Building a Healthcare System from Scratch, Part 1: Jobs

Image credit:

This series explains my framework for understanding healthcare systems. For ease of reference, I call it my Healthcare Incentives Framework. I use it to make sense of everything I see happening in a healthcare system. And one could use it to build a healthcare system from scratch. Each subsequent post builds on all the prior posts, so I strongly recommend reading them in order. For convenience, here are links to parts 2, 3, 4, 5, 6, 7, and 8.

I developed the majority of the Healthcare Incentives Framework over 5 years (2010 – 2014), during which time I was thinking/reading/studying about healthcare policy almost every spare thinking moment. I just had this insatiable drive to make sense of the evidence I was reading as well as what I was seeing in the world. It was the ultimate puzzle. Luckily, my wife was patient and convincingly interested in my thoughts, which I shared almost daily. Every time I came up against a phenomenon that I couldn’t understand or explain within the context of the greater Framework I was developing, I would struggle with it in my mind, sometimes for weeks, until I figured out what it meant and how it fit. Most of the posts on this blog came as a result of those struggles, and everything I now write about health policy is based on the foundational understanding encapsulated in this Framework.

The Healthcare Incentives Framework starts by enumerating the jobs you want a healthcare system to do for you. Initially, it seems like only two jobs: prevent disease and treat the disease that cannot be prevented. But, because incentives work differently for different kinds of prevention, I will split that job into cost-saving prevention and cost-effective prevention.

Cost-saving prevention saves more money down the road than it costs up front. For example, maybe hiring someone to visit the homes of people with really bad heart failure will prevent enough hospitalizations that it more than compensates for the salary of the person doing the home visits.

Cost-effective prevention ends up increasing total spending–that is, the money saved (if any) doesn’t outweigh the upfront investment–but the benefit is large enough to justify that investment. For example, screening for colorectal cancer costs a lot, but it catches a lot of cancers early and saves enough lives that the investment is worth it. The exact definition of what’s cost-effective and what isn’t depends on the society, but the metric used is the cost for each quality-adjusted life year you gain ($/QALY).

In addition to those three jobs related to maximizing health, and because healthcare is characterized by rare, unpredictable, potentially financially catastrophic treatment episodes, a healthcare system must also provide financial protection in the form of risk pooling.

And, finally, most people in most societies believe a healthcare system has a responsibility of providing these services even for people who cannot afford them, so equitable access (as defined by the society) can be added as the fifth and final job of a healthcare system.

The utility of enumerating these jobs will be seen in Part 2, when we will look at which parties in a healthcare system have financial incentives to perform which jobs.


Does Supply and Demand Work in Healthcare? (Updated!)

This post is an update to my popular post, Does Supply and Demand Work in Healthcare? It explains the same things, but this version is shorter and clearer.

Supply and demand carry with them a few assumptions. When those assumptions are met, supply and demand works. When those assumptions are violated, supply and demand don’t work as we would expect. Healthcare is very different from most markets in the United States, especially in a couple key ways that violate the assumptions of supply and demand. The first violated assumption is that consumers have price and quality information. The second violated assumption is that consumers actually pay the price of what they buy. There are others, but these are the biggest ones, so let’s focus on them one at a time.

Assumption 1: Consumers Have Price and Quality Information

Think about how most markets work. People spend their own money to buy things they want. And because people don’t have unlimited dollars to spend, they’re weighing the value (Value = Quality/Price) of their different options–if something is higher quality but also higher priced, they have to decide if it’s worth it for them to spend that additional money to get that additional quality.

Think about how that impacts companies competing in a market. If one company makes a really high-value product, tons of people will buy it, and the company will increase supply so it won’t miss out on all those potential sales. If another company makes a dud of a product, it won’t get many sales, and it will decrease supply. (Yes I recognize that this is ignoring many complicating factors, but those factors don’t affect the point I’m trying to make here.)

Now, let’s look at how healthcare is different . . .

When patients choose a healthcare provider, they mostly aren’t using price and quality to make their decision. So even if a hospital is super high value, it won’t generally win the market share that it should, so it isn’t building additional wards like it would if it were competing in almost any other industry. Conversely, a low-value hospital will continually get more patients than it should, and it will keep its wards open.

Taking this one step further, a hospital will have a hard time investing in value-improving innovations if it’s not going to win more market share (i.e., additional profits) as a result. And, unfortunately, our current system often financially penalizes value-improving innovations. See here for more explanation on that.

Assumption 2: Consumers Actually Pay the Price of What They Buy

As I said above, a consumer in almost any other market will think carefully before buying a really expensive good or service. “This is way more expensive than the other one. Is it better enough to be worth it?” Same goes for deciding how much of something they’ll buy; people (usually) won’t buy more of a good or service than they think they can afford.

So, in other words, when customers actually have to fork out the dough for the thing they’re buying, their demand is appropriately limited. And, when demand is limited, that in turn constrains the quantity supplied–companies don’t want to spend a bunch of money making tons of goods that are unlikely to ever get sold without drastically reducing the price.

Now, let’s look at how healthcare is different . . .

The most obvious example I could bring up is end-of-life care. Think about a patient who had a massive stroke and is now in the ICU on life support, showing no signs of life for multiple days. There’s always the slightest chance they could recover some function, so it’s not unreasonable for families to cling to that hope and keep dragging it out. . . . Except that estimates of the cost for each additional day in the ICU run around the $5,000 mark. But, the family has probably already hit the out-of-pocket max for the year, so they won’t be paying a single cent more even if they drag the ICU stay on for another few days. My point is that demand is almost unlimited in a situation like this, and the hospital is happy to continue supplying the care as long as they’re getting paid for it. I’ve written elsewhere about the problems that arise when the party choosing how much care to get is not the same party that foots the bill.

These are just two examples of broken assumptions of supply and demand in healthcare. I am not saying that supply and demand will never work in healthcare; I’m just saying that the way our system is currently organized violates some of those assumptions. I’ve also written about how to fix that.

