Predictive Model for Short-Term Symptom Relief After Organophosphate Poisoning, 138 Cases

A clinical prediction model developed from 138 adult acute organophosphate poisoning cases treated between 2018 and 2023. The model uses four predictors to forecast symptom relief within 48 hours of admission. It was created by Guoliang Huang and published on figshare under a CC-BY-4.0 license.

Use Cases

• Predicting 48-hour symptom relief in acute organophosphate poisoning cases based on baseline symptom score, admission serum cholinesterase, Glasgow Coma Scale score, and major comorbidities.
• Validating a clinical nomogram's discrimination and calibration using optimism-corrected AUC and mean absolute error metrics.
• Assessing the clinical utility of a predictive model for ICU resource allocation via decision curve analysis.
• Training and testing logistic regression models for medical risk stratification using a dataset split into training and validation cohorts.

Strengths

• Model validation includes discrimination, calibration, and clinical utility analysis, with reported AUCs of 0.930 (training) and 0.905 (validation).
• Dataset covers 138 patient cases over a 5-year period (2018–2023), providing a temporal scope for analysis.
• The study methodology is explicitly described, including LASSO regression for feature selection from 14 potential predictors.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for certain modeling tasks.
• Data may reflect temporal and institutional bias inherent to the single retrospective study design.

Provenance

Source

figshare, author Guoliang Huang.

Collection Method

Retrospective study of hospital cases.

Time Range

2018–2023

Freshness

Last updated 2026-04-28 04:29:19; freshness should be verified.

Geography

null