Several years ago, my colleagues at Los Alamos and I did a little experiment. We were interested in whether machines or humans were better at unraveling the logic of a non-intuitive problem. We chose the Monty Hall Problem as our prototype problem. The problem is loosely based on the popular game show Let’s Make a Deal hosted by Monty Hall. Stated simply, the problem is

Suppose you’re on a game show, and you’re given the choice of three doors: Behind one door is a car; behind the others, goats. You pick a door, say No. 1, and the host, who knows what’s behind the doors, opens another door, say No. 3, which has a goat. He then says to you, “Do you want to pick door No. 2?” Is it to your advantage to switch your choice?

The answer seems simple. It should not matter whether one swaps doors or not. The probability is 50% that the car will be behind either remaining door. This answer is, however, incorrect. The correct answer is that one should always switch doors. The chance of winning a car is 2/3 if one switches and 1/3 if one does not switch. This answer is so non-intuitive that the discussion on the topic made Marilyn vos Savant’s column in Parade Magazine one of the hottest columns of 1990. The great mathematician Paul Erdős refused to believe the answer until it was demonstrated to him by computer simulation.

At Los Alamos, we created our own computer simulation of the problem. We allowed the computer to randomly pick a door. The computer-generated game-show host picked a remaining door that held a goat behind it. And the computer randomly switched doors or not. The outcome was then revealed to the observer. The computer either won a car or a goat. We ran this experiment many times.

We created a second program that observed the outcome of the experiments and whether the computer switched doors or not, and then learned from its observations. We measured how quickly the program learned the correct answer. We then repeated the experiment with humans as observers rather than a computer program as an observer. We measured how quickly the human observers learned the correct answer and compared this with how quickly the computer program learned. The program was unquestionably a faster learner than the humans. We sent this result to vos Savant and it generated even further discomfort and disbelief in her readers.

Why was the computer program able to learn faster than humans? This may be related to what in psychology is called “anchoring.” “Anchoring is a cognitive bias that describes the common human tendency to rely too heavily on the first piece of information offered (the “anchor”) when making decisions.” In the Monty Hall Problem, the anchor is the belief that the probability of an outcome is inversely proportional to the number of possible outcomes. In other words, if there are three doors that a car can be behind, and the probability is random, then the car is behind any one door with probability of one chance in three. The computer program does not have the human bias. It simply looks at the evidence and makes an unbiased judgment, which is that the probabilities are not equal at all.

This suggests another popular television game show—Jeopardy. Jeopardy is a quiz show in which questions are inferred given the answers to the questions. In 2011, Watson, an artificial-intelligence computer system capable of answering questions posed in natural language, beat two former Jeopardy champions, Brad Rutter and Ken Jennings, who happened to be human. It is interesting to note that Watson had poorer performance compared with the humans in some categories such as those containing short clues containing only a few words. The machine, however, performed better than the champion humans overall. The lack of human learning shortcomings, such as anchoring bias, seems to have complemented and even enhanced our collective ability to organize and interpret information.

Watson is now working in collaboration with Memorial Sloan-Kettering in order to organize diagnostic information. Sloan-Kettering has taken on the task of tutoring Watson in cancer diagnosis. This problem may be a bit tougher than answering Jeopardy clues, but the payoff could be much more real. I quote from a March 2013 Atlantic Magazine article by Jonathan Cohn.

“The process of pulling out two key facts from a Jeopardy clue is totally different from pulling out all the relevant information, and its relationships, from a medical case,” says Ari Caroline, Sloan-Kettering’s director of quantitative analysis and strategic initiatives. “Sometimes there is conflicting information. People phrase things different ways.” But Caroline, who approached IBM about the research collaboration, nonetheless predicts that Watson will prove “very valuable”—particularly in a field like cancer treatment, in which the explosion of knowledge is already overwhelming. “If you’re looking down the road, there are going to be many more clinical options, many more subtleties around biomarkers … There will be nuances not just in interpreting the case but also in treating the case,” Caroline says. “You’re going to need a tool like Watson because the complexity and scale of information will be such that a typical decision tool couldn’t possibly handle it all.”

Perhaps machines like Watson will be an answer to the question of how to diagnose in an era of personalized medicine, in which a gene sequence costs $10, and the amount of medical information we have about any patient and his or her family is astronomical—beyond the capabilities of any group of humans without smart machines to make sense of. Good information management is cheaper than expensive medical procedures, particularly if the procedures are not backed by high-resolution information. Perhaps the type of information management that Watson represents can be key to lowering healthcare costs and improving outcomes.

