| Summary: | (1) <b>Background:</b> Patients’ comorbidities play an immanent role in perioperative risk assessment. It is unknown how Charlson Comorbidity Indices (CCIs) from different sources compare. (2) <b>Methods:</b> In this prospective observational study, we compared the CCIs of patients derived from patients’ self-reports and from physicians’ assessments with hospital administrative data. (3) <b>Results:</b> The data of 1007 patients was analyzed. Agreement between the CCI from patients’ self-report compared to administrative data was fair (kappa 0.24 [95%CI 0.2–0.28]). Agreement between physicians’ assessment and the administrative data was also fair (kappa 0.28 [95%CI 0.25–0.31]). Physicians’ assessment and patients’ self-report had the best agreement (kappa 0.33 [95%CI 0.30–0.37]). The CCI calculated from the administrative data showed the best predictability for in-hospital mortality (AUROC 0.86 [95%CI 0.68–0.91]), followed by equally good prediction from physicians’ assessment (AUROC 0.80 [95%CI 0.65–0.94]) and patients’ self-report (AUROC 0.80 [95%CI 0.75–0.97]). (4) <b>Conclusions:</b> CCIs derived from patients’ self-report, physicians’ assessments, and administrative data perform equally well in predicting postoperative in-hospital mortality.
|
|---|