“…Covariates that have been shown to influence outcomes and included in regression models were: the time from the fracture to the rehabilitation admission, admission motor FIM rating, 22 admission cognitive FIM rating, 23,24 type of hip fracture repair (hip replacement or open/closed reduction with/without internal fixation), 25 presence of a tier 1, tier 2, or tier 3 comorbidity, 26 age, 2 prehospital living arrangement (alone or not alone), 27 median household income (assigned based on residential ZIP code), 28 sex, 29 race (white or nonwhite), 30 number of patients with hip fracture treated in the facility (ie, volume), 31 county managed care penetration percent (from the Area Resource File), 32,33 rural location, 34 facility profit status (profit or not-for-profit), facility type (facility in a hospital or freestanding facility), 26 and geographic location (New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central, West South Central, Mountain or Pacific). 8 The outcomes models were estimated with and without LOS adjustment (transformed by its logarithm) to estimate the extent to which the observed outcomes may be related to LOS differences.…”