Table 2 Summary of studies presenting data on the outcome domain of effectiveness (cross-sectional studies)
Study and settingDesignNOutcome sub-domainMeasure/indicator
Ajmal46
High secure
Cross-sectional79
  • Clinical symptoms

  • GSI and RSES (patient rated)

Beer et al47
Low secure
Cross-sectional59
  • Placement appropriateness

  • Clinical symptoms

  • Percentage of patients assessed as requiring a less secure placement

  • SBS (clinician rated)

Beer et al48
Low secure
Cross-sectional68
  • Length of stay

  • Placement appropriateness

  • Clinical symptoms

  • Mean number of months

  • Percentage of patients requiring less secure care. Main reason for delayed discharge via the Royal College of Psychiatrists Research & Development questionnaire

  • HoNOS-Secure (clinician rated)

Chaplin et al49
Low secure
Cross-sectional22
  • Risk assessment

  • Incidents

  • Length of stay

  • HCR-20 median scores

  • Average per patient at 3 monthly intervals. Coded for severity using theMOAS. Median number of incidents

  • Median number of days

Chilvers & Thomas50
Medium secure
Cross-sectional
(M v. F)
77
  • Clinical symptoms

  • NAS-PI scores (patient rated)

Crossland et al51
High, medium and low secure
Cross-sectional60
  • Length of stay

  • Median number of months

Dickens et al52
Medium and low secure
Cross-sectional
16-month period
68
  • Incidents

  • Severity rated by the individual completing the form as either: near miss, minor, moderate, high or very high

  • Incidents/total bed days × 100. Average number of incidents per 100 occupiedbed days, time of incident, number of violent/aggressive incidents and total number of incidents

Esan et al4
Medium and low secure
Cross-sectional
(ASD v. non-ASD)
114
  • Length of stay

  • Discharge outcome

  • Level of supervision/discharge pathway

  • Mean and median months for both discharged and in-treatment patients

  • Number of patients with a good (move to a lower level of security) or poor (move to a higher level of security) outcome

  • Number of patients who were informal,under a MHA section, guardianship or supervised discharge

Fitzgerald et al53
Medium and low secure
Cross-sectional136
  • Incidents

  • Risk assessment

  • Number of patients involved inincident in 6-month period

  • VRAG and HCR-20

Hall et al54
Medium and low secure
Cross-sectional136
  • Treatment needs

  • Security need

  • Delayed discharge

  • Length of stay

  • Incidents

  • Clinician ratings

  • Reference group ratings of appropriate security level

  • Number of patients no longer requiring current security level, main obstacle to progress

  • Maximum and average years per level of security

  • Number of patients involved in anincident in 6 months period

Hogue et al55
High, medium, low and community
Cross-sectional228
  • Clinical symptoms

  • EPS-BRS (clinician rated)

Johnson56
Medium and low secure
Cross-sectional44
  • Clinical symptoms

  • Length of stay

  • RSES and EBS (patient rated)

  • Mean number of months

Kellett et al57
High secure
Cross-sectional45
  • Clinical symptoms

  • BSI (patient rated)

Lindsay et al45
Community
Cross-sectional52
  • Offender-like behaviour

  • Reoffending

  • Percentage of patients suspected of reoffending

  • Percentage of patients with ‘clear evidence’ of reoffending

Lindsay et al58
High, medium and low secure
Cross-sectional212
  • Risk assessment

  • Clinical symptoms

  • HCR-20, VRAG, Static-99, SDRS, RM-2000

  • EPS-BRS (clinician rated)

Lindsay et al59
High, medium, low secure and community
Cross-sectional197
  • Risk assessment

  • VRAG and Static-99

Lofthouse et al60
Rehabilitation, acute admission and residential home
Cross-sectional
5 months of data
64
  • Length of stay

  • Risk assessment

  • Incidents

  • Mean number of years

  • CuRV

  • Aggression defined as acts of physical violence,aggression, force to hurt or damage to staff,peers or environment. Included verbal abusewhich was aggressive, threatening or caused offence. Two researchers rated each incidentas: ‘aggression present’ or ‘aggression absent’. Number of patients who were aggressive in month 1 versus month 5

Mansell et al61
Medium and low secure
Cross-sectional
NHS versus private provider units
1891
  • Delayed discharge

  • Incidents

  • Percentage of patients who had completed treatment but did not have any plans to leave the service in the next month.

  • Average frequency where a patient was hurt by a patient or staff member(per patient over a 6-month period)

McMillan et al62
Medium secure
Cross-sectional
6-month period
124
  • Risk assessment

  • Incidents

  • MDT ratings per patient on risk of physical violence (scale of 0-8) and number of times patient had been violent in 6 months prior to risk assessment

  • Author coded each description based on explicit criteria and guidelines. e.g. physical violence (attempted, contact between assailant or object and victim, evidence of physical harm to victim or attendance of medical personnel). Coded from computerised hospital database

Morrissey et al39
High secure
Cross-sectional
12-month period
60
  • Incidents

  • Risk assessment

  • Clinical symptoms

  • Coded as either interpersonal physical aggression or verbal aggression/aggression to property. Further rated as low, medium or high risk of harm. Number of patients involved in an aggressive incident

  • HCR-20

  • PCL-R and EPS-BRS (clinician rated)

O’Shea et al63
Medium, low and rehabilitation
Cross-sectional109
  • Risk assessment

  • Incidents

  • HCR-20

  • Hospital records in 3-month period following risk assessment for aggression and self-harm.Coded using Overt Aggression Scale. Rated on severity (1-4). Number of patients involved in any incident

Perera et al64
Medium and low secure
Cross-sectional388
  • Length of stay

  • Delayed discharge

  • Median number of years and percentage of patients who had stayed longer than 5 years

  • Percentage of patients assessed as requiring a less secure placement

Thomas et al65
High secure
Cross-sectional102
  • Length of stay

  • Delayed discharge

  • Security need

  • Treatment needs

  • Mean and median number of years

  • Percentage of patients assessed as requiring a less secure placement and main reason for this

  • SDTN scale completed by key worker and responsible clinician

  • CANFOR-Short and CANDID-Short. Average number of needs and unmet needs

Uppal & McMurran66
High secure
Cross-sectional
(ID sample included in wider data-set)
15-month period of incidents
396
  • Incidents

  • Hospital computerised reporting system. Coded as per Department of Health: Category A: major incidents (e.g. abscond, hostage taking); Category B: serious incidents (e.g. serious assault involving a weapon, attempted suicide); Category C: untoward incidents (e.g. attempted abscond, assault without a weapon); Category D: all other incidents (minor assault and verbal abuse) Most frequent location and time of incident Percentage of incidents which were violent and which were self-harm Average monthly figure generated

  • GSI, Global Severity Index; RSES, Rosenberg Self Esteem Scale; SBS, Social Behavioural Schedule; MOAS, Modified Overt Aggression Scale; NAS-PI, Novaco Anger Scale and Provocation Inventory; ASD, autistic spectrum disorder; VRAG, Violence Risk Appraisal Guide; RSES, Rosenberg Self-Esteem Scale; EBS, Evaluative Beliefs Scale; BSI, Brief Symptom Inventory; SDRS, Short Dynamic Risk Scale; RM-2000, Risk Matrix 2000; CuRV, Current Risk of Violence; MDT, Multidisciplinary team; SDTN, Security, Dependency and Treatment Needs Scale; CANFOR, Camberwell assessment of need – forensic version; CANDID, Camberwell Assessment of Need for Adults with Developmental and Intellectual Disabilities; ID, intellectual disability.