Marc is fifty-two. For several years, his care has followed a logic that, examined step by step, is difficult to dispute.

A problem was identified. A strategy was put in place. The recommendations were followed. The parameters considered clinically important were monitored regularly. Whenever necessary, the management plan was adjusted.

Marc, for his part, did what was asked of him.

He takes his medication. He attends his appointments. He has changed several aspects of his lifestyle. He monitors certain parameters. He pays closer attention to his diet and physical activity.

And several of his results are indeed reassuring.

Yet when he is asked how he is actually doing, his answer no longer matches the story told by his medical record.

“On paper, I’m doing better. But I feel worse.”

He recovers more slowly. His sleep is less restorative. Efforts he once absorbed without difficulty now take longer to recover from. His energy has become less predictable. Several symptoms persist even though the parameters around which the original strategy was built have improved.

The first temptation is to assume that these symptoms belong to a different problem.

That is possible.

But before adding another explanation, a different question deserves to be asked:

What if the strategy is still logical, but no longer answers the clinical question the patient is presenting today?

A medical strategy is not constructed arbitrarily.

It is based on the clinical situation, biological data, risk factors, guidelines, therapeutic evidence and an assessment of the balance between expected benefits and potential risks.

When a treatment improves the parameter it was designed to correct, there is therefore a legitimate reason to conclude that the strategy is working.

But the word “working” already contains an ambiguity.

Working towards what?

Improving a biomarker? Reducing future risk? Relieving a symptom? Preventing a complication? Restoring a function? Improving the patient’s ability to live normally?

These objectives may overlap.

They are not necessarily interchangeable.

A biomarker can move in the desired direction while a symptom persists. A risk factor can become better controlled while tolerance of the overall strategy deteriorates. An intervention can continue to achieve its original biological target even as the clinical picture itself changes.

That does not mean the strategy was wrong.

It means the situation needs to be reassessed.

It would be a mistake to reason retrospectively and conclude that because Marc is not doing better, the decisions made previously must have been incorrect.

That would be too simplistic.

A decision may be entirely appropriate when it is made and become insufficient as the clinical situation evolves.

Medicine constantly faces this difficulty: decisions must be made using the information available at a particular moment, while the patient continues to change over time.

A therapeutic strategy should therefore be understood as a working hypothesis that remains continuously accountable to the patient’s clinical trajectory, rather than as a conclusion that becomes permanent simply because it was initially justified.

The coherence of a past decision does not automatically guarantee its relevance in the present.

This is precisely where chronology becomes indispensable.

If we look only at the latest result, the situation may appear satisfactory.

If we reconstruct several months or several years, another interpretation may emerge.

When did the symptoms begin? Were they present before the current strategy? Did they appear after a therapeutic change? Has their nature changed? Has their frequency increased? Has recovery after exertion changed? Did sleep deteriorate before or after the clinical worsening? Have several interventions been added successively to the point that their individual effects can no longer be distinguished?

None of these questions proves causality.

They reconstruct what actually happened.

And that reconstruction may reveal an essential discrepancy:

the marker around which the strategy was originally built has improved, but the patient’s overall condition has not followed the same trajectory.

Medicine needs measurable targets.

They make it possible to compare, monitor, anticipate certain risks and assess the effect of an intervention.

But no isolated marker can represent an entire biological system.

A glucose measurement does not summarise metabolism. Blood pressure does not summarise cardiovascular regulation. A hormone measurement cannot, on its own, describe an entire neuroendocrine axis. A symptom score does not necessarily capture a person’s actual functional capacity.

None of this diminishes the value of those measurements.

It defines their scope.

The difficulty begins when an indicator used to guide treatment gradually becomes an implicit substitute for the condition of the patient as a whole.

At that point, improvement in the number may close a question that the clinical picture is still asking.

A living organism does not remain stable because nothing changes.

It remains stable because it continually adjusts its functioning to internal and external demands.

Research on allostasis formalised this concept of stability achieved through adaptation: when a demand arises, several physiological systems may alter their activity in order to preserve essential functions.

This adaptation is physiological.

It becomes clinically relevant when its cost increases, when it must be repeatedly mobilised, when it persists after the original demand has disappeared, or when several systems must remain engaged over time to preserve an apparently stable state.

The concept of allostatic load was developed precisely to examine the cumulative cost of these adaptive processes.

It is not an individual diagnosis that can be inferred from a symptom or a laboratory result.

But it highlights an important distinction:

keeping a value within an acceptable range and restoring genuinely stable function are not necessarily the same thing.

Once a treatment plan has been introduced, it is no longer external to the patient.

It becomes part of everyday life.

Medication, schedules, monitoring, restrictions, appointments, prescribed exercise, dietary changes and self-measurement: each intervention may be entirely justified when considered separately.

But their accumulation is also a clinical reality.

Research on treatment burden has drawn attention to precisely this issue: in people managing complex or chronic conditions, the work required to follow medical care can compete with the physical, cognitive, social and practical resources available to them.

This obviously does not mean that useful treatments should be stopped.

