facilities have better June 27, 2017 outcomes? - PowerPoint PPT Presentation

Do assisted-living residents Portia Y. Cornell, M.S.P.H., Ph.D. Academy Health ARM in dementia-licensed New Orleans, LA facilities have better June 27, 2017 outcomes? Portia_Cornell@brown.edu

Specialized dementia-care units: licensing requirements  Some states issue special license for dementia care  Licensing requirements differ by state  Secured to prevent wandering  Training requirements  Staff-to-resident ratios 2

Conceptual model for specialized- care licensing  Information asymmetries between prospective resident and assisted-living community  Dementia-specialized- care license is a “signal” of underlying capacity and quality of a facility (Spence 1971)  Key insight: cost of licensure is lower for higher-quality facilities  Limitation: prospective don’t update their beliefs from experience 3

Conceptual model: information asymmetries Prospective resident AL Community Informatio n Observation • • Reputation Licensing •  Licensing requirements directly impact patient outcomes  License is a signal of deeper specialization & capacity 4

What is the effect of living in a dementia-specialized AL?  Does living in a dementia-specialized assisted-living facility, compared to a non-specialized facility, affect resident outcomes and health care use?  Outcomes:  Mortality  Hospice use  Emergency department visits  Inpatient hospital days 5

Approach  Linear model effect of residing in dementia-licensed AL  Instrumental variable analysis  Stratify by dual status and state 6

Instrument definition  Log-distance difference:  D1 = distance to nearest non-specialized AL community  D2 = distance to nearest dementia-specialized AL community  lnDD = ln(D1) – ln(D2) Regular assisted-living Dementia-licensed residence assisted-living residence With increasing lnDD, more likely to move to a dementia- licensed residence D1 D2 D1 D2 lnDD<0 lnDD>0 7

Sample  Medicare fee-for-service beneficiaries with validated residence in an assisted-living community in 2014 in the study states:  Alabama, Colorado, Mississippi, New York  Diagnosis of dementia in 2011, 2012, or 2013  Previous residence in 2011 or 2012 NOT in assisted living community State N Pct AL 498 8.71 CO 1,027 17.95 MS 283 4.95 NY 3,912 68.39 Total 5,720 8

Summary statistics Non-dementia- Dementia- licensed licensed Total Age 80.56 83.90 81.24 Female 68% 70% 69% Race=Black 8% 3% 7% Race=Other non-white 3% 1% 2% Dual Medicare/Medicaid 43% 14% 37% Inpatient days 2.49 1.11 2.21 Skilled nursing days 12.22 7.97 11.36 ED visits 1.01 0.63 0.93 Heart problems 63% 59% 62% Stroke/TIA 11% 11% 11% Kidney disease 29% 25% 29% Hip fracture 5% 5% 5% County: pct in poverty 15.60 14.19 15.32 9 County: median hh income 57026 60562 57740

Instrument strength  Probability of dementia-licensed AL increases with instrument  F-statistics >10 10

Covariate Balance Standardized Difference Change from Treatment to Instrument median_hh_inc poverty_pct hipfracture kidney stroke heart er_visits snf_days inpatient_days oth_race black female Age>=80 Age<80 Age<70 Age<60 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 Duals Instrument Duals Treatment Non-duals Instrument Non-duals Treatment 11

Outcomes Mean outcomes among study sample Duals(N=2114) Non-Duals (N=3606) Mortality (died in 2013) 0.14 0.13 Hospice (any) 0.06 0.08 Emergency Dept Visits† 0.10 0.09 Hospital Inpatient Days† 0.23 0.11 Skilled Nursing Days† 1.25 1.24 † Adjusted for observed months 12

Main Results: IV Estimates Effect of living in a dementia-specialized assisted-living residence compared to non-specialized Duals Non-Duals (N=2114) (N=3606) Mortality (died in 2013) 0.35*** 0.04 Hospice (any) 0.28*** 0.07** Emergency Dept Visits† 0.06 0.01 Hospital Inpatient Days† 0.34 0.09 Skilled Nursing Days† 0.29 -0.10 *p<0.10, **p<0.05, ***p<0.01 † Adjusted for observed months 13

IV vs. OLS results: mortality (duals) 14

IV vs. OLS results: mortality (non-duals) 15

IV vs. OLS results: hospice (duals) 16

IV vs. OLS results: hospice (non-duals) 17

Interpreting Results  Dementia-specialized communities increase mortality?  Inefficient allocation of resources  Value independence and quality of life over safety and prolonging life?  Increase use of Medicare hospice benefit suggests different priorities  Geographic differences?  E.g.: frail individuals may choose assisted living over a nursing-home special-care-unit if there is a dementia-specialized facility nearby 18

Caveats, limitations, and hedges  IV addresses patient-level selection, but NOT geographic selection  Within a dementia-licensed community with both specialized and non-specialized care, cannot identify who lives in the dementia unit 19

Thank You!  Kali S. Thomas & Momotazur Rahman, co-authors  National Institute on Aging grant (R21 AG047303; PI, K.S. Thomas)  Veterans Administration OAA Advanced Fellowship  Contact: Portia_Cornell@brown.edu 20

Selected References Carder P, O’Keeffe J, O’Keeffe C. Compendium of Residential Care and Assisted Living Regulations and Policy: 2015 Edition. 2015. Prepared by RTI International for the Office of Disability, Aging, and Long-Term Care Policy, Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services. Harris-Kojetin L, Sengupta M, Park-Lee E, et al. 2016. Long-term care providers and service users in the United States: Data from the National Study of Long-term Care Providers, 2013-14. National Center for Health Statistics. Vital Health Stat 3(38). Spence M. 1973. Job Market Signaling. Quarterly Journal of Economics 87(3):355-374. Thomas K, Dosa D, Gozalo PL, Grabowski DC, Nazareno J, Makineni R, Mor V . A methodology to Identify a Cohort of Medicare Beneficiaries Residing in Large Assisted Living Facilities Using Administrative Data 21

Instrument balance: propensity score Dementia-licensed assisted living propensity score vs. lnDD AL CO .4 .4 .2 .2 0 0 -6 -4 -2 0 2 -6 -4 -2 0 2 4 MS NY .4 .4 .2 .2 0 0 -5 0 5 -6 -4 -2 0 2 4 Propensity to move to dementia-specialized AL 95% CI Smoothed average

Main Results: IV Estimates Effect of living in a dementia-specialized assisted-living residence compared to non-specialized Duals Non-Duals (N=2114) Pct mean (N=3606) Pct mean Mortality (died in 2013) 0.35*** 250% 0.04 34% Hospice (any) 0.28*** 496% 0.07** 85% Emergency Dept Visits† 0.06 57% 0.01 10% Hospital Inpatient Days† 0.34 150% 0.09 78% Skilled Nursing Days† 0.29 23% -0.10 -8% *p<0.10, **p<0.05, ***p<0.01 † Adjusted for observed months 23