Analytics is the Driver for Useful Health Services at Philips

Although pioneer health care organizations jumped into risk sharing and fee-for-value without a good grounding in data, they have come to recognize the critical role analytics play over the past few years. Results of a 2015 survey of ACOs show that, “Nearly 85 percent of respondents report they have in place advanced analytics software to analyze disparate data sets.” The data tends to be limited (mostly claims and EHR data) and data sharing is still rare between organizations, but basic practices such as identifying high-risk patients are becoming more widely seen.

Philips has been in health care informatics for a long time, gradually building a data platform with analytics capabilities and basing more and more services on this platform. I talked to Dr. ck Andrade, Director of Product Management of Philips’ HealthSuite Digital Platform, to find how their basic analytics drive their services.

Like the ACOs mentioned earlier, Philips allows a single organization to combine and mine data of different types. Philips does not combine data from different unrelated organizations–in fact, to respect privacy, they don’t peek at user data at all. The platform is intended to aid institutions with precisely the types of data integration that are now so difficult. Now it is being incorporated by Philips into their own high-level services, showing how analytics can be a platform for building businesses.

Philips’s HealthSuite digital platform offers FHIR APIs. EHR data is read in through the vendors’ APIs when they’re available, by using the platform’s other interoperability capabilities, or through the CCD-A format. Imaging support was announced on February 18. Genomics is being pursued. Finally, device data can be taken in through several sources. Philips Device Cloud manages 8 million connected devices today, and a recently announced integration with Validic connects to data from 130 different device vendors over a wide range of protocols.

Clearly, all these data sets are interdependent. For instance, an image is of no value without the patient history that comes from an EHR.

What sorts of questions can all this data answer? The Philips platform provides a framework for aggregating data, running analytics, and exposing results through an API. The same API is used internally by Philips to develop its solutions, by customers to write apps, and by third-party developers to develop clinical solutions or packages for healthcare analytics: for instance, data scientists testing predictive models. As an example of the API’s power, it can offer access to blood glucose, wellness measures, responses to past medications, mood, and stress for diabetes patients. Healthcare organizations can run their algorithms against this data to suggest the current insulin dose and track fluctuations in glucose level.

Most customers use such information for simple interventions such as letting someone know they forget to do a reading or that the reading is outside the normal range. The platform can find patients with similar demographics and find duplicates caused by such common errors as misspelled names. A more sophisticated use of analytics would check how people are responding to medication, or how different interventions produce different effects.

The API supports both data push and data pull. Pull may be chosen for data that needs to be read several times a day.

Now Philips is enjoying the fruits of its labors by offering services based on its analytics, which are constantly getting richer. Here are examples operating at three levels of care:

  • Inpatient: Philips’s eICU tracks patients from ICU through follow-up. Tools provided by Philips take in, analyze, and form data into visualizations on dashboards.

  • Outpatient: Philips’s in-hospital and ambulatory telehealth programs are aiding transitions. Data from the inpatient EHR can be connected to data from the patient at home and from different health care providers using Philips services. Patients can upload data, using eCare Companion, allowing providers to monitor them using the same model as inpatient care. This is crucial for outpatient care, where each care coordinator might have to monitor hundreds of patients. A dashboard, eCare Coordinator, organizes critical information for the clinician. For instance, a dashboard can highlight the five people with the most disturbing trends. If a patient’s blood pressure rises even though he reports taking his medication, a clinician can prescribe a medication change.

  • Consumer health and wellness: Healthwatch uses the same analytics as other services, but for healthy living instead of recovery from illness. For instance, analytics can track different types of vital signs for people who have been identified as prediabetic. A self-management platform, which offers instant feedback as well as a look at progress over time, can measure activity and other contributors to health. Although it can run separately from any health care provider, users can share logged data with their providers.

Philips also builds some services on others. For instance, the Lifeline program for fall detection, which has been available for some time, now uses its Caresage predictive analytics for the frail and elderly. This turns Lifeline from a reactive to a predictive platform. Using analytics on a person’s frequency of falls, and patterns in their incidence, it can warn if another fall is likely.

March brought with it announcements for a whole set of new services were announced, such as for sleep and respiratory problems, for healthy seniors, and for intensive care units. I believe these advances aren’t due merely to Philips’s size and investments. They have learned how to make use of a flexible, integrated platform. It’s the direction all health providers need to head in.

About the author

Andy Oram

Andy Oram

Andy Oram writes and edits documents about many aspects of computing, ranging in size from blog postings to full-length books. Topics cover a wide range of computer technologies: data science and machine learning, programming languages, Web performance, Internet of Things, databases, free and open source software, and more. My editorial output at O'Reilly Media included the first books ever published commercially in the United States on Linux, the 2001 title Peer-to-Peer (frequently cited in connection with those technologies), and the 2007 title Beautiful Code. He is a regular correspondent on health IT and health policy for He also contributes to other publications about policy issues related to the Internet and about trends affecting technical innovation and its effects on society. Print publications where his work has appeared include The Economist, Communications of the ACM, Copyright World, the Journal of Information Technology & Politics, Vanguardia Dossier, and Internet Law and Business.