I’ll end with one other implication of all of this. Critics of “market-driven healthcare” abound because they say we’ve been trying it for a long time and look where it’s gotten us. But we actually haven’t truly tried it yet because we haven’t ever made the changes necessary to remove the barriers to supply and demand. Without explaining it fully here, I’ll assert that we can remove enough of those barriers for supply and demand to work well in healthcare. And the changes that would be required to accomplish that are compatible with any structure of healthcare system, be it a private system, a single-payer system, a fully government-run system, or whatever.

Why Drug Companies’ Medication Coupons Are Bad for the Healthcare System

Have you had the experience where you need a medication, and the brand name actually is cheaper for you because your doctor gives you a coupon for it? It’s great for you, but it’s bad for the healthcare system, and here’s why.

I have written before about the principle that is relevant to this, but it bears repeating: The party making a purchase decision must be the one who also bears the price differential between those options.

To understand why, let’s pretend you have a medium risk of heart attack or stroke in the next 10 years, so your doctor recommends you start a moderate-intensity statin medication. They’re all pretty close to equal in terms of efficacy and side effects, so the best money-saving decision would be to choose the cheapest one, right? Well, if your doctor says, “I’m happy to prescribe any of these for you. Which would you like?” You are the party who now gets to make a purchase decision. So you look at the monthly prices below (these are real prices):


But then your doctor hands you a pitavastatin $100 off coupon some drug rep from Kowa Pharmaceuticals (the manufacturer) dropped off. You, a rational person, opt for that one since it’s now the cheapest (free)!

Now the monthly cost to the healthcare system for you to be on a statin totals $0.00 (your copay) + $101.36 (what your insurer has to cover) = $101.36. That’s about 20 times more expensive than it should have been!

What just happened here? The party making the purchase decision (you) did not bear the price differential between the options. Your insurer originally set it up so that you would pay more if you chose a more expensive statin, but the coupon interfered with that.

This same situation plays out over and over every day in our healthcare system with medications and with every other health service. It’s why I keep saying that we need to make the party who makes the purchase decision the same party who bears at least some part of the price differential between the options, which leads to a value-sensitive decision. Reference pricing does it, high-deductible insurance plans do it (for services below the deductible, at least), multi-tier prescription programs do it (when they’re not being subverted by manufacturer coupons). But these, collectively, are not influencing nearly enough of the purchase decisions being made in our healthcare system! And we waste money. Even worse, the higher value options are not rewarded with market share, the lower-value options are allowed to persist as is, and the overall value delivered by our healthcare system remains much lower than it could be.

So that’s why medication coupons–and any other thing that interferes with purchasers bearing the price differential between options–are bad.

The 3 (Actually, 2) Problems with the U.S. Healthcare System

Image credit: Gary Larson

Before something can be fixed, the problem must be defined and the causes of that problem diagnosed.

The generally accepted problems with the U.S. healthcare system are that its prices are too high, its quality is too low/patchy, and not enough people have insurance coverage.

I’m going to change that a little bit. Since price and quality are the two variables that determine value, we could combine them into only two problems: suboptimal value and not enough insurance coverage.

The insurance coverage piece is kind of separate from the value piece because solving it primarily relies on government policies that subsidize the purchase of health insurance for those who wouldn’t be able to afford it otherwise. Those government policies are a (purposeful) distortion to the healthcare market. Distorting a market for a good reason is fine, but before you do it, you need to understand how the market should be structured to optimize value so you don’t accidentally ruin its potential to have higher value in the process.

And the nice thing is that value improvements will primarily come in the form of lower prices —> insurance coverage will become cheaper —> fewer people will be priced out of the market. Therefore, fixing value partially fixes the insurance coverage problem too!

Nuts and Bolts of Implementing Bundled Payments

My last post described the design of a patient engagement pilot. A central (and most challenging) part of that pilot would be bundled payments. While designing the pilot for Utah last year, I worked out some of the nuts and bolts of implementing bundled payments, which was a crucial step in designing the pilot because, to this point, no private-sector bundled payment pilot has really truly completely gotten off the ground due to huge challenges with (1) defining the bundle of services included in the lump-sum payment and (2) amending contracts (i.e., figuring out how to dispense that payment to the various participating caregivers). So, if you happen to know someone working on bundled payments, this might be a good resource for them.

1. Defining the Bundle – Guiding Principles

We need a neutral steering committee to arbitrate any impasses. Decide on this at the first meeting where the bundle definition will be discussed.

We want to minimize exclusions, which will help providers have enough volume so they can wholeheartedly redesign care processes and spread redesign costs over a larger number of patients. Good stop-loss provisions will limit their risk. As described in the payer-hospital contracting section, including good stop-loss provisions will not have a net effect on the amount payers are reimbursing.

Minimizing exclusions will mean that the risk level of patients being cared for on a bundle will vary quite a bit. In order to avoid providers shunning those higher-risk patients, we need to create risk categories after we’ve defined the bundle. A patient’s risk category should be based on criteria that are easily available up front (before the surgery). When payers and providers set bundled prices, they will set a different bundled price for each risk category.

All participating in the bundle definition decision should be very familiar with the information in the payer-hospital contracting section.

A simpler bundle definition may not be as conceptually appealing, but it will allow the bundle definition to be completed faster so that momentum will not be lost. We can use the work already done by others and make minor adjustments (if necessary).

After the episode is selected, we need to collect all the bundle definitions out there and see where they are the same and where they differ so that the bundle definition decision can be narrowed down to a limited number of specific decisions.

A guiding principle in the bundle definition decision should be to include everything a patient would need to buy the “product.” The product of a knee replacement is to have the knee returned to full function. The product of a CABG is to have a heart fully perfused again and the patient is back to normal functioning.

Including post-acute care/rehab-type services in the bundle: Reasons to include it are (1) because the patient/insurer will know what the total price of the episode (i.e., product) is up front and (2) to encourage the reduction of avoidable services.

We probably shouldn’t include pre-procedure services because it just adds more administrative complexity than it’s worth.

