Watson Health: get the facts

IBM Watson Health was created to help solve some of the world’s most pressing health challenges through data, analytics and AI. By combining human experts with augmented intelligence, IBM Watson Health helps health professionals and researchers around the world translate data and knowledge into insights to make more-informed decisions about care for their patients. In oncology, Watson is at work supporting cancer care in more than 300 hospitals and health organizations, and a large, growing body of evidence supports the use of Watson in healthcare.

Watson Health Approach

Watson Health has a unique approach to the application of data driven technology in the market. It is the combination of our data, our cloud, and our AI services that build cognitive offerings for our partners and clients.


Data alone does not provide optimal insight. We spend a tremendous amount of time cleaning, curating, and normalizing data. We collate, design, and infuse risk & severity adjustment methodologies to create actionable and useful information. This makes the data ready to be analyzed and paired with AI services to prioritize and solve healthcare challenges. We bring our analytical power to our client’s to help our clients create rich population health and value-based care programs. With proprietary models to approach challenges like cost of care and disease staging, we can find answers to complex questions like drivers of risk, predictors of healthcare expenditures, and risk adjusted mortality.


The Watson Platform for Health is specifically built for health, and it enables our partners and clients to focus on what’s important: health data analysis and discovery. It is built on the power of the IBM Cloud. In the Health cloud, we start with data protected by privacy regulations in every country where we operate. In IBM’s cloud, it is all about YOUR data. The IBM Cloud combined with the Watson Data Platform helps you to maintain ownership and helps you to protect your data. We build the IBM Cloud for AI services and cognitive technologies. Our cloud must protect your data and our cloud must enable our AI.

Artificial Intelligence

We have 80 AI Services that we and our clients can access to build offerings. For example:

  • Annotator for Clinical Data. One power of Watson is its ability to read and understand unstructured data – 80% of healthcare data is unstructured. Our natural language processing reads clinical text from any source to identify, categorize, and code medical and social concepts. This is useful in analytics with a much broader set of data, and provides users direct access to more information, and improves processes like coding and alerting.
  • Insights for Patient Data is a service that identifies the problems contained in a patient’s historical medical record, both in the structured and unstructured text. It summarizes the history of their care around those problems, and provides a cognitive summary of a patient’s record.
  • Patient Similarity identifies patients who are similar to a given patient in a clinically meaningful way and identifies a measure of clinical similarity between the patients. This allows us to create dynamic patient cohorts, rather than static patient cohorts, and enables an understanding which care path works better for a given group of patients.
  • Medical Insights enables clients to find information in unstructured medical literature to support hypotheses and to help in the discovery of new insights. Watson can read through a complete set of medical literature, like Medline, and identify the documents that are semantically related to any combination of medical concepts.

The Unmet Need

What is driving the need for Cognitive technologies in health? It is Data. Medical data is expected to double every 73 days by 2020. And, each person will generate enough health data in their lifetime to fill 300 million books. Physicians simply cannot keep up with the growing amount of information available to them.

In addition, there are growing challenges of physician shortage and physician burnout. Today, primary care physicians work on average 11.2 hours each day and 6 hours of that day is spent interacting with the Electronic Health Record. And, social determinants of health make up 70% of our health determinants. Where we live, how we eat, when we exercise, levels of stress, family support are huge factors in our overall health.

AI can make sense of the overwhelming amount of clinical data, genomic data, and social determinants of health data to find the best path for each patient.

Data Stewardship

IBM Watson Health is a healthcare company backed by powerful technology. As such, responsible data stewardship is core to our mission. We use the best data to train our solutions, we are transparent about how we train, and we believe your data is your data.

A client’s data is their own, our agreements with clients are transparent and there is no requirement to relinquish their data or their insights for our benefit. We will always be transparent about about how our AI systems were trained, the data sources that were used and how the training was completed.

Our mission is to improve lives and give hope by delivering innovation to address the world’s most pressing health challenges through data and cognitive insights. Our progress is real, and our mission gives us purpose. To have a positive impact on humanity that is sustainable for the long term.


Kyu Rhee, MD, IBM Watson Health chief health officer introduced the Watson Health Science Report in early 2018 to provide details on science and progress across the Watson Health portfolio.

The journey started in 2005 when Artificial Intelligence in healthcare was a Grand Challenge at IBM Research. That work at IBM Research was the foundation for Watson Health at our launch in April 2015 and remains foundational to our work today.

