This week I had the opportunity to participate on a lively panel at General Assembly Seattle organized by Seattle Health Innovators, and moderated by Corinne Stroum of Caradigm. Fellow panelists included Randy Wise formerly of Group Health and now at EveryMove, Ang Sun of Regence/Cambia, Lifesprite founder Swatee Surve, and Daniel Newton of Accolade.
Corrine sent us a series of great questions in advance, and we had a rich discussion and so many questions from the audience that we didn’t even get to half of them. It’s a big topic, and with payers, providers, and technologists on the panel there was a lot of opportunity for broad perspectives. There’s a discussion of having a follow-up to this panel to continue the conversation—stay tuned for more on that. The general themes of the discussion included the value of big data to influence individual health with examples like the quantified-self movement, but more generally how our ability to collect and analyze can lead to more personalized and better healthcare.
At Wellpepper, we have a lot of data to analyze. As Wellpepper CTO Mike Van Snellenberg pointed out in his Stanford MedX talk and I’ve also talked about in this paper in The Journal of MHealth, having data provides an opportunity to get answers faster than using the traditional scientific method. Rather than formulating a hypothesis, setting up an experiment, collecting data, analyzing the data, and then going back to the drawing board if your hypothesis is not born out, data enables you to ask a series of questions and get immediate and sometimes surprising answers.
The panel kicked off with the sharing of some surprising things that we’ve found from the data, ranging from which mental health tools were favored by different populations to the ability to predict hospital readmissions. In addition to finding trends from explicit patient input, we also discussed the ability to draw insight from activities including social media and mobile usage patterns. Swatee mentioned the Instagram analysis that showed color scheme on photos was a predictor of depression.
The ability to combine both passive and active patient-generated data, and draw conclusions from broad date sets these data sources can help to deliver better care – resulting in what Daniel Newton referred to as “small data.” That is, I’m going to learn as much as I can about you, and then tailor care to you, which is the approach Accolade takes.
As with any talk on tracking and data, questions of privacy came up. While all the panelists thought that there have become standard terms for people to opt-in to sharing health data, describing the use of that data was deemed important. At this point, Ang Sun from Cambia (who admitted that, as a healthcare plan, they had a heck of a lot of data on people), mused that he wished his physician knew as much about him as Google did. Generally, there was consensus that, if the purpose of the data sharing was for connecting people with the appropriate healthcare services, people would opt in.
Our panel was pretty aligned on the idea that there is big value in big data for healthcare, but that the general applications and usage are still in early days. First, there are the privacy concerns and even laws. Second, current healthcare organizations using this first generation of EMRs have limited ability to look at aggregate data for trends. However, with new technology and personalized approaches to care, we see great promise in big data and predictive analytics for healthcare.