Beware how the affect heuristic filters your view of data

3

Category : Medical Rants

The Spock in us would like to see data as hard, fixed, and totally interpretable.  The Dr. McCoy in us understands that data do not have those properties.  Nietzsche once wrote, “There are no facts, only interpretations”

In fact we always interpret “facts” in light of our biases.  Our filters come from our preconceived opinions.  If we like something, we give great value to “data” that support that belief, while we de-emphasize the negative findings.  Vice versa works also.

When you watch the debate, if you like Hiliary Clinton you will cheer her pronouncements and believe them true.  If you like Donald Trump, you will dislike her statements.

Few viewers will look at the debate dispassionately.  And thus those who claim fact checking are, in fact, not really unbiased either.  They develop data that supports their preconceived notions.

We see this in medical debates – screening for prostate cancer, age to screen for breast cancer, whether or not to empirically treat some adolescent/young adult sore throats, the value of palliative care in cancer patients, etc.

A student recently reported a conversation he had with a program director.  He asked her if her family medicine residents were exposed to direct primary care practices.  She proceeded to scold him and anyone who would even consider such a practice.

In our society we too often disregard Steven Covey’s admonition, “Seek first to understand, then to be understood.”  We will not see or hear this recommendation modeled in a debate,  We rarely see this recommendation followed in medical debates.  This recommendation requires civility, and that attribute has become much too rare.

The affect heuristic and conflicts of interest

6

Category : Medical Rants

Understanding the affect heuristic leads one to expand their understanding of the term “conflict of interest”. My simple explanation of the affect heuristic follows:

If you like something, you overestimate the benefits and underestimate the risk. The contrary holds also. If you dislike something, you underestimate the benefits and overestimate the risks.

This concept helps us understand that we should consider multiple conflicts of interest. Clearly the pharmaceutical and device industries have taken advantage of this concept. They hire very likable representatives. This representatives treat physicians very well. Physicians, nurses and office staff like the representatives. Thus, the companies are well considered. When considering a drug that they are selling or a device that they make, the recipients of the kindness attribute benefits and minimize risks for the drug or the device.

The same phenomenon occurs when the business funds research. The recipients of the research funds will have their values changed.

These values change subconsciously. We need not say that the conflict is explicit.

But there are other major conflicts. Our expertise is a conflict. For example, my thoughts about pharyngitis induce the affect heuristic. Being a diabetologist induces the heuristic. Diabetologists overestimate the value of tight control, and suppress the estimates of risks. Interventional cardiologists overestimate the benefits of stenting and underestimate the relative benefits of CABG.

Insurance company executives underestimate the benefits of expensive treatments.

The only logical way of avoiding conflicts of interest is to balance our committees and thereby balance the conflicts.

Most guideline writers and performance measurement developers would deny that they have a conflict. But we all have some conflict. We should acknowledge our conflicts and make them explicit.

We cannot avoid conflicts, thus we should consider the competing conflicts.

How the affect heuristic creates a threat to high value care

4

Category : Medical Rants

Deborah Korenstein has a brilliant commentary in the JAMA Internal Medicine – Patient Perception of Benefits and Harms: The Achilles Heel of High-Value Care

… Patients overestimated benefits and underestimated harms for approximately two-thirds of studied interventions, regardless of the type of service (eg, a screening test or a surgical treatment), clinical context (eg, care related to cancer or cardiovascular disease), or patient population. This finding may not be surprising and likely reflects a number of underlying issues, including the influence of direct-to-consumer advertising, a general societal bias toward more rather than less health care, and perhaps human nature. Nonetheless, the systematic tendency for patients to misunderstand benefits and harms is a fundamental threat to our ability to improve health care value and must be addressed.

These estimations have an easy explanation – the affect heuristic.

Patients like the idea that a test will help them, or a treatment will cure them, or an operation will improve their outlook. Therefore they will consistently overestimate the benefits and underestimate the harms. These findings have an elegant explanation.

As physicians we must learn to understand the affect heuristic. It explains why many subspecialists will favor a new treatment, even without adequate advice. Our emotions trump the data.

Perhaps this explains the classic, if the only tool you have is a hammer, then everything looks like a nail.

For Star Trek fans, Spock’s logic runs astray to Dr. McCoy’s feelings.

Unintended consequences and the affect heuristic

6

Category : Medical Rants

Why do we never consider unintended consequences? Whether we are thinking of legislation or physician led guideline panels, or governing bodies (like ACGME), the lack of consideration of unintended consequences remains mind numbing. Let me provide some examples.

Please read this articles about how the “war on drugs” has fueled the hepatitis C epidemic – War on drugs ‘driving hepatitis C pandemic’. One could also argue that this war damages more young people than the drugs themselves.

