Reliability of patient-reported complications following hip or knee arthroplasty procedures

BACKGROUND: Patient reported outcomes are increasingly used to assess the success of surgical procedures. Patient reported complications are often included as an outcome. However, these data must be validated to be accurate and useful in clinical practice. METHODS: This was a retrospective descriptive study of 364 patients who had completed their six-month follow-up review questionnaire in the Arthroplasty Clinical Outcomes Registry, National (ACORN), an Australian orthopaedic registry. Patient-reported complications following total hip arthroplasty (THA) and total knee arthroplasty (TKA) were compared to surgeon-reported complications recorded in their electronic medical records at their various follow-up appointments. Sensitivity, specificity, positive predictive value and negative predictive value were calculated. Agreement was assessed using percentage agreement and Cohen’s kappa. RESULTS: Patient-reported data from the ACORN registry returned overall low sensitivity (0.14), negative predictive value (0.13) and kappa values (0.11), but very high specificity (0.98), positive predictive value (0.98) and agreement values (96.3%) for reporting of complications when compared to surgeon-reported data. Values varied depending on the type and category of complication. CONCLUSION: Patients are accurate in reporting the absence of complications, but not the presence. Sensitivity of patient-reported complications needs to be improved. Greater attention to the clarity of the questions asked may help in this respect. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12874-018-0645-0) contains supplementary material, which is available to authorized users.