Send to

Choose Destination
See comment in PubMed Commons below

The hospital as predictor of children's and adolescents' length of stay.

Author information

  • 1Department of Psychology, Loyola University Chicago, IL 60626, USA.



To predict psychiatric hospital length of stay (LOS) for a sample of Illinois Department of Children and Family Services wards across 4 fiscal years.


A prospective design was implemented using the Children's Severity of Psychiatric Illness scale, a reliable and valid measure of psychiatric severity, risk factors, youth strengths, and contextual/environmental factors. Data were collected for 1,930 hospital episodes across 44 hospitals from fiscal year 1998 through fiscal year 2001. Youths were screened for admission appropriateness by the Illinois Screening, Assessment, and Supportive Services (SASS) program. The Children's Severity of Psychiatric Illness was completed by SASS workers upon conclusion of their crisis interviews. In addition to completing the Children's Severity of Psychiatric Illness, SASS workers reported on demographic information and LOS.


The sample of 1,930 youths was randomly split to form development (n = 983) and validation (n = 947) samples. LOS was predicted using ordinary least squares regression. Thirty percent of the variance (F(19,666) = 16.6, p < .0001) in LOS was predicted for the development sample and 22% (F(14,657) = 14.6, p < .0001) was predicted for the confirmation sample. Hospital was the largest and most consistent predictor of LOS for both samples after controlling for clinical variables. Two hospitals accounted for approximately 10% of the variance in both samples (development beta = .273, p < .01 and beta = -.169, p < .01). Two SASS agencies also consistently predicted LOS (development beta = -.134, p < .05 and beta = .102, p < .05). No consistent changes in predictors of LOS occurred over time (FY98-FY01).


These findings suggest that nonclinical variables are the primary predictors of LOS in the Illinois system of care. In addition, these variables are consistent predictors over time. Quality assurance efforts might seek to further understand potential practice pattern variations across hospitals and SASS agencies.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Elsevier Science
    Loading ...
    Write to the Help Desk