Many “illicit” drugs do damage to users. But the laws impact both the users and innocent bystanders. The laws put too many young adults in prison. The laws force a black market, and for many drugs stimulate crime. Advocating for de-criminalization is not advocating for drug use, but rather focusing on preventing complications (like infections) and on programs to help patients stop drug use.

Next read any number of blog posts on performance measures. Let me give 3 quick examples of poorly considered performance measures. The most obvious is the 4-hour pneumonia rule that led to increased unnecessary antibiotic usage. Trying to drive HgbA1c below 7% for every patient puts some patients at risk from hypoglycemia and from drug side effects. In Great Britain, paying physicians to decrease appointment waiting time induced less continuity of care (with the primary physician).

Finally, we are learning what many predicted. The work hour restrictions for residents are hampering medical education! For New Doctors, 8 Minutes Per Patient

In example after example, thoughtful critics receive minimal lip service. Each group of decision makers moved forward, “damn the torpedoes, full speed ahead”. Why?

I believe that the big problem is the affect heuristic. In short that decision making heuristic says that if you like an idea, you over value the benefits, and underestimate the harms. We must figure out how to work around this undervalued concept. We desperately need to develop decision making processes that focus on both benefits and harms and develop logical means for balancing the benefits and harms.

Unfortunately, I remain pessimistic. Decision making bodies have biases. Until they understand their biases, we will have the problem of unfortunate, unnecessary and potential dangerous unintended consequences.

The affect heuristic explains conflicting opinions in medicine

3

Category : Medical Rants

So what is the affect heuristic?  We each overweight risks and benefits depending on our feeling about the issue.  I am currently reading Thinking, Fast and Slow by Daniel Kahneman and just finished reading the chapter on this heuristic.  Intuitively I knew this, but reading the supporting research clarified how this works.

This heuristic leads to many differences in judgment.  The recent controversies over prostate cancer screening and mammography support this idea.

The committee that recommended against PSA carefully looked at the data.  They had no preconceived biases, because they had no conflicts.  Remember that conflicts include financial payments or the conflict of what you do.  The urologists have attacked this recommendation, because they do prostate cancer surgery.  Because they do the surgery they tend to overvalue the benefits and undervalue the risks.  They are not being disingenuous.   They are not being greedy.  They really believe in the benefits.  This belief is easily predicted by the affect heuristic.

The same concept explains the mammography controversy.  All women know many other women who have had breast cancer.  Some of those women had the breast cancer found by mammography.  They have heard for many years that mammography saves lives.  Therefore they look at the benefits as markedly trumping the risks.

I would argue that this heuristic explains the battles over guidelines.  

So we should strongly reconsider the constitution of guideline panels.  We should avoid experts on guideline panels, and rather substitute intelligent methodologists and generalists who can examine the data more objectively.  

Guideline panels have many explicit conflicts of interest, but we must ask whether we should include implicit conflicts of interest.

Or perhaps we cannot possibly develop objective guidelines.  Guideline panels have human beings involved, and we all have biases and affect.

Understanding why we have strong differences interpreting the same “facts”

2

Category : Medical Rants

In the past I have written about the affect heuristic.

 The affect heuristic is a swift, involuntary response to a stimulus that speeds up the time it takes to process information. Researchers have found that if we have pleasant feelings about something, we see the benefits as high and the risks as low, and vice versa. As such, the affect heuristic behaves as a first and fast response mechanism in decision-making.

Now I have learned about two related concepts.  These concepts both explain biases in reasoning – the myside bias and motivated reasoning.  Learning more about these concepts should help us better understand why many debates in medicine (and policy) seem so irrational.

The myside bias states that our opinions influence how we evaluate information.

A common error that occurs with everyday thinking is Myside Bias — the tendency for people to evaluate evidence, generate evidence, and test hypotheses in a manner biased toward their own opinions.

Finally, we should consider motivated reasoning.

The processes of motivated reasoning are a type of inferred justification strategy which is used to mitigate cognitive dissonance. When people form and cling to false beliefs despite overwhelming evidence, the phenomenon is labeled “motivated reasoning”. In other words, “rather than search rationally for information that either confirms or disconfirms a particular belief, people actually seek out information that confirms what they already believe.” This is “a form of implicit emotion regulation in which the brain converges on judgments that minimize negative and maximize positive affect states associated with threat to or attainment of motives.”

As we learn about these cognitive psychological concepts, we risk applying them to our “opponents”.  But we are all susceptible to these reasoning flaws.  This new study confirming motivated reasoning inspired a wonderful blog post that ends with this balanced and important sentence.