Watson Health made important progress:

  • Watson Health has more than 15,000 clients and partners.
  • Today, Watson Health cognitive offerings have impacted care or social services for more than 295,000 people.
  • We have 230 million clinical and claims lives in our data lake.
  • More than 50 peer reviewed publications, posters, and abstracts support Watson Health cognitive offerings and 500+ pieces of scientific evidence that demonstrate how our AI data and analytic tools are being used by clients and partners in healthcare and life sciences.
  • IBM has more than 2,500 active granted and pending U.S. patents in healthcare and life sciences. 400 of those are specific to Watson Health.

Key Study Data:

  • In January 2018, the Annals of Oncology published a study by Manipal Hospitals that found Watson for Oncology was concordant with the hospital’s multidisciplinary tumor board in 93 percent of breast cancer treatment decisions.
  • In November 2017, Acta Neuropathologica published a study by Barrow Neurological Institute that found Watson for Drug Discovery had successfully identified 5 RNA-binding proteins that had never before been associated with ALS.
  • In November 2017, the journal The Oncologist published an important study led by oncologists at the University of North Carolina’s Lineberger Comprehensive Cancer Center, who tested Watson for Genomics on 1,018 retrospective patient cases. More than 99 percent of the time, Watson agreed with the physicians, but in more than 300 cases, Watson found clinically actionable therapeutic options that the physicians had not identified.
  • In July 2017, Neurology Genetics published a study that found Watson for Genomics accurately interpreted whole genome sequencing data for a glioblastoma patient in 10 minutes vs. 160 expert human hours.
  • In June 2018, Mayo Clinic physicians presented a poster presentation at the ASCO Annual Meeting, reporting that Watson for Clinical Trial Matching boosted enrollment in breast cancer trials by 80% following implementation (to 6.3 patients/month, up from 3.5 patients/month in the immediate 18 months prior). Furthermore, enrollment was increased to 8.1 patients/month when including accruals to phase I trials with breast cancer cohorts in the experimental therapeutics program.
  • In June 2018, new data presented by Medtronic demonstrated the utility of Sugar.IQ in a real-world setting (Abstract: Real-World Assessment of Sugar.IQ with Watson—A Cognitive Computing-Based Diabetes Management Solution). The study found that people with diabetes using the Sugar.IQ app spent 36 more minutes per day in healthy glucose range compared to before they used the app. This included 30 minutes less time in hyperglycemia (>180 mg/dL) and 6 minutes less time in hypoglycemia (

Product Updates:

  • Watson for Oncology is trained in 13 cancer types.
  • Watson for Genomics supports genomic interpretation for any cancer type.
  • Watson for Clinical Trial Matching is supporting breast cancer care at the Mayo Clinic and is being expanded for investigator use.
  • Watson for Drug Discovery now features 12 new functions including an enhanced user experience, PubMed open access, and gene candidate generation.
  • Watson Health launched IBM® MarketScan® Explorys® Claims-EMR Data Set. We combine clinical and claims data to create a new longitudinal view across all settings of care, with the ability to deep dive into patient vital signs, symptoms, physician orders, and outcomes.
  • Watson Care Manager has 147,000 lives licensed. Our Government and Health and Human Services business runs 156 government programs across the globe, touching 170 million individuals with our portfolio and counting.
  • Watson Imaging Clinical Review launched in 2017, and won the 2017 “Minnie” award for Best New Software from Auntminnie.com.
  • Flexible Analytics provides action-driving insights that can help you better position your organization to optimize performance, manage population health and address compliance.
  • Watson for Benefits helps individuals understand their benefit coverage and how to navigate the healthcare system.

IBM Watson Health Scientific Update:

  • In Jan 2018, Watson Health issued the “Watson Health 100,” the first scientific update highlighting a collection of the top one hundred studies, abstracts, posters and white papers from 2017, as identified by IBM’s clinical team under the leadership of Kyu Rhee, MD, IBM Watson Health chief health officer. At the end of each quarter, the clinical team packages up a sampling of scientific evidence from Watson Health and its partners from the previous quarter. To subscribe for regular updates, email kristi.bond@us.ibm.com.
  • Watson Health 100: 2017 Scientific Update Year in Review – http://ibm.biz/BdZ4rJ
  • 2018 Q1 Scientific Update – http://ibm.biz/BdZLg2
  • 2018 Q2 Scientific Update – http://ibm.biz/BdYraN
  • 2018 Q3 Scientific Update – http://ibm.biz/BdYCZY
  • 2018 Q4 Scientific Update – https://conta.cc/2E055Ta
  • 2019 Q1 Scientific Update – https://ibm.biz/Bdz4yj


Oncology (Watson for Oncology, Genomics, Clinical Trial Matching, and Life Sciences)

Unmet need in Oncology: Discovering new genetic targets for therapeutics, interpreting massive amounts of genetic data and literature to identify actionable targets, scaling world-class oncology decision support and clinical trials access requires powerful cognitive tools to make sense of it all.

