Volume 13, Issue 3 p. 140-147

Screening Efficiency of the Child Behavior Checklist and Strengths and Difficulties Questionnaire: A Systematic Review

Erin M. Warnick

Erin M. Warnick

Department of Epidemiology and Public Health, Yale University, USA. E-mail: [email protected]

Child Study Center, Yale University, USA.

Search for more papers by this author
Michael B. Bracken

Michael B. Bracken

Department of Epidemiology and Public Health, Yale University, USA. E-mail: [email protected]

Center for Perinatal, Pediatric and Environmental Epidemiology, Yale University, USA

Search for more papers by this author
Stanislav Kasl

Stanislav Kasl

Department of Epidemiology and Public Health, Yale University, USA. E-mail: [email protected]

Search for more papers by this author
First published: 11 August 2008
Citations: 183

Abstract

Objective: Assess the screening efficiency of the caretaker-report CBCL and SDQ in community and clinical samples using published data.

Methods: PyschInfo, Medline, and EMBASE were systematically searched to identify studies with appropriate efficiency data. Estimates of sensitivity and specificity were extracted from identified studies and used to generate summary likelihood ratio estimates on which the scales were compared. Summary estimates of sensitivity and specificity were calculated with respect to a ‘true’ diagnosis to compare scales.

Results: A total of 29 and 3 studies met inclusion criteria for CBCL and SDQ respectively. Summary estimates of the likelihood ratios for domains assessed by CBCL ranged from 3.86 (2.23, 6.69) to 4.87 (2.90, 8.18); and for SDQ from 5.02 (1.61, 15.63) to 8.32 (2.72, 25.48). Heterogeneity was low. For total problems, the SDQ caretaker-report was found to be most specific (0.93, 95% CI 0.92, 0.94) and the CBCL caretaker-report to be most sensitive (0.66, 95%CI 0.60, 0.73).

Conclusions: This meta-analysis supports continued use of the CBCL and SDQ via caretaker-report in clinical and community samples. Additional research is required to determine if there is a true difference in efficiency between the two scales.