Table 3 All other study designs
Reference *If same sample as another study in list | Study design | Study Setting | Country setting | Comparison [If yes (Y), describe; no (N)] | Strategy, Whether sample size calculation was reported for non-census strategies | Ethics reporting (documented ethics committee approval; described informed consent procedure) | Participants characteristics (sample size, mean age, percent male) *Indicates gender as inclusion criteria | Inclusion criteria (excluding age criteria) | Trial status category (*1) (C = over 50% convicted; NC = over 50% not convicted; NC/A = over 50% “awaiting trial”; JI = over 50% youth justice-involved; U = unclear; NA = not applicable; NS = not stated) | Assessment instruments (diagnostic or screening tool) | Primary outcomes (p-value listed if provided in study) | Methods risk of bias score |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Eseadi et al. 2017 [42] | Pre-post | Prison | Nigeria | Y (15 treatment, 15 control group not receiving intervention) | Census | Yes, Yes | 15, NS, 100% (treatment); 15, NS, 100% (control) | BDI score ≥ 29 | NS *But 84% awaiting trial in the prison population from which the sample was selected | BDI (S) | Significant treatment by time interaction effect for cognitive behavioral coaching program on depression as measured by BDI (p = 0.000). Significant decrease from pre to post-test BDI score (p = 0.000) for the CBC group compared to control | Low |
Onyechi et al. 2017 [43] | Pre-post | Prison | Nigeria | Y (10 treatment, 10 control group not receiving intervention) | Census | Yes, Yes | 10, NS, 100% (treatment); 10, NS, 100% (control) | High scorers on CDS-12 | NS | CDS-12 (S) | After the cognitive behavioral intervention, prisoners in the treatment group has significantly lower post-intervention CDS-12 scores than the control group’s post-intervention scores (p = 0.00) | Medium |
Martyns-Yellowe 1993 [44] | RCT | Prison | Nigeria | Y (18 participants each in treatment groups receiving Flupenthixol or Clopenthixol injections) | Census | No, No | 18, NS, 100%* (Flupenthixol treatment); 18, NS, 100%* (Clopenthixol treatment) | Males; schizophrenia diagnosis; vagrant people removed from public places by law enforcement | U *Detained in prison asylum after “removed from streets” | BPRS (Brief Psychiatric Rating Scale) (S) | 57.1% drop in BPRS symptoms in the Flupenthixol group (p < 0.001) and 43.4% drop in the Clopenthixol group (p < 0.01). Flupenthixol group had better symptom reducation respsone than the Clopenthixol group (p < 0.01) | Medium |
Balogun and Olawoye 2013 [40] | Cross-sectional | Prison | Nigeria | Y (within institution comparison of high/low self-esteem and high/low emotional intelligence) | NS | No, No | 233, 31.3, 86.27% (total participants) | NA | NS | SDS Self-Rating Depression Scale (S), TMMS Trait Meta-Mood Scale (S), Rosenberg self-esteem scale (S) | Both emotional intelligence (p < 0.05) and self-esteem (p < 0.05) had a significant influence on depression | Low |
Idemudia 1998 [131] | Cross-sectional | Prison | Nigeria | N | Random, NS | No, No | 150, 27.8, 61.3% | NA | NS | API (S), MSQ/CCEI (S) | Long-term detained persons had higher mean scores of psychopathy symptoms (API), (p < 0.001), and neurotic symptoms (MSQ/CCEI), (p < 0.001), than those serving medium and short terms | Medium |
Idemudia 2007 [132] | Cross-sectional | Prison | Nigeria | Y (college students, matched for gender, youth characteristic, and age*) *However, we note that statistics show that college students have noticeably older mean age | Random, NS | No, No | 100, 17.2, 83% (detained participants); 100, 25.2, 81% (college students) | Homeless on street before prison | NS | PDS (S), MAACL-H (S) | Higher scores on the Psychopathic Deviate Scale (p < .05) and the Multiple Affect Adjective Checklist hostility subscale (p < .0001) among the imprisoned homeless group than the non-prison and never homeless group | Medium |
Ineme and Osinowo 2016 [133] | Cross-sectional | Prison | Nigeria | N | Random, NS | Yes, Yes | 212, 34.4, 86.3% | NA | NS | HADS (S), IS-HUS (S), questionnaire (S) | Participants who used psychoactive substances (questionnaire) before detention reported higher self-harm urges (IS-HUS) than those who did not use (p < .01). Participants with higher depressive symptoms (HADS) reported higher self-harm urges than those with low depressive symptoms (p < .01}. Significant interaction of prior substance use and depression (< .01) | Low |
