Improving reliability and throughput in a high-volume laboratory reporting system.
Electronic lab sample reporting tool: real-time validation and error handling for high-volume lab data entry.
- Frontend engineer · module delivery
- 2023 — 2025
- SeamHealth Group product engineering. Team size not listed.
- Web · enterprise healthcare
- Angular
- TypeScript
- REST APIs
ELSRT is a SeamHealth Group product handling laboratory data. No screenshots, code or data are shown, and the architecture is simplified.
Context
ELSRT (Electronic Lab Sample Reporting Tool) is used to record and report laboratory sample results. Staff enter results in volume, and errors carry consequences downstream.
My responsibility
I owned
- ELSRT modules for real-time validation and error handling during lab data entry
- Optimisation of the high-volume processing path
I collaborated on
- Requirements and acceptance criteria with stakeholders
- Backend services the modules submit to
The challenge
When errors are only caught after submission, or in a manual review step, staff find them late, in batches, after the context is gone. The goal was to catch errors while the person is still on the field — without slowing down people who enter records all day.
Constraints
- Accuracy
- This is healthcare data. A wrong value that passes validation is worse than a slow form.
- Throughput
- Validation can’t add noticeable lag or friction to a high-volume workflow.
- Existing system
- The work shipped as modules inside an existing Angular application.
Architecture
- Data-entry form
- Real-time field validation
- Inline error handling
- REST API
- Reporting
- Errors are reported where they happen — on the field — instead of after a batch has been submitted.
Key decisions
01Validate at entry time instead of relying on later manual verification
- Errors were being found after submission, which made verification slow and manual.
- Keep post-submission manual verification
- Validate in real time during entry
- Real-time validation adds rules to maintain in the frontend and has to stay fast. Manual verification costs staff time and catches errors late.
- Real-time validation and error handling in the data-entry modules.
- The cheapest moment to fix a value is while the person entering it still has the sample in front of them.
- Manual verification time fell by an estimated 30%.
Implementation
- Validation
- Real-time, field-level validation with specific error messages on lab data entry.
- Error handling
- Errors are reported inside the entry workflow, where they can be fixed immediately.
- Performance
- Optimised the high-volume processing path so validation didn’t slow down bulk work.
Production
- Delivery
- Shipped through the team’s Agile process, with code review and acceptance criteria for each module.
Outcome
Real-time validation moved error detection to the moment of entry. The team estimated that manual verification time dropped by about 30%.
- Manual verification time
- −30%
- Team estimate, not instrumented.
Lessons
Validation is a usability feature, not just a data-quality one. A good error message at the right moment saves more time than any later check.
Next iteration
- Instrument it. The 30% is an estimate; I’d add timing events around the verification step to measure the change properly.
- Track which validation rules fire most often — they point to confusing fields or upstream data problems.