Watson Health oncology and genomics offerings have supported oncologists’ care decisions for more than 114,000 patients and are in use at more than 300 hospitals and health organizations. Watson for Oncology is now trained on 13 cancers, and in addition to training Watson for Oncology in additional tumor types, it has also been trained in multidisciplinary treatment modalities.

Watson for Oncology

Watson for Oncology is a decision support tool that is trained by top oncologists at Memorial Sloan Kettering. Watson for Oncology ranks the treatment options, linking to peer reviewed studies that have been curated by MSK. Watson for Oncology also provides a large corpus of medical literature for a physician to consider, drawing on more than 300 medical journals, more than 200 textbooks, and nearly 15 million pages of text to provide insights about different treatment options.

Watson for Oncology now allows client institutions to add localized treatments and dosing to their system, and to identify treatments that are unavailable in their geography. Watson for Oncology intelligently finds the most appropriate articles from PubMed and medical journals using patient cohort analysis, drug class matching, and document quality models. These are matched to the patient and treatment options in seconds.

The Annals of Oncology published a full study led by oncology leaders at Manipal Hospitals in India. Their tumor board found Watson for Oncology was concordant with their own tumor board’s treatment decisions in 93% of breast cancer cases. And it was interesting in this study that when the blinded study initially compared Watson’s recommendations against retrospective treatment decisions from 2-3 years prior, the concordance was only 73%. Manipal’s tumor board re-reviewed the same patient cases manually, and the concordance rose to 93%. This is an indication that the corpus of data on which Watson for Oncology relies is staying up-to-date with the latest science. Once treatment options are provided by Watson for Oncology, we provide the following in support of the options; supportive care guidelines, drug information, Truven cost data, literature references, and dosing details.

At ASCO 2018, physicians from the Affiliated Hospital of Academy of Military Medical Sciences in China in reported a study (n=1,997), which found that disclosure of Watson for Oncology recommendations resulted in treatment decision changes in 5% of cases. Attending physicians were less likely to alter their decisions (3%) than chief physicians (6%; p < 0.001) and fellows (7%; p < 0.001). Importantly, in those 106 cases where clinicians changed their treatment decision based on Watson for Oncology, adherence to professional treatment guidelines improved from from 89% to 97% (p < 0.01).

Researchers prepare tissue samples for whole genome sequencing at The Rockefeller University, where clinical researcher Robert Darnell, M.D., Ph.D., led a study with the New York Genome Center and IBM to analyze complex genomic data from state-of-the-art DNA sequencing of whole genomes. The findings were published in the July 11, 2017 issue of Neurology Genetics, an official journal of the American Academy of Neurology. (Photo Credit: Epic Creative)

Watson for Genomics

Unmet need: Watson for Genomics identifies and supports the interpretation of genomic data and provides insights as to whether that data is actionable.

In a study published in November 2017 in The Oncologist, UNC Lineberger researchers used IBM Watson for Genomics to assess whether cognitive computing was as effective as a panel of cancer experts in identifying therapeutic options for tumors with specific genetic abnormalities. In a retrospective analysis of 1,018 cancer cases, the molecular tumor board (MTB) identified actionable genetic alterations in 99% of 703 cases, which Watson also confirmed. Using the curated Watson for Genomics gene list, researchers identified additional potentially actionable genomic information in 324 patients, 96 of whom were not previously identified as having an actionable mutation. The Watson for Genomics analysis took less than 3 minutes per case.

In July 2017, Neurology Genetics published a study that found Watson for Genomics accurately interpreted whole genome sequencing data for a glioblastoma patient in 10 minutes vs. 160 expert human hours.

Watson for Clinical Trial Matching

Unmet need: Watson for Clinical Trial Matching eliminates the need to manually compare enrollment criteria with patient medical data, making it possible to efficiently identify an individual’s potential trial options in a list ranked by relevance and eligibility.

