After Jimmy Kimmel’s impassioned critique of the Graham-Cassidy repeal bill and the canonization of the Jimmy Kimmel Test for evaluating health policy, celebrities are becoming increasingly active in the nation’s policy debates. I really wish they would stop, because they don’t know what they’re talking about, especially when it comes to healthcare, and the attention that they receive with their celebrity status makes it incredibly irresponsible for them to misinform the public. The newest entrant into the policy arena is Julia Louis-Dreyfus, famous for her roles on Seinfeld and Veep, who announced on Twitter that she has been diagnosed with breast cancer (get well, Julia).
The problem is that she then took the opportunity to plug universal health care as the solution to cancer mortality. Why is this a problem? Let’s take a quick look at how poorly the US compares to other OECD countries in breast cancer survival rates:
Oh. Well. That’s awkward. Despite not having universal health coverage, the United States is the best healthcare system for actually treating women with breast cancer and keeping them alive. That’s because contrary to popular belief, we do actually receive higher quality care than other countries. It’s important to point out that the US is a lot better than some key countries. The Commonwealth Fund’s darling “Best Healthcare System” in the United Kingdom is almost 9% behind the US in 5 year breast cancer survival rates. This is actually a massive problem, because as Julia Louis-Dreyfus correctly points out, 1 in 8 women will be diagnosed with breast cancer. That’s a full 1% of all British women who die because the NHS is failing them on this single important disease. Julia Louis-Dreyfus should study the facts before pushing such a system on American women as well.
By the way, since we’re sharing personal stories, my mother was diagnosed with breast cancer and as an uninsured poor immigrant went through chemo paying for treatment with loans. She lived. It’s certainly tough financially, but it’s not impossible, and the idea that being sick while uninsured is an automatic death sentence is yet another irresponsible falsehood.
The Sanders/Klobuchar/Graham/Cassidy CNN healthcare “debate” is a must watch for anybody who cares about healthcare reform in the US. However, it’s not important because it was filled with intelligent and critical discussion of the health care system’s problems and the smartest ways to fix them. On the contrary, it’s essential viewing because it reveals how politically dogmatic and intellectually lazy the two sides have become.
To start off, the bright point of the debate was that both sides seem to finally agree that the ACA isn’t working. When the ACA marketplace’s average deductibles increased to $6,000 and premiums increased by 17% in 2017, the true believers argued that this was a one-time actuarial adjustment as the marketplaces aligned themselves to the demographics of the newly insured population. As we get closer to 2018, this position is becoming increasingly indefensible as it’s becoming clear that exchanges are going to see another wave of large premium increases. Florida, a state that expanded Medicaid under the ACA, is forecasting a 45% increase. It shouldn’t be a surprise that the marketplaces are spiraling since ACOs, which were the primary cost saving mechanism in the ACA, have been a tremendous failure at producing any meaningful results. Importantly, this mechanism has been a failure nationally as well as locally in Massachusetts, which was lauded as a better version of the ACA because it had more robust wealth transfers and higher individual mandate penalties. This is why nobody at the debate tried to argue that the ACA is working. The Democrats praised the law for expanding coverage, but were also eager to admit that something is rotten in the state of Denmark and called out to their Republican colleagues to work together on fixing it together. And this is where the problems begin. Despite agreeing that cooperation will be needed, the ACA has warped policy thought in DC to such a degree that neither side is capable of thinking of solutions outside of the “fund the ACA” vs “defund the ACA” paradigm.
Republicans for their part are completely obsessed with pushing through a repeal bill that also controls Medicaid spending. While their hearts may be in the right place (the CMS trust is still scheduled to run out of money in 2029), this is a fight against reality. Three Senate vote defeats have made it clear that a bill that reduces future spending will not pass and that bill writers need to move into a new direction, including consideration of increasing taxes to stop the fiscal hemorrhage. The proposed Trump tax cuts, however, provide little hope for a balanced budget. Similarly, the desire to completely delegate healthcare policy to the states is also unrealistic. Nicholas Bagley wrote up a reasonable critique pointing out how despite transferring control of funds, the federal government is not going give states the ability to pick and choose federal regulations. What may have worked for welfare in the 90s probably would not work for healthcare because of the entrenched special interests that protect themselves from competition through legal barriers to entry.
