Future healthcare systems will rely heavily on clinical decision support systems (CDSS) to improve the decision-making processes of clinicians. To explore the design of future CDSS, we developed a research-focused CDSS for the management of patients in the intensive care unit that leverages Internet of Things (IoT) devices capable of collecting streaming physiologic data from ventilators and other medical devices. We then created machine learning (ML) models that could analyze the collected physiologic data to determine if the ventilator was delivering potentially harmful therapy and if a deadly respiratory condition, acute respiratory distress syndrome (ARDS), was present. We also present work to aggregate these models into a mobile application that can provide responsive, real-time alerts of changes in ventilation to providers. As illustrated in the recent COVID-19 pandemic, being able to accurately predict ARDS in newly infected patients can assist in prioritizing care. We show that CDSS may be used to analyze physiologic data for clinical event recognition and automated diagnosis, and we also highlight future research avenues for hospital CDSS.

Rationale & objective

Underlying kidney disease is an emerging risk factor for more severe coronavirus disease 2019 (COVID-19) illness. We examined the clinical courses of critically ill COVID-19 patients with and without pre-existing chronic kidney disease (CKD) and investigated the association between the degree of underlying kidney disease and in-hospital outcomes.

Study design

Retrospective cohort study.

Settings & participants

4,264 critically ill patients with COVID-19 (143 patients with pre-existing kidney failure receiving maintenance dialysis; 521 patients with pre-existing non-dialysis-dependent CKD; and 3,600 patients without pre-existing CKD) admitted to intensive care units (ICUs) at 68 hospitals across the United States.

Predictor(s)

Presence (vs absence) of pre-existing kidney disease.

Outcome(s)

In-hospital mortality (primary); respiratory failure, shock, ventricular arrhythmia/cardiac arrest, thromboembolic events, major bleeds, and acute liver injury (secondary).

Analytical approach

We used standardized differences to compare patient characteristics (values>0.10 indicate a meaningful difference between groups) and multivariable-adjusted Fine and Gray survival models to examine outcome associations.

Results

Dialysis patients had a shorter time from symptom onset to ICU admission compared to other groups (median of 4 [IQR, 2-9] days for maintenance dialysis patients; 7 [IQR, 3-10] days for non-dialysis-dependent CKD patients; and 7 [IQR, 4-10] days for patients without pre-existing CKD). More dialysis patients (25%) reported altered mental status than those with non-dialysis-dependent CKD (20%; standardized difference=0.12) and those without pre-existing CKD (12%; standardized difference=0.36). Half of dialysis and non-dialysis-dependent CKD patients died within 28 days of ICU admission versus 35% of patients without pre-existing CKD. Compared to patients without pre-existing CKD, dialysis patients had higher risk for 28-day in-hospital death (adjusted HR, 1.41 [95% CI, 1.09-1.81]), while patients with non-dialysis-dependent CKD had an intermediate risk (adjusted HR, 1.25 [95% CI, 1.08-1.44]).

Limitations

Potential residual confounding.

Conclusions

Findings highlight the high mortality of individuals with underlying kidney disease and severe COVID-19, underscoring the importance of identifying safe and effective COVID-19 therapies in this vulnerable population.