I also think we need to remind ourselves that people who disagree with us are just people. They are not demons. They have their reasons for believing what they do. They think they are right just as much as you think you are right. They don’t disagree with you because you are virtuous and they are evil. They just have a different narrative than you, and your narrative is likely just as subjective and flawed as theirs.

We should all consider these concepts as we watch and participate in policy debates.  Both sides of issues have great difficulty understanding the other side.  We often become frustrated that our perfect logic fails to convince our opponents.  They are not stupid; we are not smart; we just have different underlying beliefs.

These concepts should remind us that humans interpret all data on the basis of their underlying beliefs.  And changing beliefs is therefore very difficult.

Nietzsche famously wrote, “There are no facts, only interpretation.”  He also said, “You have your own way. I have my way. As for the right way, correct way and the only way, it does not exist“.

I hope this discussion helps us understand why many policy debates seem so irrational to both sides.  Examples might include the abortion debates, the ACA debates, the anti-vaxxers and the MOC debate.  Of course, people on both sides cannot understand the other side at all.  Until we understand the problem, our usual persuasive techniques will always fail.

Direct primary care – the answer or the problem?

5

Category : Medical Rants

Long time readers know of my fascination with the affect heuristic.  Simple stated, we overvalue the benefits of a concept that we like, and underestimate the problems 0r vice versa.

This article about direct primary care induces conflicting analyses – Here is the PCP crisis solution and it’s simple

I like the idea based on this reasoning.  Primary care in 2017 has several problems.  Both physicians and patients have dissatisfaction with direct face time.  Primary care physicians suffer high levels of burnout because the financial model requires them to see patients to quickly to do their job properly.  These quick visits likely induce physicians to order more tests and consultations than they would if they could spend more time on history and physical examination.

Direct primary care allows physicians to spend more time with patients, because they decrease their “panel size” from greater than 2000 to 800 or less.  These physicians have more time to communicate with their patients – using telephone and email.

But the panel size decrease waves a red flag for opponents of this movement.  They always ask – who will care for the patients?

When primary care physicians burnout they often totally leave their practice, often becoming hospitalists or urgent care physicians or subspecialists or retirees.  If direct primary care keeps them practicing, even with fewer patients, at least they are providing important primary care.

Currently, medical students and residents often find primary care unappealing because of the work conditions.  I often argue that direct primary care may induce students and residents to choose primary care and work with a reasonable number of patients.

This debate has no solution.  My arguments are not based on data, but rather on anecdotal observation.  I worry about primary care, because the current model often leads to more expensive substandard care.  You cannot rush visits and provide the highest quality primary care.  You must take shortcuts to shrink your visit times.

This debate is philosophically interesting and, in my opinion, a great example of that affect heuristic.  We cannot resolve this question with data, because the factors are multiple and too often you really do not understand the underlying motivations for doing primary care, or leaving primary care or moving to direct primary care.  So we will likely continue debating this issue to no clear conclusion.

But of course, I am correct.

Tips for IM Attendings – Chapter 16 – Why I like table rounds before bedside rounds

1

Category : Medical Rants

In a previous chapter, I noted that various styles can work well.  Adopting a style and perfecting it are hallmarks of highly ranked and desired attending physicians.  For 35 years I have developed a style that works.  When other attendings try this style, it may or may not work.  My learners often mention how valuable they find this style.  Their feedback has helped my develop the key features of the style.

Any teaching style should develop from clearly defined goals.  My teaching style directly follows from these principles:

  1. We want each learner to understand each patient, their diagnoses and treatments.
  2. We want to challenge each learner to grow.
  3. We want the learners to be able to hear the entire conversation and participate in that conversation.
  4. We never want to embarrass a learner in front of a patient.
  5. We want to look at images prior to seeing the patient.
  6. We want to challenge the entire team and research topics as they arise.
  7. When we go to the bedside, we want to talk with the patient, find out their concerns.  We want to make certain that the patient understands the plans for that day.
  8. When we need to have serious conversations with the patient, we should have discussed that need prior to going to the room.

Here is my formula for days without new patients.  We discuss each patient – focusing on the results of key tests, any surprises, what the consultants have written or told us and discussing today’s plans.  Often these discussion raise teaching points – so we take a few points to clarify an issue that at least one learner does not completely understand.  By the time we have left the room we all have a good idea of the day’s plans.

Then we go see each patient.  Sometimes we take more history.  Sometimes we do more examination (for example listening to the lungs to see if they have changed since the previous day).  Sometimes we educate the patient.  Sometimes we just answer their questions.

Occasionally we find something at the bedside that changes our plans.  For example today we had planned to discharge a patient with acute bronchitis causing a COPD exacerbation, but at the bedside we tested his oxygenation off oxygen.  Unfortunately, his oxygen saturation was only 87%, so we changed our plan and explained to the patient our new plan.