Democrats, on the other hand, may be in an even worse position. They quite literally have no idea what to do about the ACA. This is why Hillary’s campaign position on healthcare could be accurately summarized as “we will do stuff.” Despite agreeing with the Republicans that the ACA isn’t working, they seem to have stumbled unwillingly to the bizarre conclusion that the solution is to throw more money at the fire. The Maine and Alaska reinsurance programs Klobuchar championed at the debate is simply taxpayer funded nationalization of sick patients and their costs, and Graham was correct in criticizing it as throwing good money after bad. The other proposal from the Democrats is Bernie’s Medicare for All single payer, which not only throws good money after bad, it throws enough money to bankrupt the nation within a decade (his own words, not mine). The helplessness of the Democratic position is therefore fueling the growth of a malaise among ACA supporters. Where there was once exuberance and joy about the law’s potential, there is now a resigned feeling that this is the best that we can do. Aaron Carroll, ACA proponent and one of the panelists on NYTimes’ recent “Best Healthcare System” tournament, now says that “THERE IS NO WAY TO SPEND LESS, COVER MORE, AND MAKE IT BETTER” (his emphasis, not mine), which effectively concedes that the ACA does nothing to push the health production possibility frontier outward. Aaron Carroll is obviously wrong about this; competition and innovation have been making healthcare cheaper, broader and better for a century now. But more importantly, that statement reveals the lack of imagination that pushes the Democrats to promote more spending. In their view the only way to improve care for everybody is to spend more, so if we’re doomed to spend more then it’s better to do it now rather than later. There’s an internal logic to this fiscal suicide. In addition, because deficit spending and redistribution are critical to this vision, they are also doubling down on centralization of healthcare management in the federal government. Aaron Carroll criticizes the Republican’s state-oriented vision by claiming that none of the states have any idea how to fix this mess. It’s a strange argument to use to defend the ACA, because it makes the false assumption that the federal government somehow does know what to do and ignores the fact that the ACA itself was a watered down version of a state idea.
Ultimately the biggest problem is that neither of these boring approaches is going to solve America’s problems. The ACA isn’t what’s wrong with healthcare in the US, it just magnified the problems by adding more money into the system. Whether you eliminate it completely or empower it with more money, you’ll still be left with the same structural problems that are driving the irrational healthcare sector. There are critical market focused reforms that need to be made to price transparency, provider market entry, occupational licensing, health insurance purchasing, drug patent law and funding for behavioral health, all of which would immediately and significantly bend the healthcare cost curve. Instead we’re going to keep going back and forth about ACA funding. With so little brain activity going in Congress, the only plausible conclusion is that healthcare reform is comatose for the foreseeable future until something critical breaks.
Brilliant. The whole thing is really creepy when you think about it this way.
The passage of the AHCA through the House and the subsequent BCRA written in the Senate has made society at large anxious about the implications of these bills for the health of working Americans. Much of this fear has been fanned by politicians claiming that the bills will kill Americans. Bernie started off the parade:
Let us be clear and this is not trying to be overly dramatic: Thousands of people will die if the Republican health care bill becomes law.
— Bernie Sanders (@BernieSanders) June 23, 2017
Then Hillary jumped in:
Forget death panels. If Republicans pass this bill, they’re the death party. https://t.co/jCStfOaBjy
— Hillary Clinton (@HillaryClinton) June 23, 2017
Elizabeth Warren then accused the GOP of criminal activity:
These cuts are blood money. People will die. Let’s be very clear. Senate Republicans are paying for tax cuts for the wealthy with American lives.
— Elizabeth Warren
And then Nancy Pelosi gave an estimate of how bad the damage will be:
We do know that many more people, hundreds of thousands of people, will die if this bill passes
If I was the average citizen, I would be terrified at this point. Fortunately, I’ve been trained to think like an engineer and I don’t get scared until the data tells me that I should be scared. There are several observational reports out there demonstrating that health insurance improves outcomes, some showing no effect, and some even showing that health insurance kills people overall. This is the problem with observational studies; they are especially vulnerable to vulnerable to selection bias and confounding and therefore are inconclusive and unreliable evidence. What we really want is experiments. Fortunately, there have been two major randomized controlled trials (aka real science) of the impact of health insurance coverage on health outcomes and mortality, the RAND health insurance experiment and the Oregon Medicaid experiment. What does the data tell us then? Good news everyone, you have nothing to worry about, because health insurance doesn’t do a damn thing.
The RAND Health Insurance Experiment is a gold standard study that ran between 1978 and 1983 on a sample of 3958 people by providing participants with health insurance with a randomized level of cost sharing ranging from 0% to 95% coinsurance . At the end of the experiment, the researchers concluded that:
For the average participant, as well as for subgroups differing in income and initial health status, no significant effects were detected on eight other measures of health status and health habits. Confidence intervals for these eight measures were sufficiently narrow to rule out all but a minimal influence, favorable or adverse, of free care for the average participant.
