How Do You Handle a Disagreement With an Attending?

The Question

You will encounter this question in some form at nearly every residency interview. Common surface variants:

The behavioral form ("tell me about a time") and the hypothetical form ("what would you do") are testing the same construct. Treat them identically: answer the hypothetical with a real example anyway, or your answer has no evidentiary weight.

Why Programs Ask It

Training programs are managing a genuine tension that has direct patient-safety consequences. On one end: a resident who never pushes back lets errors propagate, defers to hierarchy over evidence, and fails patients when it matters. On the other end: a resident who escalates every disagreement as a personal challenge disrupts team function, undermines attending authority, and creates an environment where feedback cannot be given safely.

Programs have been burned by both. The ACGME's clinical learning environment standards and the broader patient safety literature both require that residents feel psychologically safe enough to speak up about concerns. Programs that interview you are legally and institutionally accountable for what their residents do and fail to do. They need to know where you fall on this spectrum before you walk onto their floors.

There is also an administrative subtext: residency programs deal regularly with grievances, HR processes, and peer conflict. An applicant who cannot narrate a mature disagreement is a potential liability in all of those domains, not only clinical ones.

What It Is Really Testing

Three distinct constructs are being probed simultaneously. Understanding them lets you architect an answer that addresses all three rather than accidentally optimizing for one while failing another.

1. Conflict resolution and emotional regulation under authority pressure

The interviewer wants evidence that you can experience a disagreement without either capitulating immediately (emotional avoidance) or escalating immediately (emotional dysregulation). The signal they are looking for is a deliberate pause between "I noticed a discrepancy" and "I acted." That gap is where professional maturity lives.

2. Patient-safety advocacy without ego

The best disagreements in medicine are disagreements about the patient, not about who is right. An answer that centers the patient's outcome—rather than your need to be vindicated—signals that your advocacy is durable and trustworthy. An answer that centers your own correctness signals that your advocacy will disappear the moment the patient is no longer convenient to the argument.

3. Ability to close the loop and learn

Programs are training you for five or more years. They need someone who can update their priors. An answer that ends with "and I was right" is incomplete. An answer that ends with "and here is what I learned about how to handle it better, or about my own knowledge gap" tells the interviewer that disagreements in your hands become learning events rather than ruptures.

Answer Architecture

Use this as a framework for constructing your own answer from your own experiences. It is not a script. The specific language must come from your actual history—interviewers probe for details, and borrowed language collapses under follow-up.

Step 1: Situate the disagreement type

Before you can decide how to respond, you need to categorize the disagreement. There are two types with meaningfully different handling:

Conflating these two types in your answer is a common error. Treating every disagreement as patient-safety-level signals poor calibration. Treating a genuine safety concern as merely interpersonal signals a dangerous deference to hierarchy.

Step 2: Name your internal process

Describe the moment you recognized the disagreement and what you did before you acted. Did you check your own knowledge? Did you look up the guideline? Did you ask a peer? This step demonstrates that you do not act on reflex and that you are honest about your own fallibility.

Step 3: Choose the right channel

Step 4: Articulate your position once, clearly

State your concern with the evidence behind it. One time. Clearly. Without hedging into invisibility and without repeating it so many times that it becomes a confrontation. "I want to flag something—I noticed X and I'm concerned because of Y. What's your thinking?" is a structure that opens dialogue rather than closing it.

Step 5: Defer and document on non-safety issues

If you have raised your concern and the attending disagrees and the decision does not put the patient at immediate risk, you defer. You may document your concern in the chart where appropriate. You do not relitigate the point at every opportunity. The medical hierarchy exists for functional reasons, and a trainee who cannot accept an overrule on non-critical decisions is not trainable.

Step 6: The safety exception

Make this explicit in your answer. There is a category of situations where deference is not the right answer—where a patient is at serious and immediate risk—and in those situations escalation is mandatory even if it is uncomfortable. Programs want to hear you articulate this carve-out because it tells them that your general deference is principled, not reflexive.

Step 7: Follow up to learn

After the situation resolves, go back. Was the attending right? What did you miss? What did you learn about how you communicate disagreement? This close is what separates a growth-oriented answer from a self-congratulatory one.

