Finding success in alternative payment models is going to require infusing clinical data into administrative claims data.
This was the key takeaway of an analysis of data from a participant in the Oncology Care Model (OCM), a value-based episodic payment model being tested by the Centers for Medicare & Medicaid Service’s Center for Medicare and Medicaid Innovation.
Researchers looked at administrative claims data and infused into it clinical data from 377 6-month-long episode payments pertaining to 210 prostate cancer patients participating at a single OCM site to see if the combination could be used to develop more useful information that OCM participants could use to help improve care and control costs.
“In order to introduce these clinical factors, what we did was divide the patients that were included in our analysis into different disease-state treatment dyads,” Anish Parikh, MD, of Icahn School of Medicine at Mount Sinai, New York, said in an interview. Dr. Parikh is a coauthor of a report detailing the results of the analysis that was published in the Journal of Oncology Practice.
“Then we did a comparative cost analysis using the OCM’s data in this framework,” he continued. “We introduced these clinical factors this administrative claims data and then ran a comparative cost analysis to see what were the contributions to cost of care for each of these dyads.”
The analysis found that the most excessive expenses in prostate cancer treatment were seen in metastatic, castration-resistant dyads containing second-line hormone therapy (ratio of observed to expected expenses [O/E], 2.66), chemotherapy (O/E, 2.09) and radium-223/sipuleucel-T (O/E, 3.01).
During the observation period, the CMS updated the payment model, and a singular change correcting for castration-resistant prostate cancer resulted in up to a 38% increase in O/E for hormone-sensitive dyads and up to a 58% decrease in O/E for castration-resistant dyads. The update improved the overall O/E for all episodes by 22%, from 1.48 to 1.15.
The analysis “really is a demonstration of how sensitive this model is to clinical factors and how necessary and important it is to include clinical data into the administrative claims data whenever possible in order to make the model more accurate and usable,” Dr. Parikh said.
The investigators reported no conflicts of interest.
SOURCE: Parikh A et al. J Oncol Pract. 2019 Feb 11. doi: 10.1200/JOP.18.00336.