The Tradeoff between Access to Community Treatment and Acute Hospitalizations of the Severely Mentally Ill Presented by: Richard C. Lindrooth, Ph.D. University of Colorado at Denver Co-authors: Anouk L. Grubaugh, Ph.D., MUSC Walter Jones, Ph. D, MUSC Anthony Lo Sasso, Ph. D., University of Illinois, Chicago B. Christopher Frueh, Ph. D., University of Hawaii at Hilo Research support: 2 RO1 HS010730-04 (AHRQ) R01 MH074151-01 A2 (NIMH) K24-MH074468 (NIMH PI: Frueh) Primary Research Questions • To what extent are emergency psychiatric admissions to acute care hospitals related to supply of: • State and Acute Beds? • Community alternatives? • To what extent are these admissions preventable? 120 125 Figure 1. Per Capita General Hospital ED Admissions, Selected Diagnoses 105 1997=100 110 115 • Utilization increased over time, especially by the severely ill. • Why has there been an increase? • Does access to community care explain regional variation in emergency psychiatric admissions? 100 • What is the role of acute inpatient care? 1996• Are acute a safety valve 1998 care beds 2000 2002 for community 2004 Calendar Year care? Psychotic Disorders Dementia Source: HCUP-SID in AZ, CO, FL, NJ, NY, WA, & WI, 1997-2005 Mood Disorders 2006 Explanations for the Increase • Increased prevalence of SMI in community • Closure of State Long-term Beds • Decreased access to community alternatives • Supply not keeping up with demand • Increased incidence • Unobserved trends Figure 1. Trends in State Hospital Beds, 1998-2005 CO FL NJ NY WA -600 -200 -800 -400 0 1995 2000 WI -600 -200 -800 -400 0 Change since 1997 -600 -200 -800 -400 0 AZ 1995 2000 2005 Source: American Hospital Association and Medicare Cost Reports Year 2005 1995 2000 2005 Data • AZ, CO, FL, NJ, NY, WA, & WI HCUP-SID 97-05 • Patients admitted through ED with primary dx: • • Comparison Admissions: • • • Psychotic disorders Mood disorders Delirium, dementia, and amnestic and other cognitive disorders Sample includes all admissions from ED to acute care hospitals. The Supply of Health Care Providers Measured spatially based on the patient’s zip code Limit sample to patients less than 45 minutes from ED • Indicate whether the closest ED has a: • Partial Hospitalization Unit or • Psychiatric Emergency Services • Measure the Health Service Area (HSA) supply of: • hospital outpatient, • freestanding outpatient • freestanding MHSA Residential Source: AHA Annual Survey (cleaned and smoothed) and Census of Economic Activity • • Methods • Multivariate Regression • Zip code fixed effects • Sample Year 1997-2005 • T =State bed per capita(1997)-State bed per capita(Time=t) • Adjust for heteroskedastic smearing and clustering by zip-code • Unit of analysis: Zip-code by year • Dependent Variable: Log per capita ED Admissions • d d zip zt = β Market zt + τ t + βT Market zt * Tst + Tst + γ z + ε zt % Change in ED Admissions related to observed bed closures: Public Payer AZ CO* FL* NJ NY* WA* WI Schizophrenia 5% 30% 50% 3% 20% 18% 8% Mood 3% 17% 33% 3% 14% 14% 5% Dementia ns ns ns ns ns ns ns *State closed significant beds between 1998-2001 Magnitude of the effect • Public Schizophrenia: 5.27 admissions per state bed (p<0.01) • Public Mood: 5.31 admissions per state bed (p<0.01) • About 11 admissions per bed closed. • No significant effect for private payors.. Explanations for the Increase • Increased prevalence of SMI in community • Closure of State Long-term Beds • Decreased access to community alternatives • Increased incidence • Unobserved trends Figure 2. Access to Psych EM and Partial Hosp, 1997-2005 CO FL NJ NY WA 0 .2 .4 .6 .8 1995 2000 2005 1995 WI 0 .2 .4 .6 .8 % Population 0 .2 .4 .6 .8 AZ Closest has nothing Closest has PH or EM Closest is Specialty 1995 2000 2005 Source: HCUP-SID & AHA Annual Survey 1997-2005 Calendar Year 2000 2005 % Change in ED Admissions with complete Access to Psychiatric Emergency Services • Public Payer, Schizophrenia • • • • CO: -18% FL: -16% NY: -18% WA: -14% • Public Payer, Mood • • • • CO: -27% FL: -24% NY: -26% WA: -20% • Public Payer, Dementia: Not significant Effect of Access to Partial Hosp. Services on ED Admissions • Public Payers– Schizophrenia and Mood Disorders: • Offsetting effect if psychiatric emergency services exist • Private Payers-- Schizophrenia • Incremental reduction in admissions • Large than for psychiatric emergency services • Not significant for Dementia Access to other services • Hospital-based psychiatric outpatient • Reduced admissions but not significant • Community-based psychiatric outpatient • About 1-3 % increase w/ state bed closure (0.5 bed) • Community-based psychiatric residential • Public schizophrenia and mood disorder patients: • About 1-3 % decrease w/ state bed closure (0.5 bed) • Not significant for private patients Results and Conclusions • Psychiatric emergency consistently reduce admissions through the ED • About 50% of the increase in admissions in the wake of a state bed closure could be prevented • Access to residential treatment facilities also consistently reduces public admissions through ED. • Access to state beds plays a large role • No effect of acute bed capacity for public patients