Healthcare Data API Performance Overhaul
47% API performance improvement for a healthcare data platform serving 500K+ daily requests.
About the client
Healthcare Data API Provider (NDA)
United States
A healthcare interoperability company providing standardised API access to clinical data for 80+ healthcare system integrators. Their platform is a critical dependency for EHR vendors, lab systems, and telehealth apps β meaning performance degradation has immediate downstream consequences for patient care workflows.
Technologies used
Results at a glance
47%
Faster API response
380ms
New median latency
0
SLA breaches post-launch
6 wk
Full remediation timeline
All three SLA breach notices were resolved within 30 days. The at-risk enterprise client renewed their contract at a higher tier. The engineering team now operates with full observability and a load testing suite that catches regressions before they reach production.
Overview
The context
Healthcare data APIs sit at the centre of care delivery workflows: slow responses delay clinical decisions, break EHR integrations, and violate the performance SLAs that are written into every enterprise contract. When latency becomes systemic, the business risk is existential.
We conducted a full performance audit of the existing API layer β from query patterns and index coverage to serialisation overhead and infrastructure configuration β and delivered a prioritised remediation roadmap with measurable outcomes at each stage.
The challenge
The Challenge
Median API response time had degraded from 400ms to over 4 seconds over an 18-month period as data volume grew 10x. The engineering team had patched individual slow queries but lacked the instrumentation to identify systemic patterns β every fix revealed another bottleneck.
Three enterprise clients had issued formal SLA breach notices. One was actively evaluating alternatives. The company needed measurable improvement within 6 weeks or faced losing 35% of ARR.
Our solution
What We Built
We started with comprehensive APM instrumentation using OpenTelemetry β giving the team, for the first time, a full trace from incoming request to database response. Within 48 hours we had identified the three highest-impact bottlenecks: a missing composite index on the core patient query, N+1 patterns in the FHIR resource serialiser, and no caching on reference data that was updated weekly but queried on every request.
We implemented multi-tier caching (Redis L1, CDN L2 for public endpoints), rewrote the serialiser with a streaming approach, added the composite indexes, and refactored the 5 highest-traffic endpoints to eliminate N+1 patterns.
- OpenTelemetry instrumentation across full request lifecycle
- Multi-tier Redis + CDN caching strategy
- Composite index redesign on core query paths
- Streaming FHIR serialiser replacing in-memory assembly
- N+1 elimination on 5 highest-traffic endpoints
- Load testing suite for regression prevention
Outcomes
The results
All three SLA breach notices were resolved within 30 days. The at-risk enterprise client renewed their contract at a higher tier. The engineering team now operates with full observability and a load testing suite that catches regressions before they reach production.
47%
Faster API response
380ms
New median latency
0
SLA breaches post-launch
6 wk
Full remediation timeline
Ready?
Want results like these?
We take on a limited number of engagements. Tell us what you're building.