It means that evaluating a strategy should not only involve asking:

“Is the patient following the treatment correctly?”

Sometimes it should also involve asking:

“What is the actual burden of this strategy, and does that burden remain proportionate to the benefit it provides?”

When a symptom persists, the intuitive response is often to intensify.

Another treatment. A different dose. Another investigation. Another recommendation. Another restriction.

That approach may be entirely appropriate.

But it is not automatically the right one.

Before adding another layer to an already complex strategy, it may be necessary to determine why the clinical objective has not been achieved.

Is the original mechanism still active? Does the initial hypothesis still explain the current symptoms? Has an intervention corrected its intended target without changing the symptom? Has another variable become more important? Is there an issue of tolerance, interaction, adherence or treatment burden?

Only after this reconstruction does intensification regain its clinical meaning.

The medical literature has long described therapeutic inertia: the failure to initiate or intensify treatment when a therapeutic target has not been reached.

But the opposite situation also deserves consideration.

A strategy can acquire momentum of its own.

It was justified yesterday, so it continues today. Every component has a historical reason for being there, yet the strategy as a whole may no longer be reconsidered in light of the patient’s current situation.

The question is not whether treatment should arbitrarily be reduced or stopped.

It is whether every component still has an indication, an identifiable objective and an expected benefit that remains relevant for this particular patient at this particular point in time.

In a situation such as Marc’s, three levels of reasoning must remain separate.

What is established: some parameters have improved; several symptoms persist; functional capacity or recovery has not necessarily followed the same trajectory.

What becomes plausible: the original target may no longer be sufficient to explain the current situation; several factors may now be contributing to the clinical picture; and the treatment strategy itself may need to be considered as one variable among others.

What remains to be demonstrated: which mechanism explains the persistent symptoms, whether any particular intervention is contributing to them, and what modification would actually produce a meaningful clinical benefit.

This distinction is essential.

Reassessing a strategy does not mean concluding that the strategy is responsible for the problem.

It means refusing to exclude it from the reasoning simply because it was originally justified.

This is probably one of the most important distinctions.

Changing a strategy does not mean that the previous physician was wrong.

It may simply mean that the context has changed, new information has emerged, the objectives have evolved, or the balance between benefits, constraints and risks is no longer the same.

Medicine that is capable of reassessing its own decisions is not uncertain medicine.

It is medicine that accepts the importance of time.

The patient in 2026 is not exactly the same patient for whom a decision may have been made several years earlier.

Biology, treatments, symptoms, constraints, sleep, activity, age, priorities and sometimes risk itself may all have changed.

The strategy must be capable of being reconsidered within that new configuration.

Clinical guidelines are indispensable.

They synthesise the best available evidence to guide decision-making and reduce unjustified variations in care.

But they do not eliminate the need for individual clinical interpretation.

As a situation becomes more complex, multiple objectives may begin to interact: prevention of long-term risk, control of an immediate symptom, treatment tolerance, sleep quality, functional capacity and the daily burden of care.

The appropriate decision is therefore not necessarily to pursue every objective to its theoretical maximum.

It is to understand their hierarchy within the patient’s actual situation.

First, it requires reconstructing the chronology before modifying the strategy.

Not only the chronology of symptoms, but also that of treatments, dose increases and reductions, dietary changes, medical events, periods of improvement and the point at which recovery began to change.

It also requires separating the objectives.

Which part of the strategy is intended to reduce future risk? Which is targeting a symptom? Which treatment has produced an objectively measurable benefit? Which intervention was added because the previous one was insufficient? Which component has never subsequently been reassessed?

Finally, it requires comparing two trajectories:

the trajectory of the measured parameters and the trajectory of the person.

When they move together, interpretation is relatively straightforward.

When they diverge, that divergence becomes clinical information in its own right.

A therapeutic strategy can be scientifically grounded, correctly implemented and still require reassessment.

There is no contradiction in that.

The original decision addressed the original situation.

The difficulty begins when its past coherence becomes sufficient reason not to question its present relevance.

The apparent failure of a strategy does not therefore necessarily mean that it should be abandoned.

Nor does it automatically mean that it should be intensified.

It means that we first need to determine what, exactly, is failing.

The marker? The symptom? Function? Tolerance? The chosen target? The original hypothesis? Or the relationship between several of these elements?

Before changing the treatment, it may sometimes be necessary to reconstruct the question the treatment was originally intended to answer.

Marc does not illustrate a case in which medicine failed because its reasoning was wrong.

He illustrates something considerably more subtle.

A strategy may remain entirely understandable when each decision is examined separately, while the strategy as a whole no longer corresponds precisely to the patient’s current clinical situation.

This is why a complex case cannot always be understood simply by asking whether each individual treatment is justified.

We must also ask what the strategy as a whole is producing, what it is no longer producing, and what has changed since it was first constructed.

The question is no longer:

“Did we follow a sound clinical logic?”

but:

“Does that logic still answer the situation in front of us today?”

The shift may appear small.

In some complex medical situations, it changes the entire interpretation of the case.