our work measurable results · proven at scale

Our work

Savant has processed more than 15 million clinical records across 30+ therapeutic areas. Explore how life sciences companies, data providers, registries, and partners use Savant to turn complex clinical records into analysis-ready, source-traceable data.

featured projects

patient journey385K notes · 1,000 TNBC patients clinico-genomics6,000+ patients · with Scipher rare-disease endpoints160K+ notes · ~2 weeks data enrichment325K+ notes · 10,000+ hours saved throughput400K notes · under a week standing registry200+ variables · recurring

Our Partners

We focus on clinical data curation and partner across the broader data ecosystem.

retrieval & linkage

Our December 2025 partnership pairs Datavant's record retrieval, tokenization, and de-identification with Savant's abstraction.

Read the announcement ↗

analytics & evidence

Announced in April 2026 — Manifold brings analytics and health-system integration to the data we curate.

Read the announcement ↗

clinico-genomics

Together with Scipher, we built the most comprehensive clinico-genomic dataset in rheumatoid arthritis — 6,000+ PrismRA-tested patients.

See the case study →

longitudinal ehr data

OMNY’s provider-led EHR network spans 175M+ patient lives. Together we showed how abstraction deepens structured oncology data — starting with a 1,000-patient TNBC cohort.

See the case study →

Case Studies

Real-world curation projects, from focused cohorts to network-scale data.

real-world data partnership

From structured EHR to the full TNBC patient journey.

~385K notes 1,000 patients

Challenge

Key TNBC details including treatment intent, BRCA results, margins, pathologic response live mostly in clinical notes, not structured EHR fields.

Approach

We abstracted ~385K clinical notes across 1,000 patients and layered the results onto OMNY’s structured oncology data.

Result

Abstraction substantially increased the availability of key clinical details, including treatment intent, BRCA results, and surgical detail, and added context such as margin status and pathologic response. Read the case study →

clinico-genomic partnership

The most comprehensive clinico-genomic RA dataset — built together.

6,000+ patients clinical + genomic integration

Challenge

Scipher wanted to pair clinical depth from unstructured records with genomic data across a large PrismRA-tested rheumatoid arthritis population.

Approach

We curated the clinical layer — analysis-ready, source-traceable — and Scipher integrated it with their genomic data.

Result

A combined clinico-genomic dataset spanning 6,000+ PrismRA-tested patients, announced publicly by both companies. Read the announcement →

for life sciences · endpoint curation

Endpoints considered infeasible to abstract at scale — delivered in weeks.

160K+ notes 450 patients ~2 weeks turnaround 3+ months therapy-date correction

Challenge

The study depended on longitudinal organ measurements and accurate therapy timelines — endpoints buried in imaging narratives rather than structured fields. Manual abstraction could not meet the timeline, and the measurement was initially considered infeasible to capture at scale.

Approach

We built the data model with the sponsor, established a source hierarchy prioritizing imaging over exam findings, and resolved relevant evidence across notes and radiology into dated longitudinal events.

Result

Delivered in ~2 weeks, while also identifying structured therapy start dates that differed from the clinical record by 3+ months.

for data aggregators · data enrichment

Clinical depth beyond structured claims, delivered from existing notes.

325K+ notes ~500 patients 10,000+ manual hours saved

Challenge

A downstream client needed treatment-pattern depth far beyond what structured claims could answer — a request that would otherwise require months of manual abstraction.

Approach

We abstracted the required treatment-pattern variables from the provider’s existing notes, mapped them into its commercial data model, and applied integrated automated and expert QC.

Result

Delivered in weeks, saving an estimated 10,000+ hours of manual abstraction.

for data aggregators · network-scale throughput

Note archives converted to data products at network scale.

400K notes 1.2M values 5M citations under a week turnaround

Challenge

Our partner’s oncology notes contained clinically valuable variables that weren’t available in its structured product — but turning them into a scalable data asset required processing at a pace manual abstraction couldn’t support.

Approach

We ran the pipeline at full scale — extraction, longitudinal resolution, and integrated automated and expert QC.

Result

Produced 1.2M structured values and 5M citations from 400K notes in under a week, delivered in our partner’s existing schema.

for registries & life sciences · standing registry

One registry, reusable across research questions.

200+ variables OMOP-aligned schema recurring refreshes

Challenge

Our partner needed publication-grade clinical depth in a disease registry that could stay current as new records accrued — without launching a new abstraction project for every research question.

Approach

We built an OMOP-aligned data model with 200+ clinical variables, ran integrated automated and expert QC, and set up a recurring refresh cadence.

Result

Delivered a 200+ variable, OMOP-aligned registry with recurring refreshes — creating a reusable data asset that has supported multiple publications.

Bring us your use case

From patient finding to endpoint curation and registry refreshes, we’ll help turn your clinical records into the data your team needs.

Request a demo