ABSTRACT
There are several established methods for comparing more than two survival curves, namely the scale-rank test or Cox's proportional hazard model. However, when their statistical assumptions are not met, their results' validity is affected. In this study, we address the mentioned issue and propose a new statistical approach on how to compare more than two survival curves using a random forest algorithm, which is practically assumption-free. The repetitive generating of many decision trees covered by one random forest model enables to calculate of a proportion of trees with sufficient complexity classifying into all groups (depicted by their survival curves), which is the p-value estimate as an analogy of the classical Wald's t-test output of the Cox's regression. Furthermore, a level of the pruning of decision trees the random forest model is built with, can modify both the robustness and statistical power of the random forest alternative. The discussed results are confirmed using COVID-19 survival data with varying the tree pruning level. The introduced method for survival curves comparison, based on random forest algorithm, seems to be a valid alternative to Cox's regression;however, it has no statistical assumptions and tends to reach higher statistical power. © 2021 IEEE
ABSTRACT
Background The optimal timing of intubation for critically ill patients with severe respiratory illness remains controversial among healthcare providers. The coronavirus disease 2019 (COVID-19) pandemic has raised even more questions about when to implement this life-saving therapy. While one group of providers prefers early intubation for patients with respiratory distress because these patients may deteriorate rapidly without it, other providers believe that intubation should be delayed or avoided because of its associated risks including worse outcomes. Research question Our objective was to assess whether the timing of intubation in patients with severe COVID-19 pneumonia was associated with differences in mortality or other outcomes. Study design and methods This was a single-center retrospective observational cohort study. We analyzed outcomes of patients who were intubated secondary to COVID-19 pneumonia between March 13, 2020, and December 12, 2020, at Henry Ford Hospital in Detroit, Michigan. Patients were categorized into two groups: early intubated (intubated within 24 hours of the onset of severe respiratory distress) and late intubated (intubated after 24 hours of the onset of severe respiratory distress). Demographics, comorbidities, respiratory rate oxygenation (ROX) index, sequential organ failure assessment (SOFA) score, and treatment received were compared between groups. The primary outcome was mortality. Secondary outcomes were ventilation time, intensive care unit stay, hospital length of stay, and discharge disposition. Post hoc and Kaplan-Meier survival analyses were performed. Results A total of 110 patients were included: 55 early intubated and 55 late intubated. We did not observe a significant difference in overall mortality between the early intubated (43%) and the late intubated groups (53%) (p = 0.34). There was no statistically significant difference in patients' baseline characteristics including SOFA scores (the early intubation group had a mean score of 7.5 compared to 6.7 in the late intubation group). Based on the ROX index, the early intubation group had significantly more patients with a reduced risk of intubation (45%) than the late group (27%) (p = 0.029). The early intubation group was treated with a high-flow nasal cannula at a significantly lower rate (47%) than the late intubation group (83%) (p < 0.001). Significant differences in patient baseline characteristics, treatment received, and other outcomes were not observed. Post hoc analysis adjusting for SOFA score between 0 and 9 revealed significantly higher mortality in the late intubation group (49%) than in the early intubation group (26%) (p = 0.03). Patients in the 0 to 9 SOFA group who were intubated later had 2.7 times the odds of dying during hospital admission compared to patients who were intubated early (CI, 1.09-6.67). Interpretation The timing of intubation for patients with severe COVID-19 pneumonia was not significantly associated with overall mortality or other patient outcomes. However, within the subgroup of patients with SOFA scores of 9 or lower at the time of intubation, patients intubated after 24 hours of the onset of respiratory distress had a higher risk of death than those who were intubated within 24 hours of respiratory distress. Thus, patients with COVID-19 pneumonia who are not at a high level of organ dysfunction may benefit from early mechanical ventilation.
ABSTRACT
The Covid-19 pandemic has only accelerated the need and desire to deal more openly with mortality, because the effect on survival is central to the comprehensive assessment of harms and benefits needed to meet a 'reasonable patient' legal standard. Taking the view that this requirement is best met through a multi-criterial decision support tool, we offer our preferred answers to the questions of What should be communicated about mortality in the tool, and How, given preferred answers to Who for, Who by, Why, When, and Where. Summary measures, including unrestricted Life Expectancy and Restricted Mean Survival Time are found to be reductionist and relative, and not as easy to understand and communicate as often asserted. Full lifetime absolute survival curves should be presented, even if they cannot be 'evidence-based' beyond trial follow-up limits, along with equivalent measures for other criteria in the (necessarily) multi-criterial decision. A decision support tool should relieve the reasonable person of the resulting calculation burden.