Medical Sorcery

Modern healthcare is akin to "witchcraft", according to Vinod Khosla, co-founder of Sun Microsystems and former Kleiner Perkins venture capitalist who recently started his own VC firm (Khosla Ventures). According to Khosla, in Do We Need Doctors or Algorithms, technology will replace 80% of physicians. Machines, which can assimilate large sets of data, can do much of the work of physicians, he argues. Khosla also believes that major disruptions in the health care industry will come from individuals outside the industry. He cites the example of Square, a revolutionary wireless payment system that allows anyone to accept credit cards. How did Square do to the payments industry what many had tried and failed to do? According to Khosla, the background of Square employees made this disruption possible: only 5% of Square's 250 employees worked in the industry.

In the last 3 years, I have worked with dermatologists, oncologists, interventional cardiologists, pain anesthesiologists, and neurosurgeons. I wonder how the roles of these physicians have changed over time, and how they will change in an era where technology plays a larger role in 1) determining therapies and 2) the therapy itself.

As expected, Khsola's 2011 comments drew wide spread discussion and criticism from the medical community.  Many felt that a doctor's intuition and therapuetic touch can not be replaced. In addition, decision support software is less likely to impact the work of surgeons and other procedure-oriented physicians. The work of surgeons has already been aided by the robotic surgery industry (i.e. Intuitive Surgical's Da Vinci robot), but it's unlikely that robots will ever be able to completely replace surgeons (at least not in the next 50 years).

Khosla is not saying anything new. Health care is already becoming more data-driven, and his comments are designed to put more focus into this area. Certain biotech drugs (like Genentech's Herceptin) are based on genetic tests to determine what type of cancer a patient has. In this case, Herceptin is prescribed for patients with Her II+ breast cancer. This type of product would probably satisfy Khosla's desire for "data-driven medicine". Herceptin is just one of many examples of drugs prescribed based on genetic testst.

While I'm confident that we can develop the algorithms to help physicians make the right diagnosis and prescribe the right treatment, the challenge will be gathering the inputs to the algorithm. The inputs will come from potentially numerous tests that patients must endure to ensure adequate information for the treatment algorithm. If some of the data is missing, the output of the algorithm will be less trustworthy and the physician will likely "go with his gut" for the diagnosis and treatment (defeating the purose of the decision support algorithm). So, the question becomes how can we make it easier / cheaper to perform these tests on the patient.  Once the data is acquired, it needs to be stored in a central location which the doctor can access to decide what to do next. The data can even be transmitted back to an implanted devices that change their treatment process based on this information (closed loop feedback).

Khosla's firm has invested in several diagnostic company focused on making data capture easier. For example, AliveCor sells a portable heart monitor that can be snapped onto your iPhone. This monitor records ECGs (Electrocardiograms) and transmits them to your doctor. Currently, AliveCor's product does not provide a diagnosis. However, in time, the device could do this as well. If it does, a significant portion of cardiologists' value would be eliminated. If it does, reimbursement for cardiologist office vists would need to be reduced drastically. In fact, there may be fewer visits to the cardiologist period. Mid-level practioners (who don't diagnosis but who are involved with the therapy) may see an increase in business or expanded roles. Even now, AliveCor's product is bad news for medical device manufacturers of in-office ECGs and for physicians who can bill for performing the ECG.

Making health care more data driven will empower patients as well. As a medical device marketing professional, I'm amazed at how little some patients know about what is happening to them and what is being implanted in them. The advice of the physician is often trusted blindly. Unfortunately, physicians themselves are not always well-educated on therapies, and may make their decisions based on a relationship with a particular manufacturer more than what is best for the patient. Patients, on the other hand, have no other incentive than to choose the therapy that makes them feel better. If patients are given more information on their condition, they may also be more motivated to follow through on the therapy as they get more quantitative feedback on key metrics.

Khosla is right. We should strive for data-driven health care diagnoses and treatments. However, just as auto mechanics have a variety of tests that can be run on cars to diagnosis the problem, the conclusions of the tests are not a substitute for hearing from the mechanic himself. And, for problems with the human body, human interaction is even more craved. Still, when more data is available, patients can play a greater role in their own health.

Innovation through Analytics

Below is a guest post from Aroon Krishnan. Aroon has consulted for a variety of health care organizations on issues such as strategy and forecasting. He is currently a Director of Global Insights for J&J.

Levitra (GSK) launched (in 2005) as a third to market drug in the ED (Erectile Dysfunction) space ( Viagra launched in 1998 and Cialis in 2003). At its peak, Levitra had about 5% market share. The ED Market remains a duopoly with Cialis and Viagra each with about 45-48% share. Why didn’t Levitra ever reach the level of traction with patients and prescribers that Cialis and Viagra received?

Analyst and GSK expected the drug to be a revolutionary billion dollar drug.

Levitra was an “entry ticket” drug:

• It offered no significant clinical benefit over Cialis (Effects of Cialis lasted for 30 hours)
• It lasted longer than Viagra (5 hours vs. 4 hours)
• It was thought to have fewer side effects than Cialis or Viagra

In the eyes of the customer, these side effects were rarely significant enough to invoke a change in brand. The marginal benefit of Levitra to Viagra was undistinguishable to patients. Thus, Levitra lost the clinical battle and never achieved widespread adoption.

The lesson from this case study is that new products can fail when their less important benefits are emphasized to patients and physicians. These benefits fail to trigger an actionable response. Being able to distinguish which benefits matter and the level of improvement desired in these benefits is the greatest challenge drug / device companies face today. In essence, innovation must focus on identifying and delivering on the unmet needs in the marketplace.

Most current approaches used to identify innovation opportunities focus on current behavior. Unfortunately, catering to current behavior does not generate breakthrough products. The typical approach involves asking KOLs, sales reps, and marketing how to improve current offerings. The results are typically incremental improvements and rarely game changing products (i.e. Levitra). GSK must have felt that they had a great product that lasted an extra hour – unfortunately, the market didn’t see it that way.

Alternatively, KOLs, sales reps and marketing say that they want “everything” which creates the ubiquitous innovator’s dilemma.

So how has this process improved in the last 10 years? Analytic organizations in medical device and pharmaceutical companies are using rigorous analytics to understand the relative importance of a laundry list of benefits of a particular device/drug/product to their customers. They have developed new metrics for quantifying the level of unmet need of these product benefits. They have then performed rigorous quadrant analysis to provide R&D with the ‘components’ of a product that must be improved and the “game-changing potential” of that improvement.

In this example below for a new COPD product, R&D is clearly able to prioritize their focus based on product strategy. The upper left quadrant shows the bare minimum a product must deliver. If the strategy is to develop a game-changer: what they must improve in is clearly highlighted. Also, focus areas which are clearly not important to their customers are highlighted. Finally, R&D can focus on high value improvement and stay away from low ROI improvements.