Abstract
Use of dialysis among those hospitalized with COVID-19 is high at around 4%. Although less than one percent of patients undergo tracheostomy and ECMO, the absolute numbers of patients who have undergone these interventions is substantial.
Introduction
Early in the COVID-19 pandemic, clinical reports on disease severity were often limited to single-center studies or small cohorts, making it difficult to grasp the true scale of resource utilization across the United States. Specifically, there was significant uncertainty regarding the prevalence of "invasive" interventions—those that bypass or replace organ functions—such as dialysis, tracheostomy, and extracorporeal membrane oxygenation (ECMO).
Understanding the use of these treatments is crucial for resource planning, as they require specialized equipment and highly trained staff. Furthermore, these interventions often represent the "last line of defense" in critical care, and survivors frequently face significant long-term morbidity. This study, a large-scale collaborative effort, provides a high-resolution look at these intensive medical interventions during the first year of the pandemic.
Methods and Data Sources
The researchers conducted a network cohort study utilizing seven distinct large-scale US databases to provide a diverse cross-section of the population. These included:
- HealthVerity: Linked medical claims and pharmacy data.
- Premier Healthcare Database: Clinical coding and hospital billing data.
- IQVIA Hospital CDM & Open Claims: Resource management and pre-adjudicated claims.
- Optum EHR & Optum SES: Electronic health records and socio-economic status data.
- VA-OMOP: Data from the national Department of Veterans Affairs health system.
To ensure consistency across these varied sources, all databases were mapped to the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM). This standardized format allowed the research team to execute the same analysis code across different systems without accessing raw, sensitive patient data. The study tracked interventions occurring within 30 days of a patient's initial COVID-19 hospitalization.
Study Population and Baseline Characteristics
The study analyzed a total of 842,928 hospitalized patients. The patient populations varied significantly across databases in terms of age, sex, and underlying health conditions, reflecting the broad reach of the study.
As shown in the table above, the cohorts ranged from the VA-OMOP database, which was predominantly male (93.4%) and older (65% aged >= 65), to databases like Optum EHR where 60% of patients were under 65. Comorbidities were prevalent across all cohorts, with hypertension and diabetes being particularly common, which likely contributed to the high rates of acute organ failure observed.
Key Findings
The Dialysis Burden
The most striking finding was the high rate of dialysis, required by approximately 4.17% (35,192) of the total patient population. This confirms that COVID-19 frequently causes acute kidney injury (AKI) in addition to respiratory distress. In patients with pre-existing chronic kidney disease, the rate of dialysis in some databases reached as high as 19.6%. Use was notably higher in male patients and those with prior renal issues.
ECMO Selectivity
Extracorporeal membrane oxygenation (ECMO), which pumps blood through an artificial lung, was used sparingly, occurring in 0.19% (1,568) of patients. Interestingly, ECMO was significantly more frequent in younger patients with fewer comorbidities. For instance, in the IQVIA Open Claims data, 0.33% of those under 65 received ECMO compared to only 0.03% of those 65 or older. This suggests a "salvage therapy" approach, where the most aggressive treatment is reserved for those deemed most likely to survive the intensive procedure.
Tracheostomy Trends
Tracheostomy (creating an opening in the neck for a breathing tube) was performed in 0.82% (6,950) of patients. Unlike ECMO, the use of tracheostomy was relatively consistent across age groups and sexes. This likely reflects its role as a standard procedural step for any patient requiring long-term mechanical ventilation, regardless of their baseline health status.
Discussion and Limitations
The consistency of results across diverse databases (claims vs. EHR) validates the findings and demonstrates the power of the OMOP CDM for large-scale public health monitoring. However, the authors acknowledge several limitations:
- Underreporting: Some interventions, particularly dialysis in the ICU, might be bundled into general "intensive care" billing codes and thus undercounted.
- Mechanical Ventilation Data: Reporting for mechanical ventilation was too inconsistent across databases to be included in the final results.
- Long-term Outcomes: This study focused on the utilization of interventions rather than long-term survival or quality of life.
The absolute number of patients requiring these life-altering interventions is substantial. Future research must focus on the long-term clinical and economic burden facing survivors who underwent these procedures.
Summary updated