Kinet-IA
Precision dosing for narrow therapeutic index drugs.
Kinet-IA is HES's flagship technology project. A platform combining Bayesian pharmacokinetics and clinical AI to individualize dosing patient by patient, embedded in the hospital clinical workflow.
Dosing well is still hard in real practice.
Narrow therapeutic index drugs — aminoglycosides, vancomycin, anticoagulants, antiepileptics, immunosuppressants — require patient-by-patient adjustment. Therapeutic drug monitoring exists, but in practice it relies on manual calculations or on tools disconnected from the hospital workflow, while real pharmacokinetic variability escapes the clinical record.
Bayesian PK plus AI, embedded in the clinical workflow.
Kinet-IA combines Bayesian pharmacokinetic models with machine learning trained on real hospital data. The platform delivers a contextualized dosing recommendation — not an isolated suggestion — with full traceability of data, model and uncertainty.
- 01
Validated population models
Curated from the pharmacokinetic literature and reviewed with clinical pharmacokineticists. One model per drug, adapted to real populations.
- 02
AI that learns from the hospital
Machine learning trained on real practice: clinical covariates, renal function, interactions, response dynamics.
- 03
Decision at point of care
Integration with the hospital's systems. The pharmacist or clinician sees the recommendation, the uncertainty, and can adjust without leaving the workflow.
From TDM request to a recommended dose, in minutes.
The loop is designed to fit into real hospital routine, not to replace it.
- 01
Patient data
Kinet-IA pulls the relevant clinical variables from the electronic record and the history of concentrations.
- 02
Bayesian estimation
The pharmacokinetic engine combines the population model with individual data to estimate patient-specific parameters.
- 03
Recommendation and uncertainty
The platform proposes a regimen, simulates expected concentrations and displays the associated confidence interval.
- 04
Clinical decision
The professional accepts, adjusts or rejects. The entire trace is recorded for audit and continuous learning.
Designed as Software as a Medical Device (SaMD) from day one.
Kinet-IA builds medical device and AI governance requirements into the design, rather than retrofitting them at the end.
- Maturity
- TRL 5–6 · validation in a relevant hospital environment
- Medical device
- EU MDR · Regulation 2017/745 · class under evaluation
- AI governance
- Aligned with the EU AI Act · high-risk category for clinical use
- Certification
- CE marking under evaluation · technical & regulatory roadmap in place
- Traceability
- End-to-end audit of data, model and clinical decision
Awarded by SEFH among 57 projects.
In 2025 Kinet-IA won the 1st prize at the VI Innovation Forum of the Spanish Society of Hospital Pharmacy. It has since been featured in generalist and specialist media.
Media reach
In the media
- Redacción Médica
- Onda Cero
- Cadena SER
- See full coverage
A clinical-technical project rooted in the hospital.
Kinet-IA grew inside HES from the clinical experience of the founding team. Technical development, clinical validation and research coordination live in the same project, not in separate silos.
Open to hospitals, scientific societies and academia.
We are adding new centres to the clinical validation process and exploring partnerships with scientific societies, public administration and the pharmaceutical industry. If you want to take part, get in touch.
Propose a collaboration