Skip to main content

Table 5 Performance of the Y-ACNAT-NO when predicting specific types of care needs

From: The development and validation of the Youth Actuarial Care Needs Assessment Tool for Non-Offenders (Y-ACNAT-NO)

 

Construction sample

Validation sample

Specific type of care needs

AUC (95% CI)

BS a

SMBS

RIS

AUC (95% CI)

BS a

SMBS

RIS

Supervision order

.767*** (.745 - .788)

.0636***

.0628

.0389

.741*** (.719 - .762)

.0616***

.0604

.0378

Total worrisome incidents

.809*** (.788 - .828)

.0789

.0805

.0003

.758*** (.736 - .778)

.0899

.0904

.0023

Child abuse

.908*** (.892 - .921)

.0458***

.0482

.0230

.823*** (.804 - .841)

.0490***

.0497

.0299

Domestic violence

.790*** (.769 - .810)

.0721**

.0744

.0051

.740*** (.718 - .761)

.0797

.0821

.0093

Domestic strife

.937*** (.923 - .948)

.0416***

.0408

.0157

.870*** (.853 - .886)

.0458***

.0442

.0220

Sexual offensive behavior

.855*** (.837 - .872)

.0734*

.0715

.0452

.846*** (.828 - .863)

.0497***

.0500

.0277

  1. aThe statistical significance of individual Brier scores was determined by calculating the Spiegelhalter’s Z statistic, with significant values indicating extreme values. Non-significant results indicate better overall performance.
  2. AUC = Area under the receiver operating characteristic curve.
  3. CI = Binomial exact confidence interval.
  4. BS = Brier score.
  5. SMBS = Sanders-modified Brier score (with the data grouped into eleven risk categories).
  6. RIS = Reliability-in-the-small.
  7. *p < .05; **p < .01; ***p < 0.001.