The test that sounds like the worst possible idea
There is one way to discover whether somebody reacts when they eat mustard.
Give them mustard.
That sentence sounds almost comically reckless in an article about avoiding allergic reactions.
Yet a properly conducted oral food challenge is the reference-standard diagnostic test for food allergy.
The crucial words are properly conducted.
This isn't:
"Go home, eat a spoonful and see what happens."
A clinical oral food challenge is a controlled medical procedure.
Typically, the suspected food is given in carefully measured, increasing amounts while the patient is monitored for symptoms. The clinical team has criteria for stopping the challenge and is equipped to treat an allergic reaction if one occurs.
There are different challenge designs. Some are open, meaning everyone knows what is being eaten. In more complicated diagnostic situations, blinded or placebo-controlled challenges can help separate reproducible allergic symptoms from symptoms caused by expectation or other factors.
The challenge answers a question the blood and skin tests sometimes cannot:
What actually happens when this person eats the food?
Why not challenge everybody?
Because an oral food challenge takes time, specialist resources and carries the deliberate possibility of provoking an allergic reaction.
If somebody has a compelling history of an immediate reaction to mustard together with strongly supportive allergy testing, deliberately feeding them mustard may add little useful information.
At the other extreme, if the history and testing make allergy extremely unlikely, a specialist may reach a conclusion without an elaborate challenge.
Food challenges are particularly valuable in the uncertain middle ground.
Perhaps the history is ambiguous.
Perhaps the test is positive but the person has never knowingly reacted.
Perhaps an allergy diagnosed in childhood may have resolved.
Perhaps avoiding the food is causing substantial dietary or social restriction and there is a genuine question over whether avoidance remains necessary.
The challenge is not performed because doctors enjoy taking risks.
It is performed because an incorrect allergy diagnosis has risks too.
Being wrongly diagnosed with an allergy isn't harmless
When we talk about diagnostic mistakes, we naturally worry about a missed allergy.
And rightly so.
But falsely diagnosing somebody with food allergy can also have significant consequences.
They may avoid foods unnecessarily.
Parents may restrict a child's diet.
Eating out becomes more complicated.
Travel becomes more complicated.
Shopping takes longer.
Food anxiety can increase.
Nutritional choices may narrow.
Families may spend more money on specialist products.
A child can begin to think of ordinary food as inherently dangerous.
That is why the goal of allergy diagnosis isn't to identify the maximum possible number of things somebody might react to.
It is to identify, as accurately as possible, what they genuinely need to avoid — and what they don't.
A good allergy diagnosis can sometimes add foods back into somebody's life.
That is every bit as valuable as removing a dangerous one.
The seductive world of "food intolerance testing"
This brings us to a very different marketplace.
Search online for food symptoms and it is easy to find commercial tests offering to analyse dozens or even hundreds of foods from a blood sample, hair sample or other specimen.
Some produce impressive-looking reports containing long lists of foods graded by colour.
Red: avoid.
Amber: caution.
Green: fine.
The presentation can look scientific.
But the important question isn't whether a laboratory can measure something.
It is whether what it measures has been shown to diagnose the condition it claims to diagnose.
Commercial food-specific IgG testing, for example, should not be confused with the specific-IgE testing used in conventional investigation of immediate food allergy.
IgG antibodies to foods can be associated with exposure and tolerance, rather than proving that the food is causing illness.
UK allergy organisations warn against using IgG food panels, hair analysis, Vega/electrodermal testing, kinesiology and several other unvalidated methods to diagnose food allergy or intolerance.
This isn't an argument that unexplained food-related symptoms are imaginary.
People can have very real symptoms that are difficult to diagnose.
It is an argument that a real symptom does not make an unreliable test reliable.
A beautifully printed answer can still be wrong
This deserves emphasis because it extends far beyond allergy.
People understandably value certainty.
A report saying:
Mustard: 87 — HIGH REACTIVITY
feels more satisfying than a specialist saying:
"Your history suggests several possibilities. We need to investigate further."
One looks precise.
The other sounds uncertain.
But precision and accuracy are different things.
A bathroom scale that always adds exactly 7.436 kilograms to your weight is extremely precise.
It is also wrong.
Medicine sometimes has to live with uncertainty because biology is complicated.
A responsible clinician would rather give somebody a qualified answer supported by evidence than a beautifully formatted number whose clinical meaning has never been established.
That principle will appear repeatedly throughout food allergy.
The most confident answer is not necessarily the best-supported answer.