Reliability and Validity Evidences of Tej Emotional and Behavioral Problem Checklist (TEJ-CL) for Child Mental Health Assessment in Nepal.

Introduction: The Tej Emotional and Behavioral Problem Checklist (TEJ-CL) was developed in Nepalese context to aid assessment of childhood emotional and problems of children. This study aimed to evaluate TEJ-CL's factor structure, reliability, and validity evidences as an add-on and symptom monitoring test.

Methods: This cross-sectional validation study included guardians of 320 children (age 5-17 years) from tertiary mental health centers in Kathmandu as referred group, along with 601 children from two schools (private and community) in Kathmandu as non-referred group. IRC was obtained ethical approval (ref: 183 (6-11-E)2/073/074 and ref: 776). TEJ-CL, an 89-item parent-reported questionnaire, served as the index test, while referral status acted as the reference standard. Factor structure, internal consistency, test-retest/cross-informant correlations and criterion validity evidence was assessed using principal component analysis, coefficient alpha, spearman's rank correlation and linear regression models, respectively.

Results: Analysis was done using 179 referred and 412 non-referred individuals based on non-missing data. Principal component analysis in referred sample reduced the number of items of questionnaire to 65 from 89 and indicated six factors: externalizing behavioral issues, anxiety/worries, upset/sadness, somatic concern, miscellaneous syndrome, and severe issues with coefficient alpha ranging from .62 to .95. As criterion validity evidence, referred children showed significantly higher scores than non-referred children across composite and factor scores, except for anxiety/worries factor. Similarly, regression analyses within the referred group demonstrated significant associations between factor scores and specific diagnoses.

Conclusions: Reliability and validity estimate of questionnaire is comparable to similar empirically based scales. Future research should focus on assessing the tool's generalizability and improving discriminatory indexes.