manager, five senior female researchers at reader, professorial or director levels, and three female researchers at lecturer or earlier career stages. One male EbE was also involved in interview data collection and analysis throughout the study. He helped to design the interviews and our approach to patients in prison. He also led the initial interviews and was involved in the qualitative analysis and the Policy Lab event. The male–female balance for both the Project Advisory and Independent Steering Groups was 50 : 50, and 25 : 75 for the Policy Lab event which was facilitated by one male individual. Diversity by ethnicity, however, was limited across each group involved in the study.
Patient and public involvement
Patient and public involvement was considered at every stage of the design, delivery and presentation of this research project. Members of the NCRI Consumer Forum who had experience of being diagnosed with cancer and of working in prison were included in designing the initial study proposal. After the project was shortlisted for funding, this element was expanded to include RDA, which has specific experience of working with EbE (people who have experience of being in prison). RDA advised on involving EbE in the co-design of the interviews with all parties, and their delivery and analysis. Once the proposal was successful in being funded, RDA engaged and supported three EbE who had experience of cancer in prison to work with us on these elements of phase 2 work package as well as to attend meetings as members of the Project Advisory Group. A paper describing the very positive benefits of this collaboration was the first study output to be published.33 The interview findings were then used to develop a ‘talking heads’ film, Cancer Cells, in collaboration with Synergy Theatre Project. This organisation works with people who have been in prison as part of their rehabilitation to develop films and plays that portray issues of justice. Once the script had been produced, EbE advised members of the research team about whether some sensitive details about taking a sample should be included in the script. Several members of the Synergy Theatre Project cast had been in prison themselves, which we believe led to the development of a film that was seen as hard-hitting but very realistic by stakeholders attending the Policy Lab. We have so far received very positive feedback about this element. The three EbE also attended the Policy Lab event and contributed to discussions, and they have been positive about their involvement. Two of the EbE were also present at the launch of its findings and the Synergy Theatre Project film Cancer Cells at the House of Lords in May 2024, where they also presented their experience of being involved in the study. No analysis of economic benefit has been undertaken, but our experience is that this kind of PPI is essential to any future study of the experiences of people in prison.