August 30th, 2023
RANDOM ACTS OF BOOK REVIEWING by Jim W. Vogele
Random Acts of Medicine, The Hidden Forces that Sway Doctors, Impact Patients, and Shape our Health by Anupam B. Jena, M.D., Ph.D. and Christopher Worsham, M.D., M.P.H. (Publisher: Doubleday) 2023, ISBN: 978-0-385-54881-6
Introduction
Anupam B. Jena, M.D., Ph.D. and Christopher Worsham, M.D., M.P.H. have co-authored an engrossing new book, RANDOM ACTS OF MEDICINE, The Hidden Forces that Sway Doctors, Impact Patients, and Shape Our Health. Published in 2023, this book lives up to its blurbs. Among commendations from best-selling authors and a Nobel Prize winner in economics, we see the name of Steven D. Levitt, coauthor of FREAKONOMICS, who says, “My favorite kind of book: smart, entertaining, and full of surprises.” I agree.
Jena and Worsham, both practicing doctors, describe themselves as “researchers steeped in the world of natural experiments.” As they explain, “natural experiments” arise by accident or chance; and, in healthcare, electronic data often allows researchers to glean insights from these “natural experiments.” That is the gist of the matter and Jena and Worsham have compiled a fascinating array of studies using statistical analysis to tell the story of these “natural experiments.” As you will see in RANDOM ACTS OF MEDICINE, natural experiments allow researchers to conduct studies involving the random or near random natural assignment of subjects where, for ethical or practical reasons, they could not generally be assigned in a controlled study. If heart attack patients happen to have different results when their health crisis occurs during a cardiology conference, perhaps the outcomes can be studied; but patients cannot be requested or ‘assigned’ to have a heart attack at any particular time.
An introduction to “natural experiments”
In an example of natural experiments early in the book, we learn that economists from Princeton and U.C. Berkeley found that families living near congested toll booths on the New Jersey turnpike experienced a decrease in premature births and low-birth-weight babies with the turnpike’s implementation of the E-Z Pass which (naturally) lead to less air pollution from idling cars.
The principle of “counterfactual” evidence is at the core of “natural experiments” and in the New Jersey turnpike example the counterfactual is inherent in the simple before and after comparison (with the authors throughout explaining additional efforts researchers take to eliminate variables and to ensure that circumstances were sufficiently random to qualify as a natural experiment). In other words, the birth outcomes from directly before the change to E-Z Pass represent what would have happened had the E-Z Pass not been implemented.
Another exemplar of a natural experiment described in RANDOM ACTS OF MEDICINE involved looking at the longevity of the winners of Olympic Gold, Silver, and Bronze medals in the early 20th century. Upon considering medal winners in the 1904 through 1936 Olympics, researchers found that the athletes winning a spot on the Olympic podium turned out, respectively, to experience longevity of:
Gold medal – 74.8 years;
Silver medal – 70.8 years; and
Bronze medal – 73.2 years.
The researchers conducting the experiment concluded that the frustration of coming in second could have an adverse effect on one’s health. As Jena and Worsham put it, with a nod to Jerry Seinfeld,* “the psychological effect of falling just short of being crowned the ‘best in the world’ can take years off your life.”
*the authors of RANDOM ACTS OF MEDICINE have a sense of humor which enhances their overall effective writing style. The Seinfeld joke is that a silver medal means, “Of all the losers, you came in first in that group . . . . You’re the number one loser.”
The authors’ admirable approach to skepticism is another reason to embrace RANDOM ACTS OF MEDICINE. Early in the book they acknowledge that the reader may be harboring some skepticism (indeed, I was skeptical regarding the study involving NFL players – I didn’t find it particularly surprising that playing professional football can shorten life expectancy). However, Jena and Worsham encourage retaining a critical viewpoint; that is good, they say, and so it is.
Are marathons hazardous to your health?
One of the more entertaining elements of RANDOM ACTS OF MEDICINE is learning the many ways in which natural experiments can arise. Chapter Five, entitled, “Are Marathons Hazardous to Your Health?” provides a great example.