When deciding on the standardized bundle, don’t forget to decide what day the 90-day window starts (on the day of discharge? on the day of the surgery?) and what to do with an in-bundle service that starts on the 89th day and continues past the end of the window (prorate it? cover the full cost of it?).

2 a) Amending Contracts Between Payers and Hospitals

This section has been prepared to simplify the process that payers and hospitals will go through to amend their contracts so that they can comply with the pilot requirement to prospectively determine prices for the standardized bundle.

To simplify the language in this guide, “hospital” is the term used to denote the primary contracting entity with a payer, although this party may be either a hospital or a healthcare system.

Instead of a contract amendment, payers and providers could opt for the less formal “partnering letter,” which defines the terms without needing to make any adjustments or amendments to existing contracts. Legal counsel can provide further clarification on this option.

Please see the hospital-caregiver contracting section for details on the process hospitals/healthcare systems and caregivers will go through to determine risk-sharing arrangements.

Step 1: Determine the bundled price

  • Using the payer’s claims history, do a retrospective reconstruction of each care episode (using the standardized bundle) that took place in the hospital over the last few years and trend those episode amounts forward to test-period prices
  • If different bundled prices will be set for different patient populations (e.g., male with BMI under 35, male with BMI over 35, female with BMI under 35, female with BMI over 35), organize those episodes into their appropriate groups and then average each group to get a test-period estimation of each group’s average price of a care episode
  • These averaged prices will be reduced in proportion to the net additional amounts the payer is expected to pay out for in-bundle services according to other provisions discussed in this guide (see Step 2 Option 2.2, Step 4, Step 5, Step 6, and Step 7) so that the average episode reimbursement level stays the same as it would be without this contract
  • Additionally, if any of the providers involved in providing care in an episode are already being reimbursed by the payer in a capitated manner, the bundled price will be reduced by the percent that historically has gone to that type of provider (this does not preclude capitated caregivers from participating in hospital-caregiver risk-sharing arrangements)

Step 2: Determine the reimbursement method

  • Option 2.1: Normal FFS with retrospective reconciliation
    • This is the simplest option, but it risks providers (hospitals and caregivers) having to pay money back to the payer if their up-front FFS reimbursements end up being more than what the retrospective reconciliation determines they should receive
    • The payer would take the role of performing the retrospective reconciliation—including using the hospital-caregiver risk-sharing agreements—to determine how much to reimburse/bill each participating provider
    • Aspects of the hospital-caregiver contracting section relating to avoiding back-billing can be ignored when hospitals and caregivers negotiate their risk-sharing arrangements
  • Option 2.2: Reduced FFS with retrospective reconciliation
    • Providers will continue to be reimbursed normally according to existing reimbursement processes, only reimbursement levels will be set to X% of normal reimbursement
    • Assumed as part of this option is that there would not be any back-billing; in other words, any reimbursement to a provider would not be required to be paid back even if that provider ended up being reimbursed more than what the retrospective reconciliation dictates and even if the payer’s total reimbursements for the care episode end up being more than the total bundled price
    • The payer would perform the retrospective reconciliation to determine how much additional money, if any, is owed to the provider group (according to the prospectively determined price), and then send that amount either to the hospital (which would make the hospital responsible for distributing that money according to hospital-caregiver risk-sharing agreements) or to the individual providers directly (which would make the payer responsible for performing the retrospective reconciliation, including taking into account hospital-caregiver risk-sharing arrangements)
    • Principles to guide the decision on where to set the reimbursement percentage:
      • No matter how high the reimbursement percentage is set, it does not affect providers’ upside risk as determined in the hospital-caregiver contracting section
      • A higher reimbursement percentage allows providers to receive more of their reimbursements earlier
      • If the reimbursement percentage is set high, it increases the likelihood of providers being overpaid if their costs for a bundle are higher than expected, which means the non-overpaid providers will be underpaid; in this situation, the reimbursement percentage acts as a cap on how much providers can be underpaid and instead causes the payer to pay out more than the bundled price, which, as described in Step 1, means the bundled price will be set lower to compensate for these additional expected outlays
      • The net effect on providers of a too-high reimbursement percentage is that it increases the downside risk of providers whose risk allocation is lower than 1 – Reimbursement Percentage and decreases the downside risk of providers whose risk allocation is higher than 1 – Reimbursement Percentage
      • Reimbursement percentages that seem to best balance all these factors are 80%, 85%, or 90%
    • Option 2.3: Lump sum paid to hospital up front
      • The hospital will be given the total bundled price as soon as an eligible episode is initiated
      • The hospital’s distribution of this lump-sum payment will be determined by hospital-caregiver negotiations
      • All providers would still need to submit claims to the payer for purposes of helping the hospital do a retrospective reconciliation to appropriately distribute the bundled payment with caregivers; these claims could either be submitted as no-pay claims or as regular claims and the payer would have to institute a regular review process to pull the ones that have already been paid for as part of the care episode
      • This option risks hospitals having to pay money back if the episode payment has to be reduced (see Step 4, Step 5, and Step 6); the payer performs a reconciliation for each episode (according to the decisions made in Step 3) to determine if any money is owed back to the payer
      • The hospital would be responsible for performing the reconciliation for the purpose of distributing the money according to hospital-caregiver risk-sharing agreements
    • Option 2.4: Monthly payments paid to hospital over course of episode
      • This is the same as Option 2.3 except the bundled price would be paid to the providers monthly over the course of the episode instead of as a single up-front payment
      • To determine what percent of the total bundled price should be paid each month, conduct an analysis of the historical percent of care episode costs incurred each month and break up the total bundled price over the episode window according to those percentages
      • Decide when monthly payments are due (e.g., on a certain day of the month, on the first day of each 30-day episode period, on the last day of each 30-day episode period)
      • This option has a lower risk of hospitals having to pay money back if the episode payment has to be reduced, but there is still some risk (see Step 4, Step 5, and Step 6) because the payer will perform a reconciliation for each episode (according to the decisions made in Step 3) to determine if any money is owed back
      • The hospital would be responsible for performing a retrospective reconciliation for the purpose of distributing the money according to hospital-caregiver risk-sharing agreements

