O21: ASSIST: Validation of a novel hybrid clinician-patient remote outcome measure for diagnosis of surgical site infection

Abstract

Abstract Introduction Current remote methods of surgical site infection diagnosis have high specificity but lack sensitivity in diagnosis. This study evaluates the acceptability, reliability and validity of a novel hybrid clinician-patient remote outcome measure (ASSIST) for detecting SSI. Methods Patients completed the Bluebelle wound healing questionnaire (WHQ) at 30 days post-surgery and additionally submitted wound image(s). Images were reviewed two independent clinicians. Internal consistency of the unidimensional scale was assessed. Inter-rater and intra-rater reliability assessments were performed. Sensitivity and specificity for SSI discrimination against face-to-face CDC diagnosis were evaluated and receiver operator characteristic (ROC) curve plotted with calculation of the area under the curve. Results 69 participants were included and no questionnaires contained missing items. The single scale showed strong Internal consistency (α = 0.9). Intra-rater reliability was good with intraclass correlation (ICC) of 0.818 (95% CI, 0.672-0.903, p<0.001). Inter-rater agreement was excellent (ICC 0.915, 95% CI 0.733-0.903, p<0.001). Sensitivity and specificity for SSI diagnosis were excellent at 91.7% and 97.6% respectively. Discrimination for SSI was high with an area under the receiver operating characteristic curve of 0.966 (95%CI 0.913-1.000). Discussion The ASSIST measure is acceptable, reliable, and valid for diagnosis of SSI. Implementation can be safely utilised in both patient-initiated follow-up or at pre-determined time points for safe post discharge surveillance.