Clinical judgement on top of CURB65-score in predicting 30 days clinical outcome in patients with community acquired pneumonia (CAP) admitted to the hospital

Background: Several scoring systems exist to determine severity and clinical outcome of CAP. Amongst those is the CURB65 score, a well established tool to guide therapy and predict mortality in these patients. We wanted to see if adding the clinical judgement of the attending physician (not ill, moderately ill, severely ill) would improve the predictive capabilities of the CURB65-score. Methods: We used retrospective data derived from the CAPISCE study, which was a randomised controlled trial in a large teaching hospital in the Netherlands to assess the benefit of corticosteroid treatment in patients with CAP. Patients were scored ill or not ill at the time of admission by the attending physician. The CURB65 score was recorded as well. Clinical outcome was scored at day 30 as either cure or failure. Results: Data from 197 patients was available for analysis. Below are the Area Under the Curves for the CURB65 score, CURB65 score with clinical judgement and clinical judgement alone. Adding clinical judgement to the CURB65 score did not show any benefit in predicting clinical outcome at day 30. The AUC went from 0.741 (CURB65 score alone) to 0.746 (CURB65 score and clinical judgement) with a p value of 0.93. ![Figure][1] Conclusion: These results show that clinical judgement at time of admission is a poor predictive value for clinical outcome at day 30. [1]: pending:yes