Stephens et al. 2006 [38] | Cross-sectional | Prison | South Africa | N | Census | Yes, Yes | 357, NS, 100%* | Males; pre-release; scheduled to be released from prison within three months after receiving intervention in parent study | U *all participants have pre-release status | Questionnaire (S) | Participants who used psychoactive substances (questionnaire) before detention reported higher self-harm urges (IS-HUS) than those who did not use (p < .01). Participants with higher depressive symptoms (HADS) reported higher self-harm urges than those with low depressive symptoms (p < .01}. Significant interaction of prior substance use and depression (< .01) | Medium |
Weierstall et al. 2011 [37] | Cross-sectional | Prison | Rwanda | N | Random, NS | Yes, Yes | 269, 33, 66% | Perpetrators of the Rwandan genocide | C *82% convicted, 18% awaiting trial | PTSD Symptom Scale-Interview (PSS-I) (D), PDS Event Scale (S), Appetitive Aggression Scale (AAS) (S) | Dose–response effect via path analysis between the exposure to traumatic events and the PTSD symptom severity (p < .001). Participants who had reported that they committed more types of crimes demonstrated a higher AAS score (p < .01), and higher AAS scores predicted lower PTSD symptom severity scores (p < .05). | Low |
Odejide 1979 [134] | Cross-sectional | Forensic ward | Nigeria | N | Census | No, No | 2158, NS, 95.9% | Psychiatric referrals | U *Referrals to hospital | Court records (NA) | 32.4% of 81 individuals with murder charges were referred for psychiatric opinion. No individuals with charges in categories of crime, including three individuals with charges of attempted suicide, was sent for psychiatric examination. Absence of mental illness in 66.6% of subjects referred for psychiatric opinion | Low |
Sukeri et al. 2016 [135] | Cross-sectional | Forensic ward | South Africa | N | Census | No, No | NA | NA | NA | Questionnaire (S) | No nurses with advanced training in forensic psychiatry. Lack of sufficient human resources. The nurse/patient ratio was 1:4. For 403 patients, 1.6 psychiatrists (1 full time),1 social worker, 1 occupational therapist, 0 occupational therapist assistants. There are 22 psychologists in all correctional centers in South Africa. None of the correctional centers have an onsite psychiatric unit | Low |
Ononye and Morakinyo 1994 [39] | Cross-sectional | Youth Institution | Nigeria | Y (50 school going children, matched for sex, age, ethnicity and educational level) | Census | No, No | 50, 14.1, 86% (youth in remand home); 50, 14.1, 86% (school-going youth) | NA | NS *Remanded youth | Carlson Psychological Survey (CPS) (S) | Thought disturbance significantly higher in youth in remand home compared to school-going youth. Antisocial tendency and self-depreciation higher among youth in remand home but not significantly. Substance abuse not significantly different between groups. (all indicated by CPS) | Medium |
Large and Nielssen 2009 [41] | Cross-sectional | Health system | International | Y (LMIC and HIC countries) | Census | No, No | NA | NA | NS | Published records in the literature (NA) | Correlation between per capita psychiatric hospital beds and prisoner numbers in the 158 countries (p < 0.01) and the subgroup of 120 LAMI countries (p < 0.01). No significant correlation within the 38 HI countries | Low |
Gaum et al. 2006 [45] | Qualitative | Prison | South Africa | N | Convenience, N | No, Yes | 10, 37.6, 50% (interviews); 18, NS, 100% (in focus groups) | Recidivists; psychological services clients | C | Interviews and focus groups | Interviews reveal a shortage of medical personnel in the prison psychiatry/psychology service. Also suggested from interviews: overpopulation in prisons may be due to rapid and dramatic political and economic changes in South Africa, coupled with the belief that crime pays and that being in prison is preferable to being jobless and homeless outside | Low |
Pretorius and Bester 2009 [47] | Qualitative | Prison | South Africa | N | Purposive, N | Yes, Yes | 3, 35–42, 0% | Women convicted of homicide of their intimate partner | C | Interview | All three participants’ interviews were indicative of PTSD and substance misuse | Low |
