Why Content Isn’t King in Healthcare (and how to build platforms that matter)

Content vs. platform is a chicken-and-egg problem. There must be an existing model of distribution before content can be important. Without the internet, blog posts are just diary entries. Without movie theaters or cable or Netflix, actors are just beautiful people who play make-believe.

Conventional wisdom, however, holds that most of the value comprises content. Distribution platforms are valuable only insofar as they have access to the best content.

This is not true in healthcare. It presupposes that we are accustomed to paying for content in healthcare. We’re not. We pay for service in healthcare. In fact, most clinical content has always been “free.” Researchers publish trial results. Medical societies publish practice guidelines. Hospitals publish (internally) treatment protocols. Health plans publish clinical guidelines. None get “paid” for that content. It is a byproduct of their business or mission.

The platforms for that content are fragmented — journals, society publications, bulletin boards, and departmental meetings. In the case of health plan clinical guidelines, often it’s “published” by exception via billing transactions, e.g., flagging a treatment as requiring prior authorization or as not fully reimbursable.

Since we pay for service in healthcare, platforms could be valuable. But, how to build them, given that content has historically been free? In fact, does free content dis-incentivize platform development? After all, what and how do you monetize?

Building a Platform

In the beginning, it is unlikely you’ll be able to charge for content. Healthcare pays for throughput (under fee for service (FFS)) and efficiency (under fee for value (FFV)). Thus, the value in platforms isn’t the content, it’s 1) the surfacing of the appropriate content at the appropriate time, and 2) workflows that are easily and repeatably acted upon. Regardless of FFS vs FFV, platforms that do this will create value for their users.

For example, the success of Omada Health isn’t it’s content. In fact, its content is a publicly available diabetes prevention protocol, called DPP. The success is due to the platform identifying people that will benefit, surfacing the right content at the right time, and keeping participants motivated and participating in the program. If Omada launches new disease areas, their success again won’t be because of the content — that’s a commodity — it’ll be because they understand the behaviors and workflows necessary to get participants to achieve success. Changing content like care protocols and treatment algorithms is difficult and time consuming. Most are established after long clinical trials. Changing the platform — to optimize participant selection and motivation, for example — is easy. Thus, in an industry characterized by slow moving evolution, platform optimization, not new content, could be responsible for near-term gains in treatment efficiency!

Other areas of healthcare, like care coordination are similar. It’s not the specific algorithm for care management that will win. Instead it’s the design of the platform. The algorithm is a commodity. Patient.io, for example, has a care management platform that can ingest and render any content. They work with many different vendors and effectively “skin” their protocols/algorithms. The default Patient.io workflow engenders great user engagement and follow-through. Still, as it has more and more users on its platform, even with different content, it gets better at the workflow design and behavioral triggers that ultimately make its product successful.

Watson Health’s AI might “discover” a new care protocol that is better than all existing published protocols. Google’s Verily might unify my wellness/tracker data, my genome, and my EMR and predict future health excursions. That’s great. That new content doesn’t hurt companies like Patient.io because the power is in the platform. While we can certainly improve upon our algorithms and protocols, most poor outcomes in chronic diseases are due to non-adherence not sub-optimal protocols. So, let’s improve the protocols, but most importantly, let’s make sure they’re followed! Creating better platforms that inspire the behavior we want is the answer.

Making Content Matter

In the beginning you can’t charge for content. The example of IBM Watson suggests that clinical protocols will always be a commodity for consumers. That won’t be true forever. Once we have platforms that can guarantee adherence (through a combination of patient selection, motivation, etc), the marginal differences between content will be important. Personalization of content — e.g., through protocols designed for your genotype — will have a lot of value.

For example, companies that are enabling in silico clinical trials, like Flatiron Health in cancer, are creating tremendous value. Patients will get better care faster. Oncologists will better understand, treat, and track their patients. Neither will incur much cost because the content (new oncology protocols) will be commoditized. On the margin they’ll pay for the platform, but the real cost will be incurred by pharmaceutical companies. Why? Because they will also gain the most. The impact on off-label prescribing and informing clinical development is worth billions!

When it comes to content versus platform, healthcare is backwards. Our legacy of paying for service while content was widely disseminated rendered it a commodity. This will not always be the case. New content will drive the personalization of healthcare. Assuming all patients are sufficiently compliant, the new content might be responsible for meaningfully better outcomes. That is valuable and will be paid. But, the assumption is big. Right now we need to focus on getting patients to adhere to their appropriate clinical protocols. That’s a job for platforms.

(“Content” refers to pathways, protocols, and treatment algorithms. It does not refer to life-sciences — biopharmaceuticals, medical devices, or diagnostics — where proprietary development is king!)