This idea can be illustrated more colorfully with the famous hot-dog vendor example from economics. Consider a beach that contains, in addition to the bathers, two hot-dog vendors located at either end of the beach. There is nothing to differentiate their products, other than their location on the beach. Bathers will purchase from the closest vendor, therefore the value or quality of the product of a particular vendor depends on the number of bathers closer to the vendor’s cart than the competition. In order to capture market share and value, each vendor moves his cart closer and closer to the other until they meet in the middle of the beach. At this time, each vendor has captured half the market share of the beach, and, since they are in the same location, their differentiation has vanished. Their dogs are now commodities, and the market is driving the price to the cost of production. This is the classic process of commoditization.

Let’s apply this to healthcare. A perfect market for a pharmaceutical company is the hyperlipidemia market. Properly treated patients with high cholesterol can live a normally long lifetime, paying the manufacturers of Crestor or Lipitor quite a lot of money for the valuable service. Since this market is large, stable, and profitable, all pharma companies will want to enter this market, which is at the center of the healthcare beach. Diabetes is a similar market. If, however, the makers of Lipitor have a patent, other pharma companies cannot enter the exact same area of the beach as the Lipitor manufacturer. The patent forces the vendors to maintain some distance on our healthcare beach. The other non-Lipitor-producing pharma companies must come up with a product that is sufficiently different from Lipitor to be patented, or they must develop drugs for an entirely different disease, say pancreatic cancer. This forces companies to innovate and provide products for the entire healthcare beach, not just the stable lucrative parts of the beach. The downside for patients is that they must pay more for Lipitor while it is on patent. The upside is that pancreatic patients may receive life-extending treatments they would not otherwise have received.

The healthcare connections and tradeoffs are not always obvious to a Medicare patient entering a donut hole and wondering why the costs of medications are what they are. See related blogs for discussions of even more connections and tradeoffs in healthcare.

1. The amount of business risk that a pharmaceutical company is allowed to mitigate and the time interval over which the risk is mitigated
2. The quality and safety of pharmaceutical products
3. The cost of the products and the efficiency in which products get into the market

These three regulatory goals can be thought of as nodes of a triangle.

In the U.S. the temporal/risk node is controlled by the Patent Office. A pharmaceutical company that develops or licenses a drug has a certain amount of time to market and sell the drug without competition. Upon expiration of the patent, competitors are allowed to sell the drug, and the price of the drug nearly always falls precipitously. The expiration of the patent can be extended under certain conditions, such as significant improvements to the drug protected by follow-on patents.

The Food and Drug Administration controls the quality/safety component of the regulatory triangle. The FDA is responsible for approval of drugs entering the market. Their decisions are based on safety and efficacy of the drug. The approval process is extensive, involving several levels of clinical trials and post-launch monitoring. In addition, the FDA regulates the advertising and promotion of the drug.

The cost/efficiency component of the triangle is quite complex. The mechanism of control is through formularies. These are lists of drugs with the requirements for compensation for each drug. The requirements for compensation for the drug are determined by a myriad of entities. About 35% of healthcare expenditures are paid through the federal government and the Center for Medicare and Medicaid Services (CMS). Private insurers, which are not regulatory entities themselves, pay for most of the remainder. State governments regulate the private insurers. State regulation of private insurers focuses on solvency of insurance companies, risk spreading, fraud and ensuring that customers are paid benefits.[i]

The interplay of these three regulatory nodes can determine the profitability of a pharmaceutical product, the patients’ access to the product, and the risk that patients incur from the product.  Pharmaceutical business is as much about navigating the regulations as it is dealing with the competition.

See related blogs.

1. a diminished margin for improvement that escalates the level of difficulty in proving drug benefits
2. genomics and other new science have not yet reached their full potential
3. mergers and other business arrangements have decreased candidates
4. easy targets are the focus as chronic diseases are harder to study
5. failure rates have not improved
6. rapidly escalating costs and complexity decreases willingness/ability to bring many candidates forward into the clinic

From Pharma Marketing Blog, Monday, December 15, 2008

The bottom line, however, is that, despite tremendous financial incentive, the current traditional R&D process has failed to deliver significant new drugs in the last few years. The dried-up pipeline has led to a patent cliff in which many branded drugs have gone off patent with little in the pipeline to replace them. While the supply has diminished, the demand for new treatments has not.

