Evaluation of Competence to Stand Trial–Revised (ECST–R)

Description and Development

Prototypical analysis with competency experts identified core representative items for three ECST–R competency scales: Factual Understanding of the Courtroom Proceedings (FAC), Rational Understanding of the Courtroom Proceedings (RAC), and Consult with Counsel (CWC). Prototypical items were also evaluated by trial judges and highly experienced forensic psychiatrists. In addition to competency scales, the ECST–R uses multiple detection strategies for validating its four Atypical Presentation (ATP) Scales: Psychotic (ATP–P), Nonpsychotic (ATP–N), Both (ATP–B; sum of ATP–P and ATP–N), and Impairment (ATP–I). A fifth ATP scale is not used to assess feigning, but masks the intent of the other ATP scales: Realistic (ATP–R).

Samples from four major studies were combined for normative data and test validation. Three samples consisted of mentally disordered offenders: (1) 100 detainees on a psychiatric unit of a large metropolitan jail, (2) competency cases including 28 pretrial evaluations and 42 inpatients in competency restoration, and (3) 56 inpatients in competency restoration. They were supplemented with 95 jail inmates and 89 additional competency referrals.

Reliability

Internal consistencies (alpha coefficients) were high for overall competency (.93) and the individual competency scales: FAC (.87), RAC (.89), and CWC (.83). Interrater reliabilities were exceptional for these competency scales: FAC (.96), RAC (.91), and CWC (.91). Even when focusing on individual competency items, interrater reliabilities remained strong (Mr = .77). In addition, most individual competency ratings remained stable across a 1-week interval with more than 90% of the ratings remaining identical. ATP scales had moderate to high internal consistencies (alphas from .70 to .87) with the exception of ATP–R (.63), which serves a different purpose. ATP scales have outstanding interrater reliabilities (rs from .98 to 1.00)

Validity

The previously described prototypical analysis provided strong evidence of content validity. For construct validity, confirmatory factor analyses were used to test various models of the Dusky prongs for the ECST–R competency items. With one cross-loading, the confirmatory factor analyses confirmed a threefactor model with high loadings (M = .72) for competency items on their designated scales.

For criterion-related validity, ECST–R competency scales demonstrated a high concordance with independent opinions of experienced forensic experts and legal outcomes. It also evidenced moderate correlations with the MacCAT–CA, despite major conceptual differences between the two competency measures. In addition, very large effect sizes were found between defendants with and without impairment on ECST–R competency scales for both the severity of psychotic symptoms (Cohen's ds from 1.95 to 2.98) and overall functioning (ds from 1.60 to 1.75).

The ATP scales were validated using a combination of known-group comparisons and simulation designs. Both suspected malingerers and simulators produced much higher scores than genuine inpatients on the ATP scales with the logical exception of ATP–R. Very large effect sizes were found for both suspected malingerers (Md = 1.99) and simulators (Md = 1.74).

...