In the United States, the uncertainties incumbent with the launch of a new drug are felt by payers and providers alike.  This uncertainty is more acutely felt for oncology therapies, where patient outcomes and affordability are primary concerns.  In clinical practice, this may mean that patients with similar clinical conditions and status may receive the new drug at different points in their disease course.

This variation in practice is a normal part of figuring out how to best use a drug in the real world. However, for the overarching health care system, this can be costly. How can this discrepancy be harnessed to more quickly gather information about the best use of therapies and then use this evidence to help providers base their decisions about treatment?

One example of an approach to solve at least part of this puzzle comes from Hackensack, NJ’s Meridian Health, an integrated healthcare system, and its collaboration with IBM Watson and Cota technology that began in the summer of 2017.1 The results of this pilot were published in the Journal of Precision Medicine in December 2020.2 Using digital classification schema (Cota Nodal Address [CNA]), 4,032 patients with some form of cancer were grouped into 1,403 unique CNA cancer disease states.

The 10 most common CNAs for each disease contained a majority of each cancer (30% breast, 23% colorectal, 48% lung). The ranges of total cost of care for identical CNAs were wide and the estimated median avoided cost per patient was $26,773.

  • Breast cancer $16,000-$88,000
  • Colorectal cancer $25,000-$222,000
  • Lung cancer $44,000-$424,000

The findings from COTA are to be embedded in the Watson Decision Support software used by Meridian Health oncologists.1 Outcomes of this phase are not yet available.

Flatiron’s proprietary EHR abstraction platform offers another avenue for real-world data collection and analysis.3 The Flatiron system appears to be a ‘bolt-on’ service to many EHR systems. The hybrid technology-based/human extracted system provides clean data for research purposes.

What are the implications for industry and patients?

Don Berwick, senior author on the COTA outcomes paper, former CMS administrator and Senior Fellow at the Institute for Healthcare Improvement sees three potential next steps in this area.4

  1. A growth in the use of registries and their databases “so the aspiration would be that the amount of information in the database grows as fast as possible because the more data we have, the faster we can learn.”
  2. Systems like COTA become more well-known in the oncology community and are used to address patient-level and population-level questions.
  3. These systems become known to patients, empowering them to become full partners in making therapy decisions that affect them, their caregivers, and families.

Clearly, all three of Dr. Berwick’s aspirations would benefit patients in the long and short run. They may also be opportunities for industry; fostering partnership opportunities with the broader oncology community so that these real-world databases may be used to generate real-world evidence that supports patient care.

Publication of such evidence may help further the knowledge and adoption of the extraction systems used but also the value of registries, thus potentially creating a virtuous circle that supports the entire health care community.


  1. Hackensack Meridian Health to pilot combined Watson for oncology and Cota Technology to power value-based clinical decision support in cancer care. OncLive. Published August 7, 2017. Accessed December 21, 2021.
  2. Pecora AL, Ip A, Wang CK, et al. The COTA Nodal Address Model: a novel digital classification system identifies variances in cancer care cost to support value based care delivery. J Precis Med. 2020.
  3. D’souza A. Flatiron Health: combating cancer with data analytics. Published July 18, 2019. Accessed December 21, 2021.
  4. Caffrey M, Berwick DM. Use of COTA Nodal Address highlights variance in cancer care cost: an interview with Donald M. Berwick, MD, MPP. Am J Manag Care. 2021;27(2):SP65-SP66.