2002 OPEN FORUM Abstracts
THE VALUE OF AN EFFICIENCY SCORE IN EVALUATING THE MECHANICAL VENTILATION PROCESS
C Haas MLS, RRT;
R Dechert MS, RRT; M Konkle MPA, RRT; L Folk RRT; P Loik RRT; L Stapp RRT; V
Stevenson RRT; A Andrews MS, RRT; C Lane RRT; S Lawrence RRT; L Young RRT; A
Benschoter RRT; K Marecle RRT; S Mack; S Gay MD
Respiratory Care Department, University of Michigan Health System, Ann Arbor
INTRODUCTION: A weakness of many studies reporting reduced duration of ventilation (DOV) is that patient severity is often not considered in the comparison. Although DOV is reduced, it may be related to patient selection, not an improved system. When severity-adjusting a predicted DOV (P-DOV) as a benchmark to compare actual DOV (A-DOV), it is desirable to use a predictive model which mathematically controls for physiologic variables. Such a tool may be useful assessing the impact of process changes, such as a new protocol. We convened an adult ventilator management performance improvement (PI) team in 1999. Following 12 months of data collection, 6 months were given to standardize a weaning protocol (across 6 ICUs) and to focus attention on many facets of the ventilation process. Following another 12 months of data collection, reduced times were observed in many segments of the ventilation process, including DOV. PI methods and preliminary results were previously reported (Resp Care 2001;46:1129). We wanted to determine whether the differences observed between the two periods were due to severity of patient illness or to an improved efficiency. Our study question was ?Are the observed differences in pre vs post DOV a result of patient selection??
Methods: A-DOV was computed from data collection sheets completed by the bedside RCP. Following IRB approval, P-DOV was obtained for the surgical ICU from an APACHE III data-base and the files merged. We also calculated a ventilatory management efficiency (EffVM) ratio (EffVM =A-DOV/P-DOV); the lower the ratio, the better the efficiency. For comparisons, p<0.05 (Student t) was considered significant.
Results: The table shows median, mean, and SD values of pre and post implementation periods for P-DOV, A-DOV, and EffVM. Although there was no difference between pre and post P-DOV or EffVM, post A-DOV was < pre A-DOV and A-DOV was < P-DOV during both periods.
|P-DOV (d)||A-DOV (d)||EffVM||p#|
|Pre vs Post||p*||0.283||0.012||0.286|
* comparing mean Pre vs Post; # comparing P-DOV vs A-DOV
Conclusions: The similar P-DOV for the pre vs post groups suggests patient similarity during both study periods, and that patient physiology was not the primary reason for the observed reduction in A-DOV. Although the efficiency of ventilation trended toward improvement over time, A-DOV was shorter than the P-DOV benchmark during the pre-period as well, suggesting an efficient system to begin with (A-DOV was only 72% of P-DOV in pre vs 62% in post period). Such an evaluation will be conducted on other populations. Because the EffVM ratio is adjusted for patient severity it may be useful in comparing patients within and between institutions.