Saturday, March 05, 2011

Essays from Ernie's students - 4

Here's the last of the four chosen essays from Ernst Berndt's MIT class at the Sloan School. As above, please offer your observations and suggest a grade in the comments below. This one was written by Ramesh Chandra. I include the entire essay, as it was the only one to include an alternative payment and patient care mechanism. There are three other selected essays.

Do AQCs solve the U.S. HealthCare System’s Woes?

1 Introduction

It is widely accepted that the U.S. health-care system is highly inefficient, with estimates of up to 30% inefficiency. The important causes of this inefficiency are considered to be: (i) fee-for-service payment methods that result in over prescription of care by doctors even when there is no significant benefit [6] (ii) lack of coordination in care delivered to the same patient by different doctors (iii) overuse of care by patients due to moral hazard, driven by the perception that more care is better care, and (iv) relatively little attention to prevention and wellness programs compared to treatment programs (also a fallout of the fee-for-service payment policies.)

One solution suggested to address these issues is to have care providers have more “skin in the game” by using global payment schemes. The idea is to allocate a fixed annual budget per patient and have the care provider manage it to deliver the care required by the patient, with bonuses awarded to the provider if it efficiently manages the budget.

In this essay, we study a specific type of global payment scheme designed by the BCBS of MA, called the Alternative Quality Contract (AQC). We look at the AQC model’s goals and analyze it to understand its limitations. We also briefly look at an alternate design, which we believe works better than the AQC in practice.

2 Overview of the AQC model

The salient features of an AQC are as follows:

1. The contract is between BCBS and provider organizations with primary care physicians (eg: HMOs, POS) for a period of five years.

2. Each patient that the provider organization provides primary care for, is allocated an annual budget. All health care expenses for the patient come from this annual budget, even if care is provided by a different organization. If the expenses are lower than the budget, the provider organization gets to keep the savings; if the expenses exceed the budget, the provider is liable for the extra cost.

3. To incentivize providers to take care of their patient’s health (and not optimize for the short term by denying care), they are paid an additional bonus of up to 10% if they hit certain health quality measures for their patients.

4. The first year’s budget is determined by the patient’s prior year’s medical expenses. The rate of increase of the budget is also negotiated based on trends.

5. Patients are unaware of whether their PCPs are part of an AQC.

The goal of the AQC is to contain the ballooning costs of health care, and it plans to achieve that by making providers more “accountable” for the costs of their primary care patients. At a high-level, it appears that the AQC’s design is aligned with its goals—Ideally, the annual budgets would motivate physicians to reduce spending on unnecessary care, while the performance bonuses would motivate them to keep their patients healthy, thereby not skimping on necessary costs.

Accountable Care Organizations (ACO) are similar in concept to the AQC, and are designed for Medicare; the main difference appears to be that the provider is not liable if expenses exceed the budget, but receive rewards for reducing costs and meeting quality improvement markers.

3 Will the AQC model work in practice?

To understand the AQC model better, let us consider a simple thought experiment. Imagine that the entire healthcare system consists of Reliable Insurance Corp. (RI) and Expert Medical Center (EMC). Let us further assume that the only illness that patients acquire is backache. EMC operates in two states, A and B, with one primary care physician and five orthopedic surgeons per state (ratio consistent with current ratios of PCPs vs specialists). Assume that 1000 patients get backache each year in each state, out of which 100 need surgery, 800 can be cured in 2 weeks by getting ergonomic chairs, and the remaining 100 are the “indeterminate” group—they can either be operated on or can be cured in 2 months by getting ergonomic chairs. A new chair costs $100, whereas back surgery costs $10K, of which $5K goes to the surgeon and $5K to EMC.

Now, to keep the example simple, assume that all orthopedic surgeons in state A recommend chairs for the patients in the indeterminate group, whereas surgeons in state B recommend surgery. Correspondingly, EMC has over the years built enough facilities to accommodate 100 surgeries in state A and 200 surgeries in state B. Thus, the total annual healthcare costs for the entire system are $3.17M, out of which $990K could be saved with no change in outcome.
Let us say RI wants to realize these savings by entering into an AQC with EMC. Would it succeed? Let us look at it from each party’s perspective. (here we assume that all parties are rational and that earned income is the motivating factor):

1. Will EMC be motivated to enter into a contract? If so, at what annual budget? Since EMC already built the facilities needed for doing 300 surgeries per year, it cannot forgo the $500K it gets for the 100 surgeries in the indeterminate group. So, it needs the annual budget to be at least $2.18M + $500K = $2.68M

2. Will the orthopedic surgeons agree to enter into the contract? If so, at what budget? The orthopedic surgeons in state B lose half their income if surgery is not done on the indeterminate patients. So, they would resist the AQC, unless their $500K of salaries is incorporated into the budget and EMC assures them that their income will not be adversely affected.