The key to starting with “table rounds” is reasonable efficiency.  We discuss the patient in depth, but with a clear focus on what is really important for the patient.

What do learners tell me?  First, many learners appreciate the relative quiet of the teaching room.  They tell me that too often in the hallway or nurses’ station they cannot hear the conversation.  They tell me that the education is less focused when not in a quiet room.  Second, interns tell me that they are much more comfortable when they have to cover their fellow intern’s patients.  They have a better understanding of the plan.  Finally, they like the efficiency that these discussions provide.

Pure card flip rounds do not work because you do not involve the patient.  Pure bedside rounds leave out some important teaching moments.

Of course I am very biased because I am describing my style.  The affect heuristic tells us that I will overemphasize the good and de-emphasize any problems with this style.

But the purpose of these posts is that you will consider these ideas as possibilities.  I do not expect you to do everything that I espouse, but rather I hope to increase your own self-awareness of your attending style.

For those who favor a different style, I invite you to write a guest post extolling the virtues of your style.

 

 

The work hours conundrum – focus on the work rather than the hours

1

Category : Medical Rants

The publication of the FIRST Trial adds fuel to the fiery debate over work hours.  Reading this study provides a variation of a Rorschach test.  A brief news report (with a biased title) on the study appeared in the Washington Post – Back to extremely long shifts for new surgeons? Study finds few negatives.  The NY Times published this op-ed – Should Doctors-in-Training Work Fewer Hours?.  The author, a current resident, frames the issue in this manner:

The right answer on how many hours residents should work may be more nuanced than we’ve been willing to accept. It isn’t the same today as it was 20 years ago, as the complexity of caring for patients and medical technology continue to evolve. It varies by subspecialty — discontinuity may have graver consequences for neurosurgery, say, than for radiology. And it hinges more on the character of work than the length of it — I’d spend twice as long at a patient’s bedside if I could spend half as long at a computer.

Ultimately, the answer may be as philosophical as it is empirical. What kind of doctors do we want to be? What kind of doctors do patients want us to be? And does what we can’t measure still matter in a profession that’s now judged and motivated by what we can?

But the problem remains much more complex than this study or most commentaries include.  What is the purpose of training?  We all go through training to best prepare ourselves for independent practice.  We know that volume matters for clinical education.  Are we compromising on volume for the sake of meeting somewhat arbitrary work hour restrictions.  As an educator do I criticize or silently praise the intern or resident who works a bit harder and longer?

As the NY Times author notes, the problem with residency training involves the work more than the hours.  Residents have to spend too much time at the computer screen.  They have too many forms to complete.  They no longer can write meaningful notes, but rather have to comply with the arcane billing inspired formula for the incomprehensible notes that we see every day.

I favor the professionalism aspect the results in devotion to the patient.  Too often handoffs do cause problems and decrease the educational value.

But then I have my biases.  How we view this problem likely involves the affect heuristic.  I worked many hours as a resident, and believe that my clinical skills improved dramatically throughout my residency.  Thus, I attribute my growth to the process we followed.  For others, the thought of 30 hours invokes fear and thus they will overestimate the dangers and underestimate the benefits of longer shifts.

This study does not resolve our feelings.  What it does tell us is that patients do not appear to suffer when residents work longer hours.

Do you love or hate the ACA?

5

Category : Medical Rants

The Affordable Care Act is a complex bill. Most objective observers can find components that they like and components that they would change. Unfortunately, few invested individuals remain objective. Why not? Why do we love or hate the ACA?

I submit that the affect heuristic rules here. I often talk about the affect heuristic. When you like something, you overestimate the benefits and underestimate the risks. The opposite is also true.

Few politicians or physicians oppose expanding coverage. The fight occurs over how we can expand coverage. If you oppose the current administration, then you will minimize the benefits in expanded Medicaid coverage and the insurance options. You will focus on the problems of guaranteeing coverage for pre-existing conditions. You will talk expansively on the unintended consequences that permeate this complex bill.

The admirers of the current administration, likewise will overestimate the benefits of various components and minimize the potential financial problems. They will gloss over the bill’s minimal focus on primary care. They will ignore the lack of attention to our growing physician shortage.

We neither need harsh objections nor admiring toadies. Rather we need both sides to objectively evaluate the strengths and weaknesses of the current bill with a goal of improving it over time.

It is the law. It is unlikely to be repealed. So why do we not work to fix it. Both sides are at fault, because both sides either love or hate the bill. Neither side is often objective.

I favor the bill generally, but see many flaws. We really do need more universal coverage. However, I fear that the insurance companies have won the day.

What do you think? Please point out the strengths and weaknesses. You cannot honestly say that everything in the bill is good or bad. Let us all move forward to improve access to health care.