Step 3: Decide on reconciliation timing

  • Decide how many months to wait after episode window has ended (to allow all episode-related claims to trickle in) before the episode will be considered eligible for reconciliation
  • Decide how frequently the payer should do reconciliations (e.g., monthly, quarterly, biannually)

Step 4: Set provisions for patients who change payers before the end of the episode

  • If doing Option 2.1 or Option 2.2, use that bundle’s Reduced Price (see below for how to calculate this) when doing reconciliation
  • If doing Option 2.3, need to get the appropriate amount back from the hospital (Appropriate Amount = Total PriceReduced Price)
  • If doing Option 2.4, the final episode payment will have to be pro-rated so that all payments for that episode add up to the Reduced Price
  • Calculating the Reduced Price:
    • First, need to determine the historical percentage of episode costs that accrue during each 30-day post-discharge period (these numbers will be known as % of Costs1-30, % of Costs31-60, and % of Costs61-90)
    • Next, need to determine the number of days the patient was no longer covered during each 30-day post-discharge period (these numbers will be known as # Not Covered1-30, # Not Covered31-60, and # Not Covered61-90)
    • Next, plug that information into the following formulas:

Reduced Price = Total Price x (1 – % Reduction)

% Reduction = (% of Costs1-30 x # Not Covered1-30 /  30)  + (% of Costs31-60 x  # Not Covered31-60 /  30)  + (% of Costs61-90  x # Not Covered61-90  / 30)

Step 5: Set leakage provisions (leakage = a patient receives in-bundle services from an outside provider during the episode window)

  • Option 5.1: Reduce the patient’s bundled price by X% of the total amount the payer had to reimburse outside providers for performing in-bundle services
  • Option 5.2: Each time the payer has to reimburse an outside provider for performing an in-bundle service, reduce that patient’s bundle price by the amount the inside provider would have been reimbursed for performing that same service (or, if the amount the payer had to reimburse the outside provider was less, then reduce that patient’s bundled price by that lesser amount instead)

Step 6: Determine complications penalties

  • If the bundle definition does not include all avoidable complications that could be attributed to services rendered as part of the care episode, a complication penalty should be implemented as an incentive for providers to work to reduce those avoidable complications
  • Determine the historical average cost to treat that complication and then use that number as a starting point to set the flat dollar amount penalty that will be used to reduce a patient’s bundled price if the patient experiences that complication
  • If the complication can only be partially attributed to services rendered as part of the care episode, that complication’s penalty should be reduced accordingly
  • This process can also be used if post-acute care is not included in the bundle definition and payers want to create an incentive for providers to minimize unnecessary referrals to post-acute care (or could instead offer a “non-referral bonus” instead of having it be a “referral penalty”), but need to make sure the pilot’s standardized quality metrics keep providers accountable for referring patients to post-acute care when indicated

Step 7: Determine stop-loss provisions to cap provider risk (not necessary if doing Option 2.2)

  • Option 7.1: An additional per diem amount for each day over X inpatient days will be added to any patient’s bundled price
  • Option 7.2: The payer will increase any patient’s bundled price by the amount of costs over X% of the original bundled price

Step 8: Decide on frequency of data sharing

  • Payers should regularly share claims data (e.g., monthly, quarterly, biannually) with providers to enable them to perform reconciliations and to help them track their episode utilization trends

2 b) Amending Contracts Between Hospitals and Caregivers

This document has been prepared to simplify the process that hospitals and caregivers will go through to create a risk-sharing arrangement for the bundled reimbursement they will be receiving from payers.

Instead of creating a contract, hospitals and caregivers could opt for the less formal “partnering letter,” which defines the terms without needing to make a formal contract. Legal counsel can provide further clarification on this option.

If any of the caregivers involved in providing in-bundle services are salaried by the hospital, please refer to Step 6 for instructions on how to adjust the processes described in Step 1 – Step 5.

If any of the caregivers involved in providing in-bundle services are reimbursed by the payer in a capitated manner, please refer to Step 7 for instructions on how to adjust the processes described in Step 1 – Step 5.

Please see the payer-hospital contracting section for details on the process payers and hospitals will go through to amend their contracts for bundled reimbursements.

Step 1: Decide which types of providers will bear risk

  • Make a list of all the provider types that will be needed to provide services included in the standardized bundle
  • For each provider type, determine if they will need to directly coordinate with the hospital and/or surgeon to reduce costs/improve quality
  • In the risk-sharing arrangement, include each provider type with which direct coordination is expected
  • For all other provider types, do not share the risk, but determine how they should be reimbursed
    • Option 1.1: Normal FFS
    • Option 1.2: Give a flat prospectively determined amount for each patient (and, if desired, can integrate provisions similar to those listed in the payer-hospital contracting section, such as stop-loss provisions)

Step 2: Determine the percent of the bundle that should be allocated to each provider type (this step is not necessary if the payer-hospital contract is using FFS or reduced FFS as the reimbursement method)

  • After subtracting out the amounts going to non-risk-sharing providers, determine the percent of the remaining amount that should go to each provider type according to historical average percentages

Step 3: Determine the withhold percentage and then do an initial allocation (this step is not necessary if the payer-hospital contract is using FFS or reduced FFS as the reimbursement method)

  • Before allocating anything to the risk-sharing providers, a percent of the bundle should be withheld until the retrospective reconciliation (described in Step 5) has been completed
  • Distribute the non-withheld portion of the bundle according to the percentages determined in Step 2
  • To avoid back-billing, each provider will be guaranteed to keep these initial allocations
  • The appropriate withhold percentage should balance the following factors:
    • A lower withhold percentage allows providers to receive more of their reimbursement sooner
    • Because providers will be guaranteed to keep their initial allocation, a lower withhold percentage also increases the frequency of providers being overpayed (especially considering the possibility of the bundle being reduced according to provisions in the payer-hospital contracting section), which means other providers will be underpaid more frequently, effectively increasing their risk to levels higher than arranged

