I have trouble understanding the recent animosity to the phrase “First, Do No Harm” as it applies to medical practice. I have always accepted it as a guide for medical practice. Way back in 2012, when Vinay and I published A medical burden of proof: Towards a new ethic, we built our arguments on Primum non nocere, which we referred to as ‘the most fundamental principle of medicine.’
In this essay, Dr. Rohlfsen articulates the argument against Primum and explains why the movement is misguided.
Adam Cifu MD
Until recently, I was unaware of efforts to abandon the practice tenet “First, Do No Harm,” a phrase often associated with the doctor’s oath. As I looked into delegitimization attempts, I was surprised not to find a cogent defense of this foundation of ethical practice. Here, I explain why abandoning this principle is a bad idea for medicine and society.
To start, let’s look at three main critiques fueling this campaign:
First, critics accurately point out that the aphorism is not part of the Hippocratic Oath, as is often assumed. The tenet originated centuries after Hippocrates, so divorcing primum non nocere from the pledge that students often recite is not unreasonable.
Second, some see “do no harm” as redundant to the principle of “non-maleficence”—one of the four pillars of modern ethics alongside beneficence, autonomy, and justice. Worse, critics argue that placing it “first " undermines the balance between these core values. I agree that this balance is important and will explain why primum non nocere reinforces each cornerstone, grounding them in the concept of fallibility.
Third, the phrase is sometimes narrowly interpreted as a command to avoid any physical act of injury even in the service of a greater good (e.g. amputating a limb to save a life). Since medicine would be impotent if this view were adopted, nobody should endorse this chronologically rigid interpretation. The phrase does demand we pause, though. And this conservatism can frustrate those who yearn for medicine to be more proactive.
If primum non nocere is so often misunderstood, why keep it? If it merely cautions against the cascades of harm medicine can cause, should it become a relic of the past?
Ironically, every call to abandon primum non nocere only amplifies its spread. The phrase does not refer to “First” as in chronology or precedence. Instead, it’s an acknowledgment of “First” as in foundational. A “first principle” is one that cannot be reduced. In this case, it speaks to our bidirectional relationship with patients, including the expectation that we are collectively held to a standard of accountability. In this sense, it’s unsurprising that the phrase spreads every time a patient is harmed. Primum non nocere is our professional contract. It is primary, foundational.
It's still curious how a “cautionary warning” became the sacred bond in our social contract. Why not, “first, maximize benefit?”
History answers this question. Medical reversals are sobering. From the horrors exposed by the Nuremberg trials to the guidelines of the Belmont Report, we learned that scientific curiosity must never outpace moral judgment. From frontal lobotomy (which won the Nobel prize in 1949) to modern cancer treatments, we know harmful practices persist for years (and even decades) before they are reversed.
Despite landmark ethical advances, confidence in health institutions is in crisis. When reflecting on threats to the social contract, I examined dozens of historical examples and found a unifying theme – fallibility. Once I centered this concept, the relevance of primum non nocere became clear. At the heart of scientific and moral progress is a process of error correction. Karl Popper said it best, “we do not know, we can only guess.” All knowledge – scientific, moral, or philosophical – is provisional. Fallibility ensures that no matter how advanced our practices or systems become, error remains inevitable. And this is true not only in medicine but in ethics itself.
Primum non nocere resonates across generations because it admits this uncomfortable truth, laying bare our failures. Despite our best efforts, we are susceptible to commission bias, misdiagnosis, and misadventures in health policy. We fall short in how we communicate uncertainty to the public. Our corrections are timidly swept away rather than boldly proclaimed.
Despite the Institute of Medicine (IOM) report in 1994 normalizing error disclosure in the culture of medicine, incremental error correction remains a foreign concept in public health. Why? Because errors are politicized. They are weaponized against rivals and minimized amongst allies. But rarely are they discussed in the town hall. At what point do we share that the road towards “zero harm” is paved with endless errors?
