Data for Clinical and Epidemiological Patterns of Acute Pharmaceutical Poisoning in Nicaraguan Adults: A Cross-Sectional Analysis
Description
This dataset contains the curated, de-identified clinical and epidemiological data of 43 adult patients ($n = 43$) hospitalized due to acute pharmaceutical poisoning at the "Carlos Roberto Huembés" University Hospital in Managua, Nicaragua, covering the period from 2020 to 2021. The primary objective of this data collection was to characterize the clinical-epidemiological profile of medicinal overdoses and evaluate the risk factors and predictors associated with clinical complications and the length of hospital stay (LOS). Data were retrospectively extracted from physical and electronic medical records of patients aged 18 years and older who met the clinical and laboratory criteria for acute poisoning secondary to pharmaceutical agents. The dataset tracks sociodemographic characteristics, toxicological profiles (specific active ingredients, drug classes, and dosage), clinical presentation structured by toxidromes, medical management interventions (decontamination, supportive care, and antidote administration), and clinical outcomes (complications, length of stay, and mortality).
Files
Steps to reproduce
Steps to Reproduce 1. Data Source and ExtractionTarget Population: The dataset consists of all adult patients (n = 43) admitted with a primary diagnosis of acute pharmaceutical poisoning at the "Carlos Roberto Huembés" University Hospital between 2020 and 2021. Data Retrieval: Data were manually extracted from physical and electronic clinical charts utilizing a standardized medical audit form designed by the research team. 2. File Setup and Software Requirements File Format: Download the master dataset file (Data25032026.xlsx) from the Mendeley Data repository. Environment Setup: To reproduce the statistical analyses and models reported in the manuscript, use an environment with R (v4.2 or higher) or Python (v3.9 or higher using the pandas, statsmodels, and scipy libraries).Missing Data Handling: Non-applicable fields or unrecorded secondary complications are stored as empty cells or specified as NaN (Not a Number) within the dataset.3. Executing the Statistical Analysis Pipeline Step A: Descriptive Summaries Demographics: Compute the mean and standard deviation for the continuous variable Age to replicate the baseline cohort age (28.6 + - 11.6 years).Frequencies: Run categorical frequency distributions on Sex, Gpharmacological, FarmacoMG (Specific Agent), and Typeofintoxication to replicate the prevalence profiles (e.g., males at 55.8%, intentional self-harm at 95.3%, and Acetaminophen at 16.3%). Step B: Multivariate Logistic Regression (Table 3 Replication) Model Specification: Fit a generalized linear model (GLM) with a binomial logit link function. Variables: Dependent Variable (Y): Complications (Binary: 1 = Yes, 0 = No). Independent Predictors (X): Time_of_Care (Continuous, representing hours elapsed to medical care) and Typeofintoxication (Categorical/Dummy coded, isolating intentional self-harm vs. accidental).Model Evaluation: Calculate the Odds Ratios (OR) by exponentiating the model coefficients (beta). To evaluate the model's predictive power for binary variables, compute Tjur’s Coefficient of Discrimination (R^2 Tjur). This will confirm the reported value of 0.026 and the lack of statistical significance for the temporal predictors (p > 0.05). Step C: Correlation AnalysisNon-parametric Testing: Execute a Spearman’s rank correlation coefficient between the continuous variables Time_of_Care and Hospitalization_Hospital_Stay (Length of Stay/LOS).Verification: This test will replicate the reported correlation strength (rho = -0.12) and its corresponding p-value (p = 0.42), confirming that delays in hospital arrival did not statistically drive the duration of confinement.
Institutions
- Universidad Nacional Autónoma de Nicaragua-LeónLeón Department, León