The Oregon Medicaid Experiment was the second randomized controlled trial, conducted between 2008 and 2010. In the study state of Oregon expanded Medicaid coverage to a random selection of 6387 beneficiaries out of a total of 12,229 eligible applicants, with the non-recipients being used as the control group. In the words of the researchers:
We found no significant effect of Medicaid coverage on the prevalence or diagnosis of hypertension or high cholesterol levels or on the use of medication for these conditions. Medicaid coverage significantly increased the probability of a diagnosis of diabetes and the use of diabetes medication, but we observed no significant effect on average glycated hemoglobin levels or on the percentage of participants with levels of 6.5% or higher. Medicaid coverage decreased the probability of a positive screening for depression , increased the use of many preventive services, and nearly eliminated catastrophic out-of-pocket medical expenditures.
Two major studies decades apart and they find the exact same thing: increased healthcare utilization, increased total healthcare spending (important finding: prevention doesn’t save money), increased financial security (and the corresponding mental health benefits), but no statistically significant effect on health status (they don’t teach you this part in school). The health insurance didn’t do anything to improve people’s health compared to those who didn’t receive insurance. Keep in mind that these are large studies involving thousands of subjects. They’re also not ‘partisan’ or biased. The Oregon study in particular was run by Finkelstein and Gruber, who worked on the ACA and would’ve loved to show a positive effect on health, they just couldn’t find one.
This is a surprising finding for most people. It’s pretty simple to establish a plausible relationship between health insurance and mortality. Person gets sick, person can’t get treatment, person dies. How can there be no positive impact of health insurance on mortality? There’s several reasons that are not immediately intuitive but are real scenarios. Consider the following:
Case 1. Man gets cancer. Man has health insurance. Man’s cancer has no treatment. Man dies. Health insurance had no effect on health.
Case 2. Man has a bad diet and doesn’t get exercise. Man gets diabetes. Man has health insurance. Doctor tells him what he needs to do. Man continues to eat poorly and not exercise. Man dies. Health insurance had no effect on health.
Case 3. Man gets cancer. Man has no health insurance. Man’s cancer has a treatment. Man partners with a charitable organization and gets his treatment paid for. Man lives. Health insurance had no effect on health.
Case 4. Man has a non-life threatening lesion on a scan. Man has health insurance, so the doctor orders a biopsy to check. Man acquires an infection during the biopsy and dies. He would’ve lived if he had no insurance and didn’t get the biopsy. Health insurance killed the man.
It must be noted how common these situations are. 1 in 25 hospital patients acquire an infection. 1 in 500 hospital patients get killed by a medical error. Medical errors are the 3rd leading cause of death in America. Let that sink in, your average internist is involved in involuntary manslaughter 1-2 times per month. On top of that the expensive stuff that kills people, we’re not even that good at treating it. Overall cancer survival is only 50-60%. We also don’t have the slightest clue on what to really do with heart disease, respiratory disease, diabetes and such, because we don’t know how to stop patients from being themselves. Raj Chetty’s megastudy from last spring (1.4 million individuals observed) showed that the biggest predictors of life expectancy wasn’t income but whether you drank, smoked and exercised. 9 out of the 10 leading causes of death in the U.S. are preventable by changing your behavior. The inconvenient truth is that no amount of health insurance is going to save a smoker from themselves, because you’re just throwing money into a fire. Health care simply does not work for people who don’t care about their health.
The sum of all this is that health insurance has some real cost/benefit tradeoffs, but none of them are to improve health or save lives, and politicians should be more careful about saying things that are scientifically untrue. If you want to save lives, then try prohibition (that went well the first time), because health insurance expansion doesn’t get you there.
(Obviously this is all in the context of marginal effects at US insurance coverage levels, your mileage may vary in other contexts)
Dr. Jerome Groopman is the Chair of Medicine at the Harvard Medical School and staff writer for the New Yorker. His 2007 book How Doctors Think seeks to inform lay readers through a collection of medical case studies about the training that physician receive and the thought process through which they gather information and formulate the diagnosis. Most importantly, however, the book emphasizes the cases in which the logical process through which the diagnosis is developed fails and the physician misdiagnoses the patient and puts their life in danger. Sometimes the diagnostic process repeatedly fails to correctly identify a condition across multiple physicians. How Doctors Think begins by presenting the case of Anne Dodge who had been misdiagnosed for fifteen years by psychiatrists, internists and dietitians as suffering from bulimia and irritable bowel syndrome until a gastroenterologist finally identified celiac disease as the cause of her weight loss. Dr. Groopman also recounts how three renowned Boston-area surgeons had misdiagnosed the hand pain he was experiencing, including one diagnosis that was not a real medical condition (pg. 170). Diagnostic errors like these provide a stark contrast to the idealistic public image of physicians as precise and evidence based practitioners.