Topp et al. 2016 [46] | Qualitative | Prison | Zambia | N | Purposive and Random, N | Yes, Yes | 79, 35.6, 100%* (detained); 32, NS, 50% (prison staff) | Detained men | C *70–100% convicted depending on facility | Interviews and focus groups | A majority of participants in prison, as well as facility-based officers reported anxiety linked to over-crowding, sanitation, infectious disease transmission, nutrition and coercion. Interviewees associated overcrowding with negative effects on both participants in prison and officers’ physical and mental health. Limited access to healthcare | Low |
Kaliski et al. 1997 [48] | Qualitative | Forensic ward | South Africa | N | Census | No, Yes | 88, 30.4, 100% | Defendants undergoing psychiatric referral | NC *Pre-trial defendants for psychiatric observation | Psychiatric record (D) | 30.7% ultimately declared mentally ill. Only 25% knew that they were to be psychiatrically examined during the 30-day period. 44.3% did not know what was to happen to them after the completion of the observation period | Low |
Dube-Mawerewere 2015 [49] | Structured health system review | Health system | Zimbabwe | N | Purposive, N | No, No | 32, NA, NA | Forensic psychiatry system stakeholders | NS | Interview | Special psychiatric institutions housed within prisons, resulting in prison-like living conditions. Lack of staff in special institutions and forensic psychiatry settings with psychiatric training. Revolving door between civil psychiatric institutions in the prison, forensic hospital, and prison | Not assessed due to study design |
Kidia, et al. 2017 [22] | Structured health system review | Health system | Zimbabwe | N | Purposive, N | No, No | 30, NA, NA | Mental health system stakeholders, excluding patients | NA | Interviews, Emerald national-level needs assessment methods | Forensic facilities were substantially under-resourced, especially shortages of psychotropic medicines and human resources. Patients lived in overcrowded holding cells with unhygienic living conditions, with high prevalence of sexual assault and HIV transmission, minimal access to psychotropic medications and psychiatric care, and little food | Not assessed due to study design |
Liddicoat et al. 1972 [136] | Tool validation | Prison | South Africa | Y (99 participants with psychopathy diagnosis and 99 without psychopathy diagnosis matched for age and IQ) | Purposive, NS | No, No | 198, NS, 100% (total participants, pooled) | Participants with and without psychopathy diagnosis | C | Questionnaire (S) | 64/150 items on the questionnaire discriminated significantly between participants with and without psychopathy diagnosis | Not assessed due to study design |
Prinsloo and Ladikos 2007 [137] | Tool validation | Prison | South Africa | Y (231 those with offense designated “high-risk” compared to 38 segregated due to history of maladjustment, disciplinary problems and other institutional infractions) | Purposive, NS | No, Yes | 269, 31.8, 100%* (total participants, pooled) | Men with offense; those designated “high-risk” | NS | SAQ (S) | The overall alpha score of the SAQ, inclusive of all the interactive subscales, is (.904) | Not assessed due to study design |
Prinsloo 2013 [138] | Tool validation | Prison | South Africa | N | NS | No, Yes | 236, 34, 100% | NA | C | Psychiatric record (D), SAQ (S) | Logistic regression model of the behavioral characteristics assessed with the Self-Appraisal Questionnaire (SAQ) shows that modeling the behavioral characteristics accounts for 61% of the variation in the dependent variable mental illness. Subscales of anger, criminal tendencies and anti-social personality have significantly higher (p < 0.05) mean scores for mentally ill respondents | Not assessed due to study design |
Bunnting et al. 1996 [139] | Tool validation | Forensic ward | South Africa | Y (50 patients designated “malingering” and 50 state patients with mental disorder or sick (State President’s Detainees) | Purposive, N | No, No | 100, NS, NS (total participants, pooled) | Psychiatric referrals and state patients | NS *Pre-trial, convicted, and referrals | Questionnaire (S) | 17/20 items on the questionnaire statistically significant based on the study sample | Not assessed due to study design |