Table 1 Summary of studies presenting data on the outcome domain of effectiveness (cohort studies)
Study and settingDesignNOutcome sub-domainMeasure/indicator
Alexander et al17
Medium secure
Retrospective cohort study
Follow-up: up to 13 years
64
  • Treatment response/outcome

  • Discharge pathway

  • Readmissions

  • Relapse in MH

  • Reoffending

  • ‘Offending-like’ behaviour

  • Level of CPA support, MHA status and CGIS

  • Level security/type of placement at discharge

  • % of patients readmitted to the same unit

  • % of patients who experienced a relapse in MH symptoms

  • Number of patients formally reconvicted, cautioned or had police contact

  • Behaviour which could be classed as an offence but did not lead to police contact

Alexander et ala,18
Medium secure
Retrospective cohort study
Discharged patients
Follow-up: 6 years
138
  • Length of stay

  • Discharge outcome

  • Discharge pathway

  • Mean and median number of days

  • % of positive (move to a lower level of security) or negative (move to higher level of security) discharges

  • Level security/type of placement at discharge

Alexander et al5
Medium secure
Retrospective cohort study
Follow-up: 6 years
138
  • Length of stay

  • Discharge outcome

  • Mean and median number of days

  • % discharged successfully (to a lower level of security) or not (to same/higher level of security)

Alexander et al19
Medium secure
Retrospective cohort study
(arson v. non-arson)
Follow-up: 6 years
30
  • Length of stay

  • Discharge pathway

  • Discharged patients only; mean and median days

  • % of discharged patients who moved to either a lower or same/higher level of security

Alexander et al20
Medium secure
Retrospective cohort study
(Discharged patients. PD v. ID v. PD+ID)
Follow-up: up to 9 years
145
  • Length of stay

  • Clinical symptoms

  • Risk assessment

  • Reoffending

  • Time to reoffend

  • Median number of days and age on discharge

  • PCL:SV pre-treatment scores (clinician rated)

  • HCR-20 pre-treatment scores (clinician rated)

  • Reconvictions at 1, 2 and 5 years follow-up. Categorised as serious/violent offending (as defined by Home Office)

  • The difference between discharge date and reconviction date for subsequent offence (median)

Ayres & Roy21
Community
Case series
(70% male)
Follow-up: up to 3 years
26
  • Cost

  • Level of support

  • Average saving cost per patient (based on hourly rate of support) and for the service over 3 years

  • Hours per annum providing both direct/indirect support

Barron et al22
Community
Cohort study
Follow-up: 9.5 months average
61
  • Clinical symptoms

  • Treatment engagement

  • Reoffending

  • ‘Offending-like’ behaviour

  • ABC and Mini PAS-ADD (clinician rated)

  • Categorised treatment use and% with history of service use

  • Number of patients who reoffended; mean number of offences per patient

  • Telephone call with carers regarding police contact

Benton & Roy23
Community
Cohort study
Follow-up: up to 3 years
113
  • Reoffending

  • Discharge pathway

  • Discharge outcome

  • Prevention of in-patient admissions

  • Number of patients arrested, reconvicted and cases dropped

  • Level security/type of placement at discharge

  • % of referrals discharged

  • Number of patients whose in-patient admission was prevented via community treatment

Butwell et al24
High secure
Cohort study
Follow-up: 10 years
Up to 278
  • Length of stay

  • Discharge pathway

  • Median and mean years, calculated per episode

  • Level security/type of placement at discharge

Day25
Medium secure
Cohort study
Follow-up: 3 years
20
  • Length of stay

  • Treatment response

  • Discharge pathway

  • Discharge outcome

  • Readmission

  • Reoffending

  • Mean number of months

  • Categorised as either good (settled and co-operative), fair (continuing lesser problems) or poor (severe problems)

  • % of patients discharged to rehab villa, community or hostel

  • Level of adjustment at follow-up, based on personal knowledge, hospital notes and liaison with involved agencies

  • % of patients readmitted to the same unit

  • Number who reoffended or returned to prison

Dickens et al26
Medium and low secure
Retrospective cohort study
Follow-up: 15 months
48
  • Length of stay

  • Clinical symptoms

  • Mean number of days

  • HoNOS-Secure: change between baseline and final rating. Quartile points used to allow for different lengths of stay (clinician rated)

Fitzgerald et alb,27
Medium secure
Retrospective cohort study
Follow-up: 2 years
145
  • Reoffending

  • Risk assessment

  • Home Office records;% of patients with both general and violent offences.

  • OGRS

Gray et al28
Medium secure
Retrospective cohort study
Follow-up: 2 years
115
  • Reoffending

  • Risk assessment

  • Home Office records:% of patients with both general and violent offences

  • HCR-20,VRAG, PCL: SV (rated once)

Halstead et al29
Medium secure
Cohort study
Discharged patients
Follow-up: up to 13 years
35
  • Length of stay

  • Treatment outcome

  • Discharge outcome

  • Discharge pathway

  • Relapse in MH

  • Readmissions

  • Offender-like behaviour

  • Reoffending

  • Mean and median months

  • Rated as either good (risks reduced, safe for discharge), some (progress made in some areas, risk remains), none (no change) or poor (worse than admission)

  • % of patients discharged

  • Level security/type of placement at discharge and follow-up

  • Recurrence of symptoms of illness or challenging behaviour

  • % of patients re-admitted to hospital

  • Behaviour which could be interpreted as an offence

  • Number of patients reconvicted

Lindsay et alc,30
Community
Cohort study
Follow-up: 13 years
62
  • ‘Offending-like’ behaviour

