Clinical Methodology

How SickPet.ai Triage Works

SickPet.ai, the veterinarian-led AI pet triage and emergency-vet locator, provides AI-assisted pet triage for educational purposes. This page explains exactly what information we collect, how urgency levels are defined, which red flags trigger emergency escalation, and how our protocols are clinically reviewed. We believe transparency builds trust, especially when health is at stake.

What Information We Collect

Every assessment begins with structured information about your pet and the situation. The more detail you provide, the more precise the urgency assessment becomes. We collect:

  • Pet profile: species, breed, age (years and months), weight, sex, and reproductive status.
  • Symptoms: a free-text description of what you are observing, plus structured fields for symptom duration.
  • Additional context: anything else you think is relevant, including recent changes in behaviour, diet, environment, or medication.
  • Location: postcode, city, and country, used to find nearby open clinics if emergency care is recommended.
  • Photo (optional): a photograph of the affected area, which can help the AI assess visible symptoms such as swelling, discharge, or coat condition.

How Urgency Levels Are Defined

Every assessment produces a five-level urgency rating. Each level maps to a recommended timeframe for action:

Red - EmergencyImmediate veterinary attention required. Life-threatening symptoms such as difficulty breathing, severe bleeding, collapse, or known toxin ingestion. Go to the nearest emergency vet now.
Orange - UrgentSee a vet within hours. Potentially serious symptoms that could worsen rapidly. Contact a vet today or visit an emergency clinic if after hours.
Yellow - SoonSee a vet within 24-48 hours. Symptoms that need professional attention but are not immediately life-threatening. Book an appointment with your regular vet.
Blue - MonitorMonitor closely at home. Mild symptoms that may resolve on their own. Watch for worsening signs and contact a vet if things change.
Green - Low ConcernNo urgent action needed. The symptoms described appear low-risk. Continue normal care and routine veterinary check-ups.

Red Flags That Automatically Trigger Emergency Escalation

Certain symptoms are recognised across veterinary medicine as immediate emergencies. If any of the following are detected in the assessment input, the system will automatically escalate to a red urgency rating regardless of other factors:

Difficulty breathing or choking
Pale or white gums (possible internal bleeding or shock)
Severe or uncontrolled bleeding
Collapse, unconsciousness, or seizures
Known ingestion of a lethal toxin (lilies in cats, xylitol, antifreeze, rat poison)
Inability to urine (especially male cats)
Severe bloating with attempted vomiting (possible GDV)
Eye prolapse or severe eye trauma
Heatstroke with body temperature above 41C / 106F
Severe burns or electrical shock
Rabbit not eating for 12+ hours (GI stasis emergency)
Snake bite or spider envenomation

How Species, Age, Weight, and History Influence Results

Species: The same symptom can mean very different things in different species. For example, a dog not eating for a day is often a minor concern, but a rabbit not eating for 12 hours is a life-threatening emergency. The AI adjusts its urgency interpretation based on species-specific physiology.

Age: Very young animals (kittens, puppies) and very old animals are more vulnerable to dehydration, hypoglycemia, and infection. The AI weighs age as a risk multiplier, meaning the same symptom may receive a higher urgency rating in a puppy or a 15-year-old cat than in a healthy adult.

Weight: Weight affects medication toxicity thresholds, anaesthetic risk, and the clinical significance of appetite changes. For toxin ingestion assessments, weight is used to calculate potential toxic dose exposure.

Medical history: Pre-existing conditions, current medications, and previous surgeries all influence risk. An animal with a history of urinary tract issues presenting with straining is treated as higher urgency than one with no such history.

How Toxins and Ingested Amounts Are Assessed

When a toxin ingestion is reported, the AI identifies the substance, estimates the potential amount ingested, and cross-references this against known toxic thresholds for the pet's species and weight. Where a toxic dose calculator is available (such as our chocolate toxicity calculator), we provide a specific mg/kg estimate.

The system always errs on the side of caution. If the amount ingested is unknown, or if the substance is considered potentially lethal at any dose for the species (such as lilies in cats), the assessment is automatically escalated to emergency. We also provide direct links to relevant poison control resources and nearby emergency clinics.

Veterinary References Used

Our triage protocols are informed by established veterinary references and guidelines, including:

  • Veterinary Emergency Medicine textbooks and peer-reviewed clinical guidelines
  • ASPCA Animal Poison Control toxicity databases
  • Veterinary Information Network (VIN) clinical resources
  • Species-specific veterinary formularies and toxic dose references
  • Published triage protocols from veterinary teaching hospitals

How Protocols Are Clinically Reviewed

Every triage protocol and urgency classification rule is reviewed by a licensed veterinarian before it is deployed. Our clinical reviewers include practitioners with experience in emergency medicine, feline medicine, and exotic animal care.

Protocols are reviewed for accuracy of symptom-to-urgency mappings, appropriate conservative bias (erring toward escalation when uncertain), and species-specific considerations. Any protocol change requires sign-off from at least one reviewing veterinarian.

How Frequently Protocols Are Updated

Protocols are reviewed on a regular schedule and whenever new clinical guidelines are published. Our editorial team monitors veterinary literature and poison control advisories for any changes that affect existing protocols. Each article and protocol page displays its last clinical review date so users can see how current the information is.

What the AI Can and Cannot Determine

The AI Can

  • Assess the likely urgency of reported symptoms
  • Identify potential emergency red flags
  • Flag potential toxin exposure risks
  • Provide breed-specific screening recommendations
  • Suggest monitoring items and warning signs
  • Find nearby open veterinary clinics

The AI Cannot

  • Diagnose a specific disease or condition
  • Prescribe medication or treatment
  • Replace a physical veterinary examination
  • Interpret laboratory test results
  • Perform surgery or hands-on procedures
  • Provide a definitive prognosis

How Uncertainty Is Communicated

When the AI is uncertain about the urgency level, it communicates this clearly in the assessment output. Rather than providing a single false-precise answer, the assessment explains the range of possible concerns and why a particular urgency level was chosen. Where information is missing or ambiguous, the system defaults to a more conservative (higher) urgency rating and advises the user to seek veterinary confirmation.

Conservative bias: When the AI cannot determine whether a situation is safe or dangerous, it always errs on the side of caution and recommends veterinary contact. We would rather flag a false emergency than miss a real one.

How Users Are Directed to an Emergency Clinic

When an assessment returns a red or orange urgency rating, the system automatically searches for nearby open veterinary clinics using the user's provided location. Results are prioritised by proximity and whether the clinic is currently open. Users can search for emergency vets at any time, regardless of assessment outcome.

Validation and Quality Metrics

We are committed to ongoing quality measurement and transparency. The following metrics are tracked and updated regularly:

Completed triage sessions

Tracked

Total assessments run since launch

Emergency escalation rate

Tracked

Percentage of assessments escalated to red urgency

Clinical pathways reviewed

Tracked

Number of symptom-to-urgency protocols clinically validated

Veterinary agreement rate

Measured

Rate of agreement between AI urgency classification and reviewing veterinarians on sample cases

Emergency case sensitivity

Measured

Percentage of true emergencies correctly identified as red or orange

Regions covered by clinic locator

Tracked

Number of countries and regions with vet locator coverage

Average assessment time

Tracked

Mean time from starting an assessment to receiving a result

Under-triage rate

Monitored

Cases where the AI rated lower than a veterinarian would have, tracked for protocol improvement

Planned validation study: A formal retrospective validation study comparing AI urgency classifications against veterinary case outcomes is planned. Results will be published on this page to support continuous improvement and clinical credibility.

Editorial and Clinical Review Responsibilities

Content authorship: Each article and protocol is authored or reviewed by a veterinarian with relevant clinical experience. Articles display the author and clinical reviewer name and credentials.

Clinical review cadence: Every published article and triage protocol undergoes clinical review at least annually, and more frequently if new evidence or guidelines emerge.

Corrections: If an error is identified in any article or protocol, it is corrected promptly and a correction note is published. Users can report errors at any time.

Methodology FAQ

Is SickPet.ai a replacement for a veterinarian?+

No. SickPet.ai is an AI-assisted triage and decision-support tool. It helps you judge how urgent a situation is and what to do next, but it cannot diagnose, prescribe, or replace a hands-on veterinary examination. For anything serious, always contact a vet.

How are the five urgency levels defined?+

Each assessment returns one of five levels, Red (emergency, act now), Orange (urgent, within hours), Yellow (soon, within 24–48 hours), Blue (monitor closely at home), and Green (low concern, routine care). Each level maps to a recommended timeframe for action.

Who reviews the triage protocols?+

Every protocol and urgency rule is reviewed by a licensed veterinarian before deployment. The lead veterinary editor is Dr. Olja Jovovich, BVSc, MANZCVS (Feline Medicine), and clinical review is overseen by Dr. Mark Edwards, BVSc, MRCVS.

How does SickPet.ai handle toxin ingestion?+

The AI identifies the substance, estimates the amount ingested, and cross-references known toxic thresholds for the pet's species and weight. If the amount is unknown or the substance is lethal at any dose for the species (such as lilies in cats), the assessment is automatically escalated to emergency.

How often are protocols updated?+

Protocols are reviewed on a regular schedule and whenever new clinical guidelines are published. Each article and protocol page displays its last clinical review date so you can see how current the information is.

Ready to assess your pet?

Start a free AI triage assessment and get an urgency rating in minutes.

Start Free Assessment

Disclaimer: SickPet.ai provides AI-assisted triage for educational and informational purposes only. It is not a substitute for professional veterinary diagnosis or treatment. Always consult a licensed veterinarian for proper diagnosis and medical care of your pet.