Measuring the improving quality of outpatient care in medical oncology practices in the United States.

Abstract

66 Background: The American Society of Clinical Oncology (ASCO) Quality Oncology Practice Initiative (QOPI) has provided a method for measurement of process based practice quality since 2006. We sought to determine whether QOPI participants show improvement in measured quality over time, and if change is demonstrated, those factors in either the measures or participants which are associated with change. Methods: 156 participant medical practice groups, which did not submit trainee data, participated in at least two collection rounds from 2006 through 2010, and reported on ≥30 patients per round from were included in analysis from a larger group of 306 participants from this time period. A database of these reports was used to evaluate trends in scoring among participants. A linear regression model, which controlled for varied initial performance, was used to estimate the effect of participation over time and evaluate participant and measure characteristics. Measures were aggregated into categories to evaluate which factors correlated with change. Results: Participants completed a mean of 5.71 (S.D. 1.84) rounds of data collection. Adjusted mean quality scores improved from 0.71 (95% C.I. 0.42 – 0.91) to 0.85 (95% C.I. 0.60 – 0.95). Overall odds ratio of improvement over time was 1.09 (p < 0.0001). The greatest improvement was seen in measures assessing newly introduced clinical information, where the mean scores improved from 0.05 (95% C.I. 0.01 – 0.17) to 0.69 (95% C.I. 0.33 – 0.91), (p ≤ 0.0001). Many measures showed no change over time. Conclusions: QOPI has gained widespread adoption and approximately 15% of U.S. medical oncologists participate. Participation over time is highly correlated with improvement in measured performance. Much of this improvement is the result of the adoption of newly introduced clinical information. Some measures show no change despite significant opportunity for improvement.