Prediction of violent reoffending in people released from prison in England: external validation study of a risk assessment tool (OxRec)
Bibliographic record
Abstract
Purpose: We aimed to externally validate the Oxford Risk of Recidivism (OxRec) tool to estimate 1- and 2-year risk of violent reoffending in people released from prison in England.Methods: We identified people released from prison using administrative data shared between official prison and police services. We extracted information on criminal history, clinical and sociodemographic risk predictors, and outcomes. The outcomes were violent reoffending at 1 and 2 years after release from prison, identified using official police data. Predictive ability was examined using measures of calibration (calibration statistics and plots) and discrimination (area under the receiver operating characteristic curve [AUC]; sensitivity, specificity, positive and negative predictive values [PPV; NPV]) for predetermined risk thresholds. Recalibration of the model was conducted when necessary.Results: In total, 1,770 individuals (median age = 33 [IQR 27–40]; 92% were male) were identified as recently released from prison. 31% (n = 550) and 43% (n = 765) reoffended within 1 and 2 years, respectively. Simple validation of the original model found a systematic underestimation of the probability of reoffending. However, after recalibration, OxRec was associated with AUCs of 0.71 (95% CI: 0.69-0.74) for 1-year and 0.71 (0.68-0.74) for 2-year follow-up. At a pre-specified threshold of 40% for 2-year violent reoffending risk, sensitivity was 77% (95% CI: 74%-80%), specificity 54% (51%-58%), PPV 56% (53%-59%) and NPV 76% (73%-79%). In addition, in the revised model, measures of calibration were good (calibration in the large was null for both time points). Conclusions: External validations of risk assessment tools for reoffending with adequate sample sizes can be conducted using linked data between prison and police services, and may require model recalibration before implementation. In this validation, OxRec had good performance on measures of discrimination and calibration. It can be considered as part of approaches to improve decision-making about risk of serious offending and the allocation of resources in the criminal justice system.
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 | 0.014 | 0.059 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".