How to Assess the Impact of Quality and Patient Safety Interventions with Routinely Collected Longitudinal Data

Abstract

Measuring the effect of patient safety improvement efforts is needed to determine their value but is difficult due to the inherent complexities of hospital operations. In this paper, we show by case study how interrupted time series design can be used to isolate and measure the impact of interventions while accounting for confounders often present in complex health delivery systems. We searched for time-stamped data from electronic medical records and operating room information systems associated with perioperative patient flow in a large, urban, academic hospital in Baltimore, Maryland. We limited the searched to those adult cases performed between January 2015 and March 2017. We used segmented regression and Box-Jenkins methods to measure the effect of perioperative throughput improvement efforts and account for the loss of high volume surgeons, surgical volume, and occupancy. We identified a significant decline of operating room exit delays of about 50%, achieved in 6 months and sustained over 14 months. By longitudinal assessment of intervention effects, rather than cross-sectional comparison, our measurement tool estimated and provided inferences of change-points over time while taking into account the magnitude of other latent systems factors.