Threshold 2: the reliability, validity and sensitivity to change of the Threshold Assessment Grid (TAG).

Objective: This study investigated the psychometric properties of the Threshold Assessment Grid (TAG), a new assessment of the severity of mental health problems.

Method: A total of 605 patients were recruited from 10 mental health adult and elderly services in London, UK. TAG ratings and other standardized definitions of severe mental illness were completed by referrers. TAG, Global Assessment of Functioning (GAF), Camberwell Assessment of Need Short Appraisal Schedule (CANSAS) and Health of the Nation Outcome Scale (HoNOS) ratings were completed by mental health service staff. Construct validation on extreme groups was investigated.

Results: Construct and concurrent validity were good. Referrer TAG scores predicted mental health team view of referral suitability, but not whether assessments were offered. Test-retest reliability was good, interrater reliability ranged from good to poor in different domains (but adequate for total TAG score), internal consistency was appropriate. Sensitivity to change requires further investigation.

Conclusion: The TAG can be recommended for use by all agencies when making referrals to mental health services.