A comparison of scoring models for computerised mental health screening for federal prison inmates
Bibliographic record
Abstract
BACKGROUND: There are high rates of mental disorder in correctional environments, so effective mental health screening is needed. Implementation of the computerised mental health screen of the Correctional Service of Canada has led to improved identification of offenders with mental health needs but with high rates of false positives. AIMS: The goal of this study is to evaluate the use of an iterative classification tree (ICT) approach to mental health screening compared with a simple binary approach using cut-off scores on screening tools. METHODS: A total of 504 consecutive admissions to federal prison completed the screen and were also interviewed by a mental health professional. Relationships between screening results and more extended assessment and clinical team discussion were tested. RESULTS: The ICT was more parsimonious in identifying probable 'cases' than standard binary screening. ICT was also highly accurate at detecting mental health needs (AUC=0.87, 95% CI 0.84-0.90). The model identified 118 (23.4%) offenders as likely to need further assessment or treatment, 87% of whom were confirmed cases at clinical interview. Of the 244 (48.4%) offenders who were screened out, only 9% were clinically assessed as requiring further assessment or treatment. Standard binary screening was characterised by more false positives and a comparable false negative rate. CONCLUSIONS: The use of ICTs to interpret screening data on the mental health of prisoners needs further evaluation in independent samples in Canada and elsewhere. This first evaluation of the application of such an approach offers the prospect of more effective and efficient use of the scarce resource of mental health services in prisons. Although not required, the use of computers can increase the ease of implementing an ICT model.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.082 | 0.140 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.004 |
| Bibliometrics | 0.006 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".