Medicare Advantage for All – A Conversation with Chuck Phelps

by Sara Khor, Yilin Chen, and Joyce Jiang

From left to right:  Brennan Beal, Joyce Jiang, Yilin Chen, Jacinda Tran, Sara Khor, Charles Phelps

The coronavirus disease 2019 (COVID-19) pandemic has shone a spotlight on the shortcomings of the current US health care system.  There is an urgent need to address the problems with healthcare access, costs, and equity.  Conversations related to health care reform have been and will continue to be front and center in the upcoming presidential election.  Many health care reform proposals have been put forward, including the expansion of the Affordable Care Act, Medicare-for-All, and Medicare expansion that maintains a role for private insurers.

Earlier this year, Dr. Charles Phelps, a renowned health economist and the author of the Health Economics textbook, visited the CHOICE Institute and gave a lecture on his proposed health care plan: Medicare-Advantage-for-All.  Under this proposal, all permanent residents will be able to choose among a wide variety of private insurance plans, similar to the Medicare Advantage program that is currently available to the older US population.  All individuals will have health care coverage, with the minimum coverage being a high-deductible health plan with a health savings account.  The health savings account can be filled up based on income levels.  High value care such as preventative medicine can bypass the deductible.

We had the honor of sitting down with Dr. Phelps to talk more about his proposal. 

Part 1 – On health care costs:

What are the most important drivers of the current cost growth in the healthcare system?

Dr. Phelps identified two main drivers of the healthcare cost growth:  introduction of new technologies and the aging population.  In the long run, he said, the introduction of new technologies, which includes diagnostic tools, pharmaceuticals, and genetic-based medicine, will increase costs.  As treatment and drugs become more targeted, they will be sold in smaller quantities for focused populations.  While these new advances will produce a lot of value, they will also drive cost growth. 

As the size of the older population continues to expand, the age distribution pyramid (where the bottom is the youngest age group and the top is the oldest) will be shaped more like a cylinder. “When you look 20 years from now, this cylinder is going to have a great big hat on,” says Dr. Phelps.  “The widest group on this cylinder… is going to be the oldest group at the top”. 

This has important implication for healthcare costs, because the current way we are financing Medicare is through the payroll tax.  The ratio of workers (who pay the payroll tax) to retired individuals (who use Medicare) is shrinking.  “At the beginning of Medicare, this ratio was 4.5,” says Dr. Phelps.  “In a couple of decades, this ratio is going to be 2 workers per retiree.  The payroll tax mechanism of paying for Medicare has to change”.

How would the Medicare Advantage-for-All Plan address these cost drivers?

In a single payer system, usually one agency (e.g. CMS in Medicare) makes the determination about which new technology to introduce into the system.  “They can make mistakes both by being too generous, letting too many things in the door, or being too stingy,” said Dr. Phelps.  According to Dr. Phelps, the current US Medicare system is too generous, as it has not built in any cost constraints, while the British National Health Service has very tight cost-effective criteria for approval of new technologies.

The Medicare-Advantage-for-All plan would allow different plans to make cost-effectiveness evaluations.  The exact details of the plan still need to be decided, but Dr. Phelps said that technologies that are very cost-effective, ICER between $50,000 to $75,000/quality-adjusted life year (QALY) gained, will be mandatory in the basic plan, and others will be left to the discretion of the insurance plan to cover it.  “People will buy insurance plans to fit their needs, just like they buy different cars with different degrees of safety and pizzazz.”

What is the potential cost impact of Medicare Advantage for all?   How is this cost impact compared to Medicare for All?

“Medicare-for-All is ghastly expensive, not because it offers universal health coverage, but because it eliminates all co-pays.”  The RAND Health Insurance experiment and the Oregon Health Insurance experiment have demonstrated that medical use will increase when there is no co-pay or deductible.  Dr. Phelps worried that the lack of co-pays will “unleash the monster in the evolution of new technologies,” referring to how moral hazard— the price sensitivity of demand for health care – may incentivize the use of low value care and technologies.  He was concerned that healthcare costs will continue to rise unless there is a very tight constraint on new technologies. 