Data is now available (particularly national data repositories) to analyze events where this wealth of information was not previously available. Thus, in the marathon study, which was published in The New England Journal of Medicine, Dr. Jena along with several others examined eleven different major U.S. marathons and the impact they may have had on people who experienced cardiac events and who happened to live near the marathon course. The study found:
- 2% of patients hospitalized for heart attacks or cardiac arrests on marathon days die within 30 days.
- 9% of patients hospitalized for heart attacks or cardiac arrests on non-marathon days die within 30 days.
Even setting aside marathons, these may be sobering odds if you’re hospitalized on any day for a cardiac event, but it is, regardless, a telling statistic. It turns out that ambulance transport times on marathon days = 18.1 minutes. On non-marathon days transport times = 13.7 minutes. This difference of 4.4 minutes is significant when transport to emergency medical care is at stake.
Thus, in cities hosting the eleven major marathons over the course of a decade, the difference in cardiac event deaths on marathon versus non-marathon days was an absolute % of 3.3%. That may not seem like much to the reader untrained in medicine (such as myself). Anyone, however, could easily relate by imagining that the 1 in 30 deaths beyond what would typically be expected were your own family member.
Without getting ‘into the weeds’ of the differences in healthcare provided, if you don’t already know the difference between the slang “stay and play” vs. “scoop and run,” perhaps it’s better not to know. I jest. To explain, “stay and play” refers to EMT or paramedics working on a patient on-site before transport, while “scoop and run” refers to simply securing the patient in a transport vehicle and immediately heading to the nearest hospital. For more on this distinction, see RANDOM ACTS OF MEDICINE.
Another informative natural experiment we read about pertains to the relative success of opt-in versus opt-out flu shot programs which is discussed in Chapter Three, “Why Are Kids with Summer Birthdays More Likely to Get the Flu?” The relative participation in programs with opt-in versus opt-out features is well-known to any attorney familiar, as I am, with class action litigation. Stated briefly, in some class action cases potential participants (class members) must submit a form to ‘opt-in’ to the case; in others, potential participants must ‘opt-out’ if they do not want to participate. You can imagine which type of case has a higher participation percentage; and the same is true, not surprisingly, with flu shot programs. In one study conducted with adults, participants in an opt-out program had a 45% vaccination rate versus those in the opt-in program who had a 33% vaccination rate.
Some heuristics and jargon
RANDOM ACTS OF MEDICINE employs enough jargon to satisfy the most ardent of logophiles. And count me in when it comes to jargon (with the exception of corporate-speak, i.e. the language of deep diving, synching and aligning is not my favorite, although I have been known to circle back). At the risk of pomposity, I might note here that my linguistics professor lauded my undergrad paper on “Oilfield Argot.” The paper should have been authentic. I had worked several summers in the oilfields of the Williston Basin in North Dakota and often saw bumper stickers in my small hometown in Eastern Montana that said, “Oil Field Trash, and Proud of It.”
Dr. Worsham mentions that, as a resident at Boston Medical Center, he participated in a weekly teaching session focused on metacognition or understanding your own thought process. The idea, he says, was to help residents “avoid cognitive biases . . . through ‘cognitive forcing strategies’” which encourage self-reflection.
RANDOM ACTS OF MEDICINE talks a lot about “cognition bias.” If you’re familiar with the work of Malcolm Gladwell and/or Daniel Kahneman, you won’t be surprised to see them both mentioned in Chapter Four, “Tom Brady, ADHD, and a Really Bad Headache.” Many of the heuristics and jargon in the book would also be familiar to anyone who has read, e.g., THE ART OF THINKING CLEARLY by Rolf Dobelli. An example would be the “representative heuristic,” which is a mental shortcut used to apply expectations to things that “appear to belong to the same category.” Or “availability bias,” which holds that the most recent thing you’ve heard or seen (and the like) may influence your next choice or decision. It’s simply ‘on your mind’ so to speak.
In medicine, cognitive biases can skew diagnoses, which is a problem for medicine and hence for all of us. Referring to UCSF professor Gurpreet Dhaliwal as a “master clinician,” which is “someone who has carefully honed the art and craft of making diagnoses,” Jena and Worsham cite Dr. Dhaliwal’s insights on how physicians may benefit from viewing the “diagnostic reasoning process” as something to be studied and mastered. That makes perfect sense to me.