The book argues that most of these errors are not due to malice or incompetence, but are a natural byproduct of the mental heuristics that physicians must use to treat patients in a timely and efficient manner. Although physicians are trained in medical schools to carefully record the patient’s narrative and consider all possible options, the operational realities of the healthcare environment such as time-consuming electronic medical records and decreasing appointment durations force physicians to increasingly rely on shortcuts that are vulnerable to cognitive errors.
Dr. Groopman concludes by urging patients to participate more in the diagnostic process by asking physicians key questions such as “Are there any other possible causes?” and “Does any of the evidence not match the diagnosis?” The purpose of these questions is to force the physician out of their diagnostic autopilot mode and give them a chance to recognize any logical fallacies they may have committed. This collaboration would ideally improve physician decision making, decrease the frequency of medical errors, and improve patient outcomes.
The transition of payment systems from fee-for-service to pay-for-performance through programs like DSRIP and MACRA will inevitably put pressure on providers to standardize processes in order to reduce variability and improve population health outcomes. How Doctors Think reveals the significant variability that exists in physician decision making caused by individual practice preferences and errors during diagnosis, which is a major challenge for large healthcare organizations trying to meet their performance targets. Because “as many as 15 percent of all diagnoses are inaccurate” (p.24), organizations that can control and reduce this variability will be increasingly rewarded by programs such as DSRIP for standardizing their care. Safety net hospital are more vulnerable to such variation because they are more likely to face the resource constraints that force physicians to reduce their time with patients, which Dr. Groopman identifies as a major cause of diagnostic error. Fortunately analyzing the types of errors that physicians make provides a framework to understand how physician processes need to be adjusted and what changes can be made in the organization to reduce the occurrence of these errors. It is therefore recommended that all providers, but especially safety net hospitals, address the problems presented in How Doctors Think by implementing three operational reforms: communication standardization, diagnosis standardization, and treatment standardization.
Dr. Groopman emphasizes that good medicine relies on effective communication between the patient and physician. This requires the standardization of physician’s communication practices to ensure that they are obtaining as much information as possible from the patient.
Physicians should be trained to ask open-ended questions when interviewing patients, which avoids leading the patient towards a diagnosis that the physician is already thinking of, and “maximizes the opportunity for a doctor to hear new information” (p.18). The physicians should also be able to interview the patients quietly and uninterrupted, as distractions can cause the physician to miss important information (pg. 75). When the patient makes a statement that conflicts with the physician’s clinical judgement, the physician should make an effort to not dismiss the patient (pg.264). To maximize trust, the physicians must explain the condition and its risks in a clear manner, but also be prepared to spend more time with the patient when it is clear to the physician that the patient is still nervous or uncertain (pg. 88). Accordingly, the hospital should make it easy for physicians to extend their time with the patient or to schedule a follow-up appointment. Physicians should also explain why they are performing any tests and specifically what they are looking for (pg. 172) to engage the patient and give them a chance to express their own opinions and concerns. Finally, the physician needs to clearly explain all possible outcomes, the positive and negative features associated with those outcomes, and the likelihood of those outcomes (pg. 173) to enable patients to make the choices most consistent with their preferences. This discussion should always be framed within the context of the condition to minimize the risk of patients fearing the treatment’s side-effects more than the disease itself (pg. 246).
Safety net hospitals should implement a diagnostic checklist that physicians must review at the end of each case to ensure that they are not committing a logical fallacy in their diagnosis. Common errors that physicians make include:
- Representativeness – assume the symptoms correspond to a standard case (pg. 44)
- Availability – diagnosis affected by ease with which options come to mind (pg. 64)
- Search satisficing – stop searching for problems once you find one diagnosis (pg. 169)
- Vertical line failure – thinking inside the box when data and symptoms disagree (pg.171)
To avoid these logical failures, after the physician decides on a diagnosis they should be required to go through the following questions that Dr. Groopman recommends in the epilogue of How Doctors Think:
- What else could it be?
- Is there anything that doesn’t fit?
- Is it possible there’s more than one problem?
- What is the patient worried about?
- Review the patient’s story from the beginning.
