Reply to Comment on: Clinical performance of existing diagnostic criteria for pneumonia in older emergency patients

Thank you to Turgut et al. for their careful review of our manuscript “Clinical performance of existing diagnostic criteria for pneumonia in older emergency patients: a prospective cohort study.” We would like to address each of their comments below.

In Table 1,¹ in accordance with the STROBE guidelines we chose to present raw data for the demographic / descriptive data without conducting post-hoc analyses or presenting p-values.² When we conduct these statistical tests, all p-values are >0.20 except race which is <0.001. Regardless of which correction is used for multiple comparisons; the interpretation of these p-values would be the same and only race would be significant.

Our goal was to describe the performance of diagnostic criteria derived and validated in other settings and to determine if their performance was sufficient for the acute care setting. The study was powered to achieve this goal.¹ We did not seek to and were not powered to rederive or to test various iterations/cutoffs of the previously published criteria. Therefore, there was no ROC curve analysis to perform. We agree that rigorous derivation and validation of a more effective rule is critical and being pursued in our ongoing work.³

We agree that long-term patient-centered clinical outcomes must be considered including the side effects of inappropriately prescribed medications (over-diagnosis)⁴ and the delay of appropriate medications (under-diagnosis).^5,6 The request for patient-centered outcomes such as these is an important but different study that we and others have attempted to answer with results consistently demonstrating potential patient harm from both types of mis-diagnosis.^7–9 Notably previous results demonstrated high diagnostic uncertainty,¹⁰ supporting the assertion that before we can affect those outcomes in a prospective study, we must ensure that we can accurately and reliably diagnose pneumonia in the emergency department. While we were not powered to report on outcomes such as mortality, we have previously published on some patient-centered outcomes such as functional decline in this study population.

Unfortunately, in older adults, accurate diagnosis of pneumonia in the emergency department remains a challenge and thus, we agree with Turgut et al., that this is an important research priority and hope that our ongoing work can contribute to improving the care of older adult emergency department patients.