3. Will the primary care physicians want to enter into the contract? In this example, PCPs won’t see any difference to their income levels and so would be fine entering the contract. However, they may ask for additional perks to “sweeten” the deal.

4. Will RI be able to realize these savings by entering into an AQC with EMC? From the above analysis, we see that in practice, RI will not be able to realize the savings. In fact, it may cost RI more, due to the administrative overheads, and because each of the other parties may expect perks for the inconvenience of getting involved in the contract.

The above example is arguably simple, but gives a flavor of the challenges involved in implementing AQC. We believe that the points made above would hold in real world scenarios. Consider the following examples:

• It is widely agreed that there is 30% waste in the system. However, the current AQC contracts that BCBS entered into start above the current spending levels. This seems to be tacit acknowledgement that this waste cannot be recovered.

• In addition, it appears that BCBS “has offered physician groups and hospitals sky-high reimbursement rates to persuade them to sign up for a new type of health care plan," which contributed to its loss in 2009.

• Some ACO providers in Medicare “want to avoid being financially responsible for patients who go outside for care”, they want to be able to “pick and choose” patients, or have asked “to limit quality measures” [2]. All of these point to providers trying hard to protect their income sources.

The above trends validate our simple thought experiment. The take-away point is that one man’s expense is another man’s income. In other words, the waste in the health care system is income to a significant portion of the providers, and it is natural that they will only enter into contracts (or try to mold the contracts so) that their income is not jeopardized. This would run against the grain of efforts to reduce waste in the system.

There are two other issues with the AQC, which we briefly discuss below:

1. The AQC, as it is currently structured, is more of a risk-transfer mechanism from the insurance provider to the care providers. This makes the provider organizations more like insurance companies. This means that they get more complicated and they need finance/insurance personnel who can negotiate AQC contracts, purchase reinsurance, manage budgets etc., This increases the already high administrative costs in hospitals. More importantly, this further detracts hospitals and doctors from their primary job—practicing good medicine.

2. Another key issue with the AQC is the glaring lack of any role for the patient in it. The fact that the AQC participation of PCPs is hidden from patients can make them distrustful of their PCPs and the AQC model, and lead to perceptions that care would be withheld from them.

The above fundamental issues can jeopardize the success of AQC in achieving its goals, and instead result in a more expensive and more complex system, with more dissatisfaction and distrust among the different parties. Given these limitations of AQC, we believe it would have a difficult time succeeding in its current form, regardless of the allocation of resources among care providers.

4 Is there an alternative?

The AQC suffers from poor system design (as does the rest of the healthcare system), as it does not account for human behavior. Given that the current health care system is a supply-driven market, it is rather difficult to negotiate with the suppliers to agree to restrict supply, unless they are promised no reduction in profits. This fundamental conflict works against the AQC/ACO model’s goal to reduce waste.

So, what is the alternative? We look to the end user/patient for a solution — one whom the AQC model ignores. We feel that the end user is likely the only party with enough wherewithal to provide the required counterbalance to excessive supply — after all, the entire system exists for the end user. But for that to happen, we need to have informed patients who have a basic understanding of healthy living, understand how to navigate the health care system, and have the confidence to not blindly go along with whatever their doctor recommends. Let’s go through the above thought experiment again with informed patients in the mix.

Imagine that our informed patients have full knowledge of all the past procedures done by the orthopedists. It would be apparent to them that the outcomes of treatments by all the orthopedists is the same, but that orthopedists in state A do fewer percentage of surgeries than those in state B. Also, our knowledgeable patients understand that surgery can have long-term ill-effects. So, when a patient is recommended surgery by a orthopedist from state B, she would solicit a second opinion from an orthopedist from state A. If the opinion differs, she may (i) ask her original orthopedist for the reason for difference in opinion, and/or (ii) use an ergonomic chair for 2 months to see whether her condition improves. Once enough number of patients realize that they can be cured without surgery, the orthopedists in state B would be “pressured by the market and the cause of science” to revise their treatment procedure, to avoid falling out of favor of their customers (or worse to avoid getting sued). This leads to much better outcome than RI trying to negotiate an AQC with EMC. Note that the care providers are still paid on a fee-for-service basis.

