How to assess the impact of process improvement interventions with routinely collected longitudinal hospital data

Diego Martinez, Mehdi Jalalpour, David Efron, Scott Levin

Research output: Contribution to journalArticle


Objective: To show how interrupted time series can be used to isolate and measure the impact of process improvement while accounting for confounders often present in complex hospital operations. Methods: Retrospective cohort study comparing the volume of operating room exit delays (OR holds) 52 weeks before and 62 weeks after implementation of a surgical patient throughput optimization program at a tertiary academic hospital. Time-stamped electronic medical records data were collected and analyzed using interrupted time-series design. Segmented regression and Box-Jenkins time series analysis were used to measure OR hold volume pre- and post-implementation, controlling for secular trends in surgical volume, downstream capacity, and the loss of high-volume surgeons. Results: A total of 8,983 surgical patients were discharged during the pre-intervention period and 9,855 during the post-intervention period. The median weekly discharge volume pre-intervention was 175 (interquartile range [IQR] 164–180), and the median bed occupancy was 86% (IQR 84.6–88.1%). The median weekly discharge volume post-intervention was 163 (IQR 150.5–169.8), and the median bed occupancy was 82.1% (IQR 78.9–84.7%). Post-intervention, there was an immediate 60% (95% confidence interval, 54–70%) reduction in the number of OR holds that was sustained over the 14-month post-intervention period.

Original languageEnglish (US)
JournalIISE Transactions on Healthcare Systems Engineering
Publication statusAccepted/In press - Jan 1 2019



  • Care delivery process
  • healthcare management
  • length of stay
  • patient flow

ASJC Scopus subject areas

  • Safety, Risk, Reliability and Quality
  • Safety Research
  • Public Health, Environmental and Occupational Health

Cite this