By TORY WOLFF

Amazon has many puzzled about its plans for healthcare. Arguably, Amazon is just as puzzled, but is – in effect — running a massive Delphi process to sort out the plan. Amazon is, after all, the Breaker of Industries, Destroyer of Margins. Allow rumors to float, hire some people, have meetings, seek a few regulatory approvals, start a vaguely missioned non-profit with other business titans. Fear and greed do the rest.

Stock prices gyrate as investors bet and counter bet on who is vulnerable, incumbent CEOs promise cooperation or competitive hostility, analysts speculate, “old hands” pontificate, and consultants send megabytes of unsolicited slide decks to South Lake Union. All that information gets exposed without any material commitment.

Disrupting the roadblocks to healthcare innovation

Proper strategic planning requires consideration of a few disruptive (if less likely) scenarios. Amazon getting into hospital supply or creating yet another benefits buying group is easy to imagine but conservative in scope. And we know Bezos thinks long-term and that profits are secondary to platform building.

We also know that the biggest target for disruption–complex, opaque, and inefficient care delivery – remains largely undisturbed despite wave after wave of innovations: ACO and bundle payment models, transparency initiatives (e.g., Castlight), private exchanges, direct contracting models, telemedicine, alternative primary care models (concierge, medical home) – all have all fallen short of their promise.

Why is that? Healthcare is a goldmine of textbook market failures thriving on (1) incentive misalignment (2) information asymmetry and (3) a system frozen by byzantine contracting and regulatory structures, highly scaled incumbents operating highly scalable processes and local market entry barriers which ensure that share shifts are glacial. Innovation is kept comfortably on the margins.

How could Amazon fix this?

Starting point: Amazon’s progressive iteration strategy and partner set

First, keep in mind how Amazon worked in the past: it started a pure commodity (books where copies are completely interchangeable), added in products with lots of different versions to choose from (like toothbrushes or shirts and used versions of books) and different suppliers to choose from (Amazon Marketplace) and finally opened the platform infrastructure to be used by other companies (Amazon Web Services). It has failures along the way but kept experimenting until it figured out what worked.

Also, keep this new joint venture in mind. Besides over 1M employees as a test bed and lure for collaborators, the three organizations are outsiders with “creative destruction” mindsets and capabilities well suited to stimulating markets in risk-ridden domains: Amazon is good at creating highly liquid consumer markets supported with efficient, transparent logistics (rapidly working towards near real-time); Berkshire-Hathaway (BH) is a huge reinsurer which, by allowing players to offload risk they cannot manage, enables value delivery on the portion they can; JPMorgan (JPM) has experience using highly liquid markets to innovate entirely new contract classes and target risk allocations by splitting up and reassembling risks in ways that attract more customers (financial engineering).

Our scenario starts with using Amazon getting started in a way much like it did with books (commodity care), then using its initial critical mass of customers to expose more complicated slices of healthcare to its markets, and then layers in BH and JPM capabilities to support product and risk model innovation which will allow share to shift faster to top performing providers. It will take many years – a decade at least – but could upend the system.

Step 1: Start with the healthcare equivalent of books with an Orbitz for minor acute

Bookshop owners wax poetic about the feel, smell and sound of cracking open a new book. Bezos recognized that this romanticism is not shared by buyers seeking bargains and convenience. Hippocratic sentimentality aside, there are several classes of healthcare that many patients are ready to shop for (and payers have been trying to get them to shop for): minor acute care (sniffles and coughs to simple fractures), basic diagnostic services (labs and imaging) and some preventative care (e.g. flu vaccines).

Amazon could start there: Focus on some specific metro regions (ones where Amazon, BH and JPM have employees), target PPO (which allow for more self-direction vs. HMO) and ASO (to avoid state regulatory frictions) and require plan administrators to create interfaces to pump network and benefit information to an Amazon Health website; cut deals with telehealth, retail and urgent care clinics, primary care and free-standing imaging centers; require them to build interfaces with their scheduling systems. And finally, employees save their benefit information into this “Orbitz” for healthcare.

