How Hematologists Think on the Consult Service

Reasoning, communication, and recalibration in inpatient hematology

Most inpatient hematology consults are not about answering the question you are asked.

They are about figuring out what the question really is, how urgent it is, and what decisions depend on your judgment right now. A page that says “platelets are falling” may ultimately lead to a diagnosis, but the consultant’s first responsibility is rarely to name a cause. It is to assess risk, define scope, and decide how fast to move.

This is the essence of consult culture.

Consult medicine lives in a space that is poorly captured by textbooks and point-of-care references. At one extreme are encyclopedic resources that list causes and algorithms. At the other are abstract discussions of clinical reasoning. Neither reflects what it feels like to be paged about a real patient, mid-service, with incomplete information and real consequences attached to delay or overreaction.

Experienced consultants do not reason in a single step. They reason in phases, each shaped by what is known at the time and by what decisions must be made next. Early thinking is fast, contextual, and deliberately provisional. Later thinking is slower, more discriminating, and willing to revise earlier judgments as new information arrives.

Crucially, recalibration does not take a single form. Across different consults, expert judgment shifts in different ways. Sometimes probability is nudged incrementally as new data accumulate. Sometimes a threshold is crossed that demands immediate action before confirmation. Sometimes the real work is choosing between two non-symmetric harms, and making explicit which risk you are accepting right now. Sometimes the decisive move is restraint, recognizing that a striking number is prompting the wrong mental model. And sometimes a treatment strategy stops working not because it was wrong, but because the physiology has evolved beyond the mechanism it relies on.

These are not differences of knowledge. They are differences of what kind of thinking the moment demands. A consult can be difficult because the cost of delay is irreversible, because harms compete, because the signal is loud but misleading, or because a reasonable plan has become insufficient. Naming the pattern does not replace judgment, but it helps clinicians recognize what they are actually being asked to do.

Importantly, this phased approach is not about learning theory or epistemology. It is about workflow. What can reasonably be inferred at the time of the page is different from what can be concluded after a peripheral smear is reviewed, a HIT antibody returns, or an ADAMTS-13 level comes back. Good consult practice respects those temporal realities rather than pretending that all information arrives at once.

This framework also reflects a practical truth of consult medicine: communication happens in more than one direction. Consultants reason internally, often testing ideas within their own team before speaking externally. Recommendations to primary teams are shaped not only by what is most likely, but by what is safest, most actionable, and most appropriate to say given the degree of certainty at that moment.

What follows in this series is not a checklist or an algorithm. It is an exemplar: a walk through real consult scenarios, showing how expert hematologists organize their thinking over time. The phases are named explicitly, not to formalize what is usually implicit, but to give language to patterns of reasoning clinicians already recognize in themselves.

The goal is not to teach you what to think, but to make visible how hematologists think when the pager goes off.