Measuring and Improving Quality in Oncology Practices

Get Permission

The seed for ASCO’s Quality Oncology Practice Initiative (QOPI®) was planted a decade ago by Joseph Simone, MD, when he contemplated the feasibility of studying a volunteer group of oncologists to measure the quality of care they provide and share those results with their colleagues. Dr. Simone’s original vision has since grown far beyond its original scope. At ASCO’s first Quality Care Symposium in San Diego, Michael N. Neuss, MD, Chair of ASCO’s QOPI steering committee, discussed the program’s first 5 years of data. Dr. Neuss is Chief Medical Officer and Professor of Clinical Medicine at Vanderbilt-Ingram Cancer Center, Nashville, Tennessee.

Biannual Analysis

“Our project was geared to identify which factors in the QOPI project correlated with improvement—whether related to practice or measures—and to see if, in fact, practices improved over time,” said Dr. Neuss, mentioning that the QOPI self-reporting program is analyzed twice a year.

“There have been approximately 50 quality measures that were stable during the first 5 years of QOPI, and we used those measures for this analysis, which included 308 practices and approximately 2,000 physicians, representing about 15% of the nation’s medical oncology workforce,” explained Dr. Neuss.

The overall quality scores were constructed based on questions from different domains, such as breast cancer, end-of-life care, colon and rectal cancers, non–small cell lung cancer, non-Hodgkin lymphoma, and symptom toxicity management. “If we look at multivariate analysis in those factors associated with improvement, the highest correlations were seen for a new clinical practice, such as the introduction of a treatment drug or a genetic test associated with a drug. And when we looked at the interaction at the new practice over time, it showed the largest corollary of improvement within any variable within our measure set,” said Dr. Neuss.

According to the group’s findings, measures seemed to fall into three distinct areas: those that showed great improvement over time, those that showed no improvement because they were already too high to improve, and those that showed no improvement despite a large opportunity to improve. “Although we saw a wide variation in baseline achievement among the practices that obscures any statistical significance to these measures [outside of the study model], I think that the underlying reality is a nice trend in improvement,” said Dr. Neuss.

Dr. Neuss next looked at whether adjuvant therapy in breast, colon, and non–small cell lung cancer was appropriately recommended to patients in the large data sets. “The level of achievement was very high, indicating that practices are doing very well in this setting. In fact, for these three diseases, the baseline achievement is more than 90%,” commented Dr. Neuss.

More Work to Be Done

“The news is not as encouraging for measures such as the appropriate documentation of smoking cessation status, the risk of infertility, and fertility preservation in patients for whom these evidence-based measures would be applicable. As we know, measuring smoking is a ‘meaningful use’ criterion, and practices do very well in documenting tobacco use. However, we see that only 20% to 30% of the time, practices make an attempt to get their patients to quit smoking,” said Dr. Neuss. He added that there is a similar pattern of poor compliance in discussing fertility preservation options with patients with cancer who are of childbearing age.

Dr. Neuss continued, “Because of the very varied initial attainment levels, we constructed a model to look at the overall achievement when normalized for baseline accomplishment. We saw that adjusted mean quality scores improved from 71% to 85%, which is significant. Also, when we looked at how the introduction of new measures did over time, we saw a very dramatic improvement,” noted Dr. Neuss. Indeed, the improvement over time was highly statistically significant (P < .0001).

Dr. Neuss used some of the experimental circumstances that led to what became known as the Hawthorne effect—in which industry workers making the same product were aggregated in communal rooms and showed increased performance—as an example of interventions that increase benefits. (The more general premise of the Hawthorne effect is that behavior during the course of a study may be the result of social situations and the subjects’ awareness that they are being observed.)

“I would submit to you that the workers got better because they were made into a team instead of individuals. And the lesson here is that when practitioners and their patients work as a team, they improve quality care together,” said Dr. Neuss.

Lessons Learned So Far

Dr. Neuss stressed that medical oncology practices have improved in some but not all measures of care, and though improvement is rapid for many measures, those with the greatest improvement are the ones measuring the adoption of new information or treatments.

“On many of the adjuvant treatments, it appears that QOPI participants are already doing well. However, measuring is not enough. We have to improve our care by using collaborative models. To that end, networks have spontaneously sprung up within the QOPI practices…. [P]ayers have been very supportive of these important quality initiatives—QOPI is truly a collaborative effort with downstream benefits for our patients,” concluded Dr. Neuss. ■