Data published at ASCO 2017 with Highlands Oncology Group and Novartis showed that Watson for Clinical Trial Matching successfully demonstrated the ability to expedite patient screening for clinical trial eligibility, reducing processing time from 1 hour and 50 minutes to 24 minutes. It also omitted 94% of non-matching patients automatically – reducing screening workload dramatically.

Dr. Tuffia Haddad from the Mayo Clinic reported at ASCO 2018 that Watson for Clinical Trial Matching boosted enrollment in breast cancer trials by 80% following implementation (to 6.3 patients/month, up from 3.5 patients/month in the immediate 18 months prior). Furthermore, enrollment was increased to 8.1 patients/month when including accruals to phase I trials with breast cancer cohorts in the experimental therapeutics program.

Watson for Clinical Trial Matching is not a simple search or basic structured data matching technology. Watson for Clinical Trial Matching leverages cognitive capabilities to look at unstructured text and derive key insights from that data. Very few providers capture all the required patient data attributes for trial matching in a structured format.

Watson for Clinical Trial Matching understands key patient attributes and how to identify them in a variety of formats, including Clinical Notes, Pathology Reports, Labs, etc., to effectively evaluate a patient against the inclusion or exclusion criteria for a trial.

Watson for Drug Discovery

Unmet need: When researchers can more quickly uncover novel patterns and connections they can accelerate discovery, which can lead to effective pharmaceuticals going to market and reaching patients sooner.

Watson for Drug Discovery helps our clients to understand what is currently known with great speed and scale, and generates evidence-based hypotheses with greater confidence. In life sciences, Watson for Drug Discovery is currently in use tackling challenging diseases such as ALS and Parkinson’s with tangible results.

The Barrow Neurological Institute used Watson for Drug Discovery to explore unidentified genes and proteins that may be linked to ALS using Watson for Drug Discovery. In a matter of months, the system rank-ordered all of the nearly 1,500 RNA-binding proteins encoded by the human genome and proposed predictions regarding which proteins might be associated with ALS. The Barrow team then examined Watson’s top evidence and found eight of the top 10 ranked proteins proved to be linked to ALS, and found WDD had uncovered five never-before linked proteins associated with ALS. In November 2017 this study was published in the journal Acta Neuropathologica.

Toronto Western, part of the University Health Network, is the first hospital in Canada to use Watson for research in Parkinson’s. Using Watson for Drug Discovery, researchers ranked a list of 620 drug candidates that could potentially be used to treat Parkinson’s. Research of the top 52 drugs revealed that 21 of the drugs identified are worthy of further study to be potentially repurposed for Parkinson’s patients. Of those, 12 had never been linked to Parkinson’s before.

Pfizer is using Watson for Drug Discovery to accelerate discovery of immuno-oncology targets. Using Watson for Drug Discovery, Pfizer is identifying novel gene sets that have not been associated with immune response.

Richard Martin, Research Scientist and Technical Lead

Meet Our Scientists

Dr. Richard Martin is a research scientist and technical lead at IBM Watson Health. For three years, Richard has been innovating and pushing the boundaries of artificial intelligence (AI) technology in Life Sciences. Dr. Martin’s research interest generally concerns how techniques from computer and information science can work together, and alongside human experts, to help accelerate scientific discovery. His vision is to build an AI research assistant that can augment the capabilities of human experts, uncovering patterns that emerge from enormous volumes of disparate data, that humans alone cannot hope to parse, let alone identify meaningful signals from.

Richard is the natural language processing team leader responsible for teaching the Watson system how to read and understand scientific literature and the language of drug discovery, by extracting and cross-referencing information about genes, drugs and diseases, and how they interact, across a wide variety of sources. Richard has invented and built new types of predictive analytics for machine reasoning upon this massive data, and Richard’s contributions to Watson’s abilities have led to impactful discoveries, such as the recent identification of five never-before-linked proteins that are associated with amyotrophic lateral sclerosis (ALS). These proteins are now potential avenues for development of novel treatments, and after further testing and validation, the findings were recently published in the journal Acta Neuropathologica.

Richard has a deep personal motivation for innovating in healthcare, rooted in his father’s kidney cancer and ankylosing spondylitis diagnoses, neither of which are presently curable. The analytics that Richard has developed enable, for the first time, the identification of common connections between diseases such as these, and the genes and drugs that are known or likely to be associated with them, providing the hope of finding new paths to discovering a cure. When researchers can more quickly uncover novel patterns and connections they can accelerate discovery, which can lead to effective pharmaceuticals going to market and reaching patients sooner.