Why Clinicians Need a New Way to Study in the Age of AI. Introducing ReviewBytes
Read more
New attending readiness goes beyond knowledge. Learn to own decisions, supervise teams, manage uncertainty, and calibrate when to act, consult, or delegate.
.webp)
New attending readiness means owning theclinical plan, the team’s execution, and the response to uncertainty—not makingevery decision alone.
Residency and fellowship already demandresponsibility. But becoming an attending changes the nature of thatresponsibility. You are no longer simply expected to make a goodrecommendation, present it clearly, and discuss it with someone more senior.Increasingly, you are the person who must say, “This is our plan,” and thenmake sure that the plan is carried out safely.
This transition can feel surprisinglydifferent from training. The medicine may be familiar. What is new is theownership of the decision, the supervision of the team, the institutionalrealities around the decision, and the consequences if things do not go asplanned.
Research on the transition to independentpractice suggests that these difficulties extend well beyond medical knowledge.New attendings report gaps involving procedural independence, practicemanagement, workload, decision-making, supervision, confidence, andprofessional development. (PMID: 39316458)
Becoming ready for independent practicerequires more than clinical competence. A new attending must learn to:
The central question is no longer simply,“Was my decision reasonable?”
It is also, “Can this team carry it outsafely, and who will recognize when the plan needs to change?”
That is the important shift from trainee toattending. You may know the medicine and still feel the weight of saying, “This is what we are going to do.” That feeling should not necessarily be viewed as aweakness. In many cases, it reflects an appropriate understanding of what thedecision now carries with it.
One useful way to think about readiness isthrough the ReviewBytes model: competence, context, and calibration.
Competence refers to what you can reliably assess, explain, and perform.
This includes the traditional things weusually associate with medical training:
Board prep that supports clinical practice, including ABIM exam prep, in-training exams, question banks, and other forms of medical upskilling, can help test and reinforce the competence component of readiness.
But competence is only one part of the picture.
Context refers to the environment in which your decision must actually work.
A medically correct plan is not necessarilya workable plan if the institution cannot safely support it.
You may know that a patient needs frequentreassessment, for example, but does your unit have the nursing coverage toprovide it? You may know which procedure would be ideal, but is that procedureavailable overnight? If the patient deteriorates, how quickly can yourinstitution provide ICU-level care, subspecialty support, or transfer?
Early-career physicians often discover thatmuch of independent practice depends on precisely this kind of unwritteninstitutional knowledge. Interviews with newly graduated emergency physicians,for example, identified important knowledge involving patient interaction,learner supervision, institutional processes, and group culture. (PMID:40833552)
Much of this knowledge is not found intextbooks.
Finally, there is calibration: howwell your confidence matches the evidence, the clinical situation, and thepotential consequences of being wrong.
This is especially important becauseconfidence and diagnostic accuracy do not always move together.
In a vignette study of 118 internists,physicians’ confidence did not fall nearly as much as diagnostic accuracy whencases became more difficult. These were clinical vignettes rather than actualbedside encounters, so the findings should not be interpreted as real-worlderror rates. Nevertheless, they illustrate an important point: feelingconfident and being correct are not the same thing. (PMID: 23979070)
Understanding clinical confidence and calibration can help clinicians distinguish feeling certain from having sufficient evidence to justify that certainty.
The attending therefore has to ask two questions:
That second question becomes increasinglyimportant when you are the final decision-maker.
Clinical decisions rarely end when youchoose a diagnosis or treatment.
The decision must move through severalsteps before it becomes safe patient care.
Start by asking:
The goal is not to generate an endlessdifferential. It is to understand what uncertainty remains and whether thatuncertainty matters.
Medicine frequently involves choosingbetween acting now and obtaining more information.
The classic threshold approach to clinicaldecision-making describes this problem in terms of testing and treatmentthresholds. The probability of disease, potential benefits of treatment, harmsof testing, harms of treatment, and consequences of delay all influence whetherthe best course is to observe, test, or treat. (PMID: 7366635)
The same general reasoning can be usefulwhen thinking about consultation and escalation.
Ask:
This is the contextual step that is easy tomiss.
“Observe overnight” sounds like a plan. Butobservation assumes that someone can observe the patient, recognizedeterioration, and escalate care.
