Hospital equipment tracking

Request, reserve, clean and repair wheelchairs on one live system, so a patient who is ready to move isn’t waiting for a chair.

Project facts

Who it’s for
Therapists, rehabilitation assistants, wheelchair technicians and site administrators in hospitals, across multiple sites and health authorities.
Stage

Stage 3 of 5: Pilot-ready

Lloyd’s role
Sole designer and developer
Year
2025 – 2026
Built with
React, Express, TypeScript, PostgreSQL, Tailwind CSS, Playwright

In short

The problem
Hospitals lose time finding, cleaning and repairing wheelchairs, and have no reliable data on wait times or repair costs.
What changes
Every request, chair and repair on one live system, with automatic escalation and a complete audit trail.
Proof so far
Scored 91/100 in an October 2026 end-to-end audit and demonstrated to hospital staff. Ready for a supervised pilot; not yet in use.
Wheelchair Tracking in 80 secondsThe problem, the journey every chair follows, and the rehab assistant’s dashboard, cut from the overview.1:20 · Captions available

Evidence so far

  • 91/100End-to-end auditOctober 2026: cleared for a supervised pilot on synthetic data
  • In the databaseSite isolationEach health authority and hospital site is separated below the app
  • Every actionAudit trailAn append-only log of who did what, to which chair, and when
  • Initials onlyPatient dataPlus two-step verification for administrators

The problem

In many hospitals, wheelchairs are tracked on whiteboards, in binders and in people’s heads. Requests arrive by phone call, text, messaging apps and email.

Chairs drift between units, wait uncleaned in hallways or go missing. When leaders ask how long patients wait for a chair, or what repairs really cost, there’s no reliable answer.

Who it’s for

  • Therapist

    Needs to
    Request the right chair for a patient and follow it to delivery without chasing anyone.
    Today
    Requests go by phone, text, WhatsApp and email, with no certainty a chair is free.
  • Rehab assistant

    Needs to
    Work one queue that says what to do next: acknowledge, find, check out, clean, return.
    Today
    Finding a free chair means walking the units and asking around.
  • Wheelchair technician

    Needs to
    Run a repair bench with a record of every fault, fix, part and vendor.
    Today
    Broken chairs are reported verbally and forgotten.
  • Leaders and site admins

    Needs to
    Plan capacity and spending from real wait times, use and repair costs.
    Today
    When asked how long patients wait or what repairs cost, there is no reliable answer.

How it works

  1. A therapist requests a chair

    A short, guided form: floor and unit, the type of need and how it should be handled. Patients are recorded by initials only.

  2. A rehab assistant acknowledges and reserves

    A three-step wizard lists only chairs that fit the patient’s size and type, at the patient’s own site.

  3. The chair is checked out and delivered

    It’s linked to the patient’s ticket, and the therapist is notified at every step.

  4. Discharge and cleaning

    After use, every chair goes through cleaning before it returns to the available pool.

  5. Repair when needed

    Broken chairs leave circulation. Technicians record the fix, parts, labour and vendor.

Screens

Captured from the app with demo data. Select a screen to see it larger.

  • Therapist: one button to request a chair
  • The request form takes initials only
  • Rehab assistant: the most urgent task first
  • Only chairs that fit the patient are offered
  • Technician: the repair bench
  • Administrator: the whole site at a glance

Main features

  • Waiting list with automatic matching

    If no chair fits, the patient joins a first-come, first-served list and is matched when one frees up.

  • Escalation built in

    Requests not acknowledged within a day notify the therapist and site administrators automatically.

  • Missing and found

    A missing chair is flagged with its last known location and routed back through cleaning when found.

  • Reports and analytics

    Median wait, requests by department and repair spend by vendor, exportable to CSV, Excel and PDF.

  • Site Team messaging

    Site chat, groups and one-to-one messages inside the platform, so staff don’t coordinate on personal phones.