In addition, patients should be encouraged to ask these questions and should be trained to do so through informational materials available to them in waiting areas and posters around the hospital. The benefit of this strategy is that it engages the patient in their health and prevents the physician from ignoring the checklist by going through it carelessly to save time.
Finally, this diagnosis verification process can be further enhanced by implementing systems of physician peer monitoring, such as radiologists reviewing a sample of each other’s slides and discussing the diagnoses that are found to be incorrect (pg. 188). This allows physicians to identify mistakes in a safe environment and collectively improve their skills. The knowledge that they are being peer reviewed also encourages physicians to more careful in their decision making.
Treatment protocols should be standardized across physicians practicing in safety net hospitals to reduce variation and ensure equity in the treatment that patients receive. Dr. Groopman highlights how two very different procedures can be believed to be the optimal treatment protocol at the same time simply because they are both championed by a prominent physician who “did it that way” (pg. 163). Physicians are also susceptible to influence from the industry to prefer treatment plans based on non-clinical incentives that may not put the patient’s interests at the forefront, which may be the case with spinal fusion surgeries (pg. 228). As much as possible, treatment protocols should be directly based on the available clinical evidence and physicians practicing at the safety net hospital should be expected to conform to them. This has the benefits of reducing variability, increasing the rate of physicians’ expertise gain, increasing NYC Health + Hospital’s ability to negotiate with insurers, and makes it easier to explain and justify treatment plans to patients.
Although How Doctors Think was published in 2007, experience with the healthcare system today quickly reveals that few of the lessons this book contains have been addressed. In fact, many providers have made the situation worse by continuing to put physicians under time-pressure without developing the physician workflows necessary to maximize patient-physician interactions or redesigning organizational processes to improve diagnostic quality. This fundamental gap in understanding between management and physicians demonstrates why How Doctors Think should be required reading for healthcare administrators.
With a 35,000 primary care physician shortage coming up in 2025, eventually the healthcare policy agenda is going to shift back to the medical education pipeline. In my previous piece on the Iranian healthcare system, I predicted that community health workers were a political dead-end because of the labor market lockout that the AMA has on the healthcare industry. I still predict that’s going to be true, which means that the conversation is instead going to focus on training physicians more efficiently. Fortunately, there’s already work being done on this front.
NYU SoM’s annual report highlights (NYU Three Year Pathway) the three-year medical program that the school launched in 2013. Back then the primary goal of the compressed schedule was to lower the overall cost of medical education and differentiate NYU SoM from the rest of the competition in the medical school marketplace, but the efficiency gains due to reduction in training must not be ignored.
The first batch of students in the compressed program are now graduating, with 15 out of 16 in the inaugural class successfully completing the degree. In addition, testing of the students shows the three year students demonstrating better knowledge and clinical skills than four year students at the end of their respective programs. This finding has enormous potential ramifications and requires additional research. Was the first three year class special, with top performing students self-selecting into the more ambitious program? I’d like to see a comparison of the three year program graduates against other program applicants who were rejected from the three year program due to space limitations and enrolled into the four year program instead. In addition, the three year students may be performing better because they’ve simply had one year less to forget everything that they crammed in MS1; it would be interesting to compare the clinical skills and knowledge of the three year graduates against their four year conterfactuals after they’ve been practicing medicine for a decade to determine which method is really better at integrating the knowledge for long term use. In the meantime, however, the three year program appears to be winning.
What’s important about the accelerated program is that it’s still not fully optimized. The majority of the program time savings is done through cutting down on the time dedicated to electives and residency interview. This is possible because the three year applicants are admitted with a guaranteed residency slot. However, the students in the three year program retain the option of switching back into the four year program,m which means that the core curriculum is still over-training physician to be able to go to any residency. The core curriculum can therefore be collapsed even further to focus only topics and rotations relevant to primary care. I would be shocked if such a degree can’t be reduced to a two year program. Could we produce a family physician in one year? Probably not. But a two year program would be a 100% efficiency gain.
Not to mention that being able to filter out primary care physicians out of the same pipeline that needs to train academic researchers and neurosurgeons would allow institutions to price the degree more appropriately, therefore further reducing the cost barrier of entry. The specialization could also allow institutions to relax their requirements for acceptance. Could we then get away with accepting people with high school degrees into this specialized primary care program, therefore collapsing the current 8 year primary care physician pipeline into a 2 year one? I don’t see why not. There’s no reason to require a Bachelor’s degree; practicing clinicians don’t use what they learn for their premed coursework and MCAT. Students in Europe start medical school straight out of high school and our kids aren’t any dumber.
The efficiency gains are enormous, but we’ll need to do the work to achieve them.