How to Automate Progeny Clinical

A pedigree is a clinical document that never stops changing

Progeny Clinical is software for genetics services: family history questionnaires, pedigree drawing, validated risk models, test ordering and results, and the analytics a genetics department reports on. It started as Progeny Genetics in 1996 and joined Ambry Genetics in 2015, which is why test ordering runs directly through Ambry with pre-populated requisitions.

Genetics is the one clinic where the person in front of you is not the whole case. A woman comes in because her sister was diagnosed at forty-one. What gets drawn is her mother, her aunts, her cousins, their ages at diagnosis, and the two relatives nobody can put a name to.

That drawing is a living document. A cousin is diagnosed, a relative tests positive, a variant is reclassified three years later, and a pedigree that was accurate when it was drawn is quietly no longer accurate now.

Genetics runs on follow-up nobody is scheduled to do

The appointment lasts an hour. What it starts can run for years.

The pre-visit questionnaire has to come back before the session is worth holding, because a pedigree drawn from a half-filled form turns counselling time into data entry. The test needs a prior authorisation, and genetic testing is the specialty where payers ask the most questions. The laboratory has its own turnaround, and a patient waiting on a result rings the clinic rather than the laboratory.

Then the part with no appointment attached at all. A pathogenic variant means every first-degree relative should be offered testing, and cascade outreach is a list somebody works when they can. A variant of uncertain significance may be reclassified years later, and somebody has to notice. A high-risk patient needs annual imaging from thirty, and the reminder has to come from the service that identified the risk.

None of that has a slot in the diary. It is done between clinics, by the same people who ran them.

Progeny Clinical

Cascade outreach with a memory Relatives offered testing after a pathogenic result, tracked by who was contacted, who replied and how many weeks ago, rather than in somebody's notebook.

Reclassifications that get noticed Variants of uncertain significance rechecked against the public classification records, so a change becomes a named patient to call.

Appeals filed before the window shuts Denied testing collected with the payer's stated reason and the closing date, sorted by which appeal deadline comes first.

Progeny calculates the risk and cannot watch the laboratory

Progeny automates a great deal of the genetics workflow, and a service paying for it should be leaning on all of that.

Patient History Questionnaires go out in a mobile-friendly format for completion before the visit, and the answers build the pedigree instead of being retyped into it. Risk Assessment runs Tyrer-Cuzick, Gail, Claus, PREMM5 and the BayesMendel models including BRCAPRO and MMRPRO against the pedigree without double entry, identifies the missing data a model needs, and saves the result to that person's record or as a timestamped PDF. Clinic notes and patient letters auto-populate from the same information, and Test Ordering places Ambry requisitions pre-populated, with results coming back in.

All of that acts on the record Progeny holds, or travels the route to Ambry it was built with.

The chasing lives elsewhere. A payer's authorisation portal, an appeal filed on that payer's own form, a laboratory outside the Ambry route with its own status page, a public variant database, a hospital's screening booking system: each is a website with a login and a layout of its own.

A family history is a caseload, not a single patient

The jobs with no slot in the diary are exactly the ones a scheduled run is good at.

Prior authorisations for genetic testing checked on the payer's portal each morning and ordered by how long they have been pending, with denials collected alongside the stated reason and the date the appeal window shuts.

Outstanding orders followed on each laboratory's status page, so a result sitting there for a week is not discovered by a telephone call from the patient. Pre-visit questionnaires checked for completion a few days out, so an incomplete one is chased before the appointment rather than during it.

Relatives identified for cascade testing tracked as a proper list: who was offered, who was contacted, who never replied, and how long ago that was. Variants of uncertain significance rechecked against the public classification records, so a reclassification becomes a named patient to contact instead of a paper somebody reads eventually. High-risk patients whose annual screening date has passed, ranked by how far overdue.

The rule is firm here, and genetics is exactly where it should be. Gathering and comparing can run unattended. Changing a record stays with a genetic counsellor or clinician, and so does any contact with a patient or a relative about their own risk: the work arrives written up, and none of it moves until one of them has approved it.

The counsellor keeps the judgement and hands over the chasing

Telling a family what a reclassified variant means is the job. Noticing that it was reclassified is not, and noticing is the part that gets missed.

WebRun is an AI agent that works a real Chrome browser, signed in the way your team signs in. It opens payer portals, laboratory status pages, public variant records and a hospital's booking system, reads what changed since yesterday, and hands back a list with names and dates on it.

It runs on your schedule inside your own private environment, a workflow can be locked to a named list of domains, and sessions are not shared between workflows. Nothing that touches a record, and nothing that reaches a patient or a relative, is sent without a qualified person approving it.

Each workflow below names the sites it opens.