Inpatient Psychiatric Care for Children and Youth: Predictors and Precipitants of Admission & Readmission

Purpose: Since the introduction of managed care into Medicaid in the early 1990s, lengths of stay for inpatient treatment have steadily declined. Some states have documented an increase in readmission rates for children under Medicaid managed care (Callahan, Shepherd, Beinecke, Larson, & Cavanaugh, 1995; Dickey, Normand, Norton, Rupp, & Azeni, 2001; Fontanella, Zuravin, & Burry, 2006). This trend has elevated concerns that shorter stays are associated with readmissions and thereby create a “revolving door.” To analyze factors associated with inpatient admissions and readmissions, authors present data from a statewide study of child Medicaid beneficiaries in one Midwestern state.

 Methods: The study used a concurrent multiple methods design. The study population comprised all children, 3-17 years old, enrolled in Medicaid during SFY 2009 (N =178,558). Quantitative analyses tested the relationship between demographic (e.g., gender, race, age, region, population density, foster care status) diagnostic (e.g., mood, disruptive, anxiety, ASD, psychotic, or other diagnosis) and service variables (e.g., prior admission, mental health service prior 30 days, polypharmacy in prior year) on admissions and re-admissions. Data sources included six state administrative databases. Multivariate logistic regression was used to analyze admission. Cox regression was used to analyze readmission. To determine precipitants of hospitalization among children aged 3-5, in-depth case record reviews were undertaken on a subpopulation of very young children hospitalized during the study period.

 Results: Among the population of child Medicaid beneficiaries, 1,912 unique children were admitted one or more times to inpatient psychiatric care in FY09 (acute and private residential treatment centers, or PRTFs). Compared to the previous year, admissions increased 7.7%, from FY08 to FY09. This increase was consistent with population trends. Children 6-8 years old (OR = 2.41, p < .001) and 9-11 years old (OR = 2.15, p < .001) were more than twice as likely to be admitted as other children. Children with a mood disorder (OR = 21.88, p < .001), a disruptive disorder (OR = 6.74, p < .001), or a psychotic disorder (OR = 6.25, p < .001) were also more likely to be admitted. Of the 1,912 children admitted to care in the study period, 43.4% were readmitted within 60 days. The PRTF readmission rate was 26.5%; the acute inpatient care rate was 50.4%. Most children who readmitted did so within 60 days. Age; mood disorder/ASD diagnoses; and prior services/hospitalization were the strongest predictors of inpatient acute readmissions. For example, young children (3-5) were significantly more likely to be readmitted from acute inpatient care than children 6-8 years old and older teens, 15-17 years old. Gender, disruptive disorder diagnosis, and region most strongly predicted PRTF readmission.

Implications: Readmission rates were high compared to other states. Additionally, the majority of children who readmitted from acute inpatient care facilities did so within 60 days of their index admission date. While not conclusive, findings suggest the plausibility of a revolving door phenomenon. Policy and practice implications of demographic, diagnostic, and service variables will be discussed, particularly with regard to very young children.