Bradford Hill listed a number of criteria that should be considered when determining whether or not an environmental exposure causes a condition, especially when faced with observational data. The author was from the University of London.
Goal: to accept a cause-and-effect hypothesis
Criteria
Description
strength of the association
frequency in exposed vs frequency in non-exposed
consistency
repeatedly observed in different situations
specificity
limited to at-risk individuals
temporality
exposure precedes onset
biological gradient
dose-response curve
plausibility
based on current knowledge
coherence
agrees (not in conflict) with generally known facts
experiment
change in exposure changes outcome; removal of exposure associated with drop in incidence
analogy
other examples of similar associations exist
None of these criteria is indisputable evidence or indispensable (sine qua non).
They serve as a starting point for discussion and further investigation.
To read more or access our algorithms and calculators, please log in or register.