Business experiments outside the traditional drug-development process have emerged to address this problem. They are taking advantage of the increased availability of information and increased abilities for communication and interaction that have occurred in the last few years. I will focus on one of these experiments, the Michael J. Fox Foundation for Parkinson’s Disease Research (MJFF), to illustrate.

The MJFF tries to open a bottleneck in the drug development process, the gap between basic research and translational research that converts scientific findings into potential cures for human diseases. Scientists tend to be a maverick lot, randomly looking under this stone and that for anything new and interesting—I can say this lovingly because I was a professional scientist for more than twenty years. The culture of testing new ideas to develop into cures is a more directed process that is driven by the need to help desperate patients. Many potentially valuable ideas do not make it across this cultural divide.

The MJFF attempts to shepherd ideas across this divide by forming teams of scientists and business people. Funding is provided to the scientists under the condition that they are subject to deadlines, milestones, and team-generated priorities. Unpromising ideas are pruned and priorities are reset periodically. The responsibility of the business people is to get the surviving ideas into the clinical-trial process. The MJFF also makes sure that the team gets the scientific tools they need to get the job done. These include biomarkers, pre-clinical models, and reagents. The foundation generates assistance and funding from the patient community. Finally, the MJFF attempts to break down scientific silos and increase communication in the scientific community by making introductions and sponsoring workshops.

Traditionally the directions of new-drug development have been driven by the pharmaceutical industry. We are now seeing that individuals through patient-driven organizations may be having an effect with MJFF being an example. It is difficult to tell from the MJFF website what has been the real impact of their efforts, but what is clear is that patient communities now have the tools to influence the supply of treatments in the healthcare market place. The new interconnectedness of the 21^st century allows not just individuals, but complementary communities of basic scientists, patients, business people, and drug developers to interact to identify and remove traditional obstacles in the delivery of healthcare.

See related blogs.

DNA Sequencing Costs. Data from the NHGRI Genome Sequencing Program (GSP)

This poses a dilemma for those people involved in the development of new treatments. How can the safety and efficacy of a treatment be determined if the treatment is designed for a single individual? Today, safety and efficacy are determined, in large part, by randomized control trials, clinical trials that typically involve hundreds of participants. The large number of patients is required to assure statistical confidence in the outcomes of the treatments. Clearly clinical trials must be designed that are able to build assurance in the predictability of the outcomes and do this with very small test populations. The challenges are at least twofold: confidence is difficult to build with small populations, and the identification and recruitment of the trial participants is quite difficult if the number of eligible participants is very small.

The second challenge, identification of trial participants, will be the topic of future blogs and involves data mining on an extensive scale. The first challenge, building confidence once a small population has been identified, may be aided by adaptive trial-design methods. In adaptive design the rules of data collection and sampling change as data is collected. Adaptive designs allow investigators to start with a small population and change the data that is being collected as data is collected. This speeds the knowledge-accumulation process, allowing for quicker decision-making.

Adaptations to clinical trials can take several forms: adaptive randomization; stopping a trial early due to safety, futility or efficacy at interim analysis; dropping the losers (or inferior treatment groups); sample size re-estimation; modifications in inclusion/exclusion criteria; evaluability criteria, dose/regimen and treatment duration; changes in hypotheses and/or study endpoints; and modifications and/or changes made to the statistical analysis plan prior to database lock or unblinding of treatment codes. Clearly this is a major shift in process.

The FDA has encouraged this approach. The FDA released a Critical Path Opportunities List in 2006 that calls for advancing innovative trial designs, “especially for the use of prior experience or accumulated information in trial design.” According to the FDA, “the purpose of adaptive design methods in clinical trials is to give the investigator the flexibility for identifying best (optimal) clinical benefit of the test treatment under study without undermining the validity and integrity of the intended study.”

Drug pipelines have been drying up for several years. The pharmaceutical industry faced a patent cliff in 2012 in which many branded products went generic. This led to significant downsizing in the industry. Personalized medicine may be a way for the healthcare industry to revamp and deliver the next level of value to patients. Before this can happen, however, several technical challenges must be overcome. Most of these challenges require improved information management. Adaptive-clinical design may be one of the technical solutions.

See related blogs.

