Untold Stories of Future Engineers: Engineering Students' Reflections in a Fiction-Based Ethics Course.

Methanol, commonly referred to as methyl alcohol or wood alcohol, is a basic alcohol widely employed in industry, serving roles in products such as windshield washer fluids, antifreeze solutions, fuel additives, and perfumes [, ]. It is produced by distilling disintegrated wood particles []. While methanol is highly absorbable through gastrointestinal track, it can also toxify individuals if inhaled or absorbed through the skin []. Methanol poisoning is mainly caused by formate, the toxic metabolite of methanol oxidation []. On the other hand, although methanol poisoning is more common in developing nations and specifically among people with low socioeconomic status, it remains a major global health issue according to statistics []. In spite of official WHO announcement for prohibition of hand sanitation with methanol, we have been observing disinfected ingestion resulted poisoning in various communities []. The mortality rate among poisoned patients varies based on different reports, ranging from 6.5% to 54% in different countries []. Therefore, management of methanol poisoned patients is very necessary. Management of methanol-poisoned patients is challenging because symptoms are often non-specific and delayed, including abdominal pain, nausea, vomiting, headache, and visual disturbances, typically manifesting 12–24 h after exposure []. This diagnostic uncertainty complicates timely clinical decision-making, and delays in care can result in increased mortality []. Moreover, the management of methanol poisoning is resource intensive and necessitates a multidisciplinary approach, involving different strategies including ICU admission, antidote administration, maintain normal pH, and alcohol elimination enhancement [, ]. Critical care resources are limited and costly, making accurate and timely triage for ICU admission essential, yet difficult to achieve in clinical practice [, ]. Currently, clinicians rely on individual judgment, standard protocols, and limited scoring systems, which can be subjective and variable. Early identification of patients at high risk for deterioration remains challenging, and mis-triage may lead to both patient harm and inefficient resource use. Machine learning (ML) and deep learning (DL) models can analyze multiple patient variables simultaneously, detect complex patterns, and provide objective, timely risk assessments, complementing clinical judgment. Predicting the need for and optimal timing of ICU admission can improve patient outcomes and optimize ICU resource utilization [].