From Repair to Preemptive Medicine

Personalization changes the timing of care as well as its precision.

Repair medicine waits for a threshold to be crossed and then intervenes. Preemptive medicine acts on a trajectory before the threshold exists, which changes three things at once. The object of intervention becomes a predicted state. The evidence base becomes a forecast rather than a finding. And the patient becomes a person who feels well, so the burden of proof for intervening in health is properly higher than for intervening in illness, and preemptive medicine will be judged on whether it respects that asymmetry.

An asymmetric burden of proof protects nobody unless something is actually required of it, and three requirements follow. Endpoints must be outcomes, not surrogates, since moving a marker in a well person establishes only that the marker moved. Net benefit must be demonstrated with the harms counted, and the harms of intervening in health include the label itself, the cascade of confirmatory testing it triggers, the treatment burden imposed on someone with no symptoms to relieve, and the years of anxiety attached to a prediction.

And any preemptive program should state its expected rate of overdiagnosis, meaning the proportion of those treated who would never have developed the disease, instead of reporting only the cases it caught. The history of screening is unambiguous on this point. Programs never asked these questions have persisted for decades on the strength of their detection rates alone.

There is a fourth consequence, and it centers on the patient rather than on the evidence. Telling a well person that a disease is coming imposes a burden the disease has not yet imposed, and how people process predictive information about their own future is poorly understood and highly variable. Some are mobilized by it, some immobilized, and some reorganize their lives around a risk that never materializes.

Preemptive medicine therefore generates a demand for a competence the profession has barely recognized. Existential counselling, the work of accompanying a person through the meaning of a predicted future, not a present illness, becomes a routine requirement. So does deeper preparation in communication, psychosocial care, cultural understanding, and human-centered decision-making. New specialties may emerge for adaptation to predictive health technologies and for the management of extended lifespan, since the psychological work created by preemption now falls to clinicians trained for neither.

A further change follows from the first three and concerns the intervention itself. The body runs on homeostasis and on positive and negative feedback loops, yet disease is conventionally caught only once the departure from normal has grown large, which forces corrections of matching magnitude: massive change met with massive correction. Much of today's drug doctrine is of that kind, a bull in a china shop by necessity rather than by design.

As measurement resolves smaller deviations earlier and interpretation improves, the therapeutic target shifts toward agents that correct slightly but accurately. That is a pharmacological change with a formative consequence, since judging a small correction against a trajectory is a different skill from judging a large one against a threshold, and the second is the only one medicine currently teaches.

Such care organizes itself around one object. Lifetime health trajectories are long-horizon health pathways integrating biological, behavioral, environmental, and social influences, managed as a continuous object instead of a sequence of episodes. The unit of clinical attention becomes the pathway and not the visit. The physician's contribution is to cohere the patient's care longitudinally across decades, recognizing that they will be neither the only participant nor, for most of the duration, the principal one.

Cohering care across decades collides with how care is now organized. Patients change insurers, employers, cities, and clinicians. Reimbursement is episodic and pays for the visit rather than the pathway. The average clinical relationship is a small fraction of the horizon over which a trajectory is managed. A model that depends on one physician knowing one patient for forty years is not a plan but a memory of general practice.

Continuity has to be carried by the record, built to hold a trajectory rather than a sequence of encounters: the pathway; its assumptions; the interventions attempted and their effect; and the reasoning behind each inflection, portable across institutions and legible to whoever inherits it. That is the interoperability requirement Chapter IV describes, and it separates stewardship as an aspiration from stewardship as a practice.