This cannot be used as single cut point as it also maximises the amount of false negatives, but it is useful as it minimises the false positives identified i. This categorisation by level of risk could contribute to identifying appropriate care pathways and, given the strength of the negative tests, may facilitate sign-off from the ACCT.
It is plausible that the respective gender-specific item sets, which resulted from the logistic regression, could form single page clinical decision aids which could be administered by any prison staff within a few minutes. The basic self-harm incidence during the six-month follow-up was The overall incidence rate recorded for males was This difference in rates would probably be expected, given the difference of study populations.
For females, the overall incidence rate recorded during follow-up was It is speculated that, when compared to males, this closer similarity of female self-harm rates is due to a higher proportion of the total female prison population also falling into the corresponding ACCT population.
The primary aim of the study was to determine whether any pre-existing instruments could predict self-harm among an ACCT population. The AUC analysis that was carried out on the candidate instruments determined that none of these performed the task adequately enough to be considered a useful aid for prison staff to utilise as part of a standardised ACCT process.
This finding has also been the case when using standardised measures to predict suicide following self-harm, where it has been warned that the use of these standardised scales, or an over-reliance on the identification of risk factors in clinical practice, may provide false reassurance that could be potentially dangerous Chan et al.
In turn, any element of risk reduction for a given individual may also reduce the probability of the final outcome occurring in the population of interest, thus interfering with any attempts to establish the predictive validity of the instruments that are being assessed. Although this issue may be present, in this instance it is unlikely to have had a major impact on the results as all study participants are from the prison-ACCT population, and are therefore already classified as being at an increased risk of self-harm.
A further potential limitation lies with the self-harm outcome data coming exclusively from prison records.
This will likely lead to an under-ascertainment of self-harm events, as some self-harm remains self-managed and unreported. This has been previously observed Borschmann et al. Although none of the pre-existing standardised instruments predicted the risk of self-harm in the ACCT population, an exploratory logistic regression revealed a set of items that may be useful when aggregated into a predictive algorithm, which could be used as a clinical decision aid to indicate risk of future self-harm.
This risk factor approach has often been used to incorporate individual risk factors into composite scales to assess for the risk of suicide following self-harm Chan et al. In a prison setting, this approach has been used for the identification of inmates that carried out suicide Blaauw et al.
A similar approach has also been utilised in order to identify self-harm self-injurious behaviour in male prisoners Lanes Barton et al. These studies produced AUC values of 0. However, both of these studies used retrospective data to classify the difference between prisoners with and without a history of self-harm, whereas the current study used prospective data to classify whether self-harm occurred among an ACCT population during an active follow-up period.
Machine learning approaches can often outperform conventional statistical predictive modelling in predicting health outcomes Song et al.
Should a machine learning approach be adopted, it would be recommended that a human aspect should remain in any final decision-making process. Some of the predictive items identified within the present study differ from those that have previously been reported as risk factors for self-harm. For example, one study focusing on female incarcerated adults reported shame, anger and child abuse as important Milligan and Andrews Although child abuse was not addressed, shame was incorporated as a question in our study, but it did not appear to be predictive of future self-harm.
This was successful at predicting self-harm with a rate of A further study identified several independent predictors for suicide including previous psychiatric service contact, history of self-harm, single cell occupation, remand status, and non-white ethnicity Humber et al.
In the present study, history of self-harm was predictive, but remand status and non-white ethnicity were not predictive of self-harm. Previous contact with a psychiatrist was predictive for males and females, but cell occupancy status was not determined.
Some of the items identified in the present study are particularly interesting. Although there are various possible explanations for these findings, it is recommended that these items are studied further within this setting. An issue with all risk factor item sets that have been derived in this way, as is the case in the present study, is that although these item sets seem to work statistically, it is likely that the identified items involve an element of capitalisation on chance within the specific dataset that is used.
Due to this restriction, it is vital that any of these risk factor items sets are revalidated prospectively. Another major issue with a lot of the scales that have been derived in this way are that they use solely retrospective data, and they are never further validated prospectively, meaning that along with the chance capitalisation, no process of causality can be assumed.
Additionally, the practical implementation of risk factor item sets may be limited for a number of reasons. The identified risk factors are often comparatively common in the populations of interest Chan et al. Another issue with the item set identified in the present study is that many of the items are static in nature.
