OneSourceIT / Healthcare IT & medical imaging

Your imaging should file itself.

If someone in your practice re-types patient names at an imaging machine, matches images to charts by hand, or chases studies that never arrived — you're paying a salary for work a system should do. I build the system: images land in the correct chart, automatically, with proof every day.

Get a fixed quote → The technical detail

10+ years in production Actively used every day in surgery centers Fixed price · no support contract

The problem

Labor is almost every practice's largest cost. Records work is where it quietly piles up.

None of this is clinical work — but in most practices, clinical staff do it every day, at every site. Here's what stops being someone's job:

Retired task · 01

Typing patient details into imaging machines

The schedule loads each machine with the right patient and exam before the visit — including walk-ins added during the day.

Was: hand-typed at the machine, for every patient, with every typo becoming a filing problem later.

Retired task · 02

Matching images to charts by hand

Images file themselves into the correct patient record the same day they're taken. Only genuinely unclear cases reach a person.

Was: someone's daily routine — and the error nobody can afford is filing into the wrong chart.

Retired task · 03

Chasing studies that never arrived

A daily report cross-checks the surgery schedule against what actually filed — so nothing slips through unnoticed.

Was: discovered weeks later, when a chart was opened and the images weren't there.

The economics are the point: a recurring payroll line becomes a one-time fixed price. Built once, it runs without a meter — no per-study fees, no standing support contract.

How it works

In plain terms

The schedule drives it

Your system hands each imaging machine the right patient and exam — automatically, before the visit.

Images file themselves

Pictures land in the correct chart the same day, with no one re-typing names or matching by hand.

Anything unclear goes to a person

If a match isn't certain, it's set aside for staff to place — never guessed. A wrong chart is the one error that can't happen.

You get daily proof

A once-a-day email shows what filed, what didn't, and what needs a look — without any patient detail in it.

Proof, not promises

Built for a real surgical practice. Still running.

Case story · anonymized

Three surgery centers, fluoroscopy C-arms, and no enterprise budget

A multi-site ambulatory surgery group in the Southwest ran fluoroscopy C-arms at three locations. Every patient's details were typed into the machines by hand; every image had to be matched to its chart by a person. The enterprise integration vendors quoted enterprise prices.

I built the loop instead: the schedule populated each C-arm automatically — refreshed every 15 minutes, so walk-ins and same-day additions were covered. Images came back, were matched against the practice's records using a tiered, never-guess matching discipline, and filed into the chart as native studies. Anything uncertain went to a staff queue with a documented workflow — including plain-English troubleshooting guides written for the people who actually used it.

Every filing was logged, every original archived with its integrity verifiable, and every morning a PHI-free status email reconciled the surgery schedule against what actually filed.

Installed in the early 2010s. Still filing images more than a decade later — through staff turnover, schedule changes, and two generations of the system itself.

10+ years

In production

3 sites

Surgery centers on one system

15 min

Worklist refresh — walk-ins covered

Same day

Images in the correct chart

2 generations

Built, then rebuilt better

Built for a clinical setting

The cautious questions, answered first

PHI stays inside your systems

The work happens inside your network and your accounts, with only the access the job needs. Even the daily status email carries no patient detail. I served as a practice's HIPAA privacy officer for two years — these workflows are HIPAA-aware by design, not by afterthought.

It never guesses

An image files automatically only on a unique, certain match. Anything less goes to a person. Filing into the wrong chart is treated as the one unforgivable error.

Documentation your staff can follow

Delivered with written guides for the people who use it — front desk to records staff — plus technical documentation any engineer you hire next can pick up.

No support contract required

Built to run unattended and hand off cleanly. The decade-old installation above has needed no retainer to stay alive — that's the standard.

For your IT team

The technical detail, when you want it

Everything above, at full fidelity — the workflow diagram and the specifics whoever vets this will want to see.

The loop below ran a live multi-site surgical practice. DICOM Modality Worklist populated GE C-arms from the practice-management schedule (an Aprima EHR); studies returned to a Conquest-based DICOM server, were matched with a tiered discipline — patient ID, then external ID, then demographics, unique hit required — and filed into the EHR as native DICOM. The same ingestion platform later absorbed vendor document feeds and two legacy-archive migrations.

Automated step

Human safety net

Proof & reconciliation

Workflow phase

DICOM Modality Worklist generation from the practice-management schedule — cron-driven and on-demand via a staff web GUI.
Tiered patient matching: patient ID → external ID → demographics, unique hit required at every tier; ambiguity always routes to the human queue.
Native DICOM filing into the EHR's attachment stack — studies open from the chart, not from a side system. Where the chart side speaks HL7 instead, the filing step adapts to the interface the EHR exposes.
Header-level date handling with tag fallbacks for modalities that populate DICOM dates inconsistently.
Duplicate suppression via content hashing, and hash-verified archival of every original study.
Daily PHI-free reconciliation against the surgery schedule — trackable internal IDs only, safe for ordinary email.
Reusable ingestion platform: the same architecture absorbed three vendor document feeds and two legacy-archive migrations.
Operational docs included: staff guides for the review queue and modality troubleshooting (e.g., changing C-arm destinations) — not just code docs.

Open source · maintained

Maintainer of `class_dicom.php`

A Composer-installable library enabling native DICOM/PACS work in PHP — reading, writing, converting and transmitting medical images. Public, referenced by other projects, and maintained to this day. The craft behind the consulting, in the open.

$ composer require rbraunm/class_dicom

View on GitHub All open-source work →

Fixed-scope offers

Ready-made ways to start

Each is a bounded project with one price, approved before work begins.

Imaging workflow assessment

A plain-English map of where images, staff time and risk actually go in your practice today — with the highest-value fixes, each carrying its own fixed quote.

Connect your machines to your charts

The full loop above, built for your practice: worklists out, images filed in, a review queue for the uncertain, and daily proof it worked — around the EHR you already run.

Rescue an aging imaging setup

An imaging integration whose vendor vanished — or whose builder did — stabilized, documented, and moved onto a foundation that will keep running.

File the backlog

A drive full of orphaned studies matched against your records and filed into the right charts — with the same never-guess discipline, and a human queue for the rest.

Tell me how images move through your practice. Get a fixed price back.

A short conversation is enough to scope most imaging work — and the quote costs nothing.

Get a fixed quote →

A note on patient data: please keep patient information out of a first message — no names, birth dates, record numbers, or images. Describe the workflow and the systems involved; if the work needs more, a secure channel comes first.