Illustration by Davanzo Design & the author

Data Empathy: Learning from health care

What the dinosaurs of health care can teach tech start-ups

Nude pictures of celebrities stolen from their own iCloud accounts. Facebook experimenting with the emotions in their users’ feeds. Google reading Gmail before their users do. Fitness trackers without privacy policies, vulnerable to security breaches, and bait-and-switch tactics to sell customers’ data. Almost every day there is a story about the gap between the expectations people have for their own data and what companies actually do with that data. To fix this gap, we first need to rethink the nature of data.

Last year, technology journalist Om Malik called for “data with empathy,” saying “data without a soul is meaningless.” He was arguing for companies to use customer data to offer better emotional experiences. Companies “should start by thinking, how do we make our customers feel special?”
Empathy is the ability to understand and share the feelings of others. Right now, data-driven companies are showing very little empathy for their customers and their customers’ feelings and expectations of their data.

Rather than make them feel special, they literally make them feel like one of a powerless number. That’s how I felt when hours after an earthquake shook Northern California, Jawbone released graphics of their users’ sleep data. I use a different tracker and live in a different city, but Jawbone’s move still stung. There’s nothing to make a customer special about seeing a graphic quantifying collective anxiety, especially one based on sleep tracker data. Jawbone’s widely circulated graphs reminded everyone that they know exactly when their users have sleepless nights, which is the opposite of empathy. Ditto Apple’s I-told-you-so statement on two-step verification in the wake of the celebrity nude picture scandal. As Janet Vertesi pointed out, Apple, Facebook, Google and scores other cloud and social media companies work hard to get their users to share data in communities, relationships, and groups, only to place the security and privacy onus almost entirely on individuals as customers. They use rhetoric like “The one place for everything in your life,” or “The app displays your data.” Is it any wonder we’ve come to think of it as the data as ours, even when for practical, commercial, and legal purposes it is not?

With Brittany Fiore-Silfvast I studied how such data might transform health care. We found that technology companies have a lot to learn about data empathy from the way the health care industry treats patient data. One of our key findings from this research seems at first blush so obvious that it is hard to believe that it has been overlooked in the design and implementation of health care information technologies. Namely, people expect data to work in very different ways. Doctors, patients, and healthcare entrepreneurs all value data in very different ways for different goals. Doctors want data that help them make good clinical decisions. Patients want data that help their stories be heard. Both want to see control and transparency in who can access data. Health care is a great place to see the importance of what Helen Nissenbaum has called contextual privacy: I may want my health data shared widely among my care team, but still closely protected from advertisers, employers, or third-parties. Context matters for privacy.

Contrast physician’s health data practices to the over-hyped slogan: Data is the new oil. For doctors, data are useful resources only when they lead to clinical decisions and actions. These are two very different sets of expectations for data. One says data have inherent value that can be mined —data mining can lead to something useful and valuable. The other expectation values data for its potential for action. As Jen van der Meer has argued, “data is not a business model,” it is only the commodity — “Action based on wisdom is the valuable resource.” In health care, data need both interpretation and connection to be useful. In other words, data is the new hydropower, always made and never simply found.

Data can be a powerful starting point for a conversation between physicians and their patients. In a MedicineX presentation with Dr. Anthony Back, we call on physicians to take up the new role of using data to create meaning with their patients. In this, we’re recognizing that the process of making data is always social. Empathy is not asking “What are we going to do with all this data?” The better, more empathic questions are data for whom, data when, and data why?

Take this example from the Collaborative Chronic Care Network or C3N, a project of Cincinnati Children’s Hospital. At the core of their philosophy is building a data ecosystem that is meaningful, valuable and useful for individual (pediatric) patients and their doctors, as well as researchers. They put co-design into the very core of their project. They view data as opportunities to build communities to help kids with chronic diseases get better. Here’s a quote they list on their website as inspiration: “Once you open the possibility that people are not only using the web as a platform to produce their own individual content, but also to pool their efforts, knowledge, and resources. . . the possibilities for what they can create are astounding.” That’s from Harvard social media theorist Yochai Benkler and it answers the key questions about data empathy: for whom, when, and why.

Data is not a generic, off-the-shelf product. It doesn’t come prepackaged, clean and ready to use right out of the box. Our personal data shouldn’t be the same as laundry soap, an interchangeable product looking for a market. Currently, we don’t have very many good bridges for data to cross the different sets of stakeholder values in health care among patients, clinicians and researchers. To do that, we’re going to need to design for data in real social settings, for real practices. And design for data empathy.

I have a few ideas for companies looking how to get there. First, we need to have transparent conversations on what data actually do as well as what different people with different sets of values think their data should do. What is valuable about data can be apparent if we have empathy for what data mean for different people. When we do that, we can start to use data for their potential to build and foster communities. We can’t get there if companies continue being tone deaf about people’s practices with data and what they value in their data.

Acknowledgements & Credits: This essay is based on a talk that I delivered at Microsoft Research New England’s Fifth Anniversary Symposium, October 7, 2013. Many thanks to Valerie Farber for coaching me through the process, to Brittany Fiore-Silfvast for research assistance and inspiration, Dawn Nafus, and Intel for funding as part of the Consumer Biosensors program.