Tuesday, July 7, 2009

Electronic fetal monitoring = FAIL?

This is what I learned in epidemiology. When deciding whether to use a test, you need to look at several characteristics, including:
Specificity: How good it is at correctly identifying negatives (does it successfully identify healthy people?)
Sensitivity: How good it is at correctly identifying positives (does it successfully identify sick people?)
Reliability: How consistent is it - will it give similar results with multiple retests? If it is subjective, do multiple assessors give similar scores?
Validity: Does it actually measure what you are trying to measure?

So how does electronic fetal monitoring measure up, according to this recent New York Times article?

Sensitivity & specificity:

...in more than 99 percent of cases, predictions based on the tracings that the baby would have cerebral palsy have proved wrong.

Reliability:

Doctors differ greatly in how they interpret tracings. In a study in which four obstetricians examined 50 fetal heart rate tracings, they agreed in 22 percent of the cases. Two months later, the same four doctors re-evaluated the same 50 tracings and changed their interpretations on nearly one of every five. Furthermore, when the baby’s outcome is already known, interpretation of the tracings is especially unreliable, the guideline report says.

Validity:

...monitoring the fetus during labor does not affect the risk of cerebral palsy, because 70 percent of cases occur before labor begins and only 4 percent result solely from a mishap during labor and delivery.


So EFM doesn't accurately measure risk to the baby and is interpreted very differently by different clinicians, and even very differently by the same clinicians at different times.

“Honestly, the technology got rolled out before we knew if it worked or not,” Dr. George A. Macones, an obstetrician at Washington University in St. Louis, said in an interview.

Oh, so maybe we should perhaps reduce routine continuous EFM?

Apparently not - just interpret it better, with three new categories of risk:

[ACOG] hopes the revised guidelines will reduce misinterpretations and inconsistencies in the understanding and use of readings on fetal monitors, although experts are not optimistic that the rate of Caesareans will drop.

I don't think it would take an expert to tell you that.

2 comments:

Amy Romano said...

This is a great breakdown. It's good to be reminded that EFM is a diagnostic test, not just some fancy feature of modern labor and delivery.

Any other test that performed this poorly would be ditched right away, and insurance would most likely refuse to pay. Thanks for the post!

publichealthdoula said...

Thanks! I think a lot of the logic is "well, it's all we have". Obviously not a great reason, yet I understand the impulse to use something, anything, even if it's not actually working the way it's supposed to.