Gov. Rick Perry of Texas championed the high-technology vision. Perry outlined how certain policies in Texas, such as tort reform and no income tax, attracted high-end medical specialists to the state. This set the state up as a destination for medical tourism that attracted very sick patients from all over the U.S. and wealthy patients from all over the world. This provided the state with a healthy income. When asked whether this process should be implemented nationally, he claimed that the federal government was not able to deliver on a vision like this. It was clear, however, that implementation at the national level would degrade Texas’ competitive advantage over the other states. When asked about non-specialized medicine and primary care in Texas, Perry was candid that many Texans were not receiving adequate primary healthcare. He gave as an example the population along the Rio Grande that had to travel several hours to access prenatal care. He mentioned that Texas was looking into some loan forgiveness for Texas medical students who were willing to spend five years in rural settings, but clearly, when he talked about this, the sparkle was missing from his eyes.

Bradford Berk, CEO of the University of Rochester Medical Center, and David Klein, CEO of The Lifetime Healthcare Companies, a healthcare funder in the Rochester area, championed the second vision. They described a collaborative system that focused on providing low-cost high-quality healthcare at a community level. This system emerged as a cooperative arrangement between large local employers, such as Kodak and Xerox, and a first-rate medical school at the University of Rochester. Of course, Kodak and Xerox are shadows of their former selves in Rochester, but the system they helped create still exists. The key observation was that, contrary to traditional economic theory, increased healthcare supply increases costs. This cynical view is that medical entrepreneurs increase the number of hospital beds, and then prescribe to fill the beds. The Rochester team decided to right-size the supply of healthcare in the community and to focus on outcomes. This was possible because of the economic motivations of the community businesses who were funding the healthcare and the cooperation of the medical school that drove the vision through its training and hiring policies. The Rochester community succeeded in its goal of providing low-cost high-quality healthcare. Of course, if an upstate New York resident needs a very exotic procedure, he or she may need to fly to Texas.

It is clear that the U.S. cannot continue its current course in which healthcare costs are more than 17% of GDP and outcomes are significantly behind the rest of the world. It is also clear that Americans are an optimistic people with a firm belief in the idea of progress, which manifests itself in technological and business innovation. The two visions for healthcare that were presented at the Forbes Healthcare Summit are almost caricatures, however. There is opportunity for creative individuals and institutions. Technology has not been applied evenly in healthcare. While the U.S. can be proud of its innovations in high-technology procedures, its adoption of information technology is not even at the level of the pizza industry. Both visions emerged as a result of economic incentives, one to increase revenue, as was the case in Texas, and one to reduce costs, as was the case in upstate New York. The question now is how do we shift the incentives to create the proper mix of outcomes, cost, and risk?

See related blogs.

In the U.S., healthcare purchasing and utilization decisions are made by three entities: the patients, the healthcare providers, and the payors. The person who uses the healthcare (patient) is not the person who orders the healthcare (provider) who is not the person who directly pays for the healthcare (payor). The healthcare transaction is much more complicated than the purchase of a sack of potatoes.

In economics, the simplest product is a commodity. The decision to buy a commodity is based solely on its price. The canonical commodity is a bushel of corn. For many purposes, such as feeding cattle, one bushel of corn is the same as another, except for how much it costs. There are buyers who would like to purchase a bushel (Demand) and sellers who would like to sell a bushel (Supply). The price of the transaction is negotiated or is the consequence of a bidding process.

Corn, however, has another property: it is planted in the spring and harvested in the fall. Its value has a temporal component. The corn seed in the spring has less value than the harvested corn in the autumn. Moreover, the buyers of the corn, e.g. the cattle producers, may have different needs for corn at different times of the year. Therefore, their willingness to pay a certain price for the corn varies throughout the year. In order to reduce the risk to both buyers and sellers, a new product, called a future was developed. A future is simply a guarantee that the corn will be sold at a certain price at a given time in the future. Buyers or sellers can purchase this future at a fixed price from the seller of the future, and the future seller absorbs the risk of price fluctuations. Since corn has been grown for many years, the future seller can calculate the risk he or she bears and can price the future accordingly. With the inception of the future, the transaction of buying and selling a bushel of corn increased in complexity from simple consideration of the price at a given time to consideration of the price now and the price at some future time.

We do not, however, make our purchasing decisions based solely on price and risk. We also base it on the quality of the particular product. For instance, a Mercedes Benz automobile is usually perceived to have higher value than a Chevy; thus we may be willing to pay more for the Mercedes than the Chevy. This is different than the time dependence of the value. This difference in value is due to workmanship, reliability, company reputation, social perceptions and other characteristics of the product. A purchasing or selling decision that takes price, time and value into account is more complex than a simple purchase of a bushel of corn at a particular time. The components of this tripolar transaction are of three types: 1) a component associated with cost or price, 2) a component associated with quality or value, and 3) a temporal component associated with risk. Risk can be mitigated by speed. For instance, if a bushel of corn could be grown in one day, there would be little need for a futures contract.

