“Nobody really cares about me, I’m not Beyonce.”
This quote was attributed to Deborah Peel, founder of Patient Privacy Rights, in a Bloomberg article about data mining of online patient conversations. Peel’s point is that though people generally understand the public nature of their online conversations, they may not realize the extent to which their conversations are being monitored, packaged for consumption, and sold. As data-enabled Internet business models continue to emerge, so too will new questions, ethical and otherwise.
Evolving Business Models
Treato, which is the focus of the Bloomberg article, is a big data company evolving one of these new business models. The company’s software scrapes tens of thousands of online patient discussions daily, aggregates that data, and then analyzes it for customers. The final product is insight into trends about how drugs are used and what problems consumers experience with them. Until recently, Treato’s customers were primarily healthcare and especially pharma. One pharma client, for example, uses Treato to understand the patient journey, particularly patient concerns as they move from diagnosis to treatment.
Now Treato is pursuing a new customer: Wall Street. Treato sends fund managers regular reports summarizing online chatter about drug side effects or prescribing trends. Wall Street interest in this area is certainly not new. Treato’s predecessors built businesses by putting investors in touch with health professionals and researchers. But the data-enabled Internet-centric approach to gathering this insight is relatively new. By using software to pluck data directly from patient conversations, Treato has access to both more and different data than has been previously available.
Enabling and Informing Patients
Though Treato’s business model is focused on healthcare and Wall Street, it also offers a free service to benefit patients. Using a simple search box on the Treato homepage, patients can search medications or conditions and be presented with related trend data. Depending on the patient’s search, the trend data might include sentiment or common concerns about a medication. And if the patient’s search has two medications, the trend data are presented in a comparison table. Patients even have the ability to view the individual patient conversations that are being incorporated into the aggregate data.
Tools like Treato are exciting because they enable patients to create new knowledge and learn from each other. Traditionally, healthcare knowledge has been passed down by authority figures like healthcare providers, while Treato and similar technologies enable healthcare knowledge to bubble up from the crowd. Now no one need be Beyonce for their perspective to matter to others.
In a recent article, Susannah Fox described her experience going down an emergency airplane slide. “The people who had really come through in the clutch were my fellow passengers — my peers — not the professionals.” Fox then related this experience to navigating healthcare:
“Call me an optimist, but I think empathy and peer leadership can spread, especially if we find ways to share what we know and allow people into our lives. I think we can handle the responsibility of knowing how important each of us can be to one another. I saw it on the plane that day, and I see it among the patients and caregivers I follow in my research.”
Challenges of New Business Models
Treato and similar companies offer exciting potential for healthcare, but their data and algorithmic accuracy can come into question. The Bloomberg article cites financial experts who are skeptical of Treato’s ability to predict Wall Street outcomes. And Google’s Flu Trends project, for instance, has been strongly criticized by academics for being inaccurate. Google recently announced that it would begin incorporating Centers for Disease Control (CDC) data into its predictions. Previously, Google Flu Trends algorithms relied solely on Google’s search data. Given the troves of proprietary data accessible to Google, this big data development is one worth noting.
Even with high-quality relevant data, the algorithms used to produce insight from that data can present difficulty. A recent Forbes article titled “The Formula: Do We Place Too Much Faith in Our Algorithms?” discussed these algorithmic challenges:
“The biggest risk of this algorithmizing of everything that Dormehl describes is the faith we all place in these experiments. Jaron Lanier articulated this in his book You Are Not A Gadget, writing that, “Turing test cuts both ways. You can’t tell if a machine has gotten smarter or if you’ve just lowered your own standards of intelligence to such a degree that the machine seems smart.” Because our experience of software is seamless (when it works!) we easily assume that the inner workings must be perfect as well.”
And then there are ethical challenges, as raised by Deborah Peel in the Bloomberg article. A common criticism of data-enabled Internet businesses is that if you are not paying for something, you are the product being sold. This criticism is perhaps an oversimplification of the issues involved, but it does highlight an important point. Data is not just data if it’s attached to people. And when data is attached to people, new ethical considerations come into play, regardless of who is paying for what. Institutional Review Boards (IRBs) are tasked with overseeing research involving humans for this very reason.
Treato’s business is primarily in mining public conversations, but similar companies are storing and analyzing far more private and sensitive material, opening up even trickier ethical questions. 23andMe, which we previously wrote about, processes the DNA of consumers who send in a saliva sample via a mail-in kit. As board member Patrick Chung told Fast Company, “The long game here is not to make money selling kits, although the kits are essential to get the base level data.” The goal, rather, is to gather data using these kits to do medical research. In the same Fast Company article, 23andMe founder and CEO Anne Wojcicki said that the company wants data from 25 million people, a scale which could open up great research discoveries. But even great discoveries must be balanced with the rights of the research participants who contribute to them, as we learned in the case of Henrietta Lacks. New business models like that of 23andMe will create new ethical questions as to what that balance should be.
Potential for Clinical Research
Given the potential of data-enabled Internet business models, it’s worth rising to these challenges. We are especially excited by the prospect of new models being applied in the context of drug development. For example, imagine how patient conversations might be mined to help us better understand the patient journey as it relates to clinical trial participation. Or, as we wrote about recently, imagine how Internet-based studies might make clinical trials more accessible in the future. We have new opportunity not only to improve the patient experience, but also to speed enrollment and the overall drug development process.
What do you think about emerging data-enabled Internet businesses? How can we seize these opportunities in clinical research? What challenges might we face?