“Follow closely as an outpatient” assumesthat the patient can obtain follow-up.
“Call me if things get worse” assumes thatthe patient understands what “worse” means.
Before finalizing a plan, ask whether thepeople, monitoring, equipment, and backup needed to execute that plan actuallyexist.
One of the biggest differences betweentrainee thinking and attending thinking is that the plan now includesownership.
Someone must know:
This sounds obvious until no one is surewho was supposed to review the potassium, call the family, follow the bloodculture, or reassess the patient after an intervention.
Every important plan should have areassessment point.
Sometimes that reassessment is yours.Sometimes it belongs to the next attending, resident, fellow, advanced practiceclinician, nurse, or outpatient clinician.
The important point is that theresponsibility should be explicit.
Structured handoff interventions illustratewhy this matters. In a multicenter study across nine hospitals, implementationof the I-PASS handoff program was associated with reductions in medical errorsand preventable adverse events. (PMID: 25372088)
The broader lesson is simple: gooddecisions can still fail when responsibility for what happens next is unclear.
The literature on new attending readinessis still developing, but several themes appear repeatedly.
A 2025 scoping review examined thetransition from graduate medical education to independent practice. Across 116articles, investigators identified 192 individual preparedness gaps.
The most frequently identified areasincluded procedural independence, practice management, and billing, whilebroader gaps involved patient care and personal and professional development.(PMID: 39316458)
In other words, residency and fellowshipgenerally spend enormous amounts of time teaching clinicians how to makemedical decisions. Independent practice suddenly asks them to do considerablymore:
This helps explain why a clinically stronggraduating resident or fellow may still feel uncomfortable during the firstmonths of attending practice. A structured approach to transition readiness after fellowship can help turn this adjustment period into specific 30-, 60-, and 90-day goals.
Supervision is one of the more complicatedparts of the transition because new attendings suddenly become responsible forcreating the learning environment that they recently occupied themselves.
In a randomized crossover trial involving22 attending physicians and 1,259 patients on an academic internal medicineservice, increased attending participation on resident work rounds did notsignificantly reduce medical error rates. At the same time, residents reportedless autonomy, and interns spoke less when attendings joined rounds. (PMID:29868877)
This does not mean that supervision isunnecessary. Nor does one single-center trial establish equivalence betweendifferent supervision models.
Rather, it illustrates the balance thatattendings must learn: provide enough supervision to keep patients safe andsupport the learner, but not so much that the learner stops thinking, speaking,or functioning independently.
A new attending may make hundreds ofdecisions in a day.
Many are small:
Others are major.
Over time, the cognitive load can become substantial. Understanding cognitive load on busy shifts can help explain why processing complex information becomes harder as competing demands accumulate.
A 2025 systematic review of 82 publicationsexamining decision fatigue in healthcare professionals found evidence ofdecision-fatigue effects in some settings, but the overall literature wasinconsistent and the concept was often poorly defined or measured. (PMID:40591577)
So it would be an overstatement to say thatevery clinician inevitably starts making worse decisions as the day progresses.
Still, the underlying experience isfamiliar: repeated decisions consume attention, and high workloads can makethoughtful deliberation harder.
For new attendings, there is an additionallayer. Many of those decisions now feel final.
I use the term final-decision fatigueto describe the strain of repeatedly resolving uncertainty while remainingresponsible for what happens next.
This is a descriptive concept rather than aformal diagnosis or established physiological mechanism.
The important distinction is that theattending is often not simply deciding what they think should happen. They aredeciding what will happen.
A fellow may say:
“I think we should discharge.”
The attending may have to say:
“We are discharging.”
Those sentences look almost identical, butpsychologically they can feel quite different.
The second statement contains ownership.
Over the course of a long day,final-decision fatigue may show up as a temptation to:
None of these behaviors automatically meansthat a clinician is practicing poorly.
But they are useful prompts for reflection.
One question I find helpful is:
“Am I choosing this because it is thebest option, or because I need this decision to be over?”
That is not a validated fatigue test. It issimply a useful pause.
One of the most useful forms of calibrationis recognizing that these three decisions have different thresholds.
Ask:
For a deteriorating patient, consultationand treatment may need to happen simultaneously. You should not delay an urgentintervention merely because you are waiting for another physician’s opinion.
