Caretrack
Web Application · Healthcare
Caretrack
Scheduling and records for multi-site clinics, with a complete audit trail built in.
6sites, one rota
Web application development for data-intensive products
Dashboards, portals and internal tools engineered for performance at scale, from thousands of records to millions.
- Dashboards
- Customer portals
- Internal tools
- Marketplaces
The challenge
Common challenges.
Internal software is often replaced because it has become too slow to use. Queries that perform well during development can degrade significantly as data volumes grow, and resolving this typically requires changes to how the application retrieves data.
Many organisations rely on spreadsheets for critical processes. These often lack validation, history and clear ownership, and their business rules are spread across formulas and individual knowledge. Replacing them begins with documenting and agreeing those rules.
Permissions become complex as portals grow. Resellers, regional managers, auditors and support staff each need different levels of access. A permission model designed at the outset keeps these rules consistent and maintainable.
What we deliver
Deliverables.
Everything listed below is handed over to you at the end of the engagement.
A scalable data model
Schema, indexes and access patterns designed for the queries your product runs and for the data volumes expected over the next three years.
Fast interfaces
Server-side pagination, filtering and sorting, streamed rendering and virtualised tables, with defined performance budgets.
A robust permission model
Roles, scopes and record-level rules enforced at the data layer, so users only receive data they are authorised to see.
Reliable integrations
Connections to your existing systems with retries, idempotency and a visible sync history.
Reporting and export
CSV exports, scheduled reports and audit trails, included as part of the core scope.
Process
Project timeline.
Typical timelines. Exact durations are confirmed during the Discover stage.
- 01
Process discovery
1–2 weeks
We work with your team to document current processes, including exceptions and existing spreadsheets, and agree the business rules.
- 02
Data and access design
2 weeks
Schema, indexes, query patterns and the permission model, validated against realistic data volumes.
- 03
Build
8–16 weeks
Two-week sprints against a database seeded at production scale, so performance issues are identified during development.
- 04
Migration and cutover
2–3 weeks
Historical data is imported and reconciled, systems run in parallel where appropriate, and a tested rollback plan is in place.
Technology
Technology stack.
Application
- TypeScript
- Next.js
- React
- TanStack Query
Data is fetched on the server, so large datasets are not sent to the browser.
Data
- PostgreSQL
- Redis
- ClickHouse where analytical
Transactional and analytical workloads are separated where needed to keep dashboards responsive.
Quality
- Playwright
- Vitest
- Seeded fixtures at scale
Test databases are seeded to production volumes so performance is tested realistically.
Why data-intensive applications slow down, and how to prevent it
Performance problems in business applications rarely come from the framework. They usually come from a small number of data access patterns that are not visible with small development datasets but become significant in production.
The most common is the N+1 query, where a page loads a list of records and then makes an additional query for each one. With a small dataset this is imperceptible; with hundreds of records it results in hundreds of database round trips.
The second is loading entire datasets and filtering them in the browser. The solution is server-side pagination, filtering and sorting, so the browser requests only the data it needs.
The third is missing or ineffective indexes. An index that does not match the columns a query filters and sorts by will not be used, and the database will scan the full table instead. Query execution plans identify this quickly.
The most effective prevention is seeding development and test databases to the data volumes expected in production. Performance issues are then identified and resolved during development.
Related work
Related projects.
Caretrack
Kiosk
Web Application · Commerce
Kiosk
A storefront and its merchant console, sharing a single product catalogue.
2faces, one system
Engagement models
Flexible ways to work with us.
Pricing depends on scope and is provided after an initial consultation.
Fixed-scope project
Replacing a known process where requirements can be defined before development.
Considerations
A separately priced discovery phase confirms requirements before the fixed scope is agreed.
Dedicated team
A portal or platform with a long-term roadmap, working alongside your internal team.
Considerations
Requires a product owner on your side who can make decisions each week.
Retainer
An existing application that needs continuous improvement and support.
Considerations
Best suited to maintenance and incremental features. Major new modules are scoped as separate projects.
Questions
Frequently asked questions.
- Our dashboard is slow. Does it need to be rebuilt?
- Usually not. In most cases the interface is sound and the issue lies in data access, such as missing indexes, N+1 queries or client-side filtering. A one-week performance audit identifies the causes with measurements and recommends the most cost-effective fix.
- Can you work with our existing database?
- Yes. We can build on your existing schema and improve it incrementally, without requiring a migration at the outset. Where the current model limits the product, we provide specific recommendations and costs for you to decide on.
- How do you handle a migration from the current system?
- Historical data is first imported into a staging environment and reconciled against the source. Where appropriate, both systems run in parallel for a period. Cutover takes place during a low-traffic window with a tested rollback plan, and you receive the reconciliation report beforehand.
- What about offline or unreliable connections?
- For browser-based tools, optimistic updates and robust error handling address most connectivity issues cost-effectively. Where users regularly work without a connection, a mobile application with local data storage and synchronisation is usually the better solution.
- Do you build the internal admin side as well?
- Yes. Administration and support tools are part of the core scope and are developed early in the project.
- How do you keep it fast as data grows?
- Test and staging databases are seeded to projected data volumes, queries are reviewed using execution plans, and performance budgets are enforced in the build pipeline so regressions are caught before release.
Let’s discuss your project.
Book a free 30-minute scoping call to discuss feasibility, budget and timeline.