Prediction Model of Acute Respiratory Failure in Patients with Acute Pesticide Poisoning by Intentional Ingestion: Prediction of Respiratory Failure in Pesticide Intoxication (PREP) Scores in Cohort Study

Abstract

Acute respiratory failure is the primary cause of mortality in patients with acute pesticide poisoning. The aim of the present study was to develop a new and efficient score system for predicting acute respiratory failure in patients with acute pesticide poisoning. This study was a retrospective observational cohort study comprised of 679 patients with acute pesticide poisoning by intentional poisoning. We divided this population into a ratio of 3:1; training set (n = 509) and test set (n = 170) for model development and validation. Multivariable logistic regression models were used in developing a score-based prediction model. The Prediction of Respiratory failure in Pesticide intoxication (PREP) scoring system included a summation of the integer scores of the following five variables; age, pesticide category, amount of ingestion, Glasgow Coma Scale, and arterial pH. The PREP scoring system developed accurately predicted respiratory failure (AUC 0.911 [0.849−0.974], positive predictive value 0.773, accuracy 0.873 in test set). We came up with four risk categories (A, B, C, and D) using PREP scores 20, 40, and 60 as the cut-off for mechanical ventilation requirement risk. The PREP scoring system developed in the present study could predict respiratory failure in patients with pesticide poisoning, which can be easily implemented in clinical situations. Further prospective studies are needed to validate the PREP scoring system.

Machine Learning-Based Prediction Models of Acute Respiratory Failure in Patients with Acute Pesticide Poisoning

“…In the case of prediction based on LSTM [10], the PPV was 0.226 and the sensitivity was 0.881. However, these respiratory failure prediction algorithms are characterized by a large measurement interval, low performance, or large number of features [4][5][6][7][8][9][10]. Our proposed algorithm demonstrated improved respiratory failure prediction within 24 h with higher PPV and sensitivity compared with those of other models.…”

“…Table 12 shows the performance of each algorithm for the prediction of respiratory failure. In recent years, respiratory failure prediction models have been developed to predict respiratory failure in COVID-19 patients based on deep learning with semi-supervised learning [4], respiratory failure based on XGBoost using clinical data [5], respiratory failure in COVID-19 patients based on LR [6], respiratory failure in ICU patients based on LightGBM [7], respiratory failure in patients with pesticide poisoning due to intentional pesticide ingestion based on LR [9], cardiac arrest and respiratory failure in ICU patients based on LSTM [10], and respiratory failure in ICU patients based on gradient boosting [8]. In the case of prediction based on semisupervised learning [4], the PPV was 0.033 and the sensitivity was 0.78.…”

“…In the case of prediction based on LightGBM [7], the PPV was 0.42 and the sensitivity was 0.80. In the case of prediction of respiratory failure in patients with pesticide poisoning based on LR [9], the PPV was 0.833 and the sensitivity was 0.606. In the case of prediction based on LSTM [10], the PPV was 0.226 and the sensitivity was 0.881.…”

“…Recent predictions of respiratory failure include predicting respiratory failure based on semi-supervised learning [4]; predicting respiratory failure with clinical data [5]; predicting respiratory failure in patients with coronavirus disease-2019 (COVID-19) [6]; predicting respiratory failure in the intensive care unit (ICU) [7,8]; predicting respiratory failure in pesticide intoxication [9]; and predicting respiratory failure with simple patient trajectories [10].…”

“…In the case of respiratory failure prediction in the ICU, the PPV was 42% and the sensitivity was 80% [7]. In the case of respiratory failure prediction in pesticide intoxication, the PPV was 83.3% and the sensitivity was 60.6% [9]. In the case of respiratory failure prediction with simple patient trajectories, the PPV was 22.6% and the sensitivity was 88.1% [10].…”