Table 4

Quantitative studies

Citation and countryParticipantsDesignAimKey findingsKey methodological strengths and limitations; quality score
Chari et al 43 (2016)
USA
Representatives from each of the 50 state department of corrections (n=45)Cross-sectional national telephone surveyTo present selected findings on the provision of healthcare in US state prisons
  • 45 of 50 States responded

  • 35 States provide hospice care exclusively on-site

  • Of this 35 to 12 have specific or reserved hospice beds

  • Of this 12 to 6 are joint hospice and long-term care units

  • Nine provide hospice care both on-site and off-site

  • ‘Most’ state that off-site care is rarely used

Strengths: Question sufficiently described, evident and appropriate design, subject group characteristics/variables described sufficiently, analytical methods justified and appropriate, well supported conclusions
Limitations: The appropriateness of the visual representation of some results is not clear
Quality score: 93%
Cloyes et al 8 (2015)
USA
All patients admitted to the Louisiana State Penitentiary (LSP) hospice programme from 01/2004 to 05/2012 (n=79)Comparative descriptive study, retrospectiveTo document characteristics of population of prison hospice patients, and to describe differences between this population and the general community
  • Average LSP patient age at time of death: 56 (SD=9.72, range 29–75)

  • Less than 16% community-hospice patients were aged 64 or younger on admission. 83% LSP patients were under 64.

  • Average time incarcerated before admission to hospice was 14.6 years

  • 1/3 LSP patients have two or more major illnesses prior to hospice

  • 41% have HIV or hepatitis

  • 60% of community admissions to hospice are for non-cancerous diseases. Only 6% LSP admissions are for non-cancer.

  • Median LSP hospice stay is 40 days compared with 19 in the community

  • 90% of LSP patients received opioids during the final 72 hours of life

  • Prison hospice patients had significantly less distressing symptoms (breathlessness, delirium, agitation) at the end of life than community based patients.

Strengths: Question sufficiently described, evident and appropriate design, subject selection/information source appropriate, subject group characteristics/variables described sufficiently, outcome well defined and robust to bias, detailed results, well supported conclusions
Limitations: Partial information on how comparison group data obtained
Quality score: 93%
Jadhao et al 44 (2015)
India
All deaths in custody from natural causes which were brought to a single hospital in India for autopsy between 01/2008 and 12/2013 (n=96)Retrospective descriptive studyTo examine mortality patterns in custodial deaths in a part of India
  • 118 deaths in custody. 96 (81.35%) of these were of natural causes

  • Of the 96, 87 (90.62%) were male, nine (9.38%) were female

  • Ischaemic heart disease most common cause of death (23.95%)

  • Pneumonia (21.87%)

  • Tuberculosis (21.87%)

Strengths: Minimal methods reported - difficult to assess strengths
Limitations: Partial reporting of question, design, subject selection and characteristics, results and conclusions. Cannot discount misclassification bias with limited information provided on how outcome (cause of death) was measured. No information on how data were extracted and analysed.
Quality score: 38%
Papadopoulos & Lay45 (2016)
UK
Nurses who have worked in a prison in England or Wales within the previous 2 years (n=31)Online surveyTo investigate views of current and former (<2 years) prison nurses with regard to end of life care being provided in UK prisons
  • 21/31 (68%) reported having some form of palliative and end of life care experience; for most this was due to a previous role as a community nurse or through a short course from a hospice

  • 12 stated their prison had a written palliative care policy, four stated their prison did not, seven were unsure

  • 23/31 provided information on their prison’s facilities:

    • 52.2% stated their prison had a hospital wing

    • 43.5% stated they had at least one nurse with palliative care training

    • 30.4% had prison volunteer carers or ‘buddies’ allocated to dying prisoners

    • 13.0% had facilities for families of dying prisoners

  • Barriers to end of life care included environmental barriers (no hospital wing, all single cells), regime barriers (perceived inflexibility, fixed visiting times) and security barriers (frequent lock-downs, failure to appreciate the reduced risk of dying prisoners causing harm.)

  • Examples of good practice include: access to specialist palliative care and specialist equipment, supportive policies (eg, named nurse, 24 hours unlocking for end of life) support (peer carers, custodial staff) and better access to families

Strengths: Sufficient participant information, well supported conclusions
Limitations: Limited depth of information on design and results. Risk of response bias due to small sample and potential for respondents to be from same establishment: cannot generalise about establishments and facilities as a result.
Quality score: 57%
Pazart et al 46 (2018)
France
All healthcare units for prisoners in France (n=190). Prison population 66 698Prospective national surveyTo assess the number and characteristics of prisoners requiring palliative care in French prisons
  • n=60 palliative care situations were identified. 10 were excluded for various reasons including consent, incomplete responses and life expectancy >1 year. Sample=50.

  • The majority of these patients were male (47:3) which is representative of the prison population as a whole

  • The estimated annual prevalence of sick prisoners requiring palliative care is 15.2 per 10 000 (CI 12.5 to 18.3). This number is twice as high as it would be for an equivalent patient in the community, or equivalent to someone 10 years their senior.

  • 33/50 requested early release on compassionate grounds.

  • 16/33 received a positive answer to this request

  • It is estimated that a further 12/50 would also have been eligible for early release on compassionate grounds, but did not request it.

Strengths: Question sufficiently described, evident and appropriate design, subject selection/information source appropriate, subject group characteristics/variables described sufficiently, outcome well defined and robust to bias, sample size appropriate, analytical methods justified and appropriate, variance reported for main results, detailed results, well supported conclusions
Limitations: More detail on physician classification/diagnosis of prisoners could improve
Quality score: 95%
Rothman et al 9 (2018)
USA
All state hospital decedents from 2009 to 2013 (n=370 831)Cross-sectional, comparative studyTo compare incarcerated and non-incarcerated decedents in California
  • 370 831 hospital decedents. 745 incarcerated, 370 086 non-incarcerated

  • Incarcerated decedents were more often male (93% vs 51% p<0.05) and younger (55 vs 73 years old, p<0.05)

  • Fewer had advanced care plan (23% vs 36% p<0.05)

  • Between 2001 and 2013, number of non-incarcerated decedents over 55 stayed at 80%, while it grew from 33% to 46%, with a peak of 55% in 2010 for incarcerated

  • Incarcerated decedents were more likely to have the following diagnoses on admission to hospital: cancer (10.2% vs 6.4%), liver disease (3.5% vs 1.4%), or mental health conditions (2.6% vs 1.1%), all p<0.05

  • On admission, incarcerated decedents were almost five times as likely to have HIV or AIDS (1.9% vs 0.4%) and 10 times as likely to have hepatitis (4.2% vs 0.4%)

  • Causes of death which were more common in incarcerated decedents included viral hepatitis (10.6% vs 1.0%), suicide (3.1% vs 0.3%), drug overdose (3.4% vs 0.4%), and homicide (0.9% vs 0.3%) all p<0.5.

Strengths: Question sufficiently described, evident and appropriate design, subject selection/information source appropriate, outcome well defined and robust to bias, sample size appropriate, analytical methods justified and appropriate, detailed results, well supported conclusions
Limitations: Subject group description limited by lack of recorded data on incarcerated decedents.
Quality score: 80%