Variable Missingness by Cohort: CXR/NLP and ECG Sub-Analysis

A 5.4 KB Excel file from 2026-04-30 by Alejandro Hernández-Arango details the percentage of missing values per variable for two clinical cohorts. It reports missingness for Cohort A (n=121, CXR/NLP analytic cohort) and Cohort B (n=27, ECG sub-analysis subset). The data highlights variables with over 20% missingness, such as img_total_ratio and HRV_LF.

Use Cases

• Assess data completeness for clinical variables based on reported missingness percentages.
• Compare data quality between a primary cohort and a sub-analysis subset.
• Identify variables requiring imputation or exclusion based on the >20% missingness threshold.
• Plan analysis for multimodal studies involving chest X-ray (CXR), natural language processing (NLP), and electrocardiogram (ECG) data.

Strengths

• Provides specific missingness percentages for two defined cohorts (Cohort A n=121, Cohort B n=27).
• Explicitly identifies variables with high missingness (>20%), including img_total_ratio (89.3% in Cohort A) and HRV_LF (11.1% in Cohort B).
• Released under a permissive CC-BY-4.0 license.

Limitations

• Row count and column-level documentation are unknown; field semantics must be inferred after download.
• The dataset is very small (5.4 KB), indicating limited scope, likely containing only summary statistics.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare, author Alejandro Hernández-Arango.

Freshness

Last updated 2026-04-30 17:49:43; freshness should be verified.