The FDA Division of Drug Experience in the early 1980's used an algorithm to evaluate the causal relationship between exposure to a drug and an adverse clinical event.
Questions:
(1) Did the adverse event have a reasonable temporal association (sequence) with use of the drug?
(2) Was a dechallenge (discontinuation) from the drug tried, and if so, what was the change in the adverse reaction?
(3) Was a rechallenge to the drug attempted, and if so, did the adverse reaction reappear?
(4) Could the adverse reaction be reasonably explained by an existing clinical condition?
A causal relationship was considered remote if:
(1) The adverse event did not show a reasonable temporal association to drug exposure.
A causal relationship was considered possible if:
(1) There was a reasonable temporal association to the drug exposure but no dechallenge from the drug was attempted.
(2) There was a reasonable temporal association to the drug exposure. When a dechallenge from the drug was attempted the observed event did not abate.
(3) There was a reasonable temporal association to the drug exposure. When a dechallenge from the drug was attempted the observed event abated. No rechallenge was attempted. The adverse reaction could be due to an existing clinical condition.
(4) There was a reasonable temporal association to the drug exposure. When a dechallenge from the drug was attempted the observed event abated. Drug rechallenge was attempted but the reaction did not reappear.
A causal relationship was considered probable if:
(1) There was a reasonable temporal association to the drug exposure. When a dechallenge from the drug was attempted the observed event abated. No rechallenge was attempted. The adverse reaction could not be explained by an existing clinical condition.
A causal relationship was considered highly probable if:
(1) There was a reasonable temporal association to the drug exposure. When a dechallenge from the drug was attempted the observed event abated. Drug rechallenge was attempted and the adverse reaction recurred.
