Yellow Card data: what can the reports really tell us?

A deeper guide to spontaneous safety reporting, signal detection and the difference between a reported event and a proven adverse reaction.

Reviewed: 22 August 2026

Bottom line: Yellow Card reports are valuable for detecting possible safety signals, but a report does not prove that a medicine or vaccine caused the event. The system intentionally accepts suspected reactions because early signal detection depends on reporting uncertainty, not waiting for causation to be established.

1. What the Yellow Card scheme is for

The MHRA Yellow Card scheme collects suspected adverse reactions to medicines and vaccines. Reports can come from patients, carers and healthcare professionals. The system is designed to identify unusual patterns that may warrant further investigation.

2. Why report counts are not incidence rates

A raw number of reports does not tell you how often an event occurs among exposed people. You also need the number of doses or people exposed, the normal background frequency of the event, the age and health profile of those affected, and whether reporting behaviour changed after publicity.

3. Temporal association versus causation

If an illness happens after vaccination, it is temporally associated with vaccination. That is not the same as demonstrating a causal relationship. Some illnesses would have occurred in the same time window even if no vaccination had taken place.

4. How a signal is investigated

Investigators may compare observed versus expected event rates, review clinical details, examine healthcare-record datasets, conduct epidemiological studies, look for biological plausibility and compare findings with other countries' safety systems.

5. Under-reporting and stimulated reporting

Spontaneous-report systems have limitations in both directions. Some genuine reactions may never be reported, while publicity about a suspected issue can sharply increase reporting. That is another reason why raw report totals cannot be treated as a direct measure of risk.

6. The correct way to use the data

Yellow Card data should neither be dismissed nor treated as proof. Their strength is signal generation. Once a signal appears, stronger analytical methods are needed to estimate risk and test causation.

Practical rule: when a social-media post says “X Yellow Card reports prove Y deaths/injuries,” ask whether the MHRA has established causation, what the exposure denominator is, what the background rate is, and whether an analytical study supports the claim.

Source trail

MHRA — Yellow Card scheme
The official UK reporting system for suspected adverse reactions.

MHRA — Interactive Drug Analysis Profiles
Report listings and MHRA cautions about interpreting spontaneous reports.

Open the Yellow Card Source Pack →