Data visualizers and creative coders take note: healthcare is ripe for disruption.
I spent last year thinking about data visualization, healthcare, and sensors. In Fall ’11, I had the good fortune of spending the entire semester charting and mapping, in 2D and code, with Feltron. As a student and intern, I’ve visualized a variety of data: NYC Data Mine sets (e.g. MTA turnstiles activity and cancer rates correlated to the location of Superfund sites across New York state); Nike+ data generated by one thousand New York City runners; and user generated data from the Web and social media (e.g. whatthebook.org). In the Spring of ’11, I dove into the healthcare domain while working on semester-long projects: Sidekick and Thrive (Fig 1.). Finally, I spent the Summer of ’11 researching spimes as a primer for my thesis work. For thesis, I will visualize personal health data tracked by devices such as the Nike+, the recently defunct Up, and the Fitbit.
Health record data generated by the healthcare industry will never be ours. Neither will the data generated by the devices implanted in our bodies. Case: graphic designer and patient advocate Hugo Campos’ unsuccessful attempt to obtain the data generated by his cardiac defibrillator. The benefits of patient ownership of health data are clear. Case: Interaction designer Kate McCurdy visualizes her history of contending with Myasthenia Gravis (Fig 2.) as a diagnosis and treatment aid to her physician. Some of the data that McCurdy visualizes is constructed from memory, yet is effective. I wonder: how much more robust would the visualizations be if her health history was generated by a device—such as the Fitbit—that tracked continuously and accurately?
If our health record and implant data will never be ours, we must sidestep the healthcare industry altogether and generate health data ourselves. As personal health trackers increase in capability and resolution, and, in aggregate, we are able derive meaning from what is collected, the demand for products serving those practicing preventive health and patients will rise, in the same manner, GPS gave rise to location- and navigation-based products (and media and so on). Those best suited for this task will be the data visualizers (statisticians and designers) and creative coders who are able to create interactive and animated, useful and beautiful visualizations of personal health data for the delivery of care and behavior change.
Last week, I created a paper prototype and conducted an eight-user test (from February 2–3) for thesis. I edited the video documentation this week and posted it on Wednesday. (Fig 3.)
The hypothesis: encoding the physical world with data can create awareness, meaning, and behavior change. The results: mixed but positive.
All users correlated their health behaviors to the changes in the tomato display: healthy behaviors had no affect on the display, unhealthy behaviors visibly ripened the tomato, creating awareness.
The display meant different things for users. For some, the display was a manifestation of the current health state of their bodies: if they exhibited unhealthy behaviors they expected the variable tomato to ripen; if they exhibited healthy behaviors, they expected the variable tomato to “un-ripen”. Others derived the intended meaning of the tomato: unhealthy behaviors accelerate the aging process, which is, of course, a uni-directional process. The results confirmed a suspicion that I previously held: some narrative (as complex as a video; as simple as a label) will need to accompany the final prototype in order to make the intended meaning of the display clear.
The tomato display was not effective in creating behavior change. No surprise, but worth establishing.
I conclude with a simple diagram (Fig 4.) that I conceived of when mentally situating this work: a series of opposing scales.