Objective Multiple organ dysfunction syndrome (MODS) is a major complication of patients with acute organophosphorus pesticide poisoning (AOPP) and is associated with high mortality. This study aimed to develop and validate a MODS prediction model for this patient population using a nomogram and machine learning methods. Methods A retrospective study was conducted on 270 AOPP patients from Linquan County People’s Hospital to establish the prediction model. Lasso regression was used for variable selection, and multivariate Logistic regression was applied for model construction. Model performance was evaluated based on discriminative ability, calibration, and decision curve analysis. Results Among the 270 AOPP patients, 129 (47.8%) developed MODS. The key predictors of MODS included heart rate, Endotracheal intubation, and blood lactic acid. The nomogram achieved an area under the curve (AUC) of 0.962 (95% confidence interval CI: 0.932–0.982). The calibration plot showed a high agreement between predicted probabilities and actual observed probabilities, and decision curve analysis demonstrated a favorable clinical net benefit of the model. Conclusion We developed a risk prediction model for MODS in AOPP patients. This model can assist clinicians in assessing MODS risk and provide a scientific basis for subsequent interventions. External validation is required to confirm the reliability of the current risk model before its clinical application.
Yu et al. (Tue,) studied this question.