By Larry Sivitz
If knowledge is power, including the power to heal what ails us, data is its lifeblood. But when the knowledge being pursued involves the body of information buried in the massive health care industry, the Web searcher is in need of a lifeblood transfusion.
Surf among the current crop of health care sites, community venues, Q&A sites and specialty journals and you’ll find the equivalent of a medical encyclopedia, or, at best, a hypertextual wikipedia. Isn’t that simply the Web 1.0 equivalent of turning paper into pixels? The Web 3.0 challenge for semantic search technologists and medical experts is to transform the silos of health care data into relevant knowledge circles informing patients about not just what works, but what works best.
All eyes are on Medify, the new search engine launching out of beta this week, as it excavates the world’s largest datamine to deliver data on health care conditions, patient experiences, clinical trials and institutions, and then filters the outcomes by patient profile. Medify is a search engine that could save lives!
The degree of datamining demanded of Medify requires the kind of talent and technology that Seattle companies have historically brought to fields like ecommerce with cost-comparison tools like NetBot, or the predictive data trending algorithms that track daily changes in global airfares a la Farecast, (another University of Washington brainchild). Next, factor in the behavioral analytics’ number-crunching of two more Seattle forebears, AdRelevance and AdReady. It is precisely these challenges and these very companies that make up the DNA of Medify co-founder and CEO Derek Streat, along with co-founder and CTO Jay Bartot. Both are leading the Medify team to do the same thing for health care data, and in a region where the last search engine re-launched from Microsoft branded itself as the “Decision Support Engine.” Great minds think alike.
To Medify, the medical ends worth achieving (healthier outcomes) justify the means (empowering patients with more relevant case studies to have more educated conversations about medical treatments with their doctors). The timing could hardly be better. Just two weeks ago, Google pulled the plug on its Google Health program designed to help users manage their personal medical records. Microsoft’s Health Vault has suffered a similar lack of enthusiasm or consumer uptake. The impression is that people do not have the interest, the confidence or the know-how to compile and manage their own medical records. Medify is seeking to bridge the chasm between real patient needs and real-world medical know-how to cut to the core issue…what are the most successful outcomes to today’s most pressing health concerns and where are they taking place?
We sat down for a ShopTalk session about medical knowledge, patient empowerment and how Medify is connecting the dots.
Seatle24x7: Derek, you’ve been a serial entrepreneur working with “big data” companies, and so has your leadership team, grappling with databases like the behavioral analytics of an AdRelevance or a Farecast. What was the lynch pin that connected you to health care?
Streat: Two things ultimately brought me to health care, at both a professional and personal level. First, we were looking for the size and scope of this kind of challenge — a data intensive space like heath care, that would tap the engineering talent we have been applying to other data-driven industries, where there were big problems to be solved. Here was data that we could point our technologies at and bring to the surface. On the personal side, I have a daughter who, when she was 2 years old, presented a serious auto immune condition ultimately resulting in a kidney transplant about 9 months ago. Very suddenly, we were thrust into a world where we faced many choices. We could be followers, relying on our doctors for every bit of information, or we could be more collaborative, informing ourselves about what avenues were available to us, and learning what options might be helpful to our daughter to share with her doctor.
By getting involved in understanding the medical environment for our daughter’s condition, we began to develop a different relationship with the doctors. We are fully convinced this resulted in better care for her.
Our family situation parallels what goes on in homes across America. 85% of the general public starts out at a Google search box, typing in scary words from their doctor. It doesn’t take long before they are up against data that seems impenetrable, or nearly impossible to understand.
Seatle24x7: Talk about vital statistics, the amount of data that is published in the medical world is truly staggering. What are the datamines you are tapping into?
Streat: We’ve concentrated our technology on the largest data research set that is available to anyone in the health care field today. It’s called MedLine and it is managed by the National Institute of Health. There are 20 million pieces of research in this data set, and it’s growing by over 5000 cases per day. The curve is steep and includes many other countries like China and India.
The data we are analyzing is unstructured and requires that we process and parse it differently. What is critical in the medical context is that we can map the data to different patient profiles. So we are looking for the specific characteristics of the people who were being tracked in different research studies. There are a number of layers – for example, demographics, treatment regimens, and the symptoms that were reported and experienced. This all gets assimilated into Medify.
Seatle24x7: So you’re doing much more than aggregating data. You’re filtering the data for particular groups of patients, matched up with the corresponding treatments?
Streat: We are looking for patterns in the data and those signals that are directionally predictive. Over time, these patterns come into sharper focus. The profiles of the patient are very important – what may work for one group or one individual will not automatically work for somebody else. By placing these findings into context we are hopeful that people will reach a better understanding of what the future might hold.
Seatle24x7: Let’s say a patient is facing a diagnosis of Melanoma, or skin cancer. Medify can inform about what modes of chemical therapy are available and which fair better for the patient’s condition and the patient’s profile?