The tripolar transaction forms the basis of what is called the “triple constraint.” A company selling a customized product to a customer will often allow the customer to choose two options from the list of good, cheap or fast, which corresponds to the three product components, value, price and time/risk. If quality or value and speed of delivery are most valuable to the customer, then the price will be a little higher. If value and cost are important, then the project may need to be fit into down production times when production would otherwise be idle. This will delay delivery. If cost and time to delivery are important, then the quality of the product may suffer. A tripolar transaction significantly increases the flexibility of the buyers and sellers.

A similar tripartite concept in healthcare is called the “Iron Triangle.” The three components of the Iron Triangle are cost; quality; and access, which plays the role of the temporal component. The three players in the healthcare market: patients, providers, and payors, can be associated with the three components of the Iron Triangle. Providers are most closely associated with quality; payors are most closely associated with cost; and patients are the recipients of access.

Unlike the triple constraint, in which the buyer is allowed to choose two of three product characteristics, healthcare reform asks that all three characteristics are satisfied: near-universal access, lower costs, and improved outcomes. This interaction is addressed in current blogs and will be further addressed in future blogs. We, however, will leave a serious investigation of the riddle of the raven and writing desk to serious scholars.

Let me contrast this with another personal anecdote. My mother, while living at home some distance from where I live, had a chronic and potentially lethal bowel infection. She had an incident in which she needed to go to the emergency department. The emergency doctor, who had no health records for my mother and who did not consult her primary-care physician, diagnosed her as having a retrovirus and sent her home. Two days later, she was admitted to the hospital with the serious bowel infection for which she was being treated by her principle-care physician and that had nothing to do with a retrovirus. She remained in the hospital for several weeks and had near-death experiences. Her life was saved by the skill of her physicians and high-technology expensive procedures.

A few weeks after her return home, she had a recurrence of the infection and went again to the emergency department, where the same emergency doctor once again diagnosed her as having a retrovirus. Once again she was sent home, and once again she was later readmitted to the hospital and treated for bowel infection. This time she came much closer to death. At one point, I was told she would not live through the night. After another heroic and expensive life-saving procedure, she did live. After a few weeks in the hospital and rehabilitation, I admitted her into an assisted-living facility.

My mother has not had a recurrence of her infection in the two years she has been in the assisted living facility. She has gained weight and is, in general, in good health for someone her age. The reason her health has improved is because, now, her health information is being better managed. A nurse has complete knowledge of her medical conditions and vital signs. Her medication is properly administered. Her doctors’ appointments are coordinated. In my mother’s case, there were dramatic and expensive technological procedures that saved her life, but her life would not have been threatened in the first place had her healthcare information been readily available to her healthcare providers. While her entire assisted-living package is expensive, it is much less expensive than the life-extending procedures she underwent because her health information was not being properly managed. The information-management systems are simply not yet in place in large areas of healthcare.

There may be many reasons why modern information management has not yet made it into US healthcare: privacy concerns; size of the problem; physician and hospital compensation designed around procedures rather than outcomes; the emphasis of healthcare training programs; and political issues. In many respects, however, online information management is an essential component of the modern lifestyle. While healthcare information management may not be as easy as pizza pie, Americans are prepped and ready for it.

1. procedures that are low-cost and effective,
2. procedures that benefit some but not all patients
3. procedures with small benefit or unproven scientific value.

Low-cost and effective procedures

These are procedures that lower costs and improve patient outcomes. The authors include in this class antibiotics; casts for fractures; and aspirin and beta blockers for heart-attack patients. This is the class of procedures that yields the lowest costs for the best patient outcomes.

Procedures that benefit some but not all patients

These procedures are effective and beneficial to a certain population of patients. An example would be angioplasty with stents for heart-attack patients treated within 12 hours of an attack. This procedure is also used to treat patients with stable angina. These procedures see diminishing returns as the costs of the procedure increase. Since many more patients are treated for stable angina than are treated within 12 hours of a heart attack, the costs increase at the expense of benefits.

Healthcare Provider by Roger D. Jones

Benefits are small or of little scientific value

These are procedures in which randomized trials indicate no benefit. Experimentation is being done in situ with these procedures.The authors give as examples vertebroplasty in which cement is used to stabilize vertebra, and intensity-modulated radiation therapy for prostate cancer. These procedures are the highest cost for the lowest benefit.