These static items refer to background and lifetime information which cannot change once the item has been affirmed. If an actual incidence of self-harm has occurred in order to trigger initiation of the ACCT, it has been suggested that a comprehensive psychosocial assessment of the risks and needs that are specific to the individual should be central to the management of these people who have self-harmed Chan et al.
The gender-specific predictive risk item sets identified in this study may be useful in this regard, as they offer the opportunity to classify three levels differing levels of risk that could be used at reception into prison.
If the risk classification was medium or high, then a further in-depth assessment could be carried out, as has been previously recommended Chan et al. Given the high negative predictive values, the predictive item sets appear to function better at screening out self-harm than screening it in. Although this is not the ideal intention, it could still help to save time and focus the limited resources that are available. Despite an apparently limited predictive power, the implementation of a screening process that is specific to self-harm could certainly contribute to an increased awareness of self-harm and mental health issues amongst prison staff.
This is consistent with other reports of a lack of staff training and policy, along with an inconsistency in response to self-harm behaviour Roe-Sepowitz This evidence leads to the critical point that an improvement in staff awareness and attitude, along with further training, are important factors which may help prevent self-harm and suicide in prisons Hawton et al.
Although this staff awareness shortfall has been identified and is being addressed, it has been acknowledged that much work remains to be done Forrester and Slade However, as this analysis was carried out post-hoc, although it is plausible that these item sets could potentially be useful, their direct predictive capacity and operational functionality remains unknown. Barton, J. Predictors of self-harm in male inmates. Journal of Criminal Psychology, 4 1 , 2— Article Google Scholar.
Beck, A. Beck depression inventory BDI. Archives of General Psychiatry, 4 6 , — The measurement of pessimism: The hopelessness scale. Journal of Consulting and Clinical Psychology, 42 6 , — Blaauw, E. Demographic, criminal, and psychiatric factors related to inmate suicide. Suicide and Life-threatening Behavior, 35 1 , 63— Bohus, M.
The short version of the borderline symptom list BSL : Development and initial data on psychometric properties. Psychopathology, 42 1 , 32— Borschmann, R.
CMAJ open, 5 3 , E Chan, M. Predicting suicide following self-harm: Systematic review of risk factors and risk scales. The British Journal of Psychiatry, 4 , — Department of Health.
Public health outcomes framework: Improving outcomes and supporting transparency. Part 1A: A public health outcomes framework for England, — London: Crown Copyright. Evans, C. CORE: Clinical outcomes in routine evaluation. Journal of Mental Health, 9 3 , — Fazel, S. Prison suicide in 12 countries: An ecological study of suicides during — Social Psychiatry and Psychiatric Epidemiology, 46 3 , — Field, A. London: Sage.
Firth, D. Bias reduction of maximum likelihood estimates. Biometrika, 80 1 , 27— Fliege, H. Three assessment tools for deliberate self-harm and suicide behavior: Evaluation and psychopathological correlates. Journal of Psychosomatic Research, 61 1 , — Forrester, A.
Preventing self-harm and suicide in prisoners: Job half done. The Lancet, , — Fruehwald, S. The relevance of suicidal behaviour in jail and prison suicides. European Psychiatry, 18 4 , — Gavin, N. Reception screening and mental health needs assessment in a male remand prison.
Psychiatric Bulletin, 27 7 , — Gray, N. Prediction of violence and self-harm in mentally disordered offenders: A prospective study of the efficacy of HCR, PCL-R, and psychiatric symptomatology. Journal of Consulting and Clinical Psychology, 71 3 , — Hawton, K. Self-harm in prisons in England and Wales: An epidemiological study of prevalence, risk factors, clustering, and subsequent suicide.
Haycock, J. Manipulation and suicide attempts in jails and prisons. The Psychiatric Quarterly, 60 1 , 85— Heinze, G. A solution to the problem of separation in logistic regression. Statistics in Medicine, 21 16 , — The ACCT approach. London: Caring for people at risk in prison.
Google Scholar. Holzinger, A. Interactive machine learning for health informatics: When do we need the human-in-the-loop? Brain Informatics. Horton, M. Assessing the risk of self-harm in an adult offender population: An incidence cohort study.
