Microsoft Donates Data Analysis Tool to Aid Research into SIDS | Healthcare Informatics Magazine | Health IT | Information Technology Skip to content Skip to navigation

Microsoft Donates Data Analysis Tool to Aid Research into SIDS

June 8, 2017
by Heather Landi
| Reprints

A team of data scientists from Microsoft Corp. donated a new research tool to Seattle Children's Research Institute, one of the top five pediatric research centers in the United States, to aid medical research into sudden infant death syndrome (SIDS).

Microsoft also plans to make the data analysis tool available to researchers worldwide.

SIDS is the leading cause of death among children from one to 12 months old. Approximately 4,000 infants in the U.S. die each year from sudden unexpected infant deaths (SUID), which includes SIDS. There has been no significant drop in such deaths since the mid-1990s.

A Microsoft co-worker's loss of a son to SIDS inspired the team of data scientists to contribute hundreds of volunteer hours to create the research tool. John Kahan, Microsoft general manager for Customer Data and Analytics, turned his focus to fighting SIDS after he and his wife, Heather, lost their only son, Aaron, 13 years ago. "My mission is to ensure that no parent experiences the pain of losing a child to SIDS, or worry that their child may be next," Kahan said in a statement. "I am incredibly moved and grateful to my team for volunteering their personal time to create this tool for SIDS research."

More than a dozen data scientists donated more than 450 hours over nights and weekends the past year, including several who are parents or are getting ready to have families of their own, as outlined in a Microsoft News Center blog post about the initiative.

The team of data scientists set out to enable researchers without deep technology skills to use big data the same way the world's top technology companies do, to more quickly identify the causes of SIDS and develop preventive measures.

Kahan's team accessed publicly available data from the U.S. Centers for Disease Control and Prevention (CDC), including information on 29 million births and over 27,000 sudden and unexplained infant deaths from 2004 to 2010. They created several machine learning and statistical models, which run on Microsoft Azure, to interpret it and crunch massive amounts of data. The data is displayed visually on Power BI, allowing researchers to click on thousands of combinations of factors — such as the child's birth order, the parents' ages and the level of prenatal care a mother received — and see how those factors might be connected.

According to Microsoft, those connections, or correlations, as scientists call them, can be key: In the early 1990s, when sleeping position was found to be correlated with SIDS, new guidance for parents significantly reduced infant deaths.

In pilot, the tool has already unveiled new correlations, including that every hour expectant mothers spend at prenatal visits results in a decrease in SIDS incidents and that each cigarette smoked by a pregnant woman is linked to an increase in SIDS deaths. The new correlations identified by the tool have been provided to Seattle Children's, where pediatric researchers will use scientific processes to see if they point to areas of further study.

“The potential of this tool to aid medical research and open new areas of exploration in identifying SIDS risk factors is both impactful and tremendously encouraging,” Nino Ramirez, M.D., director of the Center for Integrative Brain Research at Seattle Children's Research Institute, after seeing a demonstration of the system in November, said in a statement.

The data scientist team is working with Rutgers Robert Wood Johnson Medical School and Seattle Children's to develop protocols for other researchers to access the tool for research. Those from accredited medical research institutions may request free access to the tool and data by emailing Kahan's team.

Get the latest information on Health IT and attend other valuable sessions at this two-day Summit providing healthcare leaders with educational content, insightful debate and dialogue on the future of healthcare and technology.

Learn More



Study will Leverage Connecticut HIE to Help Prevent Suicides

A new study will aim to leverage CTHealthLink, a physician-led health information exchange (HIE) in Connecticut, to help identify the factors leading to suicide and to ultimately help prevent those deaths.

Duke Health First to Achieve HIMSS Stage 7 Rating in Analytics

North Carolina-based Duke Health has become the first U.S. healthcare institution to be awarded the highest honor for analytic capabilities by HIMSS Analytics.

NIH Releases First Dataset from Adolescent Brain Development Study

The National Institutes of Health (NIH) announced the release of the first dataset from the Adolescent Brain Cognitive Development (ABCD) study, which will enable scientists to conduct research on the many factors that influence brain, cognitive, social, and emotional development.

Boston Children's Accelerates Data-Driven Approach to Clinical Research

In an effort to bring a more data-driven approach to clinical research, Boston Children’s Hospital has joined the TriNetX global health research network.

Paper Records, Films Most Common Type of Healthcare Data Breach, Study Finds

Despite the high level of hospital adoption of electronic health records and federal incentives to do so, paper and films were the most frequent location of breached data in hospitals, according to a recent study.

AHA Appoints Senior Advisor for Cybersecurity and Risk

The American Hospital Association (AHA) has announced that John Riggi has joined the association as senior advisor for cybersecurity and risk.