In The Science of Can and Can’t, Chiara Marletto sheds some light, proclaiming that “Fallibilism makes progress feasible because it allows for further criticism to occur in the future, even when at present we seem to be content with whatever solution we have found.” She adds, “Any good solution to a problem may also contain some errors… Fallibilism […] tells us that errors are extremely interesting things to look for. Whenever we try to make progress, we should hope to find more of them, as fast as possible.”
Uncomfortable? Me too. Perhaps, the most illuminating criticism of First, Do No Harm, then, is how it constrains our social contract by further stigmatizing error, making it unacceptable in the pursuit of progress. In that vein, an improvement might be primum, errorem quaere (First, look for error). Far from passive or reactive, it’s inviting and engaging.
So, while primum non nocere is a loaded term, often trafficked for political gain, it represents our current professional contract with society. Let’s start where we are.
Without the acknowledgement of fallibility, the ethical practice of medicine becomes an expert class, unidirectional endeavor. With societal trust, we have something to build on. With an embrace of fallibility, progress in medicine becomes a dynamic, bi-directional, and (at least) honest conversation, rooted in error detection and disclosure.
By now, it shouldn’t be surprising that First, Do No Harm has permeated into other fiduciary sectors. Anywhere people entrust their finances, health, or basic human rights to experts wielding asymmetric power, the maxim speaks directly to the overarching fallibility of the expert class in ways that entire chapters of ethics have failed to. The sheer mimetic reach of this philosophy makes its value to society self-evident. The fact that many remember it as part of our medical oath (despite the mistranslation) speaks to its resonance.
To advocate for its outright removal is to be agnostic to societal trust.
First, Do No Harm is a humbling reminder of our flawed past, present, and future. Even the best-laid structures cannot guarantee against medical, ethical, or policy errors. This is why openness to error correction is so important. The pithy phrase will persist as long as we err, or it is replaced by something else that reinforces trust. For now, it’s what sustains our professional contract with society, one correction after another. One Belmont Report at a time.
A final point – trust is a party of two. Embracing fallibility means moving forward with a mutual expectation for error correction. Conflating primum non nocere with non-maleficence is a mistake because it creates an unrealistic expectation of error prevention. The promise of fallibility is in error correction, not in error avoidance. Fallibility does this by elevating every cornerstone of ethics, raising the bar of self-appraisal and introspection across the board. Looking for error helps us extend the invitation to be more open, but we shouldn’t be surprised if it is met with fear and mistrust. Fortunately, just as ethics is not a static endeavor, we are not a static society.
How can we make error correction an active and iterative pursuit? How can we make honest disclosure and transparent dialogue the norm? This is the only way medicine moves forward.
Cory J Rohlfsen is a hybrid internist, core faculty member at UNMC, and the inaugural director of Health Educators and Academic Leaders, which focuses on competency-based approaches to developing future leaders, scholars, and change agents in health professions education.
I agree with Cory — and would simplify his carefully crafted erudite explanation. I learned, "First, Do No Harm" over half a century ago — it simply reminds me (and those who I taught over the years) that virtually all medical treatments have potential positive and potential negative effects — and enthusiasm for pursuing a potential positive outcome without consideration (and ideally full informed consent) of the likelihood of an adverse result from such treatment ( = "harm") — is not fulfilling our pledge to our patient.
As person-as-patient my foundational principle is that I alone should decide what is the best option for me, making the inevitable tradeoffs myself, having been fully informed about them by my physician in a non-directive way that does not pre-empt/undermine my preferences under the guise of respecting some nebulous professional oath/code (or any other reason). I truly enjoy and benefit from Sensible (i.e. Bayesian) Medicine, but I can't recall the word 'preferences' ever being spoken, even when you (Adam, John and earlier Vinay) are discussing the tradeoff between the consequences of screening/diagnostic test error rates. Friendly enquiry: is this word in your (medical) vocabulary, and, if so, why does it not come naturally and routinely into your conversations and writings? Correction welcomed, of course.