“Alexa, I have a sore throat…” Options inside the Venn diagram of benefit design, network and convenient availability populate: one clinic is nearby, affiliated with the member’s primary care but no slots until tomorrow; one urgent care center is on the way to work and has immediately availability but Tier 2 so a higher co-pay. TDOC has Dr. Jones ready for a call in fifteen minutes. There’s also the ED downtown with Harvard trained doctors but that’s a $200 co-pay and current wait time is 4 hours. Once a choice is made, patient intake forms pre-populate and co-pays are collected in advance.

Market share – made fluid as lower search and transaction costs melt the constraints of habit – will shift faster in response to better value. New ideas are innovated, tested, copied: hours and capacities adjusted; more telemedicine options; experienced physicians offered as alternatives to extenders. Providers who initially opted out will revise their decision. With additional market depth, new navigation tools can be added: Customers ratings become more credible thanks to the compelling power of “n = a lot”; Amazon links with claims clearinghouses and CMS to get data on provider experience (“This provider has treated sore throats 500 times in the last 2 years”).

Step 2: Grow the market and expand care domains (Orbitz for specialty consults)

Big ASO groups – always looking for the latest gadgetry – will push plan administrators to incorporate Amazon Health into benefit designs. Indeed, those plans may not need to be pushed given medical cost improvement as members respond more fully to co-pays differentials designed to lure them to more appropriate sites of care.

Game theory’s iron logic will induce providers to expand their offer on the Amazon marketplace to care where their brands will be more compelling: primary care, more complicated procedures (e.g., colonoscopies) and, the biggest step, specialist consults[1].

Specialist consults are the front door to expensive arcs of care, often involving highly profitable procedures. The economic stakes if this kind of share becomes more fluid are much higher than minor acute care. High quality providers whose differentiated performance is not reflected in copays and availability will protest. And payers will agree. Copays will have become too crude a tool to support decision-making where clinical quality can have a much bigger impact on outcomes and costs.

Step 3: Innovate the product (from appointments to episodes)

Markets are great engines for figuring out new pricing mechanisms. High performing providers will develop bundles which show their value and payers will relax benefit designs to test uptake. Have a new chronic disease? This specialist will take care of you and all your disease-related needs for one annual co-pay. Need some fairly complex cardiac procedure? A local community/tertiary provider uses Mayo Clinic protocols and guarantees recovery in half the time or your co-pays back. Not all combinations will work, not all ideas will be allowed by plans (though Amazon’s IT support will help ensure that clunky claims and contracting systems are not the roadblock) but bad ideas will fail quickly and winning ideas copied and enhanced.

This is the point where BH will be a critical enabler: Some providers may be cautious about taking on the risk of consumer-appealing bundles (especially when these are being iterated much faster than the lengthy cycles of CMS). But the prospect of not only gaining on the risk (by managing care costs better) but also on market share will be hard to pass up. What will be needed is a partner who can help slice up the risk of any bundle, and take those risks which are beyond the providers’ control off their hands – the sort of thing a reinsurer should be great at doing.

With share shift and volume concentration among able specialty care providers, there will emerge more opportunities to specialize operating models (Herzlinger’s “focused factories”). Less capable providers – particularly those whose economics depend on complicated webs of cross-subsidization – will struggle.

Step 4: A parallel B2B market in episodes emerges – An Alibaba for Episodes

Today, big provider systems generally “make” rather than “buy” service lines. Referring care out of system is limited to very specialized, high-end care (e.g. specialty pediatrics, complicated cancer, transplant) where there was not much prospect for acquiring the physicians and equipment needed. Provider systems insource for a wide variety of reasons: better coordination on clinical strategy, less risk of losing the patient, ambiguity about cost and outcomes, cross subsidization opportunities (who gives up “knees”?)