  • Reoffending

  • Percentage of patients ‘suspected of’ reoffending

  • Percentage of patients with clear evidence of reoffending

Lindsay et al31
Community
Cohort study
Follow-up: 7 years
184
  • Reoffending

  • Harm reduction

  • Number of patients with clear evidence of reoffending

  • Decrease in the number of incidents 2 years before referral v. incident data at follow-up

Lindsay et al44
Community
Females only
Case series
Follow-up: 3 years
18
  • Reoffending

  • Number of patients with clear evidence of reoffending

Lindsay et al32
Community
Cohort study
Follow-up: 12 years
247
  • Reoffending

  • Harm reduction

  • Number of patients with clear evidence of reoffending

  • Decrease in the number of incidents 2 years before referral v. incident data at follow-up

Lindsay et al33
Community
Cohort study
Follow-up: up to 20 years
309
  • Reoffending

  • Harm reduction

  • Number of patients with clear evidence of reoffending

  • Standardised service incident report data; number of offences committed in the 2 years before referral v. follow-up reoffending data

Lindsay et al34
Mixed services
Cohort study
Follow-up: 2 years
197
  • Discharge pathway

  • Number of patients discharged and level security/type of placement at discharge, 1-year and 2-year follow-up

Linhorst et al35
USA Community service
Cohort study
Follow-up: 6 months
252
  • Reoffending

  • Treatment response/engagement

  • Local law agency data and arrests frequencies used

  • Number of patients who completed treatment

Marks36
[unpublished thesis]
Medium secure
Retrospective cohort study
Follow-up: 4 years
28
  • Reoffending

  • Length of stay

  • Discharge pathway

  • % of patients who reoffended (even if this did not lead to police contact) determined by interview with current care team

  • Mean and median number of years

  • % of patients in high, medium, low and community services

Morrissey & Taylor37
High secure
Case reports
Follow-up: 2 years
13
  • Clinical symptoms

  • Discharge outcome

  • YSQ (patient rated), IPDE and PCL:SV (clinician rated): change in mean scores

  • Number of patients remaining in treatment, discharged to medium secure or transferred to another high secure ward

Morrissey et al38
High secure
Retrospective cohort study
(ID v. other non-ID services)
Follow-up: 5 years
70
  • Risk assessment

  • HCR-20: change in mean scores

Morrissey et al39
High secure
Prospective cohort study
Follow-up: 2 years
73
  • Discharge outcome

  • Risk assessment

  • Clinical symptoms

  • % of patients who made positive (move to a lower level of security) or negative (the same level of security) progress

  • HCR-20

  • PCL-R (clinician rated)

Morrissey et al40
High secure
Retrospective cohort study
(Two cohorts: in-treatment and admission cohort)
Follow-up: up to 6 years
68 and 24
  • Length of stay

  • Incidents

  • Clinical symptoms

  • Discharge pathway

  • Discharge outcome

  • Readmission

  • Median number of years

  • Frequency per year in the first 4 years of treatment using hospital incident records. Mean number of violent incidents per patient/year

  • EPS-BRS (clinician rated) & EPS-SRS (patient rated)

  • Level security/type of placement at discharge

  • % of patients who moved to a lower level of security

  • Number of returns from trial leave

Palucka et al41
Canadian in-patient unit
Retrospective cohort study
Follow-up: up to 9 years
20
  • Length of stay

  • Discharge outcome

  • Discharge pathway

  • Reoffending

  • Median number of days

  • Number of patients discharged

  • Level security/type of placement at discharge

  • Any criminal justice involvement (even if charges were not pressed)

Reed et ald,42
Low secure
Retrospective cohort study
Follow-up: up to 14 years
45
  • Length of stay

  • Discharge pathway

  • Discharge outcome

  • Incidents

  • Mean number of weeks

  • Level security/type of placement at discharge

  • Positive (discharged to a level of lower security) or negative (discharged to a level of higher security) outcome

  • Collected from hospital incident records. Incidents at baseline (week 6 to 10 of stay) were compared to end of stay (last 4 weeks of treatment). Frequency (total number of incidents per month) was adjusted for length of stay. Change in incidents calculated per person per week

Xenitidis et al43
Low secure
Retrospective cohort study
Follow-up: up to 11 years
64
  • Length of stay

  • Discharge pathway

  • Discharge outcome

  • Incidents

  • Mean number of months

  • Level security/type of placement at discharge

  • Good (discharged to community) or bad (not placed in a community setting) outcome

  • Collected from hospital incident records. Incidents at baseline (week 6 to 10 of stay) were compared to end of stay (last 4 weeks of treatment). Frequency (total number of incidents per month) was adjusted for length of stay.

  • ABC, Aberrant Behaviour Checklist; Mini PAS-ADD, Mini Psychiatric Assessment Schedules for Adults with Developmental Disabilities; YSQ, Young Schema Questionnaire; IPDE, International Personality Disorder Examination; PCL-R, Psychopathy Checklist – Revised; EPS-BRS. Emotional Problems Scale – behaviour rating scale; EPS-SRS, Emotional Problems Scale – self report scale; CPA, Care Programme Approach; MHA, Mental Health Act; MH, mental health; CGIS, Clinical Global Impressions Scale; PCL:SV, Psychopathy Checklist: Screening Version; HCR-20, Historical, Clinical, Risk Management-20; HoNOS, Health of the Nation Outcome Scale; OGRS, Offender Group Reconviction Scale; VRAG, Violence Risk Appraisal Guide.

  • a The sample in Alexander et al18 overlaps with that of Alexander et al,5,19 and Esan et al.4

  • b The sample in Fitzgerald et al27 overlaps with that of Gray et al28

  • c The sample in Lindsay et al30 overlaps with that of Lindsay et al.3133,44,45

  • d The sample in Reed et al42 overlaps with that of Xenitidis et al.43