One Strong Worked Example

The following is an annotated model. Commentary appears in brackets to explain the function of each move. Do not recite this example—use your own story with the same structural logic.

"During my sub-internship on a medicine service, I was following a patient with a history of a severe penicillin allergy—hives and throat swelling—who had been admitted for cellulitis. The attending ordered amoxicillin-clavulanate. I reviewed the chart, confirmed the documented allergy, and pulled up the cross-reactivity literature before I said anything."

[Checking your own knowledge first is essential. It signals that you do not act on reflex and that you are honest about the possibility you might be wrong. It also means that when you do speak up, you are speaking from evidence, not anxiety.]

"I waited until we had stepped away from the bedside and asked the attending privately: 'I want to flag something about Mrs. X's allergy history—she had hives and throat swelling with penicillin. I was looking at the cross-reactivity data and wanted to make sure you'd seen it before we sent the order. What's your thinking on the allergy?'"

[Private channel is appropriate here—no immediate danger is in progress. The framing "what's your thinking" invites dialogue rather than issuing a correction. The resident is presenting data and asking a question, not declaring the attending wrong. This is the structural move that keeps the conversation collaborative.]

"The attending paused, reviewed the chart, and agreed the order should be changed. We switched to clindamycin. The attending thanked me for catching it."

[A favorable outcome, briefly stated. The resident does not dwell on being right. The point of the anecdote is not the win; it is the process.]

"Afterward, I thought about whether I'd handled the timing correctly—I'd waited until we were off the unit rather than flagging it the moment I saw the order. If she'd received the first dose before I spoke up, I'd have been too late. I decided that with allergy concerns specifically, I should treat it more like an immediate safety flag and speak up faster, even if that means a slightly less comfortable conversation in the moment."

[This closing is the most important move in the entire answer. The resident is not content with having been right. They are identifying where their process could have failed and why. This demonstrates exactly the kind of self-regulation and learning orientation that makes someone trainable over five years. It also preempts the follow-up "what would you do differently" without appearing rehearsed, because the reflection is substantive.]

One Weak Example and Why It Fails

Variant A: The Win Narrative

"My attending wanted to manage a patient conservatively and I pushed back because I'd read the literature and I knew the evidence supported intervention. I kept bringing it up in rounds, I showed the team the studies, and eventually the attending came around and we did things my way. The patient did well. I think it's important not to back down when you know you're right."

This answer fails on multiple dimensions. "Kept bringing it up in rounds" means the applicant relitigated a concern publicly and repeatedly after raising it once, which is professionally disruptive and shows poor calibration about when persistence crosses into insubordination. "Eventually the attending came around" frames the outcome as a personal victory rather than a patient-safety outcome. "I know I'm right" signals a rigidity about one's own judgment that is concerning in a trainee who has not yet completed residency. The attending is erased as a professional with relevant context and experience. An interviewer hears this answer and thinks: this person will be difficult to supervise and may demoralize attendings who give feedback.

Variant B: The Denial

"Honestly, I can't think of a time I've disagreed with an attending. I always trust the supervising physician's judgment because they have so much more experience than I do. I'm there to learn."

This answer fails differently but equally seriously. It is not credible—every medical trainee has encountered a moment of uncertainty or disagreement with a supervisor, and claiming otherwise suggests either limited self-awareness or a reluctance to engage honestly with the question. More concerning, it signals to the interviewer that this applicant either does not notice discrepancies (a patient-safety liability) or notices them and says nothing (a worse patient-safety liability). The line about "I'm there to learn" sounds appropriately humble but actually reads as avoidance. Programs are not looking for a resident who never questions anything; they are looking for a resident who questions the right things in the right way.

Follow-Up Traps

Four probes appear frequently after a strong initial answer. Each requires a distinct strategy.

"What if the attending still wouldn't listen?"

Strategy: Describe the chain of command without defensiveness—chief resident, senior resident, program director, patient safety reporting infrastructure. Make clear you know the difference between using this channel for a safety concern (appropriate, expected) versus using it to relitigate a preference (inappropriate). Name specifically that you would document your concern in the chart when warranted.

"Have you ever gone over an attending's head?"

