Low-Risk Pediatric Critical Care Patients, Are They Really a Different Population?*

    loading  Checking for direct PDF access through Ovid


In this issue of Pediatric Critical Care Medicine, Verlaat et al (1) demonstrate that complex chronic conditions (CCC) and unplanned admissions are recognizable factors associated with mortality in low-risk pediatric critical care patients.
Authors designed a retrospective cohort study based on a national registry that includes all eight of the Netherlands’ PICU. The population studied was low-mortality risk patients admitted between January 2006 and January 2012. Low-mortality risk patients were defined as those with a predicted mortality less than 1% according to either recalibrated Pediatric Risk of Mortality (PRISM) II or Pediatric Index of Mortality 2 (PIM2) scores. Independent variables were described with precision, mainly CCC, which were defined according to established criteria (2). Differences between survivors and nonsurvivors were initially explored in a univariable analysis. Associations between independent variables and mortality were then analyzed in a multivariable logistic regression model. A total of 16,874 patients, 54.8% of all Dutch PICU admissions, had a low predicted mortality risk and they constituted the studied population. From these patients, 86 (0.5%) died. Based on the univariable analysis, review of the literature, and expert opinion, seven variables were introduced in the multivariable model: age, admission outside office hours, CCC, unplanned admissions, readmissions, specialized transport, and season of admission. Three of these—unplanned admissions: odds ratio (OR) 5.78 (95% CI, 3.40–9.91); CCC: OR 3.29 (95% CI, 1.97–5.50); and admissions between April and September: OR 1.67 (95% CI, 1.08–2.58)—were significantly associated with PICU mortality.
The authors focus on less than 1% predicted mortality risk patients because they represent almost 55% of admissions and 10% of mortality (86/966 deaths). Although Dutch PICU mortality is consistent with best international standards (3), authors are still concerned about avoidable deaths and substandard quality care in this prevalent subpopulation. Studying this group of patients with the aim of improving quality of care is the major strength of this article.
However, as nothing is usually straightforward, the authors did not achieve substantial answers. They found two risk factors, CCC and unplanned PICU admissions, that are neither preventable nor modifiable; and a third risk factor, season of admission, that is difficult to understand. Therefore, based on these factors, we readers cannot infer actions to try to improve survival in low-risk patients and we are kept in the usual position of tolerating uncertainty (4). But, of course, uncertainty generates questions.
Are low-risk patients really a different population with distinctive factors associated with death? CCC has been independently associated with mortality in the overall PICU population (5). Unplanned admissions, a PIM2 score component, have also been established as a risk factor of death in the general PICU population (6). Thus, recognizable factors related to death in low-risk patients are likewise associated with mortality in the any-risk PICU population. This suggests that low-risk patients are probably not different from the general PICU population.
Why are low-risk patients with unplanned admissions and CCC more likely to die? There is agreement that mortality is significantly associated with physiologic dysfunction (3). This hypothesis has been proved right with the accurate estimation of mortality risks in different populations through the quantification of physiologic status using predetermined physiologic variables. This is the basis of PRISM and PIM scores, which for decades have effectively estimated the risk of dying in PICU (7, 8). Thus, low-risk patients with unplanned admissions or CCC probably have a higher risk of experiencing a deterioration of physiologic status that consequently increases their chance of death.

Related Topics

    loading  Loading Related Articles