France's first private medico-administrative data infrastructure. From protocol to publication in 8 to 12 months. Built on the SNDS (CNIL Deliberation No. 2025-006 of 27 January 2022), data already extracted and structured. CNIL sandbox active since May 2021. Enriched by the Victoria pipeline and our phenotyping algorithms.
Real-world evidence in months, not years. The data is already extracted: after CESREES/CNIL approval, analyses start immediately.
SNDS data already extracted and structured by the Victoria pipeline. After CESREES/CNIL approval, no waiting for CNAM extraction: analyses start immediately.
99% of the French population covered. DCIR + PMSI + Covid vaccine data. 266 variables selected for maximum granularity, reduced to 161 variables (2–3%) in the Magellan data lake — privacy by design.
Clinityx manages the entire regulatory pathway (CESREES, CNIL). Privacy by design: data minimisation, entropy control, inference risk control, noise injection ±10 patients. Only aggregated and anonymised indicators are delivered. Minimum threshold of 11 individuals per aggregate.
Standard access to the SNDS takes approximately 23 months. With Magellan, data is already extracted and structured: after CESREES/CNIL approval, analyses start immediately without waiting for CNAM extraction.
Magellan leverages SNDS data to provide national coverage of 65 million beneficiaries, with deep historical data and individual-level granularity.
All beneficiaries of the general French health insurance scheme. Anonymised longitudinal data with individual follow-up.
10-year rolling window of healthcare consumption. Longitudinal follow-up of patient pathways with continuous updates.
All reimbursed drug dispensations (community and hospital). ATC classification, dosage, treatment duration, adherence.
Public and private hospital stays via PMSI linkage. Primary and associated ICD-10 diagnoses, CCAM procedure codes.
All active LTCs with ICD-10 diagnoses. Identification of chronic conditions and comorbidity tracking.
Clinityx proprietary algorithms for cohort identification, treatment lines, comorbidity scores and care pathways. 266 variables selected from the SNDS, 161 in the Magellan data lake.
Magellan provides access to all SNDS reimbursement data flows, structured and enriched by our phenotyping algorithms.
| Category | Detail | Available |
|---|---|---|
| Demographics | Age, sex, region, municipality (IRIS), health insurance scheme, CMU-C, deprivation index | ✓ |
| Medications | ATC 7 levels, INN, dosage, form, number of packs, dispensation date, prescriber | ✓ |
| Consultations | General practitioner, specialist, act type, CCAM code, date, reimbursed amount | ✓ |
| Hospitalisations | GHM, ICD-10 diagnoses (DP/DR/DAS), CCAM procedures, length of stay, facility type | ✓ |
| LTC & Pathologies | ICD-10 codes for active LTCs, start date, identified chronic conditions | ✓ |
| Biology | Medical biology acts (NABM codes), date and amount. No quantitative results. | ✓ |
| Medical devices | LPP (list of products and services), delivery date, amount | ✓ |
| Phenotyping | Clinityx proprietary algorithms: cohort identification, comorbidity scores, treatment lines | ✓ |
| Deaths | Date of death (linked CepiDC), cause of death (when available) | ✓ |
Magellan applies a three-step data minimisation strategy, in line with privacy by design principles and CNIL recommendations.
Magellan enables the full range of pharmacoepidemiological studies, from disease burden to post-listing studies.
Estimation of disease frequency at national scale. Temporal trends, regional variations.
Patient trajectory analysis: diagnostic delays, treatment sequences, switching, augmentation.
Economic and clinical burden of diseases. Direct costs, comorbidities, preventable hospitalisations.
Comparison of therapeutic strategies under real-world conditions. Propensity score, IPTW, survival analyses.
Detection of safety signals. Treatment-associated risks, adverse events, active comparator design.
Cost-effectiveness, budget impact, post-listing studies for HAS/CEPS dossiers and price negotiations.
From protocol to manuscript, we manage every step with expertise and transparency.
Validation of the research question and available data.
2–4 wks
Protocol writing and preparation of CNIL dossiers.
8–12 wks
Extraction and qualification of data from the SNDS.
4–6 wks
Data preparation and analyses according to the validated protocol.
8–12 wks
Report writing and support for scientific publication.
4–6 wks
Every Magellan study is documented on our transparency portal. Here are some examples of projects conducted on the Magellan EHR database.
Study of antifibrotic treatment persistence in idiopathic pulmonary fibrosis. CNIL authorisation DR-2024-183.
Assessment of long-term outcomes after transcatheter aortic valve implantation (TAVI) in the French population.
Analysis of vaccination coverage, vaccination schedules and post-vaccination follow-up for Covid-19 using SNDS data.
Medium-term quantification of the reoperation rate for complications and recurrence after BSU implantation across the entire French population.
The processing engine that transforms raw SNDS data into structured, actionable variables. Peer-reviewed and published methodology.
Deduplication, validation, management of DCIR/PMSI overlaps and billing splits.
Production of actionable entities: CCAM, ICD-10, LPP, Medications (with DDD), UM, Patients, Consultations, Biological Acts, LTC, GHM. 266 variables selected.
4 families of automated cleaners. End-to-end auditable transformation chain. Validation on reference cohorts (70M Magellan patients).
Data never leaves the secure enclave. Progressive minimisation: 233 → 161 variables. Maximum retention: 10-year rolling window.
SNDS processing pipeline for pharmacoepidemiological research. Published in the Journal of Health Informatics and indexed on PubMed.
DOI 10.1177/14604582251318250 · PubMed PMID 39913942266 selected variables → 161 variables · Querying · Analytics · Aggregated indicators · Privacy by design
A secure 5-step pipeline, from data extraction to results delivery.
Extraction of raw SNDS data via the permanent access authorised by the CNIL.
Victoria pipeline: cleaning, deduplication, linearisation and structuring into 266 actionable variables.
Storage on certified HDS infrastructure (cegedim.cloud). Pseudonymised data, restricted access.
Extraction of a project-specific dataset. Only the variables required by the protocol are accessible.
Aggregated and anonymised indicators. Re-identification risk analysis at end of project to guarantee anonymity.
Comparison between the Magellan pathway and standard SNDS access. Same regulatory scope, but a decisive time saving.
| Criterion | Magellan (Clinityx) | Standard pathway |
|---|---|---|
| Time to results | ~7 months | ~23 months |
| CESREES / CNIL approval | Required — managed by Clinityx | Required — your responsibility |
| CNAM extraction | Immediate — data already extracted | 14 months |
| Pre-extracted & structured data | ✓ 266 variables | ✗ |
| Automated pipeline (Victoria) | ✓ Published & peer-reviewed | ✗ |
| Beneficiaries covered | 65M (99% of population) | 100% (if authorised) |
| Data window | Rolling 10 years | On request |
| French HDS hosting | ✓ cegedim.cloud | CASD / Other |
| Regulatory support | ✓ End-to-end (CESREES, CNIL, Semaphore) | Self-managed |
| Privacy by design | ✓ Aggregated only (threshold 11) | Access to individual data |
| Transparency portal | ✓ Semaphore | ✗ |
| CNIL-authorised projects (2025) | 35+ | — |
Contact our team to assess the feasibility of your project on the SNDS.