Step 4: Decide what percent of the risk each provider will bear

  • Each at-risk provider must decide what percentage of the risk to bear
    • If the hospital decides to bear 70% of the risk and the total reimbursements for an episode add up to $1,000 less than the bundled price, the hospital will receive $700 of that excess; alternatively, if the total reimbursements for an episode add up to $1,000 more than the bundled price, $700 of that overage will come from the hospital (how this is done is explained in Step 5)
  • Option 4.1: Use the same percentages as the payment allocations from Step 2
  • Option 4.2: Use the percent of savings that would be attributable to the work of each provider type
    • If we think that 30% of the potential savings could be generated by surgeons’ efforts to lower costs, surgeons would bear 30% of the risk

Step 5: Agree to a process for performing retrospective reconciliations

  • If the hospital is required to perform retrospective reconciliations to allocate money according to hospital-caregiver risk-sharing arrangements, follow the same schedule as the one the payer uses for its reconciliations (see Step 3 of the payer-hospital contracting section)
  • Here is the recommended process for calculating the reconciliation payment that should be given to each risk-bearing provider (this process works the same regardless of the reimbursement method used by the payer-hospital contract):

Reconciliation Payment  = Total Deserved Reimbursement  – Amount Already Paid

Total Deserved Reimbursement  = Episode-related Claims  + Risk-sharing Amount

Risk-sharing Amount  = Risk %  x (Reduced Bundled Price  – Total of All Risk-sharing Providers’ Episode-related Claims)

Reduced Bundled Price  = Net Bundled Price Received from Payer  – Total Amount Already Paid to Non-risk-sharing Providers

Net Bundled Price Received from Payer   =   Total Bundled Price   +/-   Adjustments Made During Payer’s Reconciliation Process

  • If any reconciliation payment is calculated to be negative (meaning the provider was already paid more than the Total Deserved Amount), that provider will not be back-billed; instead, the following adjustment formula can be used to adjust the reconciliation payments for all the non-overpaid risk-bearing providers:

Adjusted Reconciliation Payment  = Reconciliation Payment  – Adjustment

Adjustment  = Adjusted Risk %  x Total Amount Overpaid to Risk-bearing Providers

Adjusted Risk %  = Risk %  / (100%  – Sum of Risk Percentages of Overpaid Providers)

  • In the event that a provider’s Adjusted Reconciliation Payment is calculated to be negative, repeat the adjustment formula after performing the following adjustments:
    • Set that provider’s Adjusted Reconciliation Payment to zero
    • Subtract that provider’s Reconciliation Payment from the Total Amount Overpaid to Risk-bearing Providers
    • Add that provider’s Risk % to the Sum of Risk Percentages of Overpaid Providers

Step 6: Changes to Step 1 – Step 5 for salaried caregivers

  • Step 1: If the salaried caregiver will not bear risk, no reimbursement changes are necessary—the caregiver will not be included in this contract
  • Step 2
    • If the payer-hospital contract is using FFS or reduced FFS as its reimbursement method: This step is not necessary
    • If the payer-hospital contract is using lump-sum payments paid up front or monthly as its reimbursement method: The bundle percentages allocated to salaried caregivers will be added to the employing hospital’s (for use in paying their salaries)
  • Step 3
    • If the payer-hospital contract is using FFS or reduced FFS as its reimbursement method: This step is not necessary
    • If the payer-hospital contract is using lump-sum payments paid up front or monthly as its reimbursement method: The hospital will contribute the appropriate amount of its allocated bundle percentage, which, as described in the changes to Step 2 above, includes its bundle percentage and the salaried caregiver’s bundle percentage
  • Step 4: No changes necessary; the salaried caregiver can still be allocated a risk percentage just like any other caregiver (refer to the changes to Step 5 below for an explanation of how this will work)
  • Step 5: Perform reconciliation calculations for salaried caregivers the same as for other caregivers except for the changes described below
    • The salaried caregiver’s Episode-related Claims and Amount Already Paid will be set to $0 and both will be added to the employing hospital’s amounts instead
    • In cases where the Total of All Risk-sharing Providers’ Episode-related Claims > Reduced Bundled Price, the salaried caregiver’s Risk % will be set to 0 and will instead be added to the employing hospital’s Risk %
    • The effect of these adjustments is to shift salaried caregivers’ downside risk to their employing hospitals but to preserve salaried caregivers’ upside risk so that they will share in the savings when they coordinate with other providers to lower costs

Step 7: Changes to Step 1 – Step 5 for caregivers that are reimbursed by the payer via capitation

  • Step 1: If the capitated caregiver will not bear risk, no reimbursement changes are necessary—the caregiver will not be included in this contract
  • Step 2
    • If the payer-hospital contract is using FFS or reduced FFS as its reimbursement method: This step is not necessary
    • If the payer-hospital contract is using lump-sum payments paid up front or monthly as its reimbursement method: The payer will subtract the bundle percentages allocated to capitated caregivers from the bundled price before giving it to the hospital (because the payer has already reimbursed the capitated caregiver for the services that will be rendered as part of the bundle)
  • Step 3
    • If the payer-hospital contract is using FFS or reduced FFS as its reimbursement method: This step is not necessary
    • If the payer-hospital contract is using lump-sum payments paid up front or monthly as its reimbursement method: Capitated caregivers will not be asked to contribute any of their capitated payments into the withhold, which means they will have no downside risk, but see the changes to Step 5 described below for an explanation of why an unequal upside/downside risk allocation for capitated caregivers is reasonable
  • Step 4: No changes necessary
  • Step 5: Perform reconciliation calculations for capitated caregivers the same as for other caregivers except that the capitated caregiver’s Amount Already Paid and Episode-related Costs should both be set to equal the amount the payer initially subtracted from the bundled price due to the caregiver being capitated
    • The effect of these adjustments is to make the bundle costs attributed to capitated caregivers always be the same, which means their direct costs will not place a financial burden on other providers by contributing to any overages; but, because upside risk will still be allocated to capitated caregivers, they will have an incentive to coordinate with other providers because they will get to share in the savings generated

A Patient Engagement Pilot

If I could do one pilot to show how healthcare needs to change, this is what I would do . . . (See my next post for the nuts and bolts of how this could be implemented.) (P.S. I know this isn’t of general interest even to people interested in health policy, but I wanted it to be available to the world, so here it is.)