Strategy: If yes and the reason was patient safety, own it directly and describe the outcome. If no, explain the situations where you would do so and why you have not yet needed to. Do not treat "going over someone's head" as inherently transgressive—the chain of command exists precisely for situations where direct conversation fails. An applicant who would never use it under any circumstances is not safer; they are more dangerous.

"Tell me about a time you disagreed and turned out to be wrong."

Strategy: This is a pure intellectual honesty probe. Have a real example ready. The ideal answer describes raising a concern, being overruled, following up, and discovering that the attending's reasoning was sound—or that you had a knowledge gap you subsequently closed. The payoff of this answer is not the humility performance; it is demonstrating that you follow up at all, which most people do not.

"How would you handle this differently as a senior resident?"

Strategy: Shift the answer from individual advocacy to system stewardship. As a senior resident, your role includes modeling disagreement norms for interns and students, creating an environment where junior trainees feel safe raising concerns to you, and understanding that your disagreements now carry more weight and therefore require more precision. This answer shows that you are thinking beyond your own progression.

Identity Variants

International Medical Graduates

Programs interviewing IMGs—particularly those trained in systems with steep hierarchical cultures—sometimes carry an assumption that IMGs will be passive under authority pressure. This assumption is worth actively disproving without appearing to perform against stereotype.

The practical guidance: choose an example that demonstrates advocacy clearly, where you did raise a concern and the outcome was better for it. Do not choose an example that emphasizes deference, even if it reflects your actual training culture. If your prior training genuinely offered fewer opportunities for this kind of advocacy, say so directly and describe how your US clinical rotations changed your approach. Programs value that arc of adaptation; it is honest and it answers the underlying question about your function in their environment.

Avoid framing your example in terms of "in my country we were taught to defer, but here I have learned to speak up." This framing, while sometimes authentic, can inadvertently reinforce the assumption you are trying to address. Instead, center the clinical reasoning and the patient—the same frame that works for everyone—and let the example speak.

Visa-Dependent Applicants

Applicants on visa sponsorship, particularly J-1 or H-1B tracks, may internalize a constraint that conflict with institutional hierarchy threatens their training position and therefore their visa status. This internalized constraint can produce exactly the deferential affect that interviewers sometimes interpret as a patient-safety liability.

The answer calculus does not change because of visa status—programs expect the same advocacy standards from all residents regardless of immigration situation, and most programs are well aware that residents in sponsored visa tracks have the same patient-safety obligations as anyone else. If anything, being explicit in your answer that you understand the professional obligation to speak up—regardless of the personal discomfort involved—addresses this concern before it becomes one. Choose examples where you did advocate, not examples where circumstances made it easier not to.

Verify current requirements directly with ECFMG/Intealth and official sources for your application year.

Old Graduates and Career Changers

Applicants returning to medicine after a significant gap, or transitioning from a prior career with its own authority structures, face a specific variant of this question: the interviewer may wonder whether prior experience as an authority figure—a practicing physician in another country, a senior professional in a different field—makes you resistant to re-entering a trainee role.

The risk in your answer is not that you will sound passive; it is that you will sound rigid. An anecdote from a prior career where you were right to push back can work, but only if you clearly contextualize the role differential and describe how you understand that the attending-resident relationship is different from peer disagreement. The strongest version of this answer acknowledges that re-entering a supervisory hierarchy after having been senior carries its own adjustment, describes that adjustment honestly, and demonstrates that you have made it.

Applicants With a Prior Difficult Program History

If you left or were asked to leave a prior program, or if your application contains a gap or a program change that will prompt questions, this interview question becomes substantially more charged. The interviewer is almost certainly using it to probe what happened.

The fundamental rule: do not use this question as an opportunity to relitigate a prior conflict, assign blame to prior supervisors, or offer unsolicited explanation of your history. If you do, you have answered a different question than was asked and you have confirmed the concern rather than addressed it. Answer the question as asked, with an example that demonstrates maturity. If the prior program situation comes up directly elsewhere in the interview—and it may—address it there with factual brevity and forward orientation.

If the disagreement in your example happened to occur during a prior program that ended badly, consider whether a different example serves you better. The goal is to demonstrate the construct, not to revisit territory that carries risk.

Couples Match Applicants

No meaningful variant here. The question and optimal answer architecture are the same regardless of couples match status.