Century Health

Century Health's Datasets platform provides analysis-ready real-world clinical data registries built in partnership with specialty providers across multiple therapeutic areas. The registries are designed to capture longitudinal clinical depth not typically available in legacy data sources, and are delivered in OMOP CDM format to ensure compatibility with existing analytics environments. Each dataset includes variable-level traceability back to source data, supporting transparency and auditability in research workflows.

A core component of the platform is CHARM (Century Health Retrieval and Abstraction Model), a proprietary extraction and abstraction system that processes unstructured clinical data — including physician notes, imaging reports, and procedure records — to surface disease-specific variables that would otherwise remain inaccessible. CHARM provides full traceability from abstracted variables to their source documents, underpinning the data quality of each registry.

Therapeutic Area Coverage

• Respiratory (Asthma): Captures structured variables including smoking status, allergy test results, oral corticosteroid bursts, ED visits and hospitalisations, biologic agent and start date, ICS/LABA and LAMA regimens, and comorbidities. AI-abstracted variables include FEV1, FEV1/FVC ratio, eosinophil count and percentage, and FeNO. Supports analysis of T2 vs. non-T2 biologic response, biologic switching patterns, early biologic impact on OCS dependence, and exacerbation risk by phenotype.
• Respiratory (COPD): Structured variables include smoking status, OCS bursts, inhaled therapy regimen, ED visits and hospitalisations, biologic agent and start date, and comorbidities. AI-abstracted variables include FEV1, FEV1/FVC ratio, and eosinophil count and percentage. Supports analysis of exacerbation risk by GOLD stage and phenotype, eosinophil-guided biologic response, triple therapy adoption and outcomes, and OCS dependence and cumulative burden.
• Renal (Chronic Kidney Disease): Structured variables include CKD stage (1–5/ESKD), eGFR and eGFR slope, UACR and UPCR, SGLT2 inhibitor use, RAAS agent use, MRA use, immunosuppressive therapy, and comorbidities. AI-abstracted variables include glomerular disease subtype and biopsy findings (chronicity, percentage crescents). Supports analysis of SGLT2 and finerenone real-world eGFR outcomes, GN subtype progression and treatment response, pathways to dialysis and transplant, and proteinuria reduction by therapy class.
• Ophthalmology (Diabetic Retinopathy): Structured variables include DR severity stage, DME status, visual acuity, CST and macular volume, anti-VEGF agent and injection date, systemic diabetes medications, and comorbidities. AI-abstracted variables include Central Subfield Thickness, Macular Cube Volume, Cube Average Thickness, Cystoid Macular Edema, and intraretinal and subretinal fluid. Supports analysis of NPDR-to-PDR progression predictors, anti-VEGF response and treatment intervals, systemic disease control and visual outcomes, and DME development and resolution patterns.
• Ophthalmology (Age-Related Macular Degeneration): Structured variables include AMD subtype, visual acuity, anti-VEGF agent and injection date, and comorbidities. AI-abstracted variables mirror those in the DR registry, covering CST, macular cube volume, cube average thickness, cystoid macular edema, and intraretinal and subretinal fluid. Supports analysis of predictors of dry-to-wet conversion, anti-VEGF treatment intervals and switching, geographic atrophy progression by subtype and location, and fibrosis and vision loss trajectories.
• Neurology (Alzheimer's Disease): Structured variables include CSF and blood biomarkers (Aβ42, tau, p-tau, NfL), APOE genotype, anti-amyloid therapy and start date, cholinesterase inhibitors and memantine, and comorbidities. AI-abstracted variables include MMSE, MoCA, ADAS-Cog, CDR and FAST staging, neuropsychiatric symptom scores, disease subtype and severity, ARIA incidence, and amyloid and tau PET findings. Supports analysis of cognitive trajectories by biomarker profile, real-world ARIA rates and management, anti-amyloid therapy persistence and discontinuation, and early vs. late treatment initiation outcomes.
• Neurology (Multiple Sclerosis): Structured variables include JCV antibody results, infection events, lymphocyte counts and safety labs, encounter frequency and visit cadence, and comorbidities. AI-abstracted variables include estimated EDSS, relapse events, MRI results and progression, DMT agent with initiation and discontinuation reason, DMT prescription date, infusion reactions, concomitant medications, and patient-reported symptoms. Supports analysis of DMT switching patterns and outcomes, high- vs. moderate-efficacy therapy sequences, relapse frequency and disability progression, and real-world DMT discontinuation and persistence.
• Immunology (IBD): Structured variables include disease subtype and behavior, biologic use and start date, JAK inhibitor use, immunomodulator use, corticosteroid use, CRP, ESR, fecal calprotectin, surgical procedure type and date, and comorbidities. AI-abstracted variables include Mayo Endoscopic Score, SES-CD score, Harvey-Bradshaw Index/CDAI, disease location and extent, and stricture, fistula and abscess presence. Supports analysis of biologic and JAK inhibitor sequencing patterns, drug monitoring and dose optimization, disease behavior progression and surgery risk, and endoscopic vs. clinical remission rates.
• Metabolic (MASH): Structured variables include ALT, AST, GGT, bilirubin, FIB-4, APRI, HbA1c, fasting glucose, lipid panel, BMI and weight, GLP-1 agonist use, SGLT2 inhibitor use, and comorbidities. AI-abstracted variables include fibrosis stage, steatosis severity, NAS score, FibroScan results (kPa and CAP score), MRI-PDFF, and ascites and signs of cirrhosis. Supports analysis of fibrosis progression predictors, GLP-1 and SGLT2 real-world response, metabolic comorbidity burden and outcomes, and treatment initiation timing and patterns.

Data Format and Compatibility

• All registry datasets are delivered in OMOP Common Data Model (CDM) format, enabling integration with existing analytics environments without requiring data transformation.
• Variable-level traceability is maintained back to source data across all datasets, supporting audit and validation requirements.
• Data is sourced from community provider networks, capturing real-world clinical practice patterns across specialty settings.

Security, Privacy, and Compliance

• The platform is HIPAA-compliant and SOC 2-certified.
• Clinical data undergoes secure handling and state-of-the-art de-identification prior to use.
• Data is encrypted both in transit and at rest.
• Access controls include role-based access controls (RBAC) and multi-factor authentication (MFA) to restrict data access to authorised users.
• Security operations include continuous monitoring, vulnerability assessments, and disaster recovery protocols.