Sepsis is often misdiagnosed, or just missed. That may change soon.

[Astronuc]

TL;DR Summary

Each year in the United States, sepsis kills more than a quarter million people—more than stroke, diabetes, or lung cancer.

That may soon change. Back in July, Johns Hopkins researchers published a trio of studies in Nature Medicine and npj Digital Medicine showcasing an early-warning system that uses artificial intelligence.

Doctors Still Struggle to Diagnose a Condition That Kills More Americans Than Stroke
https://www.theatlantic.com/health/...telligence-diagnosing-early-detection/671755/

Each year in the United States, sepsis kills more than a quarter million people—more than stroke, diabetes, or lung cancer. One reason for all this carnage is that if sepsis is not detected in time, it’s essentially a death sentence. Consequently, much research has focused on catching sepsis early, but the condition’s complexity has plagued existing clinical support systems—electronic tools that use pop-up alerts to improve patient care—with low accuracy and high rates of false alarm.

That may soon change. Back in July, Johns Hopkins researchers published a trio of studies in Nature Medicine and npj Digital Medicine showcasing an early-warning system that uses artificial intelligence. The system caught 82 percent of sepsis cases and significantly reduced mortality. While AI—in this case, machine learning—has long promised to improve health care, most studies demonstrating its benefits have been conducted using historical data sets.

Given such complexity, over the past decade, doctors have increasingly leaned on electronic health records to help diagnose sepsis, mostly by employing a rules-based criteria—if this, then that.

One such example, known as the SIRS criteria, says a patient is at risk of sepsis if two of four clinical signs—body temperature, heart rate, breathing rate, white-blood-cell count—are abnormal. This broadness, although helpful for catching the various ways sepsis might present itself, triggers countless false positives. Take a patient with a broken arm: “A computerized system might say, ‘Hey, look, fast heart rate, breathing fast.’ It might throw an alert,” says Cyrus Shariat, an ICU physician at Washington Hospital in California. The patient almost certainly doesn’t have sepsis but would nonetheless trip the alarm.

. . . .

It's important to clean wounds (breaks in the skin or inside surface), even if they appear minor.

Interesting application of AI to help diagnose patients.

 

[BillTre]

That's interesting. I didn't know it was a issue.

I used to work for an eye bank collecting corneas and eyeballs from from well documented dead people. Those with sepsis didn't qualify since you don't want to introduce disease with tissues that get transplanted into someone.
By looking at the bodies you could get an idea of whether they had sepsis or not (along with medial notes on the patient). Sepsis would look like a intense blush on the skin and was usually patchy in different areas of the body. In addition, we would always get blood samples from the deceased person which would be screened for a number of diseases in the eye bank lab. Apparently there is not a good sepsis blood test.

When I was working on a ship once a guy got infected blood (bacteria in blood), which (as far as I know as a non-medical person) is distinct from sepsis (bacteria in tissues) and also very serious. His blood vessels were distinctively very bright red, which I never saw with sepsis infected people.
They helicoptered him right out of there as soon as they got the ship in range.

 

[Godot_]

Ummm...

Many - if not most - of those sepsis deaths (often ascending urosepsis) happen in the course of hospital treatment. Often in patients that are very sick to begin with, venerable, have an immune disorder, or are temporarily immunocompromised (like during oncologic chemotherapy or a high-dose corticosteroid treatment for diseases like rheuma). It's a sad aspect of the workload doctors and nurses have to handle, that not-so-rarely the symptoms are registered rather late...

Also, hospital-acquired germs have the nasty tendency to be multiresistant - which means that the standard first line antibiotic therapy often is mostly ineffective, and only after two days, when the antibiogram is there, an effective antibiotic will be administered. Also, these tend to be darn expensive, daily treatment costs being in the four-digit-range in EU and US.

 

[Astronuc]

Sepsis: Woman, 24, dies weeks after getting flu symptoms
https://www.bbc.com/news/uk-wales-65191845

The family of a 24-year-old woman who died of sepsis just weeks after developing a sore throat have said they want others to be aware of the symptoms of the infection.

Bethannie Booth from Merthyr Tydfil said goodbye to her family and even planned her own funeral before being put into a coma.

After discovering red bumps on her face, Bethannie called NHS 111 and was told it was probably acne.
. . .

After going to Royal Glamorgan Hospital in Llantrisant, Rhondda Cynon Taf on 5 March, Bethannie discovered she had strep A and a collapsed lung.

She then developed sepsis, a life-threatening condition where the body's immune system overreacts to an infection and starts to damage tissues and organs.

Bethannie was transferred to Guy's and St Thomas' Hospital in London where she spent two-and-a-half weeks on an ECMO machine, which takes over a person's breathing using an artificial lung.

After showing good progress, Bethannie was transferred back to the Royal Glamorgan Hospital where she re-developed sepsis and died on 31 March surrounded by her family.

 

[Laroxe]

Ummm...

Many - if not most - of those sepsis deaths (often ascending urosepsis) happen in the course of hospital treatment. Often in patients that are very sick to begin with, venerable, have an immune disorder, or are temporarily immunocompromised (like during oncologic chemotherapy or a high-dose corticosteroid treatment for diseases like rheuma). It's a sad aspect of the workload doctors and nurses have to handle, that not-so-rarely the symptoms are registered rather late...

Also, hospital-acquired germs have the nasty tendency to be multiresistant - which means that the standard first line antibiotic therapy often is mostly ineffective, and only after two days, when the antibiogram is there, an effective antibiotic will be administered. Also, these tend to be darn expensive, daily treatment costs being in the four-digit-range in EU and US.

According to the CDC @87% of cases occur before hospitalisation, but all the things you mention are recognised as risk factors but it gets confusing when we recognise sepsis as an abnormal immune response and drugs that suppress this response, like steroids are an important part of the treatment. As with most serious infections it is usually the very young and the very old who are most at risk. Recently, many of the people admitted to ITU with Covid 19 met the criteria for Viral Sepsis.
I think your comment about the workload of health professionals is important, the speed at which sepsis causes changes really needs fairly continuous monitoring to avoid missing important red flags. I thought the idea of A.I. was interesting but might simply be part of the AI hype. I suspect any form of continuous monitoring that identified key symptoms, whether informed by machine learning or A.I. or not is what would make a difference in these cases. In at least one of the articles they describe alerts leading to an immediate clinicians assessment, this is similar to many other forms of monitoring and its already the case that many medical monitoring devices are becoming increasingly multifunctional.