Advances in Connected Health from the MIT Media Lab

By Joseph C. Kvedar
08:24 AM

I’ve always been a fan of the accomplishments of the folks at the MIT Media Lab.  Their accomplishments have been legion, and just a few admirable highlights are:  their ability to continually attract large industrial sponsorships; their open IP model; their collective ability to look 10-15 years out and capture those innovations that will be mainstream at that point; and not least their efforts to add technologies that will aid in the repair of our broken health care delivery system.

I’m excited to say we have a big Media Lab focus at this year’s Symposium.  Two of their current giants, Sandy Pentland and Rosalind Picard, will be speaking  as well as a Media Lab graduate, Tim Bickmore, who is doing amazing things with his program at Northeastern.

These three have something in common that excites me.  They are exploring the boundaries of technology in its ability to substitute for humans in the caring process.

Sandy has been prolific in commercializing his work, first by founding Cogito, a company whose goal is to pick up your mood state based on a 10 second voice print.  Subsequently, he launched, which is founded on the basis that all of those electronic bread crumbs you leave behind each day (GPS data, who you text, who you call, where you spend money and what you spend it on, etc.) can be analyzed to come up with a unique health behavior print that can aid in both public health applications and in encouraging you to improve your own health.  I recruited Sandy to be on the Scientific Advisory Board of our Center’s spin out, Healthrageous, because I value his perspective on this topic of disparate data collection and analysis.

Roz has been working for years (she was Tim Bickmore’s adviser) on ‘affective computing.’  She is also a founder of Affectiva, a company whose technology can objectively assess your emotional reaction to stimuli via a wrist sensor and face-recognition technology.

Tim’s team continues to impress with their implementation of Relational Agents, software agents that are designed to exhibit relationship-building behaviors.  His most recent achievement is showing that, strategically used, these software agents can participate in health care delivery and that patients prefer them to doctors and nurses.

So how do they all fit together?  As the demand for health care services (largely driven by lifestyle-related chronic illnesses) skyrockets, we don’t have enough providers to meet the need and it doesn’t make sense to try to train or import them.  We’re also trying to control costs, 60-70% of which are related to human resources, so adding more will not solve that problem.  We MUST adopt solutions that increase self-care and spread providers across larger groups of patients.  However, for a number of cogent reasons, we can’t abandon our roots as a caring profession to pull this off.

Think of it another way.  To get the most out of our providers we will need to look beyond the traditional office visit.  Expecting patients to meet one-on-one in a physical location creates a true bottleneck for health care delivery.  However, using today’s technology, while I can get a near continuous read of your weight, blood pressure, blood glucose, activity level and some data on your sleep, I can’t put those physiologic measurements in any kind of emotional context. That is arguably one of the most important aspects of what the doctor does in the office exam.  Systems like Cogito and Affectiva are the beginning of providing the emotional sensing to enable the doctor to get a full picture of your physical and emotional state, in a time and place independent manner.

The other side of the office visit is caring for the patient.  There are myriad of data showing that a patient that feels cared for will do better and that the strength of the patient-provider relationship is important in health outcomes.  So how do we extend our providers across greater numbers of patients while preserving that caring bond?  Undoubtedly it will be a challenge. But Bickmore’s work showing that patients prefer relational agents in certain care settings is an encouraging first step to getting us there.

As I said, the Media Lab is leading the way in creating these technologies and showing their early phase potential. We at the Center for Connected Health are pleased to have a strong showing from their faculty at this year’s Symposium.


Joseph C. Kvedar, MD is the director of the Center for Connected Health (, a division of Partners HealthCare in Boston. Connected Health is focused on developing new methods of delivering quality patient care outside traditional medical settings. This post appeared at The cHealth Blog.