Concordance between parent reports of children's mental health services and service records: The Services Assessment for Children and Adolescents (SACA) Article

Hoagwood, K, Horwitz, S, Stiffman, A et al. (2000). Concordance between parent reports of children's mental health services and service records: The Services Assessment for Children and Adolescents (SACA) . JOURNAL OF CHILD AND FAMILY STUDIES, 9(3), 315-331. 10.1023/A:1026492423273

cited authors

  • Hoagwood, K; Horwitz, S; Stiffman, A; Weisz, J; Bean, D; Rae, D; Compton, W; Cottier, L; Bickman, L; Leaf, P

authors

abstract

  • The concordance between parent reports of children's mental health services and medical and administrative service records were assessed in a field test of the Services Assessment for Children and Adolescents (SACA) interview instrument. Service use reports from primary caregivers, usually mothers, for their child's emotional or behavioral problems were compared against inpatient, outpatient, and school records in St. Louis, one of the pilot sites for the Multi-Site Study of Service Use, Need, Outcomes and Costs in Child and Adolescent Populations (UNOCCAP). A global "any use" service variable, comprised of inpatient, outpatient, and school reports, yielded an overall service use concordance kappa of .76 between parent reports and records. Parent reports of inpatient hospitalization services using the SACA yielded the highest agreement with medical records, with kappa statistics of 1.00 for use of any inpatient hospital care and for medication use. Parent reports of specific inpatient services concurred with medical records more moderately, yielding kappas from .50 to .66. Reports of any outpatient mental health services yielded variable rates of agreement, with kappas ranging from .67 for any use of outpatient care, to .66 for medication use, to negligible kappas for specific treatments. Parent reports of school services were weakly related to records for most services, except for moderate agreement (.48) on placement in special classrooms for emotional or behavioral problems. Family burden or impact discriminated more powerfully than other variables between respondents who concurred with records and those who did not. © 2000 Human Sciences Press, Inc.

publication date

  • January 1, 2000

published in

Digital Object Identifier (DOI)

start page

  • 315

end page

  • 331

volume

  • 9

issue

  • 3