So, how can RI get its users to be more informed? Here’s one potential design: RI can encourage the formation of a health advisor organization, whose sole charter is to guide consumers through the health care process and keep them healthy. RI can compensate the health advisor with a minimal fixed fee per user and tie the majority of compensation to the health care savings of its users (with compensation spread over several years and weighted by the health of its users to avoid short term optimizations). A health advisor itself does not perform any procedure and does not offer medical advice. Its users are not bound to take procedural advice it offers—the only leverage it has with its users is mutual trust. Also, to avoid conflict of interest, it cannot obtain revenue from any other sources. Given these constraints, the health advisor can only increase its revenue by (i) chasing down major waste in the system (ii) finding ways to mitigate this waste while maintaining user health (or even better, improving it, as in the case of avoided surgeries) (iii) gaining user trust, and (iv) convincing them to do things that are good for their health, which also happen to reduce waste. We can imagine several ways in which the health advisor organization can do this (collect effectiveness metrics on healthcare providers; gather scientific studies on effectiveness of various procedures; suggest users take second opinions; run ads to convince customers to practice prevention; build health IT systems to take care of customers—reminders for pills, for appointments, preventive care etc.,) If we reflect on the the above thought experiment again, it is easy to see how these steps of a health advisor would map to the actions of informed patients.

If the guidance of health advisors convinces customers to reduce consumption of health care services, that acts as a market signal to providers to throttle expansion of unnecessary services, achieving the desired result of reducing waste.

5 Concluding remarks

In this essay we analyzed AQC, a global payments proposal to contain health care costs, and show that its design suffers from a fundamental challenge — convincing suppliers in a supply-driven market to reduce their revenues. We outlined an alternate consumer-centric mechanism that instead focuses on educating and guiding consumers on how to be healthy, while at the same time reducing consumption of unnecessary medical care, thereby reducing demand. We believe that a consumer-driven approach provides the right counterbalance to the supply glut. The limited size of this essay does not permit a more detailed discussion of this approach, and the rationale behind the arguments against current approaches; however, we end with a
note on the wider applicability of the consumer-driven approach, by pointing out how it applies to the Indian healthcare scenario. The health care system in India is also increasingly suffering from a supply-driven approach. Modern hospitals are outfitted with expensive equipment, and doctors are prone to prescribing procedures to put this equipment to use. Furthermore, the fee-for-service approach leads to more focus on treatment than on prevention. Patients are typically timid, rarely question their doctor’s analyses, and do not generally seek second opinion. All this, even though most patients pay for their medical expenses out of their own pocket. We believe that a health advisor organization, as proposed above, would also work well in that scenario.


Anonymous said...

Allowing for the fact that this was the only essay presented in its entirety, I give this one an A; for thorough analysis, scientific approach insofar as possible for this subject, and original thinking.


Paul Levy said...

Dear Anon 10:44 (this morning),

Sorry, but I am not publishing your comment because your choice of a word ends up being an ad hominem remark about a person. I do not know if you meant it that way or not, but I don't publish things like that.

Keith said...

You could decrease overall costs without decreasing profit margins to the various players. The model is essentially like what the goverment has done in the past with agriculture, where some farmers are actually compensated to leave their fields fallow in an attempt to regulate production. Instead of decreasing the supply of corn, they are attempting to reduce the number of very expensive procedures and tests of questionable or marginal benefit.

The orthopedist could be paid the same even if he/she does not perform as many surgeries. It is all the inherent costs that can be saved from the decision not to perform an expensive surgery that will go into some bonus pool.

Specialists are expensive not only because of what they need to be paid, but often because their decisions to utilize their skills often cost alot more in ancillary costs. Take away the incentive to use these expensive resources (hospital operating rooms, MRI scanners, new fancy pacemakers, etc. etc.), you will get some real savings. That is why the eventual result of health care reform that is being pushed is one in which the physician is salaried with no incentive to perform higher cost procedures to boost his/her pay as is currently the case. This may actually be the goal of the ACO, but this idea does not sit well with docs, so we are instead being fed this concept of better communication and integration as the reason for ACOs.

Mayo is the model the goverment wonks like and are ultimately shooting for and conceivably is the system they tout as most efficient. While communication is excellent amongst the specialist at Mayo, the main reason they can be more efficient is that the proceduralists are not incentivised to do more.

Thomas said...

I like the emphasis on the end-user and the concept of Health Advisors. I believe that if you spoke to any primary care physician, that is how she would see herself. I think that we need to recognize that we need to continue to utilize more PA's, NP's, etc and move physicians into the consultative/oversight/procedural areas. Removing the incentivized FFS system is important, but the key is changing patient expectations and behaviors; as this writer points out, better pt education is fundamental.

Anonymous said...

Small fact check: AQCs are between BCBS and physician groups, not hospitals. The "hospitals" that may have thrown off the essay's author are PHOs. I'd direct readers to this Health Affairs article: