☠️ Poison Control Center Triage Decision Tree

Initial Call Assessment — Building the Exposure Picture in the First 90 Seconds

Every one of the roughly 2.1 million human exposure calls answered by US poison control centers each year starts the same way: a Specialist in Poison Information (SPI) — a nurse, pharmacist, or physician with AAPCC certification — runs a structured intake to reconstruct exactly what happened before any recommendation is made. The quality of this first two minutes determines whether the rest of the triage tree even points in the right direction.

• ~2.1M: Human exposure calls/yr (AAPCC NPDS annual report)
• ~72–75%: Managed without ED referral (on-site / non-healthcare-facility)
• 1-800-222-1222: National hotline (routes to regional center by ZIP)
• ~55: US poison centers (covering all 50 states, 24/7/365)

The structured intake dataset every call must capture

The specialist works through a consistent checklist regardless of caller panic level or substance, because incomplete data at intake is the single biggest cause of downstream mis-triage:

Agent identification: • Exact product/brand name — read directly off the container label if available • Active ingredient(s) and concentration — a "cleaning spray" could be dilute quaternary ammonium or concentrated bleach • Formulation — immediate-release vs. sustained/extended-release tablets change the entire kinetic picture • NDC number or UPC lookup when the label is ambiguous or in a foreign language

Exposure parameters: • Amount — number of tablets/pills, volume swallowed, estimated mouthful for children (a "taste" vs. a full swallow) • Route — ingestion, dermal, ocular, inhalational, or parenteral; each has distinct toxic thresholds and first-aid needs • Time since exposure — critical for interpreting symptoms and for deciding whether levels will be meaningful • Whether the ingestion was witnessed, and reliability of the reporting adult

Patient factors: • Age and weight — pediatric mg/kg dosing is the backbone of most threshold calculations • Underlying conditions — renal/hepatic disease, cardiac history, pregnancy • Current medications — potential for interaction or additive toxicity

Presenting symptoms at time of call: • Mental status, vital signs if known, vomiting, rash, respiratory effort • Symptom onset relative to exposure — early symptoms after a "delayed toxin" ingestion can be reassuring and misleading at once

This intake is documented in real time directly into the center's call-management system, which feeds NPDS, so the data captured here is also the raw material for national surveillance later in the case lifecycle.

Substance Identification & Toxicity Database Lookup — Turning a Product Name Into a Threshold

Once the agent is identified, the specialist turns to Poisindex (IBM Watson Health / Merative Micromedex), the toxicology reference used by essentially every poison center in North America, to pull the specific toxic-dose thresholds, mechanism of injury, and expected clinical timeline for that exact substance — because "how much is too much" varies by orders of magnitude across agents.

• 150 mg/kg: Acetaminophen acute threshold (single ingestion, adult/child)
• >20 mg/kg: Elemental iron symptomatic (severe toxicity risk >60 mg/kg)
• Emergency: Button battery ≥20mm (esophageal burn within 2 hrs)
• ~1–2 mL/kg: Ethylene glycol toxic dose (of concentrated (95%) solution)

Representative agents and the thresholds that drive the call

Poisindex monographs give the specialist agent-specific decision points rather than a one-size-fits-all rule:

Acetaminophen (APAP): • Acute single ingestion >150 mg/kg (or >7.5 g in an adult) triggers the need for a 4-hour post-ingestion level plotted on the Rumack-Matthew nomogram • Deceptive because patients are often asymptomatic for the first 24 hours while hepatotoxicity develops silently — this is the classic "delayed toxin" red flag

Iron: • Elemental iron (not tablet weight) >20 mg/kg is the symptomatic threshold; >60 mg/kg risks severe GI hemorrhage, metabolic acidosis, and shock • Radiopaque tablets may be visible on abdominal X-ray, useful when history is unreliable

Button/coin batteries: • A lithium cell ≥20mm lodged in the esophagus can cause a caustic tissue burn within 2 hours via generated current, not leakage — this is a same-day-imaging emergency regardless of symptoms • The National Battery Ingestion Hotline coordinates with local centers for these cases

Ethylene glycol (antifreeze): • Toxic dose ~1–2 mL/kg of concentrated fluid; metabolized to glycolic and oxalic acid causing an elevated osmolar/anion gap acidosis and renal failure • Fomepizole (alcohol dehydrogenase inhibitor) is time-critical — delay materially worsens outcome

Hydrocarbons (gasoline, lamp oil, lighter fluid): • Toxicity is dominated by aspiration risk from low-viscosity products, not systemic absorption — even a small aspirated volume can cause chemical pneumonitis • Because of this, emesis induction is contraindicated regardless of ingested volume

Each lookup also returns expected clinical course and antidote availability, which is what stage 3 risk stratification and stage 4 disposition are ultimately built on.

Risk Stratification — Grading Exposure Against the Poisoning Severity Score

With a threshold in hand, the specialist categorizes the exposure into a working risk tier. The internationally recognized Poisoning Severity Score (PSS), developed by IPCS/EAPCCT/EC, grades outcome from None through Fatal, and US centers use an equivalent minimal/potentially-toxic/clearly-toxic framework layered with dose-relative-to-threshold, symptom trajectory, comorbidities, coingestants, and — critically — how reliable the history actually is.

• No symptoms: PSS 0 · None (dose below toxic threshold)
• Mild, transient: PSS 1 · Minor (resolves without treatment)
• Pronounced/persistent: PSS 2 · Moderate (symptomatic treatment needed)
• Life-threatening: PSS 3–4 · Severe/Fatal (ICU-level care or death)

Inputs that move a case up or down the severity ladder

Risk stratification is not a single lookup — it is a synthesis of five inputs re-evaluated as new information arrives during the call:

1. Dose relative to toxic threshold: • The ratio of reported/estimated dose to the agent's known toxic threshold is the single strongest predictor • Ratios below ~1× threshold trend toward PSS 0–1; ratios several-fold above threshold push toward PSS 2–3

2. Symptom severity and trajectory: • A single episode of vomiting after a mild ingestion is reassuring; altered mental status, seizure, or hemodynamic instability at any dose is an automatic upgrade • Trajectory matters as much as a snapshot — worsening symptoms on a follow-up call escalate the tier even if the initial call looked minor

3. Comorbidities: • Renal or hepatic impairment prolongs clearance of many agents and lowers the effective toxic threshold • Cardiac history raises concern for agents with membrane-stabilizing or QT-prolonging effects

4. Coingestants: • Polydrug or polysubstance exposures (common in intentional overdoses) can have additive or synergistic toxicity that no single-agent threshold captures • Alcohol coingestion changes absorption kinetics and clinical picture for many agents

5. Reliability of the history: • Witnessed pediatric exploratory ingestions with an intact container and known concentration are high-confidence • Unwitnessed ingestions, missing pill counts, or a distressed/unreliable reporter push the specialist toward the more conservative (higher) risk tier by default, since the true dose cannot be bounded

The output of this stage is a working PSS-equivalent grade that becomes the primary input to the stage 4 disposition decision.

Disposition Decision — Home Observation vs. Emergency Department Referral

This is the decision the entire call has been building toward: can this patient be safely watched at home with a scheduled callback, or does the case need an emergency department — or an ambulance dispatched right now? US poison centers manage roughly three-quarters of calls without any ED referral, but the remaining quarter is identified through an explicit red-flag checklist rather than gut feeling.

• ~72–75%: Home observation rate (of all human exposure calls)
• 1–24 hrs: Scheduled callback window (agent-dependent recheck)
• Nomogram: APAP 4-hr level rule (Rumack-Matthew treatment line)
• Immediate: EMS/911 dispatch (for airway/circulatory compromise)

Red flags that override a "home observation" default

Home observation is the default disposition when dose is sub-threshold, the patient is asymptomatic or minimally symptomatic, and the caregiver is reliable — but any of the following findings overrides that default immediately:

Clinical red flags: • Altered mental status — lethargy, agitation, or unresponsiveness beyond baseline • Respiratory depression or significant respiratory distress • Cardiac arrhythmia, syncope, or hemodynamic instability • Seizure activity of any kind

Situational red flags: • Unreliable caregiver — unable to monitor the patient, follow instructions, or call back if symptoms change • No reliable access to emergency transport if the patient decompensates at home

Agent-specific red flags — delayed-onset severe toxicity: • Acetaminophen: even an asymptomatic patient with a supratherapeutic dose needs a timed level and nomogram interpretation — waiting for symptoms means waiting for liver injury • Sustained/extended-release formulations (e.g., SR verapamil, XR bupropion): peak toxicity can be delayed 12–24 hours past what an immediate-release ingestion would suggest, requiring extended observation or ED-based monitoring rather than a short home callback window • Agents needing a specific antidote or lab-confirmed level not available at home (iron levels, osmolar gap for toxic alcohols)

When no red flag is present, the specialist sets a scheduled callback — typically within 1–4 hours for fast-acting agents, up to 24 hours for slower ones — and gives explicit worsening-symptom instructions that would trigger an immediate 911 call rather than waiting for the scheduled check-in. This callback loop is itself a monitoring tool, not just an administrative courtesy: it catches the subset of cases whose trajectory changes after the initial call.

A patient who ingested a supratherapeutic acetaminophen dose can look completely well at hour 2 and still be on a trajectory toward serious hepatotoxicity by hour 36 — which is exactly why disposition for delayed-onset agents is driven by dose and timed labs, not by how the patient looks on the phone right now.

Case Closure, Outcome Tracking, and the Public-Health Surveillance Function of Poison Data

A poison control call does not end when the specialist hangs up. Every case is documented into the National Poison Data System, callback protocols are executed until outcome is known, and — when a case is referred out — the center coordinates with EMS and the receiving ED. Aggregated across ~2 million calls a year, this dataset becomes one of the country's few near-real-time public-health surveillance systems for chemical and drug exposures.

• ~2.1M+: NPDS records/yr (human exposure cases logged)
• Near real-time: Data upload frequency (centers upload continuously to NPDS)
• CDC, FDA, CPSC: Surveillance partners (signal-sharing for hazard alerts)
• Outcome-coded: Case closure (no effect → death, standardized taxonomy)

From individual callback to national early-warning system

Closure and follow-up:

• Every open case carries a scheduled callback until the patient is asymptomatic or has reached definitive care, at which point the case is closed with a standardized outcome code (no effect, minor, moderate, major, death) • Referred cases are tracked through coordination with EMS and the receiving emergency department so the center's record reflects the actual clinical outcome, not just the referral decision • Every exposure — home-managed or ED-referred — is entered into NPDS, maintained by AAPCC, which aggregates data from all US poison centers essentially in real time

Quality and surveillance function:

• Outbreak detection: NPDS has flagged clusters such as contaminated products, mislabeled supplements, and regional cleaning-product misuse spikes days to weeks before they would surface through other public-health channels • New drugs of abuse: emerging synthetic cannabinoids, novel psychoactive substances, and shifts in kratom- or nitazene-related exposures are often first visible as a rising trend in poison center call data before formal epidemiological studies catch up • Product safety hazard signals: single-load laundry detergent packets, hand sanitizer exposures during periods of increased use, and button battery injuries have all been characterized and acted upon (labeling changes, CPSC standards) using NPDS trend data • Federal partners — CDC, FDA, and the Consumer Product Safety Commission — routinely draw on NPDS for signal detection precisely because poison centers are one of the only touchpoints that captures exposures regardless of whether the patient ever sees a clinician

The individual triage decision made in stages 1–4 of a single call is, in aggregate with millions of others, what makes poison control data a genuine national sentinel surveillance system rather than just a call log.

Because NPDS captures cases managed entirely at home alongside those referred to EDs, it sees a far broader slice of real-world exposures than emergency department or hospital-discharge data alone — which is exactly why it can detect an emerging hazard while the affected population is still mostly calling a hotline rather than showing up in an ED.