Abstract
This article provides a formal correction to a significant multinational study published in PLOS Digital Health. The correction addresses errors in the original publication's demographic visualizations and data privacy methodology, specifically regarding the "S1 Methods" and the comprehensive demographic breakdown in Figure 2.
Introduction
While COVID-19 is primarily recognized as a respiratory illness, clinical evidence has increasingly shown that it can trigger significant neurological complications. Understanding the global scale of these issues requires harmonizing data across different countries, healthcare systems, and coding practices. This study characterizes the demographic and clinical profiles of patients who develop neurological conditions during COVID-19 hospitalization, categorizing them into no neurological conditions (NNC), central nervous system (CNS) diagnoses, and peripheral nervous system (PNS) diagnoses.
Methodology
The research utilizes a multinational cohort study design, leveraging the Consortium for Clinical Characterization of COVID-19 by EHR (4CE) framework.
- Data Integration: The study harmonizes Electronic Health Record (EHR) data from diverse international sites across the US, Europe, and Asia.
- Stratification: Patients are grouped by neurological status: NNC, CNS (e.g., stroke, encephalitis), and PNS (e.g., nerve damage).
- Privacy Preservation: A key component of the methodology is the obfuscation process. To protect patient confidentiality while maintaining data utility, the researchers applied specific counting and reporting thresholds to the EHR data.
Key Results and Insights
The primary contribution of this correction is the updated demographic visualization, which provides a comprehensive breakdown of the cohort across various international sites.
Global Neurological Impact
The data demonstrates that CNS and PNS diagnoses occurred across every participating healthcare system, regardless of the country or local demographic makeup. The cohort includes tens of thousands of patients, such as over 23,000 at APHP in France and over 10,000 at UPITT in the US.
Demographic Trends and Outcomes
- Age Distribution: Patients with CNS and PNS diagnoses often skew toward older age ranges (70-79 and 80+) compared to the NNC group.
- Adverse Outcomes: Nested pie charts reveal that a higher proportion of patients with neurological diagnoses (particularly CNS) fall into "Severe" and "Deceased" categories. These patients also consistently show higher rates of hospital readmission across most international sites.
Limitations and Future Directions
As an EHR-based study, results depend on the accuracy and consistency of clinical coding across different systems. Additionally, while obfuscation methods protect privacy, they can lead to "Obfuscated" data points in smaller sub-cohorts, slightly limiting granularity for rare conditions. Future research is needed to determine long-term neurological impacts (Long COVID) using this multinational framework as a baseline.
This post summarizes a correction to the article with DOI: 10.1371/journal.pdig.0000484.
Summary updated