Developing a Clinical Platform that a $32B Medical Device Company Built its Global Community Health Mission on

Medtronic Labs needed clinical software that worked without internet, on a $30 Android phone, across nine countries and compliance frameworks. Ideas2IT built and scaled SPICE from the ground up into a Gates Foundation-backed global platform that has since screened 642,000 patients.

Client

Medtronic LABS

Industry

Healthcare

Service

Data Engineering

Compliance

HIPAA · FDA 21 CFR Pt 11

Team

14 engineers + 2 architects

01 Challenge

Two-thirds of NCD deaths happen in countries where reliable internet and functioning health infrastructure are the exception, not the rule. Medtronic Labs needed SPICE to run on a $30 Android phone without connectivity, across nine countries and compliance frameworks, on one codebase.

02 Solution

Ideas2IT built offline-first as the architectural spine, with on-device risk algorithms and a sync layer that never lost a patient record on reconnect. When nine countries outgrew a monolithic architecture, the team rebuilt SPICE on Spring Boot and Kubernetes so new government integrations could ship without destabilizing the rest of the platform.

03 Outcome

642,000 patients screened, 122,000 lives measurably improved, and $1.7M in grants from the Gates Foundation, McGovern Program, and World Diabetes Foundation. Patient retrieval is 60% faster, concurrent capacity tripled, and deployment time dropped 50%.

Phase 01

Foundation & Core Platform (2019-2021)

Building SPICE 1.0: offline-first, open-source, deployed
across 5 countries

SPICE had to work offline before it could work anywhere else. Ideas2IT built the architectural spine first: a local SQLite store, an on-device CVD and Red Risk algorithm engine, and a sync layer that reconciled records without data loss on reconnect.

On that foundation, the clinical layer took shape:
  1. Dynamic forms adapted to health worker workflows with no medical training required.
  2. Risk scoring ran on-device, surfacing CVD and Red Risk flags without a connection.
  3. A patient dashboard gave community health workers actionable clinical data in the field.

SPICE 1.0 deployed across five countries and was released as open source in 2021, meeting funder requirements while holding up to public scrutiny.

This Phase Produced
  • SPICE 1.0 (offline-first mobile)
  • CVD + Red Risk algorithm engine
  • SQLite local data store + sync
  • Telecounseling module
  • Government interoperability layer
  • SPICE 1.0 open-source release

Phase 02

Microservices & Global Scale (2022–2023)

Rebuilding the platform so nine countries could scale without breaking each other

By 2022, the Gates Foundation program in Sierra Leone, BRAC in Bangladesh, and Virtua Health in the US were each adding compliance requirements a monolithic codebase could not absorb. A fix for Kenya required a full platform release. A module for Bangladesh risked touching clinical logic in Ghana.

Ideas2IT rebuilt SPICE on Spring Boot and Kubernetes, with a HAPI JPA FHIR Server as the canonical data layer:
  1. New country configurations became targeted service deployments instead of full platform releases.
  2. Deployment time dropped 50%. Concurrent user capacity tripled.
  3. SPICE rolled out to Sierra Leone, Bangladesh, and the US without disrupting existing deployments.
This Phase Produced
  • SPICE 2.0 microservices platform
  • FHIR HL7 government integrations
  • Open-source data pipeline
  • SPICE Bangladesh + SPICE US Patient App for Bhutan
  • Kenya Tiberbu NCD+ integration

Phase 03

AI, ML & Data Platform (2023–2024)

Turning five years of patient data into a system that predicts who needs care before they drop out

By 2023, the question shifted from collection to intelligence. The McGovern Program commissioned predictive models to identify patients likely to drop out of care before they did.

Ideas2IT built an end-to-end CDC pipeline on AWS, replicated as a fully open-source stack for organizations outside AWS:
  1. Predictive models gave health workers the ability to intervene before patients dropped out.
  2. NCD coordinators with no data science background could query cohort data in plain language via AWS Bedrock and OpenAI.
  3. The platform was no longer just recording care. It was guiding it.
This Phase Produced
  • Predictive enrollment + retention models
  • Open-source Airflow / Spark pipeline
  • AWS Bedrock GenAI integrationt
  • CFR reporting portal (MOH 740)
  • Shruti device integration
  • Superset migration from QuickSight

60% faster patient information retrieval with a threefold increase in concurrent users and a 50% reduction in deployment time through a smart microservices approach. Their expertise speaks volumes about optimizing efficiency.

Ruchika Singhal

President, Medtronic LABS

The Outcome

Five years. One engineering team. A platform that three global funders trusted enough to expand.

Category Metric Description
Performance 60% Faster patient information retrieval at point of care
Scalability 3× Increase in concurrent users the platform could support
Efficiency 50% Reduction in deployment time via microservices
Global Reach 642K+ Patients screened across 9 countries
Lives Improved 122K+ Enrolled patients who completed NCD care
Grant Funding $1.7M+ Gates Foundation, McGovern Program & World Diabetes Foundation
The technical outcomes were the result of architectural discipline held across five years. The Gates Foundation, McGovern Program, and World Diabetes Foundation funded SPICE as infrastructure, backing a platform with the clinical credibility and open-source posture to earn it. A 12-person pilot became an 80-engineer global platform without a single country deployment breaking. That is what an embedded engineering team looks like when the foundation is built right.