Can the patient's memory of the timing of pain events replace chart notes?

The reliability of the patient's memory as to when pain remitted and relapsed was studied ten years after knee replacement. Pain was chosen because it is the most frequent spontaneous patient complaint, and the remission and relapse points in time are clinically important. The standard for the comparison was information extracted from the medical records. The patient's information on the time of remission of pain could not be trusted. For relapse there was a fair but insufficient association (Spearman's correlation coefficient + 0.51).