Consultation should not be treated asevidence that you are not ready to be an attending.
The real question is:
Could another clinician’s expertisematerially improve this decision?
Consultation is particularly reasonablewhen:
Good consultation also requires a clearquestion.
“Can you see the patient?” is often muchless useful than:
“I am concerned about X. We have done Y andZ. My question is whether you think we should proceed with A or continueobservation.”
Delegation also needs calibration.
Do not delegate simply because somebodyholds a particular training title.
Ask whether this particular person hasdemonstrated the ability to perform this particular task, in this particularcontext, with the appropriate backup available.
That may mean allowing a senior residentsignificant independence in one situation while supervising closely in another.
Good supervision is dynamic.
Many new attendings struggle withsupervision because they remember how much autonomy mattered during residencyand fellowship, but they now also understand how much responsibility theattending carries.
A practical opening question is:
“What is your plan, what worries you, andwhat would make you call me?”
That short conversation tells you severalthings.
You learn:
Supervision should then change with thesituation.
The goal is neither absence nor control.
It is graduated responsibility.
Another major change for new attendings isdocumentation.
The medical record serves many purposes. Itcommunicates with clinicians, records clinical reasoning, supports billing, andmay later be reviewed by patients, payers, administrators, or attorneys.
Electronic records have made documentationeasier in some ways and much more complicated in others.
A systematic review of copy-and-pastepractices found that copied material may contribute to internalinconsistencies, outdated information, error propagation, and note bloat,although direct evidence linking copy-and-paste to patient harm remains limited.(PMID: 28074211)
The goal is therefore not to write thelongest note.
It is to write an accurate one.
For consequential decisions, considerwhether the note clearly answers four questions:
If patient preferences affected thedecision, document that as well.
And, of course, never document anexamination, discussion, or clinical fact that did not actually occur.
Patients generally do not need a lecture onevery theoretical possibility.
They need to understand what is happening,what the reasonable options are, what you recommend, and where uncertaintyremains.
A useful way to frame this is:
“I recommend this approach because of whatwe know so far. There is still some uncertainty, so let me explain thealternatives and what would make us change course.”
That language does two thingssimultaneously.
It provides leadership while acknowledginguncertainty.
Shared decision-making does not mean thatthe clinician simply lists options and asks the patient to choose. Theclinician still provides expertise and a recommendation, while helping thepatient understand the trade-offs and express what matters most to them.
A 2024 Cochrane review of 209 randomizedtrials found that patient decision aids improved patient knowledge, riskperception, and participation in decision-making. (PMID: 38284415)
The larger point is that ownership of adecision does not eliminate patient autonomy.
It should make the discussion clearer.
One of the easiest mistakes for a newattending is assuming that the same medical plan works the same way everywhere.
It does not.
A plan that is perfectly reasonable at alarge tertiary academic center may be unsafe in a hospital without the same:
This is why new-attending on boardingshould include much more than access badges, passwords, and compliance modules.
A useful institutional map should include:
Transition research suggests that theseinstitutional and cultural processes are among the important forms of tacitknowledge early-career clinicians must acquire. (PMIDs: 39316458, 40833552)
Learning the local culture, however, doesnot mean blindly adopting every local habit.
Some habits exist because “that is how wehave always done it.”
Those still deserve scrutiny.
One of the most difficult parts of becomingan attending is accepting that sometimes a thoughtful, defensible decision willstill lead to a poor outcome.
Sometimes a complication occurs despiteappropriate care.
Sometimes the diagnosis becomes clear onlylater.
Sometimes the system fails.
And sometimes a genuine medical erroroccurs.
Those situations should not all be treatedas the same thing.
The first priority after an adverse eventis the patient.
If an error has occurred, clinicians shouldwork within institutional processes to address the patient’s immediate needs,communicate what is known, report the event appropriately, and participate inefforts to understand what happened.
Communication around medical errors isitself a clinical skill. Literature in this area emphasizes transparentcommunication, acknowledgment of harm, explanation of next steps, andorganizational support for both patients and clinicians. (PMID: 33280965)
New attendings should also know where toobtain help.
That may include:
Being the attending does not meannavigating every adverse event alone.
