Boston Children’s emerges as digital health innovator and accelerator

Pediatric hospital focused on tools that leverage social media data as well as artificial intelligence, clinical decision support and predictive analytics.


As one of the largest pediatric medical centers in the country, Boston Children’s Hospital has gained a reputation as a nationally ranked healthcare organization. But, it’s in the area of digital health technology that the provider is setting itself apart.

The pediatric hospital has developed a host of online and mobile applications aimed at both clinicians and patients. The driving force behind the technology development is John Brownstein, chief innovation officer at Boston Children’s Hospital, whose job is to help shepherd promising technologies through the commercialization process and to promote a culture of innovation within the organization.

A pioneer in the field of digital epidemiology, Brownstein manages a 50-person research team focused on digital health apps that leverage social media data as well as artificial intelligence, clinical decision support, and predictive analytics.

“We’ve doubled down on machine learning and data science, which are going to be so critical to the future of care that we now have an incredible team of engineers and data scientists that are really thinking about how to best use our data, how to extract it, and make it more available,” Brownstein told the Machine Learning in Healthcare: Industry Applications conference on Wednesday in Boston.

Also See: UCSF bolsters digital health, clinical informatics faculty

To address the problem of overcrowding in the emergency department, Boston Children’s is piloting the Prediction of Patient Placement (POPP), an advanced, real-time forecasting tool that can predict the likelihood of admission from the ED within 10 to 20 minutes following triage.

POPP is meant to transform a static predictive model into a real-time capacity prediction dashboard, giving clinical staff visibility into incoming admissions from the ED by predicting likely admissions before the staff have completed their evaluation by taking information from the initial triage. By proactively and intelligently managing the flow of patients into the ED, the hospital hopes to achieve shorter wait times and improved outcomes, according to Brownstein.

In the area of chronic disease management, Boston Children’s has also created a web-based platform called TriVox Health that enables providers to track and manage their patients by providing remote monitoring, real-time analysis, and tracking of patients’ disease symptoms and response to therapy over time.

TriVox Health enables clinicians to use electronic surveys to gather data remotely from multiple responders, including patients, parents, school personnel and ancillary physicians, and view the responses in a timely manner using graphical, tabular and natural language summary formats.

While data silos continue to plague healthcare and data accessibility and quality are significant challenges to the industry’s adoption of machine learning, Brownstein contends that providers can more easily leverage machine learning-based approaches to tap into the vast amount of social media data for the purposes of public health monitoring.

“This led to the creation of a system that we developed called HealthMap, which is a global tracking system,” he said. The system uses disparate data sources such as online news aggregators, eyewitness reports, expert-curated discussions and validated official reports for disease outbreak monitoring and real-time surveillance of emerging public health threats.

Developed by a team of researchers, epidemiologists and software developers at Boston Children's, HealthMap monitors, organizes, integrates, filters, visualizes and disseminates online information about emerging diseases in nine languages through an automated process that is updated constantly. As a result, the system is leveraged by public health organizations globally, including the World Health Organization, Centers for Disease Control and Prevention and the United Kingdom’s Health Protection Agency.

HealthMap has created several spinoff web-based public health surveillance informatics projects, including Flu Near You, which provides local health reports on influenza outbreaks, and HealthMap Vaccine Finder, a free online service where users can search for locations that offer immunizations. Originally, Google created a Flu Vaccine Finder in response to the 2009 H1N1 influenza pandemic, but it later transitioned it to HealthMap, which then morphed into the HealthMap Flu Vaccine Finder.

In a related project, a HealthMap Foodborne Dashboard leverages social media data to help public health departments improve foodborne illness surveillance and reporting. In real time, the dashboard enables public health officials to identify and geo-locate food poisoning complaints within their local jurisdictions. The platform is currently used by multiple local and national public health agencies.

Likewise, Thermia is a website administered by HealthMap that serves as a decision support framework for fevers and associated febrile illnesses, enabling patients and families with support tools to better manage and treat fever and sickness. The website collects anonymized data that contributes knowledge of health trends in specific areas.

Similarly, Boston Children’s is working with a company called Buoy on a symptom checker that leverages artificial intelligence to deliver a personalized and more accurate analysis of symptoms. An algorithm analyzes thousands of real world data points drawn from the same medical literature physicians study.

“A chat box interacts in a conversational way,” according to Brownstein, who likens the experience to chatting with a doctor.

Another project called MedWatcher tracks and reports side effects for drugs, medical devices, and vaccines. Users of MedWatcher are notified of any government safety alerts and have access to information about side effects that consumers have experienced from medical products.

“With natural language processing, you can actually come up with a very clear understanding of drugs and adverse events,” said Brownstein.

Created in collaboration with the Food and Drug Administration, the MedWatcher system was run by a Boston Children’s spin-off informatics company called Epidemico, until it was bought by management consulting, technology and engineering services vendor Booz Allen Hamilton in 2014.

More for you

Loading data for hdm_tax_topic #better-outcomes...