  • Built for health authorities

    Each authority is its own tenant and each site is isolated, enforced in the database. Every action lands in an append-only log.

Product decisions

  1. Record a named unit or room, not GPS

    Instead of capturing a GPS fix at every status change.

    Indoors, GPS can’t tell a floor or room apart, and it collected staff location data with no operational value. It was built, tested, then retired.

  2. Patient initials only

    Instead of full names and contact details on the request form.

    The team needs to match a chair to a patient, not identify them. The server now refuses full names, and the audit treated accepting them as a serious defect.

  3. Self-hosted sign-in on the app’s own database

    Instead of a vendor identity service, or deciding later.

    It runs anywhere the app runs, from test hosting today to Canadian cloud hosting for real data, with no third party holding staff identities.

Left out on purpose

  • Patient transport routing. The form collected trip details that rehab assistants never saw, so it was removed until there is a real workflow for it.
  • Merging “discharge” and “send for cleaning”. It would save a tap but merge two audit events and change how turnaround is measured, so it waits for a deliberate decision.
  • A celebration animation on submit. Removed in favour of a plain confirmation that suits a clinical setting.

What changed along the way

  • GPS capture retired. Testing showed indoor GPS couldn’t resolve a floor or room; a pre-filled unit or room replaced it.
  • The separate Cleaner role removed. Cleaning moved to whoever has the chair, from the post-use tab, rather than a dedicated role.
  • Backend rebuilt for Canadian hosting. Real hospital data needs Canadian hosting and no vendor identity service, so the server moved to a standard stack with the same data rules and unchanged screens.

How it saves time

  • Patients get chairs sooner, because no request can be quietly forgotten.
  • Every chair and every action is accountable.
  • Repair and usage data turns equipment planning into evidence rather than estimates.
  • Less time spent searching for chairs and chasing requests by phone.

These are intended outcomes. Measured baselines haven’t been taken yet, and figures in the screenshots come from demo data.

How success would be measured

MeasureTarget or resultStatus
Time to process a requestUnder 5 minutes, from about 15Not yet measured
Chairs that can’t be accounted for90% fewer, with every move traceableNot yet measured
Staff using it90% of intended users within 3 monthsNot yet measured
Availability99.9% uptimeNot yet measured

More video

Overview for decision-makersThe problem, the chair’s journey, each role’s screens, messaging and how the platform keeps sites separate.6:54 · Captions available
Demo walkthroughOne request followed through a demo hospital: requested, reserved, delivered, repaired, cleaned and returned to service.4:52 · Captions on screen

How it’s built

Lloyd designed and built the platform as its only developer, from the chair lifecycle and role permissions to the database security, QR workflows and the test suite. The idea came from his years as a rehabilitation assistant, where finding a chair was part of every shift.

Lloyd Dela Cruz, Founder, DYOL Digital Labs. About Lloyd

Interface
React, Vite, React Router, Zustand, Tailwind CSS, Chart.js, QR scanning
Server
Express 5, TypeScript, zod validation, Better Auth with two-step verification for admins
Data
PostgreSQL with row-level security; every status change runs through locked-down database functions
Quality
Vitest and Playwright, plus a scored end-to-end audit before any pilot

Where it is now

  1. Concept
  2. Working prototype
  3. Pilot-readyThis project
  4. Public beta
  5. In use

Demonstrated to hospital staff, not yet in use at any site. An October 2026 end-to-end audit scored it 91 out of 100 and cleared it for a supervised pilot on synthetic or de-identified data. Use with real patient data waits on the steps below.

What’s next

  1. Single sign-on with the health authority’s identity team
  2. Privacy impact assessment, security review and Canadian hosting
  3. An independent penetration test
  4. A supervised pilot at one site on synthetic or de-identified data

A free 20-minute conversation, then an honest assessment. You’ll know what it costs and how long it takes, in writing, at a fixed price.

Or email lloyd.delacruz@outlook.com. We reply within 24 hours.