Researchers may not have measured meaningful outcomes in studies that compare treatments, or the ones they have measured may not matter to your service user.
For example, studies often use outcomes that are easy to measure, such as whether or not a patient coughed while eating, as a substitute for outcomes that may be most important, such as pleasure gained from eating.
However, the effects of treatments on those substitute (surrogate) outcomes (coughing when eating or drinking) often do not provide a reliable indication of the effects on outcomes that are important (like pleasure from eating). Similarly, short-term effects may not provide a reliable indication of long-term effects.
REMEMBER: Ask whether the outcomes that are important to your service user have been measured in fair comparisons of treatments, and look out for surrogate outcomes.