Recall Bias in Retrospective Assessment of Preoperative American Shoulder and Elbow Surgeons Scores After Reverse Total Shoulder Arthroplasty

These authors sought to determine whether patients who undergo RTSA are susceptible to recall bias and, if so, which factors are associated with poor recollection.

72 patients completed the American Shoulder and Elbow Surgeons (ASES scores) Standardized Assessment Form at the time of preoperative assessment. Patients were contacted at a minimum of 24 months after surgery to retrospectively assess their preoperative condition.

Patient assessment of their preoperative shoulder condition showed poor reliability (intraclass correlation coefficient = 0.453, confidence interval, 0.237-0.623).

Greater preoperative shoulder ASES scores were associated with a greater difference between preoperative ASES scores and recall ASES scores.

The greatest part of the difference was due to a greater amount of preoperative pain recalled two years after surgery in comparison to the amount of preoperative pain reported before surgery. In other words patients retrospecitively recalled an exaggerated amount of preoperative pain.

Comment: Patient satisfaction after surgery is based largely on the perceived improvement after the procedure. This study indicates that the perceived improvement may be greater than the actual improvement, especially for patients with less actual pain prior to surgery.

This finding also confounds to a degree the "anchor method" for determining the minimal clinically important difference (MCID) for a procedure (see this link). The anchor method asks each patient to rate his or her shoulder as “worse,” “unchanged,” “better,” or “much better” relative to the recalled preoperative condition. The MCID is the minimal difference in the preoperative-to-postoperative outcome measure between patients describing the treatment outcome as “worse” or “unchanged” compared to those describing the treatment outcome as “better.” This study suggests that in this calculation, more patients may be placed in the "better" group than would be justified by their actual postop vs preop scores.

In conclusion, we need to recognize the existence and effects of recall bias in which patients may perceive more improvement than what actually occurred.

You can support cutting edge shoulder research that is leading to better care for patients with shoulder problems, click on this link.