Risk calculators are used routinely to decide who is offered preventive treatment. Understanding how they are built explains what family history contributes and why the output is not a prediction.
The output is a population frequency
A calculator is built by following a large group over years, recording characteristics at the start, and modelling which of them predicted later events.
The resulting equation converts an individual's measurements into the frequency of events observed among people in that study who shared those characteristics.
It therefore describes a group, not a person. A stated risk of one in ten means a tenth of similar people experienced the event, not that a tenth of this person will.
Family history stands in for what was not measured
Relatives share genetic variants, and they also share diet, activity, smoking patterns, income and environment, all of which affect risk.
Family history captures both at once without distinguishing them, which is why it remains useful even where measured risk factors are already in the model.
Its inclusion adds information precisely because it summarises causes the calculator has no direct measurement of.
Definitions are narrow for a reason
Most calculators count only first-degree relatives, and only events occurring before a specified age, rather than any relative with the condition.
Restricting to early events is what separates an inherited tendency from disease that would be expected in older age regardless of family.
Broader definitions add noise, since a condition common in the general population appears in most families without indicating anything about a particular one.
Transportability is the main limitation
An equation derived in one population may perform poorly in another with different disease rates, diets and ethnic composition.
Calculators are therefore recalibrated for local use, and applying one outside its intended population can misstate risk in either direction.
This is why different countries use different tools for the same clinical question, and why the same person can receive different figures from each.
Thresholds are policy, not biology
Where treatment is offered above a stated risk level, that level was chosen by weighing benefit, harm and cost across a population.
Nothing changes biologically at the boundary, and someone just below it differs negligibly from someone just above.
Which is why guidance treats the score as one input to a conversation with a clinician rather than as a decision in itself.