Clinical decision support
AI-powered alerts that surface relevant patient history, flag potential drug interactions, and highlight abnormal trends, inside the visit, while you're working.
How it works
Clinical decision support is two on-demand, advisory-only tools on the patient chart, not a background watcher. Nothing runs automatically as you type; you ask a specific question and get a specific answer, and nothing here ever writes to the patient's record.
- Drug interaction check: type a drug you're considering, and Chartling checks it against the patient's current active medications for clinically meaningful interactions (species-specific sensitivities like NSAIDs in cats or MDR1 breeds and ivermectin are considered).
- Suggest differentials: enter the presenting complaint (and vitals, optionally), and Chartling proposes a short, prioritized list of differential diagnoses with a rationale and suggested next step for each.
Every result carries a confidence badge and the same fixed disclaimer: advisory only. Verify against current references and your clinical judgment before acting. Each check consumes one AI action against your plan.
These tools never block signing, never auto-populate the SOAP, and never write anything on their own. You read the result and decide what (if anything) to do with it.
Drug interaction check
Type a candidate drug name and click Check. Chartling pulls the patient's species, breed, and current active medications, cross-references your formulary's own interaction/contraindication notes where they exist, and returns:
- An overall risk level: none, minor, moderate, or severe
- Each interacting pair, with its severity, the concern in plain language, and a concrete recommendation (monitor, adjust dose, avoid)
- A confidence level for the check as a whole
Requires the rx.write permission, the same one that gates writing a prescription.
Suggest differentials
Enter the presenting complaint and, optionally, vitals. Chartling returns up to five differentials ordered by likelihood given the patient's signalment, each with a rationale and the next step that would confirm or rule it out. This requires only patient.read, so any staff member who can view the chart can use it, not just the prescribing vet.
How results are grounded
Both tools are scoped to one patient at a time. The model only sees that patient's species, signalment, active medications, and (for differentials) what you typed for the complaint and vitals. Weight is always passed to the model in kilograms regardless of your clinic's display unit, since dosing reasoning is mg/kg.
CDS only knows about medications actually recorded in Chartling. If a patient is on something prescribed elsewhere that isn't in their Meds tab, the interaction check can't see it. Enter any known external medications first.
CDS is available on Intelligence and Practice tiers (feature ai.cds). It is not available on Operations-only plans.
FAQ
You ask it something. There's no ambient monitoring of the SOAP as you type, no standing alerts panel, and nothing that flags overdue vaccines or long-term trends on its own. That's a different surface (see patient risk scoring on the Patients page for the closest thing, a summary layer that's also informational-only). CDS itself is the two tools above, used when you choose to use them.
No. Nothing about these tools touches the sign/finalize flow. They're a read-only lookup you can run (or not) at any point while charting.
The AI action itself is metered against your plan's usage the same as any other AI call. There isn't a per-check audit-log entry recording which staff member ran which check.
It only knows about medications recorded in Chartling. If a patient is on something from another clinic that isn't in their Meds tab, that interaction won't be detected. Enter any known external medications there first.