  • Remove barriers to patients being able to choose the highest-value providers, which will reward those providers financially for being higher value and encourage lower-value providers to innovate to improve their value as well


  • Initially, the pilot will only focus on a single care episode (for example, hip replacements)

Two Core Design Principles of the Pilot

  1. Uniform incentives for providers
  2. Engage patients to use quality and price when choosing among providers

How These Principles Will Be Implemented

  • All payers (public and private) will be encouraged to participate inasmuch as regulation allows (see the note on Medicare and Medicaid participation below)
  • Payers and providers will participate in determining the care episode to be tested, what services will be included in that care episode, and which quality metrics will be reported by all participating providers
  • Providers will report the standardized quality metrics, which will be made available for patients to view on (this website needs to be a widely known one-stop-shop for patients to get information on healthcare providers’ quality)
  • Payers and providers will amend contracts (if needed) to reimburse for the care episode via episode-based pricing; capitated contracts could also work, but there are issues*
  • Payers will adjust cost-sharing requirements for the care episode to allow patients to bear all or part of the price differential between providers (by having patients pay less out of pocket when they choose lower-priced providers and/or by having them pay more out of pocket when they choose higher-priced providers) (for example, reference pricing or different copay tiers)
  • Payers will also have mechanisms to assist patients who need help understanding these plan benefit changes

Predicted Impacts of the Pilot

  • Short-term effect: More patients will choose higher-value providers (lower prices and/or higher quality) because they will have price and quality information and will have cost-sharing incentives to use that information when they make decisions among providers
  • Long-term effect (will have a more significant impact on value): Providers will be assured profit increases via increased market share if they can innovate to raise their value relative to competitors, so provider-led value-improving innovation efforts will increase
    • To be able to observe this long-term effect, we need to choose a care episode that is complex enough and variable enough in its cost and outcomes that there is a significant opportunity for care redesign to improve value (e.g., hip and knee replacements, coronary artery bypass grafts)

Why Broad Participation Is Key

  • Without the majority of patients being engaged in this way, providers’ potential market share rewards for value improvements may be insufficient to spur innovation efforts
  • If payer reimbursement policies are not uniform, providers’ value-improving innovations may result in reimbursement reductions (and, therefore, opposing incentives) from non-participating payers
    • e.g., a provider redesigns post-surgical knee replacement care and reduces readmissions within 90 days by 50%; that provider will still receive the same amount of reimbursement from participating payers, but it will lose a substantial amount of revenue from non-participating payers, and this revenue reduction could be enough to dissuade providers from doing these care redesigns

Pilot Evaluation

  • Data from an all-payer claims database will be used to track the average total price of the care episode in the market
  • The pilot’s standardized quality metrics will be used to track to average quality of care

A Note on Medicare and Medicaid Participation

  • Regulations on cost sharing for publicly insured patients will limit the ability for public payers to amend cost-sharing requirements; but, for the sake of providers having uniform incentives, public payers are encouraged to still participate by at least amending contracts with providers to reimburse them in the ways described above
  • Public payers can be rewarded for pilot participation because they can set prospective prices at historical averages minus 2%, and providers will most likely be willing to accept the slight price reduction because it will be more than compensated for by the fact that they will get to keep all the savings they generate through their cost-lowering care innovations

* I won’t get into them now, but the issues have to do with the dilution of incentives caused by one party doing a lot of work to innovate and then having to share the benefits of that innovation with everyone in the organization. Would you do extra chores if your parents split the extra allowance you earned between you and all your siblings?

What Is an ACO? What Is a Medical Home? What Is Bundled Billing? What Is P4P?

Image credit:
Image credit:

Our healthcare system is currently in experimentation mode–we are trying thousands of experiments to figure out how providers can be rewarded for “value instead of volume.” All the new terminology and reimbursement ideas accompanying these reforms can be hard to keep straight if you’re not steeped in this stuff every day, but guess what? There aren’t actually that many different ideas being tried; there are just a bunch of the same ideas being tried in various combinations. First I’ll describe those four basic ideas, and then I’ll show how they are the building blocks of all the main payment reform experiments out there.

Quality bonus: Give a provider more money when he hits performance targets on whatever quality metrics are important to the payer.

Utilization bonus: Utilization metrics and quality metrics are not usually separated, but they should be. Here’s the difference: improved performance on a quality metric increases spending; improved performance on a utilization metric decreases spending. They both improve quality, but they have different effects on total healthcare spending. So, for example, ED utilization rates and readmission rates would be considered utilization metrics. And childhood immunization rates and smoking cessation rates would also be considered utilization metrics because they tend to be cost saving. Insurers love giving providers bonuses on utilization metrics because they are stimulating providers to lower the amount being spent on healthcare.

Shared savings: If a provider can decrease spending for an episode of care (which could be defined as narrowly as all the care involved in performing a single surgery or as broadly as all the care a person needs for an entire year of chronic disease management), the insurer will share some of those savings with him.

Capitation: The amount a provider gets paid is prospectively determined and will not change regardless of how much or how little care that patient ends up receiving. Again, this could be defined narrowly, such as all the care involved in performing a single surgery (in which case it’s actually called a “bundled payment”), or it could be defined broadly, such as for all the care a person needs for an entire year.

By the way, did you notice that shared savings and capitation are almost the same thing? The only difference is who bears how much risk. In shared savings, the risk is shared, which means that if the costs of care come in lower than expected, the insurer gets some of the savings and the provider gets some of the savings. In capitation, the provider bears all the risk, which means that if the costs of care come in lower than expected, the provider gets all of the savings.