How to interpret this table: No single method prepares someone for every aspect of independentpractice. The intervention needs to match the actual readiness gap.
One of the more important ideas in medicaleducation is that graduation should not be treated as the moment learning stops.
The transition from residency or fellowshipinto attending practice is simply another stage of training, except that the learning is now much more self-directed.
That makes upskilling especiallyimportant.
The key is to match the learning method tothe type of problem.
Use:
These tools are particularly useful for identifying and repairing knowledge gaps. One practical approach is treating missed questions as learning data, separating knowledge gaps from reasoning errors, misreads, guideline confusion, and confidence mismatches.
Review how your institution actuallyhandles:
No amount of board prep can teach you how your hospital functions at 2 AM.
Review difficult decisions with colleagues.
Ask:
This is different from simply asking, “WasI right?”
Sometimes the diagnosis was ultimatelywrong, but the reasoning was appropriate given the information available.
Sometimes the diagnosis was ultimatelycorrect, but the reasoning was poor.
Those are very different learning problems.
Perhaps the biggest adjustment in becomingan attending is realizing that uncertainty does not disappear when trainingends.
What changes is your relationship with it.
As a trainee, uncertainty often prompts thequestion, “Who should I ask?”
As an attending, the better questionbecomes, “What does this uncertainty require me to do?”
Sometimes the answer is to act.
Sometimes it is to wait.
Sometimes it is to ask for help.
Sometimes it is to give someone else roomto lead while you remain responsible for the safety of the process.
That is what becoming the final decision-maker really means.
Readiness means more than knowing themedicine. A new attending has to combine clinical competence with anunderstanding of the local practice environment, appropriate help-seeking,effective delegation, clear communication, and reliable follow-up. Confidencealone is not enough. (PMIDs: 39316458, 23979070)
No. Independent practice does not meanpracticing in isolation. Another clinician should be involved when theirexpertise could materially improve an unfamiliar, uncertain, orhigh-consequence decision. The important thing is to ask a clear question and communicatethe urgency. (PMID: 39316458)
In this article, final-decision fatiguerefers to the strain of repeatedly resolving clinical uncertainty while alsocarrying responsibility for what happens next. It is a descriptive term ratherthan a formal diagnosis. The broader scientific literature on decision fatiguein healthcare remains mixed. (PMID: 40591577)
Direct involvement should generallyincrease when the patient is unstable, the task is unfamiliar or high risk, theclinician performing it has not demonstrated sufficient capability, or theconsequences of an error are substantial. Local supervision requirements shouldalways be followed. (PMID: 29868877)
The note should accurately describe theassessment, remaining uncertainty, reasoning behind the current plan, relevantpatient preferences, and the plan for reassessment or follow-up. Examinations,conversations, or findings that did not occur should never be documented.(PMID: 28074211)
No. Board prep and exam prep are useful forstrengthening medical knowledge, but independent practice also requiressupervision skills, communication, leadership, institutional knowledge,workflow management, and effective follow-up. (PMID: 39316458)
The name ReviewBytes reflects ourbelief that medical learning should be clear, focused, and built for the modernlearner. Review speaks to scientifically grounded learning methods thatimprove retention and recall, while Bytes reflects both bite-sizedlearning and a technology-forward educational experience.
We chose ReviewBytes because itcaptures the way we think learning should work: evidence-based, efficient, andthoughtfully designed. The name brings together proven review methods withmicrolearning and AI-powered innovation.
Yes. Many learners search for ReviewBytes as a variation of ReviewBytes, and both refer to the samebrand and mission.
Absolutely. The “Bytes” in ReviewBytesis a nod to bite-sized learning—breaking complex medical concepts into smaller,easier-to-review pieces—while also reflecting our tech-forward approach.
“Bytes” reflects two ideas:bite-sized learning and a modern, technology-forward approach to education. Itcaptures both accessibility and innovation.
Yes. Some people hear or search for ReviewBytesas “review bites.” While the spelling is different, the meaning alignsclosely with our mission of smarter, more focused medical learning.
⚠️ Educational disclaimer: This article is for educational purposes only and is notpersonalized medical advice or a substitute for local institutional policies,supervision requirements, or professional regulations. Individual clinicalconcerns should be addressed with an appropriate healthcare professional.
.webp)
