Conditions which can result in error when using a clinical prediction rule:
(1) use of a poorly designed prediction rule with methodological problems
(2) use of a prediction rule in a situation for which it was not designed
(3) use of a prediction rule in a group which does not meet inclusion criteria
(4) use of a prediction rule in a group for which it has not been validated
(5) use of a prediction rule in a situation where the pre-test probability is significantly different from its development group
(6) failing to execute the rule properly, including errors in data or assumptions
(7) misuse of the rule's output
(8) imprecision in estimates
Additional problems that can arise:
(1) becoming distracted while executing the rule, rather than focusing on the patient or problem at hand
(2) over-reliance on the rule, rather than using clinical judgment (being blinded)