Letter by Porath et al Regarding Article, “Influence of Inpatient Service Specialty on Care Processes and Outcomes for Patients With Non–ST-Segment Elevation Acute Coronary Syndromes”
To the Editor:
We agree with Roe et al1 that it is difficult to retrospectively evaluate the appropriate triage of patients with non–ST-elevation acute coronary syndromes. Therefore, we believe it is important to stress the fact that the findings of the present study should not be interpreted to indicate that admission to a cardiology service is causally linked to improved patient outcome. In fact, the article raises the important question of whether it is possible to ascertain from a retrospective study the causes for the observed differences in outcome of non–ST-elevation acute coronary syndrome patients admitted to a medicine service or to cardiology.
As is obvious and acknowledged by the investigators, triage decisions were made by the admitting physicians on the basis of their overall assessment of the patient. As shown in many previous studies, patients admitted to cardiology were younger and healthier than those admitted to medicine.2 The investigators attempted to adjust for these differing characteristics, but we believe that this adjustment was seriously inadequate. Although the authors did adjust for the differing cardiovascular comorbidities, they did not account for other noncardiovascular comorbidities such as cognitive impairment, chronic lung disease, cancer, and acute infection. Presumably, these data were not available in the database, but their absence seriously weakens the adjustment. Furthermore, the authors did not adjust for laboratory abnormalities such as the level of albumin or creatinine, which can certainly affect prognosis, and did not use more comprehensive case-mix tools such as Charlson’s scoring system.3,4 It also is possible that patients not admitted to cardiology were more functionally dependent and less eligible for or less interested in invasive therapies such as early coronary intervention. For example, there was a difference in “do not resuscitate” status between patients admitted to medicine versus cardiology. These differences in comorbidities and functional status also could affect the ability to use certain medications recommended in clinical guidelines.
Although the authors tried to overcome these difficulties by using a propensity score, this tool also is limited by the absence of the above baseline characteristics.
Therefore, it appears clear that admission to a cardiology service is associated with a higher rate of guideline-recommended management and improved outcome. We believe, however, that causality is very uncertain and could be assessed only in a prospective, randomized trial.