Criteria for determining if a treatment for pain works.

Claims that a treatment works are hollow unless qualified in terms of: in what respects, by how much, how often, and for how long. Essential co-requisites for improvements in pain are improvements in function, psychological distress, and use of health care. Validated instruments are available for these outcome measures. Mean scores and p-values are not informative. Categorical data are required to reveal by how much a treatment works and how often. In order to provide a full picture, outcomes need to be followed until they plateau. Readers of studies should not rely on what authors claim. Instead, readers should demand comprehensive, transparent data on outcomes so that they can decide for themselves if a treatment works to their satisfaction.