Streat: That is the path we are on. And we take it one step further. You can look at all of the people who have Melanoma who have been studied. And out of the millions of cases that are in the system, you can start to further refine that set to find people that are more like you, by age, by gender, by region and so on. For instance, you may want to look at a segment that had a history of other conditions, a possible cause or a side effect or a particular treatment regimen. As you’re filtering down, you’ll see the number of patients getting smaller and smaller, until you get to a group that is small enough to be relevant, and still large enough to have data that is conclusive to your particular situation.
Seattle24x7: What kind of data are you aggregating? Do you mean clinical trials?
Streat: It’s actually well beyond just clinical trials, which is what makes this data set so interesting. For instance, the biggest clinical trial site right now is clincaltrials.gov. There are about 90,000 clinical trials represented there. In Medify, we have stored nearly 20 million pieces of research in a database that is growing by over 5000 a day. It includes clinical trials. But let’s imagine there’s a doctor at Seattle Children’s Hospital who has only one patient where he observes an odd reading associated with a particular treatment that is being administered for a particular condition and nobody’s ever heard of this outcome happening before. He decides to write an article about it. We have compiled that as well.
There are 30 to 50 million people in this country that are spending a lot of time online searching for answers and attempting to compile this information. The question for us is once you have the information, what can you do with it? The resources that are available to users today are inadequate. I can get overly-generalized, high level stuff from Web MD, but you end up reading through 50 different ads.
If I try to navigate inside of MedLine itself, and the NIH database, it’s very difficult to gain actionable knowledge. By presenting this information to consumers in n easy-to-read visual form, I can see at a glance where my medical treatment exists in the bigger picture.
Seattle24x7: Your Graphical User Interface presents results in an interesting way. The presentation resembles a cross between a Venn diagram and a Google Plus Circle?
Streat: We’ve placed a lot of emphasis on rendering the data in a visual format. For example, if my child has autism, I can look at treatment options for his autism in a visual format that places it in the context of the most popular or relevant drug therapies available. The graphic indicates that this protocol falls flat in terms of its effectiveness. In the next diagram, I can visualize where the non-drug therapies break out. For example, for autism specifically, it looks like music therapy has proven to be productive in reducing symptoms and it gives me another alternative treatment to try before immediately falling back on antidepressants.
Once you start finding alternatives or therapies that might be complimentary to the current standard of care the doctor may have prescribed for you or your loved one, you can, at the very least, have a more informed and collaborative conversation with your doctor about additional options.
Medify includes features that allow users to be alerted to changes. I may “pin” one particular therapy to a record and Medify will notify me about the changes in this treatment, or if the treatment falls below the effectiveness index relative to another.
Seattle24x7: Do I understand correctly that Medify will also track performance and outcomes at health institutions like hospitals and clinics?
Streat: Yes, Medify can rank institutions based on their effectiveness. So, I might see that University of California is only a 2 hour flight away, and they’ve actually seen more patients like my child. I might choose to use that new found insight to reach out to my social networks and allow them to push it to their networks to see if anybody has a connection at the University of California that might help cut through some red tape and help me get a second opinion. We view this “insight discovery engine” if you will, as a catalyst to helping a patient or a caregiver become more empowered and informed.
Seattle24x7: Health care reform has been among the most consequential and controversial issues of recent years. Do you see the information you are gathering on what is effective and what isn’t, including which institutions are the most successful and efficient will yield insight into which hospital or type of therapy is the most efficient or cost-effective?
Streat: We do! We didn’t set out on a path to change the two trillion dollar healthcare system in this country. But we’re already finding corollaries to the broader health care system based on what we’re doing. For example, a health care center can market all day long and, buy a bunch of billboards to try and convince me that they are the best cancer center in the country. But when I look at the Medify data I see that the MD Anderson Cancer Center at the University of Texas in Dallas is the gold standard and is at the cutting edge. That’s real data that is hard to ignore!
Another area of interest stems from understanding cost options and cost containment. If a treatment is costly and has not produced results, that is a decision to be aware of in terms of the cost of time and money involved.
Seattle24x7: How much will it cost to tap into Medify as an end-user?
Streat: We’re launching as a free service to users. We think the public is fed up with clawing through 50 different ads to read the same old text they would find elsewhere. We’re adopting a more transparent and open approach. If, and only if, a user wants to get more information from a service provider we’ll provide a path for the user to go on and do that.But our data will always be 100% unbiased.
We said to ourselves, let’s not go out there and create another company that’s going to try to change behavior or game the system. Let’s build a simply superior service to what’s out there right now. So far, we’re finding engagement levels that are much better than we thought possible and it’s going very well.
At the end of the day, we understand this is one big mountain peak to climb and in the final analysis it all rolls down to where the patient is standing. So we are focused on user experience first and foremost. Otherwise, it’s just pushing the same rock up the same hill. Medify thinks it’s time to think differently! [24×7]
Visit Medify at http://www.medify.com
This interview was recorded and automatically transcribed using Cogi, the amazing time-saving solution for interviews, meetings and brainstorming in the office or on-the-go! Contact Cogi today for more info!
