The authors reply
Our contention, however, is that a positive result in the screening for potential delirium using various instruments now appears to have become the recommended diagnostic test for delirium (3), rather than the subsequent clinical assessment as described by Drs. Traube and Silver (1) in their correspondence. Our literature review addresses this point in particular—the message is that if one is solely using the screening instruments to determine epidemiology in the PICU, then there is the risk of under- or overestimation in case prevalence because of the overlapping “diagnostic” domains with other entities such as sedative/hypnotic IWS (2). (We have outlined sensitivity and specificity of “diagnosis” against en face psychiatric assessment in our article (2), and the reader can review the findings by age). This lack of precision (i.e., screening as the “diagnosis” vs en face assessment as the diagnosis of delirium), along with inherent overlapping domains in the screening instruments for delirium and IWS, means that there is the potential for circularity when the screening instrument is then used to identify, for example, sedative/hypnotic agents as a potential modifiable factor (4, 5). What is being modified? Perhaps it does not matter, just as long as it is. However, treatment for straightforward IWS without delirium would involve replacing some of the agent(s) previously withdrawn, while these agents might be avoided or minimized in delirium.
We believe that epidemiology has a role in moving our field forward, and hence, we contributed the review to “Pediatric Critical Care Medicine,” not to be contrary, but with the aim of focusing thought on an issue that Drs. Traube and Silver (1) have also recognized (6). Therefore, our conclusion remains the same as we wrote in our report (2): “We have drawn attention to the overlap between the symptom/sign scales used for assessing IWS, delirium and anticholinergic toxidrome not to comment on individual patient care, but rather to point out issues that need to be addressed in more widespread epidemiologic research of these symptom syndromes. Indiscriminate implementation and interpretation of any of these scales may result in misdiagnosis. Although each of these scoring systems may serve as useful tools, we suggest interpretation in the full clinical context, and interventions administered on a case-by-case basis.” If we are to learn from epidemiology in this particular area, then we still need robust case definitions that can be used efficiently in large populations.