In a way, Medicare-Advantage-for-All is similar to Medicare-for-All if everyone’s health saving account is filled up.  When determining how much should be in the health saving accounts, Dr. Phelps said that “it will be an experiment through time to trade off the risk bearing and cost control,”  adding that essential medicines and services, such as birth control or insulin for diabetes, can completely bypass the deductibles and co-pays. 

Right now, Dr. Phelps said it is unclear how the different program parameters should be set in order to balance costs, new technology introduction, and equity considerations, but he is trying to devise an instrument that would allow users the flexibility to tweak parameters over time.

How about administrative costs?  Dr. Phelps said that single payer plans, like Medicare-for-All, definitely reduce administrative costs. He argued that this administrative cost, although pricey, is giving us choice, and is controlling the costs by negotiating with providers about how much to pay for things instead of having fixed fee schedules.  “Our society really values choice,” Dr. Phelps added.  A one-size-fits-all plan takes away choice.  Using an analogy in car shopping, Dr. Phelps said having just one healthcare plan is like saying: “You can buy any car you want, as long as it is a Honda Accord.”

Instead of simply comparing the administrative costs between a single-payer government-funded plan and Medicare-Advantage-for-All plans, Dr. Phelps emphasized the importance of taking into account the welfare loss and tax distortions that arise from the increase of taxation in a single-payer government plan. “Every dollar of tax we collect distorts the economy in some fashion…If you raise income tax, you change the labor supply.”

Part 2 – On choices and health technology assessment (HTA):

If there are many Medicare-Advantage-for-All health plans and each plan uses a different threshold, how do consumers make informed decisions about which plans to choose?

“Advisors will emerge,” said Dr. Phelps. Like the advisors for buying automobiles or financial advisors for the stock market, Dr. Phelps believed that a service for advising people on healthcare insurance and utilization will develop, and so would competition.  These advisors could be independent or affiliated with big health plans.  The emergence of these advisors, he added, will “depend on having access to electronic medical records that people can share.”

What role will HTA play?  What role would an organization like the Institute for Clinical and Economic Review (ICER) play? 

 Dr. Phelps felt that the current way of conducting cost-effectiveness analyses is incomplete.  Individuals who want to maximize their own utility do not necessarily think about everything that the society thinks about.  While the QALY is a very important component of the overall value index, Phelps argued that the value index should include other things that are not necessarily built into a single individual’s utility, such as the fear of contagion and equitable distribution of health services.  “Policies about Ebola, Zika, and now the coronavirus, are not made on the cost-effectiveness of vaccines.  People would pay anything to get a coronavirus vaccine right now.  The fear of contagion dominates public discourse and public policies.  That is not captured in cost-effectiveness analysis.” 

Ideally, Dr. Phelps says, there will be competition among HTA organizations to produce estimates with quality control by the government, similar to how the FDA has control over the new drugs that come on to the market.  “I can see different insurance plans, [especially] some of the large ones, creating their own HTA shops.  ICER would continue and offer [assessments] to smaller insurance plans.” 

Part 3 – On equity:

Under Medicare-Advantage-for-All, multiple plans with varying premiums and deductibles mean that people who are better at navigating the system (e.g. more able to afford an advisor) or can afford the higher premiums may get more comprehensive care, potentially creating a situation where there is differential access and differential outcomes across the population based on education or wealth.  How should we think about this?

“A plan that has absolutely equal access to health care…is imaginary”.  Dr. Phelps argued that as long as people have different incomes, there would be differential care.  Even under the British National Health System, those who can pay more can seek additional or better care with private insurance.  In Ontario, Canada, where there is universal health coverage, some Canadians opt to purchase insurance to buy medical care in the U.S.  One thing that the Medicare-Advantage-for-All plan can guarantee, Dr. Phelps added, is minimal level of access to quality care for everyone.  The minimal plan could include an independent advisor.

How will international students be included in the plan?

International students are often in the U.S. for a few years.  They are not permanent residents, but they are also not temporary visitors.  It will be important that there are ways for these individuals to get access to the health plans.  Some suggestions that Dr. Phelps had were to charge individuals an actuarially-based fee to join the plans, or to negotiate bilateral exchanges with different nations. 

Part 4 — On political economy: 

What are some of the potential political pushbacks related to this proposal?

The country is very heavily divided between those who want a single-payer plan (e.g. Medicare-for-All) and those who want to continue with private insurance.  Those who are proponents of single-payer plans will not like Medicare-Advantage-for-All because it continues to use, as a central feature, and quite deliberately, private insurance plans.  In this political climate, Dr. Phelps said, “My forecast would be…that there is zero probability that a Medicare-for-All plan would pass through the congress.”

“The high deductible health plan creates some anxiety among some people…for two reasons”, he continued.  “One is that they say it discriminates against the poor.”  Dr. Phelps said he can completely eliminate this concern by filling up the health savings accounts on an income-related basis.  Another concern is that people in the high deductible plans will stop using the care that they need.  A short run fix, Dr. Phelps explained, is that all highly-valued services and medicines, like diabetes medications, bypass any deductible and copayment.   

“The long run fix is to vastly repair our vulnerable K-12 education system,” said Dr. Phelps. Higher education helps people navigate the highly complex health care system.  There is also a very steep education gradient on lifestyle choices that are bad for your health, like tobacco smoking and binge drinking.

A Brief Introduction to Infectious Disease Modelling: Simulating a Hypothetical COVID-19 Outbreak

By Enrique M. Saldarriaga

Introduction

The COVID-19 pandemic has boosted the interest for mathematical models of infectious diseases. In this entry, I will briefly introduce some of these models and provide an R-code to simulate an outbreak of COVID-19.

These models synthesize multiple sources of information into equations that aim to model the evolution of a disease and make predictions. When used correctly, they can be incredibly powerful tools to explain a very chaotic and complex reality, to evaluate policy options to inform decision-making, to understand hidden mechanisms that drive an epidemic, and others.

Infectious diseases do not occur in isolation in each person.  They are transmitted through contact with a pathogen. Thus, there is a need to understand the mechanisms for a susceptible person to establish effective contact (i.e. contact that results in a transmission; sexually transmitted disease is a good example) with someone who is infected with that pathogen. On the population level, disease prevalence is considered a risk factor for the incidence: the higher the proportion of people living with a disease, the higher the likelihood that an infected person gets in contact with a susceptible person. This relationship between incidence and prevalence can be characterized using dynamic models. Here, the probability of getting infected is determined by the probability of contact with an infectious person (or animal in case of diseases transmitted by vectors, like malaria), which is given by the prevalence. A contact resulting in an infection is called a susceptible-infected effective contact.

Infectious and non-communicable disease models have substantial similarities: both can be compartmental or agent-based (microsimulation), as well as deterministic (static transition probabilities) or stochastic (transition probabilities are random draws of a specified distribution). In any case, the decision about which model to use is determined by the scope, purpose of the analysis, and many times, the target audience for results dissemination.

In the following section I will describe compartmental, deterministic, closed-cohort models. In a closed cohort model, we assume no deaths or births, but the population remains constant over time.

Model Types

The Susceptible-Infectious (SI) Model. This is the most basic infectious disease model. It is characterized by two state variables or compartments: Susceptible (S) and Infectious (I). Here we model one transition, and once all susceptibles are infected, the epidemic is over (no deaths in this model). The transition is driven by the transmission coefficient. This is a very important concept because regardless of model type, this parameter determines the rate at which people get infected. It is usually denoted by lambda, λ, and it is the product of the infectivity or probability of transmission per contact (ρ), the contact rate at a given period (c), and the prevalence of infected (I/N; where N  is the total population): λ = c * ρ * I/N. At any point in time, and for all model options, the number susceptible decreases by λ.

