1033 — Longitudinal Changes in Observed Minus Expected Length of Stay (OMELOS) Associated with the FY07 Flow Improvement Inpatient Initiative (FIX)
Glasgow JM (Center for Research in the Implementation of Innovative Strategies in Practice (CRIISP), Iowa City VAMC), Kaboli PK
(Center for Research in the Implementation of Innovative Strategies in Practice (CRIISP), Iowa City VAMC)
Quality improvement and system redesign activities are increasingly common in healthcare. A prominent VA example was the Flow Improvement Inpatient Initiative (FIX), a nationwide collaborative to help facilities reduce bottlenecks, delays, waste, and errors associated with patient encounters. Our objective is to understand the impact of the FY07 FIX collaborative on medical inpatient length of stay.
Continuous piecewise linear regression of 5 years of discharge records (FY05 – FY09) from 126 VA facilities including 2,255,568 discharges. The main outcome measure was observed minus expected length of stay (OMELOS) based on a FY07 risk adjustment. Analyses focused on identifying the pre-FIX trend in OMELOS (FY05 – FY06) at each facility and then determining whether there was any change in the facility OMELOS during (FY07) or post-FIX (FY08-FY09).
Mean LOS and (OMELOS) in days for each of the successive five years was 5.40(1.73), 5.31(1.62), 5.21(1.52), 5.44(1.83), and 5.32(1.70). A piecewise regression model for OMELOS showed an improvement in OMELOS of 0.17 days/year prior to FIX, a worsening by 0.25 days/year during FIX, and once again an improvement of 0.04 days/year post-FIX (p < 0.001 for all). When evaluating individual facility performance, results varied greatly: 7 facilities (5.5%) had an OMELOS that was negative or statistically equivalent to zero at the beginning of FY05 and exhibited very small changes during the study period. Of the remaining 119 facilities, 36 (30%) showed statistical improvements in their OMELOS either during or after FIX, while an additional 57 (48%) maintained their OMELOS over the 5-year study period. The final 26 (22%) had an increasing OMELOS throughout the study.
This analysis of FY07 FIX suggests that success with the initiative varied extensively across the participating facilities. While some facilities showed an ability to make efficiency improvements, others either could not achieve or maintain improvements.
FIX is a complex program that likely cannot be evaluated by a single metric. However, this analysis shows that not all facilities respond the same to a national initiative. Future research should examine how facility characteristics influence their ability to achieve and maintain improvements, as well as evaluate the potential benefits of targeting outliers for quality initiatives.