Watson Health Scientific Update
At IBM Watson Health, we help health professionals and researchers around the world translate data and knowledge into insights to make more-informed decisions about care for their patients – and we are committed to do this through data, analytics and AI.  
 
In the first half of 2018, IBM and our partners have published more than 160 pieces of evidence highlighting scientific progress and achievement across Watson Health AI and real-world evidence solutions, as well as health and life science advances from IBM Research.  As our science continues to evolve, we are pleased to share the early impact of how we support health professionals in data-driven and evidence-based decision-making by advancing research, real-world evidence, and AI in healthcare.
 
This report highlights scientific progress from the second quarter of 2018 and spotlights some of the new data that was recently published in peer-reviewed medical literature and at major scientific events like ISPOR, ASCO, ADA and more.  It also features an in depth look at one of our most recent pieces of evidence and profiles one of the authors from that poster presentation.
 
We hope you will enjoy!  
Meet the Scientists: Gigi Yuen-Reed
Gigi Yuen-Reed is a Principal Data Scientist at Watson Health. Gigi’s primary responsibility is to combine machine learning and predictive solutions, with knowledge of healthcare data and domain expertise across all markets in Watson Health. Her work ranges from architecting the AI algorithm, to collaborating with customers to get early market feedback, to enabling the engineering team in product development and ultimately, making machine learning and AI algorithm meaningful and useful to our customers.  Some specific examples of her work include managing rising healthcare cost through better insights in practice patterns and treatment efficacy, improving population health through optimizing patient engagement and customizing the care plan. Check out the ‘Behind the Data’ section below featuring a recent poster presentation co-authored by Gigi.
 
Why Gigi is interested in this field of work, in her own words
“Data science is one of the hottest fields in the market, and the development in healthcare is growing at a lightning speed. However, a lot of these advancements are trapped at academic and industry research labs. My passion is to make these one-off studies and experiments a reality to our customers, so that patients, providers and care takers can benefit from the transformative power of AI in healthcare at-scale. This journey is full of ups and downs, and is truly rewarding. It challenges me to be innovative and practical, risk-taking and methodical, all at the same time.” 
Behind the Data: Understanding Medication Adherence
Medication nonadherence poses a significant public health problem. While the volume of healthcare data is growing exponentially, the healthcare community continues to seek to derive meaningful insights to positively impact care delivery. Until recently, healthcare claims and clinical patient data have not been linked in a way that can consistently deliver insight around resource utilization and patient characteristics and outcomes. Prescription filling and adherence to prescribed medications cannot be examined by claims or electronic medical record (EMR) data individually. 
 
In this new poster,  “Prescription Fill Rats for Acute and Chronic Medications in Claims-EMR Linked Data”   – presented at the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) 2018 in May – researchers studied the medication filling behavior of patients by leveraging the IBM® MarketScan® Explorys® Claims-EMR Data Set. This new offering represents the integration of claims and clinical data from two key IBM Watson Health acquisitions, Truven Health Analytics and Explorys. 
 
The Claims-EMR Data Set links de-identified patient-level claims data found in the MarketScan Commercial and Medicare Supplemental Databases with de-identified data for the same patients found in the IBM Explorys electronic medical records (EMR) dataset. This data set reflects real-world treatment patterns and costs by tracking patients as they travel through the healthcare system, offering detailed information about all aspects of care. Data was used from 134,434 patients with prescriptions for antibiotic, antihypertensive, or antidiabetic agents documented in the EMR were evaluated to determine patient characteristics and prescription fill rates as measures of medication adherence. 
 
The overall fill rate following an order was 73%, 74%, and 76% for antibiotics, antihypertensives, and antidiabetics, respectively. Data also showed that African Americans were negatively associated with overall medication fill rates. Prior treatment with the same class drug had positive association with fill rates in patients receiving antihypertensives or antidiabetics in contrast to its negative association in patients receiving antibiotics. This study also found that significant proportions of patients did not fill their prescription for antibiotics, antihypertensives, or antidiabetics. Medication adherence is important for controlling acute or chronic conditions, and these findings indicate a need for further work to identify contributors to nonadherence and targets of intervention.
Data + Clinical Evidence Highlights – Q2 2018
Watson Health AI Offerings & Capabilities
 
Manuscripts

2018 IEEE 15th International Symposium on Biomedical Imaging , April 2018

2018 IEEE 15th International Symposium on Biomedical Imaging , April 2018

Scientific Reports , May 18, 2018

Neuropharmacology,  June 1, 2018 

Abstracts and Posters 

2018 ASCO Annual Meeting , June 2018

2018 ASCO Annual Meeting , June 2018

2018 ASCO Annual Meeting , June 2018
 
2018 ASCO Annual Meeting , June 2018
 
2018 ASCO Annual Meeting , June 2018

2018 ASCO Annual Meeting , June 2018

American Diabetes Association’s 78 th Scientific Sessions , June 2018
 
Watson Health Real-World Evidence
 
Manuscripts

Diabetes Research and Clinical Practice , May 23, 2018

Health Services Research,  May 30, 2018
 
Cardiovascular Drugs and Therapy,  May 31, 2018

The American Journal of Medicine , June 12, 2018

American Journal of Perinatology , June 19, 2018

Abstracts and Posters

18 th International Conference on Integrated Care (ICIC18), May 2018

ISPOR 2018 , May 2018

ISPOR 2018 , May 2018

2018 ASCO Annual Meeting , June 2018

2018 ASCO Annual Meeting , June 2018

American Diabetes Association’s 78th Scientific Sessions , June 2018
 
AcademyHealth Annual Research Meeting, June 2018
 
IBM Research
 
Manuscripts

PLoS One, April 9, 2018

Abstracts and Posters
 
Public Health Genomics , April 4, 2018
Coming Up…
Stay subscribed for our Watson Health Scientific Update and follow our journey to help healthcare stakeholders accelerate innovation and tackle some of the world’s biggest health challenges with data, analytics and AI. It was great to see many of you at this year’s ASCO Annual Meeting AHIP and the American Diabetes Association’s 78 th Scientific Sessions; we hope to see you at some of these industry events happening this year, including:

-         Medicaid Enterprise Solutions Conference (MESC) , August 13-August 16, 2018 
-         National Association of Medicaid Program Integrity (NAMPI) , August 26-29, 2018 
-         Chinese Society of Clinical Oncology (CSCO) , September 23, 2018
-         19th World Conference on Lung Cancer (IASLC) , September 23rd-26th 
Contact Us:
For additional information, or to receive the full list of studies published in Q2 2018, please contact Kristi Bond at kristi.bond@us.ibm.com .
IBM Watson Health| 75 Binney Street, Cambridge, MA