Okay, now that I have listed out those four ideas, take a look at the popular payment reforms of the day . . .

Medical Home

General idea:

  • Give a primary-care provider a per member per month “care management fee” (in addition to what he normally gets paid) for providing additional services (such as care coordination with specialists, after-hours access to care, care management plans for complex patients, and more)
  • Also give the primary-care provider bonuses when he meets cost and/or quality targets

Breaking down a medical home:

  • A care management fee is actually a utilization bonus (because the net effect of the provider offering all those services is to avoid a lot of care down the road)
  • A bonus for meeting quality targets is either a quality bonus or a utilization bonus depending on the specific metrics used
  • A bonus for meeting cost targets is shared savings

Accountable Care Organization (ACO)

General idea:

  • Give a group of providers bonuses when they lower the total cost of care of their patients (but the bonuses are contingent upon meeting quality targets).

Breaking down an ACO:

  • A bonus for lowering the total cost of care is shared savings
  • When a bonus is contingent upon meeting quality targets, that means it’s also a quality or utilization bonus (depending on the specific quality metrics used)

Pay for Performance (P4P)

General idea:

  • Give a provider bonuses when she meets quality targets.

Breaking down P4P:

  • This is either a quality or utilization bonus (depending on the specific quality metrics used), but it tends to be utilization bonuses because insurers especially like when providers decrease the amount of money they have to fork out

Bundled Payment/Episode-of-care Payment

General idea:

  • Give a group of providers a single payment for an episode of care regardless of the services provided.

Breaking down bundled payment:

  • A bundled payment is a narrow form of capitation

There you have it. They are all repetitions of the same ideas but combined in different ways.

How to Change How Prices Are Set in Healthcare

Image credit:
Image credit:

In my previous post, I described the three ways prices can be set in healthcare: administrative pricing, bargaining power-based pricing, and competitive pricing. I also bemoaned the fact that the prices paid to providers by private insurers are determined more by relative market share than by anything else* . . . but this post explains how we can change all that.

I see two possible pathways from bargaining power-based pricing to competitive pricing, so here they are.

First Pathway

Let’s pretend a colonoscopy clinic is super innovative in how they do things, and they eventually are able to lower their costs by 20%. This is great for them because the prices they are paid by insurers has stayed the same (remember, those prices are determined primarily by relative market share, not costs), so now they have a really solid profit margin. And yet, the managers of this clinic still aren’t satisfied because they have excess capacity they want to fill.

One day, the managers come up with an idea. They say, “Up to this point, we’ve always maximally leveraged our bargaining power with insurers to win the highest prices we possibly can, but what if we do something radically different? What if we offer to lower our prices by 10% in exchange for the insurers putting us in a new, lower-copay tier? This would induce way more of their policy holders to choose us for their colonoscopies, so we’ll fill up our excess capacity. And, according to our calculations, our increased volume will more than make up for the lower prices. So everyone wins! Our profit increases, the insurer saves money with the lowered prices, and the patients are happy because their copays are less as well.”

Soon, some of the clinic’s competitors would figure out why their volume is starting to drop, so they would probably find a way to offer lower prices to get put into that lower-copay tier as well. Competitors who can’t or won’t lower prices will slowly lose market share until they, too, are forced to either lower price or improve quality enough to convince patients that going to them is worth the extra money.

Voila! Competitive pricing.

I have a friend who manages a large self-insured employer’s insurance plan, and I asked him what he would do if a clinic came to him offering lower prices in exchange for steering more employees to it. He said as long as the provider can show that quality won’t go down with the additional volume and that wait times for appointments won’t increase, he’d probably go for it.

Now, of course this wouldn’t work with every kind of healthcare service. I purposely chose a non-emergency service that already has pretty straightforward pricing. But as a priori quality and pricing information becomes more available, more services will be candidates for this pathway to competitive pricing.

One other point: Hospitals generally do a horrible job of cost accounting (they’re just such complex organizations!), so they usually have no idea if a proposed price reduction will still be profitable or not. Thus, they’ll be left behind in this game until they start to develop better cost accounting methods. If they have some foresight, they’ll start fixing that now.

Second Pathway

An insurer is despairing the fact that many of the providers in the region have combined into a single price-negotiating group, so now the insurer is stuck paying way higher prices than before. But then some health policy-savvy managers figure out a solution. They say, “Let’s implement reference pricing for a bunch of non-emergency, straightforward services. Let’s start with colonoscopies. This is how it works. We’ll tell our policy holders that we’ll put $1,200 toward a colonoscopy (the “reference price”). If a policy holder chooses a provider who charges more than that, they will pay everything over that price. A few clinics in the area offer prices that are lower than $1,200, so policy holders will still have a few options if they don’t want to pay a dime. But, (and here’s the best part) the price the providers in that huge price-negotiating group forced us to accept is $3,000, so they’re definitely going to lose a lot of volume, which will probably force them to lower their price.”

Soon other insurers jump on the reference pricing bandwagon and higher-priced providers who are losing tons of volume will be forced to price competitively.

In conclusion, shifting to competitive pricing is not immediately possible with most healthcare services. But the way to make more healthcare services amenable to competitive pricing is to improve a priori quality and pricing information: quality information needs to be standardized and more relevant to the factors patients should be considering when they’re choosing between providers, and the full price of an episode of care needs to be available beforehand so patients can compare them apples to apples. Only after these changes happen will we be able to rely more on competitive pricing, which, most importantly, will do more to stimulate value-improving innovations in our healthcare system than almost anything else.

* I also complained about how administrative prices don’t encourage (and actually stifle) innovation toward higher quality and lower prices. Check out the Uwe Reinhardt quote in this blog post and then think, “Uwe must have been reading Taylor’s blog.”