The Susceptible-Infectious-Recover (SIR) Model. In addition to susceptible and infected, the SIR model includes the recovered (R) compartment. R includes people that were infected and overcome the disease. The rate of transition is given by the inverse of disease duration, also known as the recovery rate (γ). Some diseases confer immunity (e.g. measles) after infection, but others do not. To capture this, a SIRS (susceptible-infected-recovered-susceptible) model would be more appropriate and allows those who don’t develop immunity to transition back to susceptible.

The Susceptible-Exposed-Infectious-Recovered, (SEIR) Model. This model adds an exposed (E) compartment. Exposed are all persons who have been infected but are not yet symptomatic, and more importantly, not yet infectious. Infectious persons are the only ones capable of spreading the disease, hence, an accurate count of them is very important. When using a SEIR model, the transition between S and E is given by lambda (λ) and the transition between E and I is given by the inverse of the latency or incubation period (σ).

seri

COVID-19 Outbreak Example

I am going to simulate a COVID-19 outbreak using a SEIR model, depicted in the figure below. All parameters have been obtained from the MIDAS Network repository – an excellent and publicly available compilation of COVID-19 parameters.

Let’s model the transitions between compartments considering 1-timepoint increment:

seir2
By taking the partial derivative of these equations with respect to t, we obtain the changes in every compartment at any point in time:

f2

With this in mind, let’s go to the R-code to see how to implement the simulation.

COVID-19 Example Results

We model an outbreak for 1 year, using the following parameters: c * ρ = 1.5, σ = 1/4.2, and γ = 1/20, for a population of 1 million where 1 persons were already infected. The following image describes the outbreak.

p1

We can see a very steep increase in the number of infected, which peaks at 625,095 infections on the 37th day of the outbreak. As it is often pointed out, this rapid increase in cases can overload health systems, reducing the possibility of many people to access care.

How can we flatten the curve? One intervention to contain the COVID-19 pandemic was to increase the physical distance between people. The objective was to reduce the probability of an effective susceptible-infected contact. In modelling terms, this would directly reduce c * ρ  and therefore λ.

The following image shows the results of reducing c * ρ  to 0.6 instead of 1.5.

p2

The peak of infection occurs later, on day 65, at a lower count as well: 550,446. This is an example of how effective behavioral changes can be to reduce the severity of an outbreak.

In this example we changed only one parameter. But one thing that amazes me about infectious disease modelling, is that (almost) every parameter driving the outbreak is susceptible to change given the right intervention. You can now use the R-code to see how variations in other parameters affect the outbreak and think about what kinds of interventions might produce such changes.

Suggested Readings

Vynnycky, E. & White, R. G. An introduction to infectious disease modelling. (Oxford University Press, 2010). BookSite

Garnett, G. An introduction to mathematical models in sexually transmitted disease epidemiology. Sex Transm Infect 78, 7–12 (2002).

Kretzschmar, M. Disease modeling for public health: added value, challenges, and institutional constraints. J Public Health Pol 41, 39–51 (2020).

Dr. Mark Bounthavong’s Talk on Formulating Good Research Questions

by Enrique M. Saldarriaga and Jacinda Tran

On April 16, 2020, the ISPOR Student Chapter at the University of Washington hosted a webinar on how to formulate good research questions featuring Dr. Mark Bounthavong, PhD, PharmD, MPH. He discussed aspects of compelling research questions, shared his formulating process, presented best practices, and provided recommendations for students at all stages of their career.

Dr. Bounthavong is a graduate of the UW CHOICE Institute and a prolific researcher with several years of experience in HEOR. He currently serves as a health economist at the VA Health Economics Resource Center and a Research Affiliate at Stanford University, and his research interests include pharmacoeconomics, outcomes research, health economics, process and program evaluations, econometric methods, and evidence synthesis using Bayesian methods.

Our UW Student Chapter thanks Dr. Bounthavong for his insightful presentation and hopes our fellow researchers find this recording of his presentation to be a helpful resource.