More heuristics: “Left digit bias” refers to the common knowledge that a price of $7.99 encouages more sales than $8.00, even though the monetary difference is just 1/8th (.125) of 1%. The “win-stay/lose-shift” heuristic refers to the tendency of humans to stick with a winning approach in a situation or game. Although it can be useful at times, the win-stay/lose-shift approach can also be counterproductive if what you learned in a prior situation isn’t necessarily applicable to a new problem.
Two related concepts likely familiar to many of us are “anchoring bias,” which refers to hewing closely to a starting point, and “confirmation bias,” which refers to our tendency to endorse new information that confirms our initial conclusion while ignoring or de-emphasizing new information that does not do so.
As Jena and Worsham point out, many mental shortcuts can be useful, such as using “cumulative experience” which can help us do our jobs more effectively. Nonetheless, “debiasing” – or making yourself aware of cognitive biases that could adversely affect decision-making – is also a key to growth. One of the ways to debias is the use of “cognitive forcing strategies,” which, as in Dr. Worsham’s residency teaching sessions, are processes that require us to “reevaluate [our] own line of thinking.”
I should note that Jena and Worsham state early in Chapter Two that “We won’t bother you with too much jargon in this book . . . .” I took them at their word when they said that on page 13 but by page 200 I was reconsidering this promise. Not in a negative light, mind you – as mentioned, I’m a word freak. And by the way, there’s another title for you, WORD FREAK, a book about hardcore Scrabble players authored by Stefan Fatsis (who became a ranked expert Scrabble player by the time he finished writing the book and who also once, at age 43 and in the course of writing another book, was allowed to take the field in a preseason game suited up as a football placekicker with the Denver Broncos).
Besides all the useful jargon RANDOM ACTS OF MEDICINE also includes lots of . . . random (and interesting) facts. Some are related to medicine; others are not. For example, we learn that the highest percentage of hang-ups while on hold occur when relaxing music is played, while the lowest percentage of hang-ups occurs when jazz is played. Joey Chestnut eats a hot dog every 8 seconds for 10 minutes straight, i.e. 76 hot dogs in 10 minutes at the big competition on Coney Island. Pathologists are nicknamed “the doctor’s doctor.” And the “tincture of time” refers to watchful waiting.
Cardiac surgeons and Used-Car Salesmen?
Okay, there is no close connection between the work of cardiac surgeons and the work of used-car salesmen (I’ve worked with cardio-thoracic surgeons and I definitely would not suggest there are close similarities between their work and that of used-car salesman – nonetheless Chapter Eight of RANDOM ACTS OF MEDICINE is entitled: “What Do Cardiac Surgeons and Used-Car Salesmen Have in Common?”). However, both professions use certain heuristics, e.g. to sell used cars and/or to evaluate patients. Used-car salesmen use left-digit bias as a sales tool (a car with 39,900 miles is typically viewed as more valuable than a car with 40,100 miles, even though the ‘life expectancy’ of these vehicles with only 200 miles difference in wear and tear is minimally different). Likewise, a similar heuristic may be used when emergency department physicians encounter patients who are in their 39th year versus their 40th year. Here a study has found that the 40-year old is disproportionately more likely to be checked “for evidence of heart attack using a blood test called troponin.” Why? Left-digit bias may be the answer according to a study cited in Jena and Worsham’s book.
There are many more such intriguing studies included in RANDOM ACTS OF MEDICINE. Does a surgeon having a birthday near the date the surgeon performs a surgery affect outcomes? You’ll have to read the book to find out!
All these interesting studies, but what do they have to do with physician employment contract reviews? Eventually, well into RANDOM ACTS OF MEDICINE, I did encounter a reference to “value-based payments,” a concept that squarely overlaps with my work as a physician contract review attorney. Whether I am working on physician contract reviews with Oregon, Washington, Montana, or California physicians, I often see “value-based” components in their employment contracts.
