Predictors of Length of Hospital Stay Using a National Sample of Emerging Adult Patients with Psychotic Disorders

Background and Purpose: 

Schizophrenia is a disabling disorder with typical onset during emerging adulthood.  Emerging adults (EAs) with schizophrenia have impairment in important developmental domains including: residential/financial independence, educational attainment, and romantic relationships. EAs have the lowest rates of insurance coverage and consequently the lowest rates of behavioral health care access. The 2010 Patient Protection Affordable Care Act (ACA) could reduce barriers to behavioral health services for EAs, thereby increasing community care and reducing acute psychiatric care. However, little work has been done to document pre-ACA trends in acute psychiatric use by EAs.  This study will document regional differences in rates of EA acute psychiatric care use, an important behavioral system indicator, in the years 2002 to 2011 in a nationally representative sample. We will also examine the relationship between patient-level characteristics, insurance source, and hospital-level characteristics on the length of hospitalization among EA patients. These analyses will provide important context for understanding how implementation of ACA affects acute psychiatric care use in EAs. 

Methods: 

EA patients, aged 18 – 29, with a primary diagnosis of psychotic disorders (n = 149,521) were drawn from the 2002 – 2011 Nationwide Inpatient Sample (NIS). Generalized estimating equations (GEE) with negative binominal distributions and log links were used to estimate regression models. The dependent variable was length of stay; independent variables included patient-level characteristics (age, sex, race, and problem severity), payer source, and hospital-level characteristics (geographic region, location, control or ownership, teaching status, and bed size).  The nationally representative survey design and census data were used to construct rates of hospitalization for EAs in four U.S. geographic regions.

Results: 

Mean rate of hospitalization was 140 per 100,000 age-adjusted population. Rates differed dramatically by region with the Northeast having a rate of 219 per 100,000, while the Midwest, South and West had rates of 167, 112, and 122 per 100,000, respectively. Longer hospital stay was associated with being Asian and Native American versus white; higher versus lower severity of illness; and use of Medicaid and Medicare versus private insurance. Hospital characteristics associated with longer hospital stay included larger versus smaller bed size, urban versus rural location, Northeast region versus all other regions, and public versus private hospital ownership. 

Conclusions and Implications: 

Findings suggest strong regional and payer source differences in the use of hospitalization to treat mental illness. Particularly interesting are the higher rates and longer lengths of hospitalizations among EAs in the Northeast. These findings have implications for social work policy and practice.  With ACA implementation and expanded Medicaid, social workers may see increased acute care use for psychiatric disorders among EAs. Policy considerations include responding to regional differences in length of stay.  Future research should investigate the appropriateness of acute care service use from an overuse (Northeast) and underuse (South) perspective.  Furthermore, public payers and hospitals had longer hospital stay in these restrictive settings, suggesting either underfunding of public community care resources or excessive cost cutting by private providers.  These analyses raise important questions for understanding how the ACA affects EAs’ acute care use.