Health Services Research
Measurement
Oral Presentation
Chih-Ying Li, PhD, OTR
Assistant Professor
University of Texas Medical Branch
Galveston, Texas
Amol Karmarkar, PhD
Professor
Virginia Commonwealth University (VCU) & RTI International, Virginia
Allen Haas, MS
Statistician
University of Texas Medical Branch
Galveston, Texas
Yong-Fang Kuo, PhD
Professor
University of Texas Medical Branch
Galveston, Texas
Kenneth Ottenbacher, PhD, OTR, FACRM
Professor, Director
University of Texas Medical Branch
Galveston, Texas
Objective : To examine the associations between co-calibrated self-care and mobility scores and subsequent hospital readmission.
Design : We conducted secondary data analysis using Medicare claims data (2013-2014). Functional scores (self-care and mobility) at admission were co-calibrated into a 0-100 scale across post-acute care (PAC) settings and categorized into four quartiles (Q1-Q4). Survival analyses were used to examine the trends of hospital readmission based on the admission functional scores at the initial PAC setting. Hazard Ratios (HRs) of the subsequent hospital readmission were estimated based on the co-calibrated admission functional scores at the initial PAC setting, adjusting for PAC utilization and patient-level covariates (socio-demographics, condition severity, acute length of stay, Elixhauser comorbidities, disability, chronic condition, dual eligibility).
Setting : Inpatient rehabilitation facilities (IRF), skilled nursing facilities (SNF) and home health agencies (HHA).
Participants (or Animals, Specimens, Cadavers):
Medicare beneficiaries aged 66 or older on Fee-For-Service plan and were initially discharged to one of three PAC settings (IRF, SNF or HHA) after acute hospital stay (three primary impairment conditions: stroke=143,277, total hip [THA] and total knee [TKA] arthroplasty=512,577, and hip/femur fracture=125,167).
Interventions : None
Main Outcome Measure(s) : Ninety-day hospital readmission after acute hospital discharge.
Results : For all three primary impairment conditions, higher admission self-care and mobility function were associated with lower risks of the subsequent hospital readmissions (e.g., for self-care in stroke, compared to Q1 (the lowest quarter of self-care function), risks of the subsequent hospital readmissions decreased gradually from Q2, Q3 to Q4 (HRs= 076 [0.74-0.79] for Q2, HRs= 0.72 [0.69-0.75] for Q3 and HRs= 0.64 [0.61-0.66] for Q4). Survival analyses showed four distinct patterns of hospital readmission based on admission co-calibrated self-care and mobility scores across all impairment conditions, except for elective THA and TKA (three instead of four readmission patterns based on admission mobility scores for both).
Conclusions : Variations in the subsequent hospital readmissions were attributable to co-calibrated admission functional score at the initial PAC settings. High self-care and mobility scores at admission were less likely to have subsequent hospital readmissions across three included impairment conditions. We also found current functional scores were not sensitive enough to distinguish patients’ mobility status into four levels for the THA/TKA conditions. Practitioners at PAC settings should enhance self-care and mobility function to prevent future preventable hospital readmission.