Practice transformation at scale through a microsystems quality improvement (QI) approach.

e18667 Background: How oncology providers should implement practice transformation for value-based care is unclear, particularly at scale. Organizational size enables efficient “top down” approaches, but also presents challenges such as physician engagement. Dis-economies of scale can be acute in oncology due to physician autonomy and coordination costs. We hypothesized that organizational change based in sense-making models that enhance physician engagement and use a decentralized, iterative microsystems approach will enable practice transformation to scale. Methods: Florida Cancer Specialists & Research Institute (FCS) is a physician led 250-oncologist statewide practice, with regional variation in disease state/mix, patient cohort, etc., making a purely top-down approach to organizational change infeasible. FCS prototyped a transformation strategy starting in June 2017 based on sharing interpreted data with physician and executive leadership. Later implementation directly engaged physicians in a microsystems quality QI strategy focused on regional performance. Interventions targeted disease, health service utilization, location, and individual physicians. Performance was evaluated using data from Medicare’s Oncology Care Model (OCM) and assessed using the one-sided risk target (4% below benchmark). We analyzed 70,239 performance period (PP) episodes at FCS across 35,116 patients. Results: In the pre- intervention period (90% of PP1 episodes, completed by June 2017), FCS was 5.8% above target. Performance was 10.9% above target for the remainder of PP1 (10% of PP1 episodes), then improved to 0.3% above target in PP2 and PP3, and below target by 0.9%, 0.8%, and 0.75% in PP4, PP5, and PP6. Early QI efforts focused on performance in lung cancer, which was 2.5% over target in PP1; it improved to 2.1% under target in PP6. Later regional QI sessions targeted cancer, utilization and providers. Pre-intervention, all 18 regions were above target; by PP6, 11 out of 19 regions were below target. Relative to the pre-intervention period, per-episode inpatient costs increased by 12.1% for the remainder of PP1 and increased by 4.3% and 1.3% in PP2 and PP6; inpatient costs decreased in PP3, PP4, and PP5 by 3.8%, 2.4% and 4.8%. Conclusions: Practice transformation in oncology can achieve scale through models of organizational change that foster physician engagement. Data, when clinically contextualized, is a foundational tool in the sense-making process. Scale can develop through an additive microsystems approach in which QI units are de-centralized, accountability is defined, and iteration becomes part of organizational culture. [Table: see text]