Platform
The Mantrana technology platform
APIs for AI. One platform connects hospital data, through open healthcare standards, to every stage of a patient's care and the apps and platforms that use it.
- IHE
- DICOM
- SMART on FHIR
- HL7 FHIR
Powered by Mantrana
Mantrana enables these capabilities, so that research teams can use them as a public good.
EHR data sources
- HIS
- Pathology lab
- Radiology lab
- Wearables
- Pharmacy
- Claims
Swipe to see all three stages
Stakeholder layer
- NHA
- SHA
- EHCP
- DHU
- PO
Layer
Enrolment stage
14 parameters
Treatment stage
25 parameters
Claim / utilisation stage
18 parameters
Data
Eligibility criteria and demographic details
Clinical records, hospitalisation data
Claims, utilisation, payment history
Service
Verification, registration, e-Card
Pre-authorisation, admission, treatment
Claim processing, settlement
Modes
Trust / insurance / mixed
Trust / insurance / mixed
Trust / insurance / mixed
Linguistic
Multilingual support
Regional communication
Multilingual claims
Protection
Authentication data of the beneficiaries
Data security, compliance, confidentiality
Privacy, financial security, accountability
Data mining
Awareness pattern discovery, behavioural pattern mining, beneficiary segmentation based on services
Sentiment analysis, opinion mining, anomaly detection in service delivery
Fraud detection during claims processing, utilisation and payment trend analysis
Healthcare providers
Third-party apps
Two-way
- Care provider app
- Care consumer app
External data consumers
Over HL7 FHIR
- Research platforms
- Diagnostics platforms
Interoperability and integration services
- IHE
- DICOM
- SMART on FHIR
- HL7 FHIR
Link
- Python script
- Interface engine
- Interface engine + data store
- DICOM connector
On-premise
- HIS
- PACS / RIS
- Lab
- Local IT systems
Platform capability · In progress
A healthcare databank for researchers
The team is architecting a healthcare databank that will give researchers access to datasets of many types, from one place.
for the Sri Sathya Sai Healthcare Databank, which links biospecimens with clinical, imaging, multiomic and long-term outcome data.
Open to collaboration
Join the databank project
Mantrana is open to collaborations, and invites research scholars, M.Tech graduates and MS graduates to join the team and be part of this multi-year project.
- Research scholars
- M.Tech graduates
- MS graduates
Multi-year project
Databank pipeline · Illustrative
From sample and record to research-ready data
Samples and data in 10
- Participant consente-Consent, scope, tiers
- Clinical recordsEMR, diagnoses, medications, procedures
- Biospecimen accessionLIMS, barcode, aliquots, freezer
- Pathology and lab reportsDiagnosticReport, Observation
- ImagingDICOM: retinal, cardiac, radiology
- Raw sequencing readsFASTQ: WGS, WES, RNA-seq
- Proteomics and metabolomicsmzML, mass-spec runs
- AMR isolate dataCulture, AST, isolate genomes
- Longitudinal follow-upOutcomes, recurrence, survival
- Participant identityPseudonymised ID, held separately
Pipeline 12 steps · 4 phases
-
Phase 1
Ingest and register
- 01 Consent checkEvery record and sample gated by consent scope
- 02 PseudonymiseIdentity split from data; study IDs issued
- 03 Accession and QCBarcode, sample integrity and chain of custody
-
Phase 2
Harmonise clinical data
- 04 Map to standardsSNOMED CT, LOINC and ICD-10 into OMOP and FHIR
- 05 Link sample to phenotypeSpecimens joined to episodes, imaging and outcomes
-
Phase 3
Genomic and omics processing
- 06 Read QC and alignmentFASTQ to BAM against the reference genome
- 07 Variant callingSNVs, indels, CNVs and SVs into joint VCFs
- 08 AnnotationClinical significance and Indian population frequencies
- 09 Expression and proteomicsTranscript and protein abundance matrices
- 10 Pathogen genomicsAMR genes, lineages and phylogeny for isolates
-
Phase 4
Curate and release
- 11 Feature storeVersioned, analysis-ready datasets for AI and ML
- 12 Governed releaseTiered access, output review, provenance logged
Research-ready out 5
- Analysis-ready cohortsOMOP tables linked to the live sample inventory
- Variant and genotype datasetsAnnotated VCFs and genotype matrices for association studies
- Multiomic matricesExpression, proteomic and metabolomic profiles per sample
- AMR and pathogen datasetsResistance genes, lineages and outbreak clusters
- AI training setsDe-identified, labelled multimodal data with provenance
Data lake behind the databank
- Service layerFor research and innovation Secure workspaces, governed access, AI and diagnostics pipelines
- Storage layerTo approved standards OMOP and FHIR-aligned clinical, imaging and multiomic data
- Ingestion layerData ingestion Clinical records, LIMS, imaging and sequencing feeds
Illustrative; inputs vary by programme and study. Pipeline and data lake as published by the Sri Sathya Sai Biobank and Research Hub.
How it fits together
Reading the platform
-
Connect hospital data
Records flow in from hospital information systems, pathology and radiology labs, wearables, pharmacy and claims, over HL7 FHIR and DICOM.
-
Speak open standards
An interoperability layer built on IHE, DICOM, SMART on FHIR and HL7 FHIR links on-premise HIS, PACS/RIS, lab and local IT systems.
-
Follow the care journey
Six layers, from data and services to language, protection and data mining, organise information across enrolment, treatment and claims.
-
Serve apps and research
APIs feed care provider and care consumer apps two ways, and share data with research and diagnostics platforms.
Work with us
For collaboration, internship coordination, and institutional engagement, please connect through the official Mantrana channels.