Note: Dr. Bounthavong has authorized the publication of his talk in this post.

The Utility Function, Indifference Curves, and Healthcare

By Brennan T. Beal

Impetus For The Post

When I first learned about utility functions and their associated indifference curves, I was shown an intimidating figure that looked a bit like the image below. If you were lucky, you were shown a computer generated image. The less fortunate had a professor furiously scribbling them onto a board.

https://opentextbc.ca/principlesofeconomics/back-matter/appendix-b-indifference-curves/

A few things were immediately of concern: why are there multiple indifference curves for one function if it only represents one consumer? Why are the curves moving? And… who is Natasha? So, while answering my own questions, I thought sharing the knowledge would be helpful. This post will hopefully provide a better description than maybe most of us have heard and by the end you will understand:

  1. What indifference curves are and what they represent
  2. How a budget constraint relates to these indifference curves and the overall utility function
  3. How to optimize utility within these constraints (if you’re brave)

For the scope of this post, I’ll assume you have some fundamental understanding of utility theory.

Click here to link to my original post to continue reading.

Expected Loss Curves

By Brennan T. Beal, PharmD

The Second Panel on Cost-Effectiveness in Health and Medicine recommends model uncertainty be reflected by displaying the cost-effectiveness acceptability curve (CEAC) with the cost-effectiveness acceptability frontier (CEAF) overlaid (more on this can be seen here). However, on top of being relatively difficult to interpret, these graphical representations may miss a crucial part of decision-making: risk.

A risk-neutral approach to decision-making would mean choosing a strategy that is most likely to be cost-effective despite what one stands to lose economically when the strategy is not cost-effective. Though, we know that decision-makers are often not risk-neutral. With this in mind, selecting a strategy based solely on the probability of being cost-effective could expose a decision-maker to unnecessary risks. It is not always the case that the most likely to be cost-effective is truly the optimal decision; notably, the optimal decision should be thought of as the strategy with the lowest expected loss.

Consider the following example:

Let us suppose that you want to compare two strategies (Strategy A and Strategy B) to see which will be optimal for your company. Your head statistician informs you that Strategy A will be cost-effective 70% of the time and in the 70 times out of 100 that it is cost-effective, you stand to gain $5 dollars each time (i.e., you lose $0 each of those 70 times). She then proceeds to tell you that for every time you are wrong (30% of the time) you stand to lose $100. Your expected loss would be $30 (30% of the time losing $100). With that in mind, you also calculate the expected loss for Strategy B. Turns out it is only $7! ($7 is arbitrary for the sake of example).

In this example, Strategy B would be favored on the CEAF given that it has the lowest expected loss but the CEAC would have shown it to be less likely. So, having the CEAF at least informs us what strategy is optimal, but we are still left with a relatively confusing picture of cost-effectiveness.

Below are three hypothetical distributions of the incremental net benefit (INB) of Strategy B when compared to Strategy A. Simply stated, the INB curves can be thought of as the probabilities of monetary outcomes when comparing strategies.

brennan figure 1

This example is described in greater detail in my recent blog entry on the topic of expected loss. For each distribution above, Drug B is considered optimal as it has the lowest expected loss in each scenario. However, in situations where the mean and median have opposite signs (such as in the case of the right skewed blue curve above, mean INB of $90 vs. a median of -$75), only considering the most likely to be cost-effective will not provide a decision-maker with the optimal decision. For the blue curve above, Drug B has a lesser chance of being cost-effective (46%), but an expected loss of $271 vs. $361 for Drug A.

Expected loss curves (ECLs) account for the probability that a strategy is not cost-effective and how drastic the consequences are in those scenarios. The ELC represents the optimal strategy at each willingness-to-pay threshold and provides a much clearer picture of risk for more informed decision-making.

In my full blog entry, I cover:

  1. An in-depth explanation of ELCs;
  2. A working example of the mathematics and associated R code;
  3. and an interactive example at the end so you can see for yourself