The Effects of Framing: How We Influence Our Patients
We usually need time to consider whether an article belongs on Sensible Medicine. That was not the case for this article, which combines a baseball metaphor with a meditation on doctor-patient communication. I decided we needed to post it before I finished reading it. I hope you like it as much as I do.
Adam Cifu
If you watch a lot of baseball, you probably understand catcher framing. But for those who are unfamiliar with the concept, it is the catcher’s ability to receive a pitch located just outside the strike zone (an area nicknamed the Shadow Zone) in a way that convinces the umpire that it was actually a strike.
Since his Major League debut in 2023, San Francisco Giants catcher Patrick Bailey has been baseball’s best framer. In 2024, he led all catchers by converting just over half (52.5%) of non-swing pitches in the Shadow Zone into called strikes. This saved an estimated 16 runs for the Giants over the course of the season; the league average is zero. This brief highlight video will show you just how good Bailey is at influencing umpires and stealing strikes from opposing batters.1
Some baseball fans dislike catcher framing because it is a subtle, but intentional, attempt to deceive the umpire. Moral questions aside, catchers who excel at framing will continue to provide their teams a crucial tactical advantage — at least for now. In 2026, MLB will likely implement the Automated Ball-Strike (ABS) system, which uses high-speed tracking cameras and computer modeling to determine whether pitches are strikes. Some have advocated using these “robo-umps” for every pitch, but it appears that for the moment, MLB will only consider their use in a limited challenge system.
Physicians influence our patients in a very similar way. We, however, do so inadvertently. A 1982 study by McNeil, et al in the New England Journal of Medicine demonstrated the effects of framing on the treatment decisions made by patients. They observed that how a potential intervention was presented impacted whether a patient would accept it. In their hypothetical example, patients were much more likely to favor surgical treatment for lung cancer over radiation if the outcomes were framed as the probability of survival instead of mortality:
Mortality frame:
“Of 100 people having surgery, 10 will die during treatment, 32 will have died by one year, and 66 will have died by five years…”
Survival frame:
“Of 100 people having surgery, 90 will survive during treatment, 68 will have survived by one year and 34 will have survived by five years…”
The numbers were statistically equivalent, but if they referenced survival, patients were more likely to choose surgery. If mortality was cited, patients more often chose radiation.2 Later studies have replicated these findings.
This, of course, has important implications for clinical practice. Physicians are ethically obligated to obtain informed consent, but the way we do it may manipulate patient preferences and thus threaten their autonomy. As you can see, we must be particularly thoughtful in these conversations.
Consider the following vignette:
A 55-year-old male presents to his PCP. He recently underwent a coronary CTA for chest pain and was found to have non-obstructive coronary artery disease. He has hypertension, hyperlipidemia, and a BMI of 32. He has been considering a GLP-1 receptor agonist to help him lose weight, but has been reluctant due to concern for potential side effects.
Semaglutide has established efficacy for weight loss. Recently, it was shown to reduce the risk of adverse cardiovascular events in obese adults without diabetes. The PCP has several options for how to deliver that information:
The busy clinician: “Semaglutide is a great idea — it will reduce your risk of a heart attack.”
The pharmaceutical rep: “Semaglutide will lower the chance of a cardiovascular event by 20%!”
The evidence-based medicine guru: “Semaglutide has been shown, in a large, randomized trial, to reduce the risk of cardiovascular events — like a heart attack, stroke, or cardiovascular-related deaths. When compared to a placebo, these events occurred less often in study participants taking semaglutide by 1.5%. In other words, for every 67 patients treated with semaglutide over 3.3 years, one cardiovascular event was prevented.”
Those who choose the first option likely do so because of the time pressures of modern ambulatory practice. Many of us will discard option two since it quotes relative risk reduction. Pharmaceutical companies and the media tend to prefer this frame because it exaggerates the medication’s effect. Still, clinicians less well-trained in critical appraisal may find this statistic easy to pull out of their lab coat pocket.
But judging these approaches is not the purpose of this essay. I have taken each of these in my 14-year career as a family physician. I believe that empowering patients with helpful information to make decisions is an ideal to strive for. I also appreciate the value of the “number needed to treat” for conveying study outcomes to patients, so I prefer example three.
I think most patients in this case example would agree to semaglutide. But if I take the findings of the McNeil study seriously, I am forced to acknowledge that the patient may make a different choice if I frame it this way:
“For every 67 patients treated with semaglutide over 3.3 years, 66 will not see a difference in their chance of a cardiovascular event.”
Or this way:
“While semaglutide does help some, for 66 out of 67 people, there will be no change in their risk of a major cardiovascular event over 3 years.”
It’s easy to believe that some patients, hearing that version might think twice about starting semaglutide. Moreover, if you overheard a colleague making this statement, you might reasonably conclude that they were trying to talk the patient out of taking it. Although statistically true, the negative frame is unnerving. I imagine that this is because most physicians would support the addition of semaglutide in this patient—and why not—the risk-to-benefit ratio is favorable. But how discomforting is it to know that your words could so easily tip the scale in the opposite direction?
I don’t believe that any of us seek to manipulate our patients. But this should make us reflect on how we conduct shared decision-making. For example, how do you frame choices around cancer screening or vaccines3 — which, for any given individual, are statistically unlikely to benefit them? As the authors suggested in NEJM back in 1982, perhaps it makes sense to present the information in both positive and negative frames. If the preference remains the same, you can be confident that this is what the patient truly wants. If not, further conversation is warranted.
Just as some baseball fans hope to rid the game of the hoodwinkery perpetrated by wily catchers on umpires, most physicians will feel uncomfortable knowing that they may also exert subconscious influence on their patients’ decisions. Unfortunately, it cannot be helped because human communication will always be prone to the effects of framing. The best we can do is to hold fast to ethical principles and remain mindful of our influence as we seek to inform our patients — until, of course, the AI “robo-doc” comes to replace us too.
is a practicing family physician and associate professor at Wake Forest University School of Medicine. On his Substack, Statcast & Stethoscopes, he explores what baseball can teach us about making better decisions in medicine.It also reveals the level of nerdiness that baseball fandom can reach.
The researchers also found that despite their medical expertise, physicians were similarly susceptible to these framing effects.
The RSV vaccine is widely recommended for adults over 60. According to this meta-analysis, the number needed to vaccinate to prevent one case is between 417 and 500, depending on the definition of “lower respiratory tract disease” you choose.
Photo Credit: Chris Robert
 | A guest post by
|
Framing is yet another example of why it is important to have a consistent source of primary care. My patients, through years of talking with me, have noticed how I frame things and the manner in which I tend to evaluate and present evidence. If it is not congenial to the manner in which they think, they drift into other doctor's schedules and if it is congenial, they say things like, "I didn't want to do what my specialist said until I talked with you."
The manner of framing is one of the many things that get folded into what a patient means by "I trust my doc."
I recently wrote my honors these on this topic and one of the interesting things about this from a biomedical ethics standpoint is that these framing biases bring informed consent into question. Since a patient’s consent can be solely dependent upon how the procedure or medication is described, the patient is not actually consenting to the intervention but to the way in which it is described. Hypothetically speaking, there is someone who consented under the survival rate description that would have dissented if given the mortality rate description. Therefore, if consent is dependent on the way in which the procedure is described and not the procedure itself, the patient’s consent might actually be invalid (if we are taking the common view of consent here). Also to note: I am not commenting with any medical expertise so I don’t have much boots on the ground experience here, just some philosophical research. Anyway, I found this article to be very interesting and enjoyed it!