Variation in adherence to QOPI core measures based on tumor type.

Abstract

280 Background: QOPI measures fall into one of three categories: (1) core, (2) disease-specific (breast, colon, lung, non-Hodgkin lymphoma, gynecologic cancers), or (3) domain-specific (symptom control, end-of-life). For each data collection period (DCP), participating sites choose to submit data in at least one disease- or domain-specific module. Charts are identified and abstracted for the selected module(s) based on eligibility criteria. The same charts are also abstracted for core measures. Our group hypothesized that the case mix resulting from choice of module(s) would impact performance for a subset of core measures. Methods: The MD Anderson Regional Care Centers have participated in QOPI over nine DCPs from Fall 2009 to Spring 2014. Unexplained variation was identified in staging documentation (core measure 2) and chemotherapy intent documentation (core measure 10). For each DCP, QOPI chart-level data were reviewed. Adherence for each measure was tabulated and stratified by tumor type. Due to small sample sizes within each DCP, data were pooled and analyzed with descriptive statistics and chi-square testing. Results: Over nine DCPs, stage and chemotherapy intent were documented in 89.1% and 81.3% of charts, respectively. There was a significant association between tumor type and documentation of stage (χ2(4) = 30.4, N=727, p <.001) and chemotherapy intent (χ2(4) = 157.5, N=534, p <.001). Documentation of stage and chemotherapy intent was highest for breast (100%, 93.6%) and colorectal cancers (92.7%, 92.1%) and lowest for NHL (71.8% 32.8%). Conclusions: Observed variation in documentation of stage and chemotherapy intent was primarily due to tumor type. Reasons for this observation are myriad and likely include factors related to the providers, the practice, the measures, and differing complexity of tumor types. This variation in quality scores by tumor type (driven by module selection) could have significant implications in today’s pay for performance environment. [Table: see text]