Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
We appreciate the concerns raised by Dr Houck and colleagues and by Dr Sanders. As we state in the "Comment" section of our article, "Sound clinical judgment clearly supports timely blood cultures, oxygenation measurement, and antibiotic administration in the treatment of patients hospitalized with CAP." We fully expected to demonstrate that such processes of care would lead to better patient outcomes in our study. Our failure to demonstrate such a relationship prompted an exploration of methodologic problems common to all observational assessments of medical interventions. In particular, we found evidence of "confounding by indication"1; ie, physicians appeared to be most aggressive with their sickest patients. As a result, the relative odds of a poor outcome were generally higher among patients receiving rapid diagnostic assessment and antibiotic therapy (although most of these elevated odds ratios did not achieve statistical significance). These effects persisted after adjustment for severity of illness, likely indicating residual confounding by disease severity. Particularly striking was our observation that patients for whom no blood culture was performed did better than those whose blood was cultured. Presumably, patients who appeared less ill prompted less diagnostic activity.
Dedier J, Singer DE, Atlas SJ. Dilatory Standards Are Below the Level of Care—Reply. Arch Intern Med. 2002;162(7):844-845. doi: