Most decision support stops at the diagnosis and suggests a consult. The harder question is the one after that: is this drainable, by whom, and will draining it actually fix anything?
Pogosh CDS encodes those decisions as routing triggers — the size of the collection, whether a safe percutaneous window exists, whether the retrograde route is feasible, whether the patient is septic. It is the information the proceduralist would ask for on the phone.
The question is rarely "is there a collection" — it is whether this one is drainable, by whom, and whether drainage will actually achieve source control. Pogosh CDS encodes the threshold and the access question together.
When an obstructed system needs decompressing there is usually more than one way in, and the right one depends on how sick the patient is. That is a judgement normally made by phone.
Bleeding that has failed the first-line service is a distinct routing decision, and the trigger is usually a specific imaging finding rather than a vital sign.
These are the cases where the routing decision and the escalation decision are the same decision.
Not everything interventional is an emergency. A large share of what a proceduralist actually does is access and tissue, and those decisions get missed by decision support that only thinks in diagnoses.
It is not complete. 49 conditions carry procedural routing today. Coverage is deepest in drainage, decompression, embolization and vascular intervention — the areas the author practices — and is still expanding. Other specialties are thinner.
The thresholds are not universal. They are stated per condition because that is how they are used: 2.5 cm for a soft tissue collection, 3 cm for intra-abdominal. Anyone presenting one global number is oversimplifying.
There is no outcome validation. These pathways reflect one physician's practice and the published guidance behind it, not a trial. A clinical validation partner program is beginning. Until then, treat every recommendation as a starting point you overrule freely.
It does not decide. Pogosh CDS names the service and the trigger. Whether this patient gets the drain is a conversation between clinicians, and it should be.
Pogosh CDS was created by a practicing interventional radiologist at GigHz LLC. That is offered here as an explanation rather than a credential. The reason the corpus knows that a sub-2.5 cm collection can be aspirated for culture but will not achieve source control — or that a septic obstructed kidney is a nephrostomy rather than a stent — is that those are decisions the author makes on call. That kind of knowledge does not usually reach a decision support system, because the people who hold it are not usually the people building one.
No. The procedural routing spans interventional radiology, urology, general surgery, gastroenterology and vascular surgery, and several conditions name two services together because the right operator genuinely depends on the patient and the institution. An obstructed infected kidney names both urology and IR. A perforated appendicitis with a large abscess names both general surgery and IR.
The trigger. Generic decision support tends to suggest a consult once a diagnosis is made. Pogosh CDS states the finding that makes the procedure appropriate — the size of the collection, whether a safe percutaneous window exists, whether the retrograde route is feasible, whether the patient is septic — because that is the information the proceduralist is going to ask for anyway.
No, and Pogosh CDS does not present them as such. A soft tissue collection is flagged at 2.5 cm or larger for catheter drainage; intra-abdominal collections such as periappendiceal and diverticular abscesses are flagged at 3 cm or larger. Thresholds are stated per condition rather than as one global rule, because that is how they are actually used.
As of August 2026 the corpus carries procedural routing across 49 conditions. Coverage is deepest in the areas the author practices — drainage, decompression, embolization and vascular intervention — and is still expanding. We would rather state the current coverage than imply it is exhaustive.
Pogosh CDS was created by a practicing interventional radiologist at GigHz LLC. That matters here less as a credential than as an explanation: the reason the corpus knows that a sub-2.5 cm collection can be aspirated for culture but will not achieve source control, or that a septic obstructed kidney is a nephrostomy rather than a stent, is that those are decisions the author makes on call.
Paste a collection, an obstruction, a bleed. See whether the trigger matches how you would have called it.