Addressing Barriers in Health Equity in Cancer with Technology

Share this post:

Like many doctors, I became a physician to care for those who needed it most.

That motivation is central to my whole career. It’s why I worked in community health centers for underserved communities in Baltimore and Washington, DC, and pushed initiatives to eliminate health disparities when I worked at the National Institutes of Health. And it is what drew me to my work now with IBM Watson Health, where our ambition is to use artificial intelligence (AI), data and analytics to scale and spread medical knowledge across oncology, genomics, life sciences, providers, governments, and more, helping physicians make more informed decisions to provide the right care for those who need it most.

We are by no means there yet. But with growing adoption of our oncology technologies in health systems around the world, we’ve come to see the potential that this new technology has in making a real difference in the lives of patients and the physicians who care for them.

I want to share a little about what we continue to learn at IBM Watson Health, and why we are so optimistic about the future. But first, we must concede that we — all of us working in medicine — still have much to do to bridge the abiding chasm in care for people with cancer, which is one of the leading causes of death worldwide.

In one year, 2015, 8.8 million people died from cancer across the globe. That’s equivalent to every man, woman, and child living in the San Francisco Bay Area. That’s a devastating statistic but what’s even harder to learn is how unevenly this devastating burden is felt with nearly 70 percent of those deaths in the poorest of countries.¹ Some may say to themselves, “Well of course that’s true.” But as caregivers we can’t help but be troubled by the idea that two people with the same disease will have very different prognoses simply based on where they live or the color of their skin.

And each country has its own unique hurdles. In Brazil, for instance, where about 57,000 women were diagnosed with breast cancer last year, more than half did not receive that diagnosis until it had advanced into later stages. India too, where the demands are much greater and the resources even more limited, is often late in diagnosing cancer, and inconsistent in offering care.^[1] In eight countries, many of them in Africa, there were no clinical oncologists at all, to provide the expert decision-making needed for treatment decisions. And in more than two-dozen other countries, such as nine in Asia, there are oncologists, but not enough of them to support an ever-growing patient population as cancer rates continue to rise.³

But even in developed countries such as the United States there are persistent disparities in health outcomes for racial and ethnic minorities, lower-income families, and people in rural communities. Although there have been decreases in cancer death rates nationwide, there is a slower reduction in cancer death rates in rural America compared with urban America.⁴ Even something as simple as the distance a patient must travel to get to an oncology clinic can play a role in results.

So, geography, income, insurance status and race, as well as many other factors, contribute to the vast disparities in detection and outcomes for cancer patients. It is something that many of us know and are frustrated with because we have made incredible medical advancements and breakthroughs for treating cancer, but those advances are not reaching all patients who could benefit from them. Right now, where you live, the color of your skin and the ability for you to pay for your care can make all the difference in whether you survive a cancer diagnosis.

What could IBM Watson Health possibly do about all that?

More than you think.

IBM has always excelled at finding innovative solutions to hard problems, and Watson Health is following in that tradition. Bridging disparities in oncology care is a hard problem, but what we’re learning as we’ve worked with hospitals and health organizations across the globe, is that even with such a complicated issue as this, it is the small improvements that sometimes can make all the difference.

While the AI that powers Watson has enormous potential in the field of medicine, right now what we offer in oncology is access and a tool to help provide decision support, and that helps clinicians to make those decisions with more confidence, speed, data, and evidence. This is important to both physicians and patients because even in the most remote parts of the globe, we can put the most relevant information available for making critical medical decisions within reach for those who are providing the care. Watson can save doctors and multidisciplinary/molecular tumor boards time by helping them sort through treatment options for their patients.

A doctor with access to Watson has at his or her fingertips access to evidence-based, personalized treatment options and an up-to-date knowledge base, one that is regularly being updated as new findings are added to the literature. Most parts of the world do not have access to the kind of cancer specialists found at the leading cancer institutes in places like New York or Los Angeles, but we can leverage the knowledge from oncologists at those institutions. At hospitals where Watson is being utilized, AI can help to scale and spread that experience to give community-based doctors who are required to treat a wide range of cancers, by giving them access to specialized expertise when they make treatment decisions.

Like the best innovations, the AI that is part of Watson Health is but a tool designed to assist clinicians in making more informed medical decisions and help doctors and hospitals improve the efficiency of caring for patients.

e have seen some proof points on this — with numerous, peer-reviewed abstracts and studies published in medical literature and featured at major scientific/medical conferences — and there will be more to come as we continue to publish information on these scientific advancements being made with our clients and partners around the globe. For physicians to use an AI system, it will be essential that they understand the science behind it. Watson Health is committed to advancing and sharing the science of Watson Health in the areas of research, real world evidence, and AI. Here are some scientific results that we are sharing at this year’s ASCO Meeting:

• Artificial intelligence data showing: Increased enrollment in clinical trials, greater accuracy in identifying mutations to guide treatment, and reducing variability in care and adherence to guidelines:
  • Abstract: 6550: A poster presentation in which Mayo Clinic will report data demonstrating that Watson CTM drove an 80% increase in breast cancer clinical trial enrollment over an 18-month period, and further expansion of the program drove a 130% increase in trial enrollment.
  • Abstract e24254: Watson for Genomics matched bioinformatics molecular tumor board’s manual analysis of mutations in 43% of lung cancer cases. However, in the 57% of cases, Watson for Genomics found 1.54 additional mutations, on average, that the bioinformatics molecular tumor board had missed (n=115).
  • Abstract e18566: A study on concordance, decision impact and guidelines adherence using artificial intelligence in high-risk breast cancer showed that disclosure of WFO recommendations resulted in treatment changes in 106 or 5% of cases. The adherence rate among the 106 cases where decision changes were made also improved adherence rate to treatment guidelines from 89% to 97% (n = 1997 breast cancer cases).
• Real-world data showing cost savings for breast and prostate cancer patients:
  • Abstract 1067: Evaluation of Watson Health MarketScan®claims data, a collaboration with MGH Cancer Center, Dana-Farber, Novartis and IBM Watson Health found that everolimus-based treatment was associated with lower total (= $6,462 less PPPM) and breast cancer-related (= $5,706 less PPPM) costs compared to chemotherapy-based treatments in patients with HR+/HER2- metastatic breast cancer following CDK 4/6 inhibitor treatment.
  • Abstract 18893: Study using Watson Health MarketScan®claims data compared the annualized costs of hospitalization for pre and post initiation of ENZA and ABI in chemotherapy naïve mCRPC patients (n= 3,351) in the U.S. Annualized hospitalization costs for ENZA were 25.1% lower compared to ABI patients, providing insights on real-world mCRPC management with appropriate therapy to minimize hospitalization costs.

We are in the early days in using this technology, however, we believe it has and will continue to improve over time, as Watson Health learns and more data and knowledge is pulled into the system. Working closely with key stakeholders and trust brokers in health systems like physicians will help us to continue to improve these AI systems. We are just scratching the surface of what this technology can do, but we are already making a difference.

IBM Watson Health is committed to leveraging our AI to scale and spread expertise and medical knowledge so that patients and populations who need it most can access the best care and we can reduce and, ultimately eliminate health disparities and achieve global health equity.

1  World Health Organization Cancer Fact Sheet, February 1, 2018.

2 The Union for International Cancer Control, , Responses to Global Cancer Control Leaders, March 29, 2018

3 Journal of Global Oncology, Global Survey of Clinical Oncology Workforce, February 8, 2018

4 Centers for Disease Control and Prevention, , New CDC Report shows deaths from cancer higher in rural America, July 6, 2017

[1]