The compensation physicians earn by means of these “value-based” bonuses or formulas is always minimal compared to productivity compensation or base compensation. Nonetheless, many employers offer “value-based” elements in their compensation plans (regardless of whether the employer may be a little reluctant to detail what is included, and/or the specific formula used, often because of their complexity). Jena and Worsham have a few words about incentives that may not be directly related to physician contract reviews, but which are of relevance to the healthcare business and healthcare professions. Suffice it to say that value-based payments can have the potentially untoward effect of doctors spending more time charting and less time in direct contact with their patients. I will leave it to Jena and Worsham to explain this a bit further, which they do in their book.
What Makes Better Doctors?
And beyond physician contract reviews, to what greater end are all these studies? If you’re reading this, you’re certainly aware of the outsized role healthcare plays in our national economy in the United States and in our personal lives. The amount spent annually on healthcare in the U.S. is $11,582 per person. This is 17.7% of our gross domestic product. While these hard numbers tell a tale, as RANDOM ACTS OF MEDICINE puts it, “examining the role chance plays in medicine can [also] contribute to the health of our patients and the well-being of our communities.”
Near the end of the book, the authors turn their narrative attention to the question of what makes “better doctors.” Specifically with regard to hospitalists, are the better doctors younger (under 40) or older (over 40 but the book focuses on over 60)? Using 30-day mortality rates (which seems a rough measure, but it appears to be a standard and/or established benchmark), Jena and Worsham share some findings. Without spoiling the outcome of the research entirely, it appears that older and younger physicians are approximately equally effective IF the older physicians continue to see patients with high frequency and also take the opportunity and time to learn about new techniques and drugs. That is the case with hospitalists. With surgeons, we see experience playing a greater role in terms of effectiveness. I will let Jena and Worsham explain further, in proper context and terminology.
Before closing, the authors take note of gender and diversity in the medical professions. With regard to gender specifically, you’re probably aware of the fact that women earn less for the same work pretty much across the board in all industries. The authors also provide examples of explicit and implicit bias against women in medicine. A reader may or may not be surprised to learn that women MDs seem to be spending more time with their patients than are men (are women more caring than men, or more patient than men or something else?).
Another potentially important question addressed in RANDOM ACTS OF MEDICINE is: “Do ‘better’ schools graduate better doctors?” The answer is, “No,” to the extent this can be measured. As the authors point out, attempting to do so is subject to countless biases, including debate over how so-called “better” schools are ranked. Furthermore, they say, anyone who obtains a medical license is apt to be so accomplished, having passed through so many “checkpoints,” that one wouldn’t expect major differences “at least as those differences in education pertain to patient outcomes.”
Conclusion
Just as you will often find that your physician employment contract requires you to satisfy your CME obligations, attorneys are required each year to take a certain number of hours of Continuing Legal Education (“CLE”). Over the years I have taught many CLE courses (including presentations on employment contract law to members of the Oregon State Bar Association). I sometimes question whether I will learn anything at a seminar, whether attending or presenting. On occasion, I admit feeling that continuing education seminars are a bit of a chore. Yet I always learn something at every seminar I attend or teach. Often quite a bit. While this is not exactly an epiphany, I find it encouraging that one can always learn, oftentimes by doing things that may not be your favorite endeavors.
RANDOM ACTS OF MEDICINE concedes that, “There are limits to what natural experiments can (so far) tell us” but also reminds us that “chance occurrences change the course of our lives all the time.” Indeed, they do. And will reading RANDOM ACTS OF MEDICINE make you a better doctor? Neither specifically nor directly is it likely to do so. Yet reading this excellent book may well be a useful cognitive forcing strategy (something which can be more enjoyable than it sounds). In other words, reading this book will almost certainly make one pause to reflect for a moment about one’s own thought processes and approach to decision-making. That reflection or cognitive debiasing could lead to improved outcomes in our work and in our daily lives.
When I saw RANDOM ACTS OF MEDICINE at my favorite bookstore (Powell’s City of Books in Portland, Oregon), and noticed that it was full of statistical analyses of “natural experiments,” I wondered if it would be a dry read. To the contrary. Jena and Worsham’s work is captivating, not dry. To hearken back to Steven Levitt’s dust jacket blurb, to encounter a book this intelligent yet also this entertaining is an excellent find! I highly recommend RANDOM ACTS OF MEDICINE, the book.
