Evaluating Quality Improvement Interventions: Strengthening Causal Inference with Observational Data
- Author(s): Prasad, Priya April
- Advisor(s): Gonzales, Ralph
- et al.
As innovations in healthcare delivery systems and electronic medical records (EMR) data capture develop, the methods employed to evaluate interventions disseminated to improve the quality and efficiency of patient care should evolve as well. Through my dissertation work I have explored methods and strategies to employ when developing a responsible healthcare quality improvement (QI) evaluation in three separate domains.
The first chapter of my dissertation focuses on assessing the effect of an intervention to identify and manage sepsis through a retrospective cohort study at the University of California San Francisco (UCSF) Medical Center. The analysis revealed that the UCSF sepsis bundle was associated with a 31% decreased risk of in-hospital mortality across hospital units (adjusted incidence rate ratio 0.69, 95% Confidence Interval (CI) 0.53, 0.91) and the adjusted number needed to treat was 15 (CI 8,69). In the second chapter of my dissertation, I present a review of the causal inference framework in the setting of interrupted time-series analysis and apply it to the formation of an analysis plan for a multidisciplinary intervention implemented at UCSF Medical Center to decrease the use of packed red blood cell transfusions. The third chapter of this work focuses on developing a metric to measure timely access to ambulatory specialty care in a cohort of nearly 60,000 UCSF primary care patients. I explored associations between population-level weekly ambulatory specialty care access defined three ways and the rate of population-level weekly poor health outcomes. Based on unadjusted Poisson analysis, there were correlations identified between poor outcomes and timely access to care in some specialties and the results provide a springboard for future exploration of metric performance and adjustment.
As a body of work, my dissertation illustrates the breadth of the field of healthcare QI, providing evidence to support the continued evolution of robust methods for evaluation of interventions that will improve the quality, safety, and value of healthcare delivered nationally and globally.