Pain scores in the management of postoperative pain in shoulder surgery.
PURPOSE: Recent public opinion has focused on the lay perception that postoperative pain control is inadequate. The Joint Commission on Accreditation of Healthcare Organizations has attempted to address this issue by mandating patient-reported numeric pain scores as part of postoperative pain assessment. No studies exist to validate these scores in orthopaedics. The purpose of this study was 3-fold: (1) to determine which preoperative variables were associated with elevated postoperative pain scores, (2) to identify whether pain scores correlated with more traditional signs and symptoms of pain, and (3) to identify whether patients with a high pain score (HPS) affect perioperative and postoperative complication rates. METHODS: We retrospectively evaluated 105 patients who underwent routine arthroscopic rotator cuff repair over a 12-month period. Correlations and linear regression analysis were used to assess factors associated with postoperative pain scores. Logistic regression analysis was used to construct a model of preoperative, demographic, surgical, and postoperative variables associated with HPSs. RESULTS: Wide variability existed in pain scores despite the similarity of the surgical stimulus. What subtle differences that were present in the surgical procedures had no bearing on pain scores. No correlation existed between pain scores and any physiologic correlate of pain. Pain scores were correlated with Workers' Compensation insurance, preoperative antianxiety medications, antidepressant medications, preoperative narcotics, and alcohol use. HPS patients had a longer recovery room stay (P < .05) and a higher incidence of postoperative nausea than those with a normal pain tolerance. HPS patients also had a significantly delayed return to work, as well as significantly lower Simple Shoulder Test scores at 6 weeks, although the final shoulder scores were similar to those in patients with a normal pain tolerance. CONCLUSIONS: Patient-reported pain scores have been mandated, but they vary widely among patients undergoing virtually the same procedure and correlate poorly with physiologic manifestations of pain. Pain scores tend to correlate with other types of variables, most of which would be expected by the experienced surgeon. The exclusive use of pain scores in postoperative pain management should be limited to prevent complications. LEVEL OF EVIDENCE: Level IV, prognostic case series.