All Research Contributions
  • Conference Poster
  • Engineering Contribution

ApapTox — Web-Based Clinical Decision Support for Acetaminophen Toxicity

A web-based, rule-driven clinical decision support system for acetaminophen toxicity, presented as Poster #67 at NACCT 2026. I designed its software architecture and built the application around physician-provided clinical logic.

My role · ApapToxLead Software Engineer & System Architect
Conference
NACCT 2026 — North American Congress of Clinical Toxicology
Format
Scientific Poster
Poster number
#67
Date
October 23, 2026
Location
Golden State Ballroom, San Diego
PosterThe original NACCT 2026 poster, as submitted. Sajjad Fani is credited in the author line (affiliation: Medical Toxic, Inc.).

Overview

ApapTox is a web-based, rule-driven clinical decision support system developed to assist clinicians with the evaluation and management of acetaminophen (APAP) toxicity — including atypical cases that may require assessment beyond a conventional nomogram.

The Rumack–Matthew nomogram guides most acute, single, oral ingestions with a known time. Delayed or unknown-time presentations, repeated supratherapeutic ingestions and non-oral routes need a more structured application of toxicology principles. ApapTox classifies the presentation, routes it to the matching pathway, and returns explainable guidance on N-acetylcysteine (NAC), serial labs, monitoring and escalation.

My Engineering Contribution

  • Designed the complete software architecture of the application.

  • Developed the web application end to end.

  • Analyzed the scientific literature and clinical management guidelines the system is built on.

  • Transformed guideline-based decision processes into structured clinical flowcharts.

  • Integrated the physician-provided Python clinical decision logic into the web application.

  • Implemented the interactive clinical assessment and decision workflows.

Attribution

  • The initial Python clinical logic was prepared by a collaborating physician, using AI assistance.
  • My responsibility was turning the medical requirements and the physician-provided logic into an integrated, functioning web application, together with the overall software architecture and development.
  • The clinical treatment guidelines are published consensus guidance. I did not author them, and I did not medically validate the system’s recommendations.

How ApapTox Selects the Pathway

The application is organized as a rule-driven decision pipeline. Clinical inputs are collected once, the presentation is classified by route, episode and time, and exactly one pathway is applied before shared safety checks run over every case. Each recommendation is returned with its rationale, so the clinician can see why the system reached it.

  • Clinical inputs — ingestion pattern, dose, time from ingestion, serum APAP and key labs (ALT/AST, INR, pH, lactate, creatinine).

  • Classification — route, episode and time decide which pathway applies.

  • Acute oral, known time — Rumack–Matthew nomogram (4-hour line) to initiate or withhold NAC.

  • Unknown time or late (> 24 h) presentation — a lab-guided approach, with NAC initiated or continued and reassessed.

  • Repeated supratherapeutic ingestion (RSTI) — history, APAP and ALT/AST in place of the nomogram.

  • Non-oral, unclear or unusual route — risk assessment and toxicology evaluation, with no single algorithm.

  • Shared safety and reassessment — high-risk features (altered mental status, co-ingestions, severe acidosis, ECTR/ECLS considerations) and serial labs.

  • Explainable output — NAC guidance, serial labs, monitoring, escalation and the rationale behind each.

Conference Recognition

ApapTox was presented as Poster #67 at the North American Congress of Clinical Toxicology (NACCT 2026) on October 23, 2026, in the Golden State Ballroom, San Diego. The results below are reported as printed on the poster.

  • A prototype was created, with lab-guided pathways integrated in one clinical interface.

  • Case-based testing across common and complex scenarios (unknown time or late, repeated, non-oral) demonstrated consistent, guideline-aligned recommendations.

  • The system generates patient-specific NAC recommendations with monitoring and reassessment, and lets users review the rationale for each recommendation.

  • Expert review and user testing refined the tool and its interface; broader validation and integration are named as further development.

Scientific Poster

“A Novel Web-Based Decision Support System for Acetaminophen Toxicity Management”

Credited on the poster

  • Omid MehrpourMD, FACMT · Florida Poison Control Center, University of Florida
  • Farshad M. ShiraziMD, FACMT · Florida Poison Control Center, University of Florida
  • Sajjad FaniMedical Toxic, Inc.

Names and affiliations as printed on the poster, except that the poster’s misspelling “Mehtapour” is corrected to Mehrpour. The author listing of the corresponding abstract in the published NACCT 2026 abstract collection is governed by that publication; this page does not claim that I am an indexed author of the journal abstract.

Evidence Gallery

  • PosterThe original NACCT 2026 poster, as submitted. Sajjad Fani is credited in the author line (affiliation: Medical Toxic, Inc.).
  • PhotographPoster #67 on display at NACCT 2026, with Dr. Omid Mehrpour beside it. The photograph documents the physical presentation of the poster.
  • Application screenshotThe production ApapTox web application — start screen at apaptox.medicaltoxic.com.

Official References