A liquid market for specialty episodes could eliminate some of these reasons: the market-driven definition of bundles will make a more modular approach to care collaboration easier; provider systems will a lot more visibility into their relative performance; and, the incentives for focused factories to hand back the patient at the end of the episode (vs. stealing them for other care needs) will be very strong.

Thus, a B2B market for specialty care could emerge among providers operating HMOs (in parallel to the B2C market for PPO). Population health managers (HMOs or other models) negotiate with focused factories for portions of care where they have stronger performance. The PCP would refer patients to recommended specialists as usual, but instead of generally guiding the patient towards in-system specialists, might refer more to third party providers and thereby benefiting from better value delivered.

At this point in the scenario, JPM can become a critical enabler: Risk tracking would get complicated as an accountable provider is taking risk at the population level, then handing out targeted portions of that risk to focused factory providers. JPM’s know-how in building markets around complicated derivatives deals could become important to making this aspect of the B2B market work. Derivative contracts set up complicated chains of financial flows across multiple parties in response to highly customized event triggers all of which are settled via true-ups across parties at specified periods of time. The process could look very similar for healthcare providers.

Step 5: Private exchanges become real benefit design innovation engines

As accountable providers – whether they are today’s integrated delivery systems or clinically integrated networks or emerging national ambulatory systems (like a matured OptumCare) – revise service portfolios and operating models to take advantage of focused factories, they will have more options to vary the value proposition they offer patients: different combinations of focused factories for specific conditions, a primary care panel weighted with expertise in certain conditions, concierge features, etc.

There’s no reason prospective patients shouldn’t participate in the cost of those features through private exchanges (where members are offered a defined contribution towards selecting among various benefit structures). When a member first signs up for HMO coverage, they could be offered choices: a standard model and then extra features and pricing for specific accountable providers. Want care provided exclusively by the academic medical center and medical school team? Willing to have your health and care managed by the local OptumCare team but with access to Cleveland Clinic for cardiac care? You will see the impact on your premiums and co-pay structures.

And with that, the Amazon will have gone full circle from market making in commodity acute care to exposing more and more slices of delivery system operating models and cost structures to market forces, thereby tearing them down and supporting their reassembly into more efficient models winning on more liquid markets, to changing the way healthcare coverage is itself purchased.

Fully integrated provider systems as today’s department stores

You may not find this scenario particularly feasible. We are not fully convinced either and continue to puzzle about how it might work. But many objections seem to depend on differing assumptions about patients (e.g. “patients do not want to shop for healthcare”) or Amazon (e.g. “they won’t wait x years to make money”) or providers (“branded systems won’t play”) each of which can be challenged (for example, patients are already being asked to shop via high deductible plans but lack the tools and don’t see the value to doing so, Amazon can make money with a lot of supporting transactions – their recent deal to sell Amazon’s brand of OTC medications made by Perrigo will complement any foray into scheduling minor acute visits).

One assumption which is not challenged – that there is a lot of waste in healthcare delivery driven by lack of clarity on what the market really wants (which needs a real market with real pricing for everyone to figure out) and by lack of opportunity to capture share if you deliver it – is what defines the size of the opportunity.

An Amazon model could start small but will become a powerful engine fed by adding clinical adjacencies. Shifting share and more focused factories will expose operating models and cost structures–layered with complicated cross-subsidies – to market forces. As those cross-subsidies are removed, the value for the Amazon enabled shopper will be attractive. And entrepreneurial energy and capital will be attracted by the opportunity to gain share. The adoption curves on new technologies and processes will be faster, raising the ROI on many ideas.

I suspect many imagine some version of this scenario happening at some point — just well beyond any reasonable planning horizon. Amazon is moving up the timeline. A lot of questions which might have seemed academic and deferrable a year or so ago should be put back in the front of the queue: what value are we creating? Where are we differentiated? What do we really need to own? How can we become both leaner and more nimble and innovative? How can we win in a more transparent, fluid, value-seeking market?

Many book store owners, department store executives, consumer electronics retailers wish they had good answers to those questions a few decades ago.