Hemostasis Today

June, 2026
June 2026
M T W T F S S
1234567
891011121314
15161718192021
22232425262728
2930  
Ken Nepple: Postpartum Anemia Is Not Defined by Transfusion
Jun 19, 2026, 11:28

Ken Nepple: Postpartum Anemia Is Not Defined by Transfusion

Ken Nepple, Clinical Professor and Associate Chief Health Information Officer at University of Iowa Health Care, shared Andrew Lundquist‘s post on LinkedIn:

“Personally I wouldn’t jump to ‘never had it’ regarding postpartum anemia diagnosis. This ‘AI study’ from BCBS doesn’t survive the peer review of one person (this dude).

For background, way back in the day when I was first learning about clinical documentation improvement as young staff physician, we looked at several months of data and our urologic surgery service had no document anemia.

The insurance companies never reminded us to document that. I learned that acute blood loss anemia was based on clinical picture, estimated blood loss, lab of H/H, and management and we moved to appropriate documentation.

Fast forward to this BCBS analysis which says ‘cases coded for acute posthemorrhagic anemia, a serious condition that typically requires interventions such as blood transfusions. However, many patients coded with the diagnosis never received those treatments.’

Flat out, not true.

Clinically, patients can have significant blood loss, for example drop hematocrit from 40 down to 22, and then not get a blood transfusion if stable.

Especially if young healthy mothers without medical comorbidity.

Even the BCBS base data says that 6.8% of patients had postpartum anemia and blood transfusion was 1.0%, so transfusion is not ‘typical’.

The BCBS ‘study’ chose to ignore lab values and estimated blood loss, as suspect not available to them. Analysis is taking a very limited and clinically invalid evaluation of this problem.

Not to get blunt, but some of these patients would have had physician witnessed pools of blood and clot quantified as estimated blood loss, then anemia with documented low H/H, and management.

The physician has literally had their blood on their hands, monitored anemia with lab values, made a decision about transfusion versus no transfusion, and guided the patient through their postpartum admission.

And the insurance company says ‘nah, no anemia’.

Free medical advice to BCBS that anemia isn’t defined as transfusion.

The AI headline is easy to chase, but we need to stay grounded in fair analysis.”

Andrew Lundquist, Clinical Director at Nabla, Chief Medical Officer and Clinical AI advisor, shared a post on LinkedIn:

“AI wrote a diagnosis into thousands of women’s charts. Most of them never had it.

Here’s what Blue Cross Blue Shield found when they pulled tens of thousands of maternity records.

At the hospitals that adopted AI coding tools fastest, the rate of patients coded for acute posthemorrhagic anemia (serious bleeding after delivery)went from about 4% to over 12% in under three years.

At the hospitals that didn’t, it sat still.

If that many new mothers were actually hemorrhaging, you’d expect the transfusions to climb too. They didn’t.

So auditors went and read the charts at the hospital with the worst spike.

Fewer than one in five cases met the diagnosis.

The bleeding wasn’t real.

The diagnosis is. It’s in her record now, and it doesn’t come out.

It’ll be there the next time she’s pregnant, the next time an underwriter pulls her history, the next time a clinician opens the chart and makes a call based on something that never happened to her.

I’ve done the thing that lets this happen.

Tuesday afternoon, fourth note of six, the system drafts it, it looks right, I sign. I’d be lying if I said I read every line every time.

So would most people who do this work.

That’s the part nobody wants to sit with.

The tool didn’t fail in that study.

We did. The software offered.

A human accepted. And a document that’s supposed to be true about a person quietly stopped being true.

A chart isn’t a bill. It says: this is what was true about this pateint, on this day.

The only reason to keep a clinician in the loop is that someone has to carry the weight of what gets written down.

Make the machine ten times faster and ten times more certain…it doesn’t move that weight an inch.

Your name is still at the bottom.”

Stay updated with Hemostasis Today.