Health Technol Assess, 18 64 , 1. Humber, N. A national case-control study of risk factors among prisoners in England and Wales. Social Psychiatry and Psychiatric Epidemiology, 48 7 , — Version Armonk: IBM Corp. Kroenke, K. The PHQ Validity of a brief depression severity measure. Journal of General Internal Medicine, 16 9 , — Lanes, E. Identification of risk factors for self-injurious behavior in male prisoners. Journal of Forensic Sciences, 54 3 , — Lohner, J.
Risk factors for self-injurious behaviour in custody: Problems of definition and prediction. International Journal of Prisoner Health, 3 2 , — Lovibond, P. The structure of negative emotional states: Comparison of the depression anxiety stress scales DASS with the Beck depression and anxiety inventories. Behaviour Research and Therapy, 33 3 , — Metz, C. Basic principles of ROC analysis. Paper presented at the Seminars in nuclear medicine. Milligan, R. Suicidal and other self-harming behaviour in offender women: The role of shame, anger and childhood abuse.
Legal and Criminological Psychology, 10 1 , 13— Ministry of Justice. Guide to safety in custody statistics. Morgan, J. Self-reported suicidal behavior in juvenile offenders in custody: Prevalence and associated factors. Crisis, 25 1 , 8— Ness, J. Waters, K. Alcohol use and misuse, self-harm and subsequent mortality: An epidemiological and longitudinal study from the multicentre study of self-harm in England.
Emergency Medicine Journal , 32 10 , — Crown copyright. Self-harm: longer-term management CG National Institute for Health and Clinical Excellence. O'Leary, W. Custodial suicide: Evolving liability considerations. Psychiatric Quarterly, 60 1 , 31— Perry, A.
Detecting and predicting self-harm behaviour in prisoners: A prospective psychometric analysis of three instruments. Social Psychiatry and Psychiatric Epidemiology, 44 10 , — A new psychometric instrument assessing vulnerability to risk of suicide and self-harm behaviour in offenders: Suicide concerns for offenders in prison environment SCOPE.
Screening tools assessing risk of suicide and self-harm in adult offenders: A systematic review. Prisons and Probation Ombudsman. Annual Report London: Crown copyright. Learning from PPO investigations: Risk factors in self-inflicted deaths in prisons. Public Health England. Quinlivan, L. Scales for predicting risk following self-harm: An observational study in 32 hospitals in England. BMJ Open, 4 5 , e Rasch, G. Probabilistic models for some intelligence and attainment tests.
Copenhagen: Danish Institution for Educational Research. Roe-Sepowitz, D. Indicators of self-mutilation: Youth in custody. Sansone, R. The self-harm inventory SHI : Development of a scale for identifying self-destructive behaviors and borderline personality disorder.
Journal of Clinical Psychology, 54 7 , — Saunders, K. Attitudes and knowledge of clinical staff regarding people who self-harm: A systematic review. Journal of Affective Disorders, 3 , — Shaw, J. A screening questionnaire for the detection of serious mental illness in the criminal justice system. Suicide by prisoners national clinical survey. The British Journal of Psychiatry, 3 , — Slade, K. Applying the cry of pain model as a predictor of deliberate self-harm in an early-stage adult male prison population.
Legal and Criminological Psychology, 19 , — Smith, H. Inmate self-injurious behaviors: Distinguishing characteristics within a retrospective study. Criminal Justice and Behavior, 37 1 , 81— Song, X. Comparison of machine learning techniques with classical statistical models in predicting health outcomes.
In Medinfo pp. Stata statistical software: Release Tiffin, P. Rise of the machines? Machine learning approaches and mental health: Opportunities and challenges.
Self-harm among UK female prisoners: A cross-sectional study. Journal of Forensic Psychiatry and Psychology, 20 5 , — Ward, J. A participatory action research methodology in the management of self-harm in prison.
This research was carried out using in-depth qualitative interviews with members of staff and prisoners across six prison establishments in England and Wales.
It explores staff and prisoner views on the ACCT process, and identifies possible improvements. To help us improve GOV. It will take only 2 minutes to fill in. Cookies on GOV. UK We use some essential cookies to make this website work. Accept additional cookies Reject additional cookies View cookies. Hide this message. Home Crime, justice and law Prisons and probation. Research and analysis overview: The Assessment, Care in Custody and Teamwork process in prison: findings from qualitative research.
From: Ministry of Justice Published 22 March Explore the topic Prisons and probation.
0コメント