DOI: 10.1097/CCM.0b013e31828fd714

,  

,  

PMID: 23863242

Issn Print: 0090-3493

Publication Date: 2013/08/01


Print

Reducing the Risks of Intrahospital Transport Among Critically Ill Patients*

Teryl K. Nuckols

    loading  Checking for direct PDF access through Ovid

Excerpt

Six weeks into internship, my classmate and I were having breakfast together after rounds. We heard a code blue called overhead; it was in the CT scanner. “Death begins in Radiology,” she advised me with as sage a look as an intern can visage. “It was been proven that clinical deterioration often begins there or en route.” She herself had needed to call anesthesia to emergently reintubate a patient in the scanner just days earlier—we trained before duty-hour limits and dedicated transport nurses.
Despite her penchant for making dramatic statements, my classmate, now friend, was right. Several studies have shown that moving critically ill patients around the hospital, particularly to remote and sparsely staffed diagnostic units, is associated with adverse outcomes, including death (1). Yet, the challenge for researchers is untangling whether being transported causes complications, symptoms related to severe illness induce a need to be transported, or severe illness leads to both a need for transport and complications. This chicken-and-egg issue is called, in research parlance, endogeneity.
In this issue of Critical Care Medicine, Schwebel et al (2) report how intrahospital transport (IHT) is associated with the risk of several complications among more than 1,700 ventilated ICU patients. This appears to be one of the largest and most rigorous studies to date on the risks of IHT, never mind that it focused on ventilated patients. The authors attempted to address the chicken-and-egg issue using propensity scores, that is, by matching transported patients with controls who had similar likelihoods of being transported but were not (2). The sample size is several-fold larger than previous analyses. The authors examined a wide range of complications, not just respiratory ones. Some complications had not been previously linked to IHT (1, 2).
Propensity score matching is widely accepted as a means of optimizing, although not guaranteeing, the comparability of patients who do versus do not receive some exposure or treatment. The only substantially better designs might be a randomized controlled trial or an instrumental variables analysis. However, institutional review boards would be unlikely to approve of randomly assigning transportation to emergent diagnostic tests, and finding a suitable instrumental variable seldom proves feasible, for technical and logistical reasons.
In this analysis, Schwebel et al (2) created a statistical model that predicted each patient’s likelihood of being transported, that is, the propensity score. Next, they matched each transported patient with several controls using the propensity scores and days in the ICU (before the day of transport). To check whether the matching procedure produced similar cases and controls, the investigators compared their baseline clinical and demographic characteristics. The two groups were similar but transported patients were more likely to be intubated, have arterial and central catheters, and have higher Sequential Organ Failure Assessment scores. IHT and complication events were taken from a prospective clinical monitoring and event reporting database. Complications were considered IHT associated when they occurred on the day of transport or shortly thereafter. Unfortunately, the investigators were unable to review charts to make clinical judgments about the relationship between individual complications and IHT. To achieve their main objective, the authors created a statistical model that compared the predicted risks of various IHT-associated complications between transported patients and controls (and adjusted for the baseline differences).
Schwebel et al (2) found that almost 30% of ventilated patients were transported while in the ICU, 94% of the time to the CT scanner. Compared with controls, transported patients were significantly more likely to develop bleeding, deep vein thrombosis, pneumothorax, atelectasis, ventilator-associated pneumonia, hypoglycemia, hyperglycemia, and hypernatremia on the day of transport or shortly thereafter.
    loading  Loading Related Articles