Date of Award

Spring 5-2020

Degree Name

Doctor of Philosophy (PhD)

Advisor(s)

PAUL ROWAN, PHD

Second Advisor

STEPHEN H LINDER PHD

Third Advisor

MICHAEL D SWARTZ, PHD

Abstract

Cancer care is changing rapidly. Understanding of the increasing subtypes of cancer and exponentially increasing therapeutic interventions are unprecedented due to the rapid pace of scientific discovery and clinical innovation. This immense change within the field, lends itself to quality control initiatives, especially among general oncology providers who see a wide array of cancer types as general oncologists will see many different tumor types, and most of which have several potential treatment choices that have grown over time. Evidence-based pathways are an effective way to nudge quality control by presenting choice architecture at the point of care to facilitate guideline compliance among a wide array of therapeutic choices. This evaluated the impact of a clinical decision support system (CDSS) tool, a "nudge" within the electronic health record among a network of oncology providers. This study examined the results of its implementation across 9 statewide practices over 6-month interval. We evaluated the effects of the CDSS on regimen compliance with value pathways across practices, within practices, and the influence on physician compliance with value pathways across the study interval. SAS 9.4 software was used to evaluate the hypothesis using multi level modeling. Across the 29,926 regimens included in the study, the CDSS tool significantly impacted compliance to evidence based pathways. By applying a multi-level logistic regression model to the entire cohort, and segregating the levels as patients as level 1, doctors as level 2, and practices as level 3, the post CDSS implementation odds ratio of compliance to evidence based pathways was 1.48 (1.25;1.76). When we segmented the cohort by practices, the majority of individual practices had a significantly higher likelihood of evidence based pathways compliance after implementation of the CDSS tool with odds ratios of 1.60 (1.33;1.94), 1.13 (0.88; 1.45), 1.39 (1.08; 1.79), 1.85 (1.53; 2.24), 1.76 (1.32; 2.36), 1.71 (1.38; 2.11), 1.23 (0.96; 1.57), 1.37 (1.12; 1.67) and 1.46 (1.30; 1.63). In addition, each oncologist’s compliance was evaluated and, while we did not demonstrate a statistically significant improvement in compliance with the limited number of evidence based pathways prescribed by each oncologist with implementation of the tool, the number of regimens by oncologist was very low. Using McNemar’s test we did find that the percentage of oncologists who reached an individual benchmark of 75% compliance was significantly higher with implementation of the CDSS tool: among the 560 physicians included in this study, 327 (58%) were at or above a benchmark of 75% compliance prior to the CDSS tool and 402 (72%) achieved that benchmark after implementation of the CDSS tool (p<0.001). In conclusion, implementation of the CDSS tool can be a successful mechanism to increase compliance to evidence-based pathways overall, and within most individual practices. In addition, physician compliance to benchmark performance of 75% compliance with evidence-based pathways can be improved by implementing a CDSS tool embedded within the EHR.

Share

COinS