PARIS AI, the Polypharmacy Advanced Risk Intelligence System, is a clinical decision-support suite in three parts: medication risk, acute events and recovery.
engine, read in both directions
components across the patient journey
clinical domains
new devices or data collection
Each component is a product in its own right; together they cover the patient journey from medication risk, through the acute event, to recovery.
Risk intelligence for people on several medicines: one actionable risk tier with a defensible recommendation.
How long before a patient crosses a clinical threshold, with the confidence attached, across fourteen specialties.
Recovery tracked across eleven rehabilitation pathways, with setbacks identified before they occur.
PARIS AI AEP answers one question for one patient, in advance: how long before their condition crosses a clinical threshold, and with what confidence. It reads the observations hospitals already record, and is configured to each specialty. The first in product form is Sepsis and Critical Care, delivered as PARIS AI Sepsis AEP.
Warns before a septic patient's physiology crosses the thresholds that define septic shock and organ dysfunction, and supports the timing of resuscitation bundles, vasopressor de-escalation, antibiotic stewardship and step-down from intensive care.
Productised as PARIS AI Sepsis AEPWarns of haemodynamic deterioration in patients with cardiogenic shock, decompensating heart failure and post-operative instability, so that vasopressor initiation, mechanical support and step-down decisions can be planned rather than reacted to.
Warns ahead of glucose and the physiological markers of endocrine crisis reaching treatment thresholds, supporting earlier intervention in diabetic ketoacidosis, thyroid storm and adrenal crisis, and safer glycaemic management and discharge on the wards.
For cancer patients, whose immunosuppression masks the usual warning signs of deterioration. Gives early trajectory warning and the stability evidence needed to discharge or step down immunocompromised patients with confidence.
Warns before blood pressure, temperature and related parameters leave the target ranges set by neurocritical guidelines, so that neuroprotective intervention can be timed ahead of the crisis.
Warns of progression through the stages of acute kidney injury, the approach of dangerous potassium levels and the need for renal replacement therapy, and extends to longer-horizon prediction of chronic kidney disease progression.
Warns of coagulopathy, falling platelet counts and haemorrhage before they reach the thresholds that define haemostatic collapse or transfusion need, giving time to prepare blood products and escalate specialist care.
Warns ahead of the next severity boundary in respiratory failure, so that proning, ventilator adjustment and escalation can be planned, and confirms readiness for extubation and step-down.
Warns before the markers of hepatic decompensation reach the criteria for transplant listing, encephalopathy escalation or organ-failure grading, giving transplant coordinators, hepatologists and intensivists the timing that current scoring systems do not provide.
Warns of the progression of haemorrhagic shock and coagulopathy from the moment of arrival, supporting earlier activation of transfusion protocols, the timing of damage-control surgery and the selection of patients for aortic occlusion.
Drug concentrations follow trajectories in the same way as vital signs. The application warns before a drug level reaches a toxic or subtherapeutic threshold and pairs that with the patient's physiological course, adding a temporal layer to the dosing software already in use.
Obstetric emergencies have precisely defined intervention thresholds. The application warns before a woman's blood pressure, haemodynamics or laboratory markers reach them, giving the obstetric team a preparation window for treatment, delivery decisions, haemorrhage management and sepsis recognition.
The most time-compressed setting in the portfolio. Provides multi-parameter trajectory warning before in-hospital cardiac arrest, during emergency department waits and in overdose and mass-casualty scenarios, and an objective trajectory layer for escalation under Martha's Rule.
Because the platform works from routine bedside observations and needs no laboratory input, it can run on the equipment every facility already has. Covers the infectious, maternal, neonatal and paediatric conditions behind most preventable mortality in these settings, with an offline deployment architecture for tablet and handheld use.
We welcome clinical partners across the suite and its domain applications. Further detail on any application is available on request.