An attempt to reduce costs and improve outcomes

Portions of the Accountable Care Act (ACA) attempt to shift payments for procedures to payments for patient outcomes. The hope is that physicians’ focus will move to preventative processes. The ACA envisions a new class of clinician; one that is charged with keeping patients out of the hospital by improving patients’ health practices. For instance, lifestyle coaches may be effective in helping diabetics improve their diets, monitoring and testing so that they have fewer diabetic emergencies. This keeps the patients out of the emergency departments and surgical theaters. In this picture of healthcare, generalists become dominant players.

The change may propagate to other parts of the structure. If patients are kept out of the hospital through preventative measures, then the importance of the hospital as a facility is reduced. The portion of the National Health Expenditure (NHE) spent on hospital stays is currently 31%. Home care, currently at 3%, may become a more important component.

It is still too early to determine if the ACA will have an effect on improving patient outcomes and reducing costs. Future blogs will examine other methods for reducing healthcare costs, including benefits from gaining insights from Big Data. So stay tuned.

This simple equation breaks down when applied to the healthcare market. A person at the end of his or her life may be willing to spend an entire life savings on a few months of extended and more comfortable life. The demand for those healthcare services that can satisfy this desire is effectively infinite. Traditional economic theory would say that the supply would meet the demand and the price of healthcare products and services would become effectively infinite.

Healthcare differs from simple commodity economics in another key aspect. For simple commodities, the user of the commodity is the buyer and the supplier of the commodity is the seller. In healthcare in the U.S., there are three players rather than two. The person who prescribes the product or service is the provider; the person who uses the product or service is the patient; and the person who pays for the product or service is the insurer.

Why is the cost of healthcare not infinite? In many countries, such as the U.K., the cost is regulated. Americans, however, tend to prefer market solutions to government regulation. Market solutions have not been effective at controlling costs in the United States. Healthcare costs are increasing at a 6.7% annual rate—much higher than the current inflation rate of 1.7% and the current GDP growth rate of less than 2%. In 2010, healthcare costs of $2.6 trillion consumed 17.9% of the US GDP. This is the highest rate in the world. At the same time, the quality of care fell behind most industrialized countries. Healthcare costs in the U.S. are rising to meet the demand. About one-quarter of Medicare outlays are for the last year of life.

Healthcare in the U.S. is not in a state of equilibrium in which supply meets demand. Supply is lagging behind demand. What is holding back supply? One answer is the pace of invention. Over the last half century a number of life-extending procedures have been developed, from open-heart surgery to advanced chemotherapy. A common property of these procedures is that they are expensive. It has taken time to develop these procedures. As more expensive procedures are invented, healthcare costs will rise to meet the demand. Unfortunately, these costs are becoming a larger proportion of GDP, which adversely affects the ability to pay for other activities that make life worth living.

Medicare and Medicaid were established in 1965 to manage the healthcare of the elderly and poor. In the early ‘70s Health Maintenance Organizations (HMOs) emerged as the dominant decision maker. These are organizations controlled by payors that can employ physicians directly or indirectly and that control the flow of patients among those physicians. In 1982, Diagnosis-Related Groups (DRGs) were developed to capitate the costs of hospital treatments. This enhanced the decision-making focus of the payors. Then, providers banded together in Planned Provider Organizations (PPOs) to create enhance their negotiating power of the providers with respect to payors. Point of Service Plans (POS) evolved as hybrids of HMOs and PPOs, thus setting up a pick-two process in which the payors and providers are favored over patients. In 2010, the Affordable Care Act (ACA) provided for the creation of Accountable Care Organizations (ACO) to shift the focus to the patients, specifically to basing reimbursement on patient outcomes. The complexity of insurance has increased with time. Several experiments have been done; none have been very successful.

As we can see from technology innovation and the changes in payor dynamics over the last half century, healthcare economics in the U.S. is not in a stable state of supply/demand equilibrium. In fact, if equilibrium of supply and demand existed, it would make the cost of healthcare so high that Americans would not be able to spend their income on anything except healthcare. Equilibrium exists, but it is not based on anything as simple as traditional supply-and-demand economics. One might expect an equilibrium involving three entities rather than two. We already see the beginnings of this tripartite equilibrium. Patients represent demand, providers represent supply, and payors are the third and regulating party in the triad.  Stay tuned. The dynamics of this three-way interaction will be the subject of upcoming blogs.