Amazon's recent entry into the healthcare world has left a lot of people speculating as to the aims, goals, and ultimate profit motives. The best take I've seen is Ben Thompson's at Stratechery - his analysis is spot on. However, since he is coming at this from outside of healthcare, he frames it through his aggregation theory instead of through the structure of the healthcare industry.
I believe that Amazon is going to become a Third Party Administrator (TPA), first for themselves, then for others. Most large employers take on the costs of healthcare themselves, while outsourcing management of the plan to a traditional health insurance company that acts as a TPA for the employer. From Wikipedia:
A third-party administrator (TPA) is an organization that processes insurance claims or certain aspects of employee benefit plans for a separate entity. ... This can be viewed as "outsourcing" the administration of the claims processing, since the TPA is performing a task traditionally handled by the company providing the insurance or the company itself. Often, in the case of insurance claims, a TPA handles the claims processing for an employer that self-insures its employees. Thus, the employer is acting as an insurance company and underwrites the risk. The risk of loss remains with the employer, and not with the TPA.
Many pundits are dismissively comparing Amazon's effort to the early days of Kaiser Permanente, concluding that this initiative is no threat to the traditional healthcare landscape. A careful observer of Amazon knows that they won't stop this service with themselves, and once they've proven out the strategy and economics they will aggressively expands.
Traditionally, a TPA takes a percentage cut of the claims that it processes. This creates opposing incentives with regards to serving their employer-customers. These customers want to reduce their healthcare spending, while the TPAs want to increase their revenue (which is directly linked to that spending!). Enter Amazon, where Jeff Bezos famously said, "your margin is my opportunity." Amazon is not afraid of reducing the spend from any one customer because Bezos believes that they can operate at a lower cost basis and gain market share to such an extent that what Amazon losses in margin they'll make up in volume. But how can Amazon lower the cost basis of a TPA by an order of magnitude? Enter interoperability. Interoperability is the ability for health care systems to share data programatically.
Despite all of the hype, there has been very little practical advancement in interoperability to date. Why? Lack of incentives. There doesn't exist an ROI for a provider to invest in systems that make it easier for a customer (patient) to switch providers. Nor for an EHR company. Many health systems are "achieving interoperability" by standardizing all of their hospitals, physicians, and facilities on the same EHR. That's not int_er_operability, that's int_ra_operability! So why don't the health insurance (and thus TPA) companies invest here? The amount of time a patient stays with a health insurer is notoriously small - I've heard average customer lifespan of 18 months, and annual churn rates of 20 - 30%. The short-termism of the health insurance industry doesn't provide incentives for long-term investments.
The small gains we have gotten in interoperability have thus far been produced by government efforts. When there is a structural failure in a marketplace it makes sense for the government to step in to enact positive change. Unfortunately, the government's intervention hasn't had as big of an impact as many had hoped, hence the continuing struggle for interoperability.
Amazon, through it's optimization of market share over margins, finally has the incentives right to drive interoperability. Their timeframe is better aligned to the incentives of interoperability than those of the insurance companies. By turning the TPA business into an API layer of healthcare ("Amazon Health Services" aka "AHS"), Amazon can dramatically reduce the cost structure of the activities performed by a TPA. Using a go-to Amazon strategy of being their own first customer, they are going to build their TPA using their own employees (plus those of JP Morgan and Berkshire Hathaway). Amazon will first experiment on the operational and financial model of their internal TPA without the pressures of going to market. But once they have it right, they will open up their TPA to all employers. This parallels their strategy in order market segments: after years of building up their internal shipping and logistics business, Amazon is now launching Ship With Amazon to other businesses.
But why would a healthcare provider invest in Amazon Health Services' APIs? For providers, the number of employees participating in this initiative, and thus potential customers, will finally provide this incentive. Insurance companies have needed as many providers as possible in their networks to sell to the largest number of employers, so aren't incentivized to enforce things like interoperability as the risk of excluding a provider. Amazon doesn't have to attract other employers at this phase, so they can exert pressure on providers via threatened exclusion, thus driving the adoption of their AHS APIs.
Amazon is going to restructure healthcare by competing on a completely different dimension and set of incentives than the existing industry players.
Jacob Alcauskas is the Co-founder and COO of Superconductive Health. Superconductive Health focuses on dramatically accelerating the value of data in healthcare. Our team is a rare blend of two crucial skill sets: deep domain knowledge of healthcare data, and cutting-edge expertise in data engineering, automation, and machine learning. We believe that faster, higher quality exchange of information, both within and between organizations, is a lever point for healthcare — a key enabler to improve patient outcomes, lower costs, and accelerate research.
Superconductive Health is pioneering tools and process that dramatically accelerate the mundane—but critical—work of building and maintaining the context-rich data pipelines that enable research, analytics, workflow, interoperability, and AI. Our Accelerated ETL service offerings and Data Checkpoints product suite have the power to dramatically cut the overhead cost of data work in healthcare and life sciences. Agile data practice, deep domain knowledge, and proprietary workflow tools make us the best data engineers in healthcare.
 I am going to refer to the Amazon, Berkshire Hathaway, and JP Morgan consortium as Amazon throughout this article. This is easier stylistically and I believe that Amazon is steering the ship here, strategically. However, the inclusion of JP Morgan and Berkshire Hathaway in this deal are very instructive and show that Amazon understands the existing industry structure - as mentioned by Thompson, those two are the pieces of the puzzle that will truly allow Amazon as a TPA to happen.
Without Berkshire, Amazon may be locked out of the reinsurance market (and thus unable to expand this to other employers). Without JP Morgan, they may be locked out of the financial markets. ↩︎
 Which gives them an incredible entry point to later extend into other pieces of the industry, such as PBMs, Mail-order Pharmacy, Specialty Pharmacy, etc.↩︎
 Moving to a systematic healthcare world was recently discussed by David Chapman in a great post over at meaningness.↩︎
 Every health insurance company, and esp. those with large TPA businesses should be put on notice by this initiative. Every provider should cheer it - they finally have the incentives to streamline much of their business, but only if they are able to transact externally programmatically. PBMs, RCMs, and other "efficiency-driving" middle men should also be worried because Amazon is headed there next. Amazon is going to replace the traditional TPA and along the way finally user in an era of healthcare interoperability.↩︎