The Three Different Ways We Could Set Prices in Healthcare

Image credit: AP Photo/Damian Dovarganes
Image credit: AP Photo/Damian Dovarganes

Out of the three general ways we could set prices in our healthcare system, one is best. Too bad we’re using the other two.

First, I’ll describe each method:

  1. Administrative pricing: This one is very straightforward. The government says, “For procedure A, healthcare providers will be paid X dollars.” Usually the methods for coming up with that dollar amount are sophisticated and rely on the best available data, but not always because they are subject to various political influences and government budgets.
  2. Bargaining power-based pricing: This one is easiest to explain using an example. Think of a small town with only two family medicine docs. One, Dr. Awesome, treats 90% of the town’s residents; the other, Dr. Mediocre, treats the other 10%. All patients are insured by one of four different private insurers, each of which has approximately equal market share. Now think of Dr. Awesome sitting down at the bargaining table with one of the insurers to decide on prices. He says, “If you don’t pay me at least Medicare rates times 1.4, I won’t accept your insurance. I’m serious, I won’t accept anything less.” And the insurer says, “Hey, that’s a horrible deal, but if we stop covering care you provide then most of our policy holders in your town will just switch to one of our three competitors and we’d lose out on even more profit!” Now think of the conversation between Dr. Mediocre and that same insurer. Dr. Mediocre says, “Pay me Medicare rates times 1.4.” And the insurer responds, “No. We’ll pay you Medicare times 0.8. If you say no and we don’t have you listed as a provider in our network anymore, that’s okay because only a tiny percentage of our policy holders are your patients. And we know that you don’t have many patients, so you can’t afford to risk losing 1/4 of them by saying no to the price we offer.” Relative market share between the two parties is the primary determinant of bargaining power, so a bigger market share means you can get a better price.*
  3. Competitive pricing: This is the method used to determine prices in almost every other industry. Here’s basically how it plays out: Competitor A says, “Everyone knows that our product has similar quality to our competitor’s product, so we can’t price it higher than theirs without sacrificing quite a bit of market share. We could sell it for less than theirs to win more market share, but then the price is perilously close to our costs, so we’ll have to do some math to see what the profit-maximizing price/market share combination is likely to be.” Note the one huge condition that is required for this to work: Potential customers must be able to compare the price and quality of all their options, which is starting to happen more and more as better quality information is starting to become available and as prices are becoming more transparent.

Our healthcare system currently relies primarily on number 1 (think: Medicare and Medicaid) and number 2 (think: private insurers and providers setting prices with each other). But which method is best?

If you want to have the lowest possible prices, administrative pricing is the obvious best choice. But that’s only for the short term (as you’ll see), and it does nothing to encourage quality improvement unless you start getting into the treacherous area of performance incentives.

The only thing I’ll say about bargaining power-based pricing is that I don’t like it. I’d rather not have prices that are totally unrelated to costs or quality and instead are determined by relative market share.

Now let me tell you why I like competitive pricing so much. I want our healthcare system to deliver better value right now (Value = Quality / Price), and even more than that I want that value to go up over time as providers and insurers innovate in ways that allow them to decrease prices, increase quality, or both. Competitive pricing is the only method that provides an incentive for competitors to innovate because it rewards the highest-value offerings with increased market share and profit. The other two options don’t do that, which seems like a pretty big downside, don’t you think? I’d be willing to forgo short-term super-low administratively set prices in favor of stimulating innovation that will improve value way more over the long term.

In my next post, I’ll explain how we can shift from bargaining power-based pricing to competitive pricing.

* Do you ever hear those arguments that if public insurers lower their prices any more, providers will just raise their prices for private insurers? Well, now you know why those arguments are mostly hogwash. Providers are already leveraging their relative market share to get as high prices as possible from private insurers, and getting paid less by public insurers doesn’t change that relative market share.

How Backward Integration Is Starting to Fix Healthcare Delivery

A good friend just told me about Montana’s state-run clinics that are only for state employees. Going to the clinic is free for state employees, which means the state is paying for everything. And yet, despite paying for everything, the clinics are doing such a good job of managing diseases that the state is actually saving more money than it’s spending on the clinics.

I’ve talked about the importance of cost-saving prevention before, but my point in describing this example is to illustrate a growing trend in healthcare–a trend that is largely unrecognized, but is starting to fix healthcare. So let’s break it down.

Think of the Montana state government as a company. This company, just like most companies, has suppliers that sell it critical inputs it needs to perform its services. And one of the most important suppliers to this company isn’t obvious: healthcare providers. Think about it–they are supplying the healthcare that keeps employees productive, which is surely a critical input.

And here’s the interesting thing about the relationships companies have with their suppliers: if the supplier’s product is too expensive, or isn’t good enough in some way, companies will sometimes just take over the production of that critical input themselves. This is called “backward integration.” Think of all the ways employers are backward integrating into healthcare, whether it’s having their own salaried physicians or working closely with providers to redesign care processes; they’re all variations on the same theme.

But employers aren’t the only ones with a supplier-buyer relationship with healthcare providers. Insurers depend on providers to supply the healthcare they are guaranteeing to their customers. So are insurers backward integrating as well? YES. Any time an insurer joins up with a provider, it could be seen as an attempt by insurers to backward integrate (ahem, ACOs). And insurers are also going crazy trying all sorts of hands-off approaches to backward integration (if it’s hands-off, can it still be called backward integration?) with things like pay for performance, bonuses for starting medical homes, and probably hundreds of other experiments I’ve never heard of. They are all attempts to exert some degree of control over the unsatisfactory supplier. Or, in other words, to fix healthcare delivery.

So, I guess we could say that employers and insurers are fixing healthcare delivery. Strange, isn’t it?

[Update: This is good and all, but there are only so many innovative things a single provider can develop, which is why an even better (system-wide) solution would be to do the following: get patients to choose the highest-value providers, which then rewards those highest-value providers with market share, which creates an incentive for all providers to be innovating to win more patients. This idea is expounded more in other posts on this blog. Anyway, in the meantime, this backward integration thing is a great alternative.]

%d bloggers like this: