A patient presents with headache. The history and physical examination do not reveal the cause. You take a second look at the vital signs. Oral temperature is 99.3. You have the nurse check the rectal temperature. It is 101.9. What is the sensitivity and specificity of oral temperature for fever?
There are numerous studies on this topic, and unfortunately, the populations studied are far too heterogenous to truly give a range for sensitive and specificity. In some studies the sensitivity is as low as 47%.(Jensen J Adv Nurs 1994)
Correctly identifying a fever can change the workup in certain presentations. During residency I saw a patient with a 99 degree temperature elevation and back pain. The attending physician was sharp to perceive this and rectal temperature turned out to be 101.5. We eventually diagnosed epidural abscess.
Another example might be delirium in the elderly. I recall a nursing home patient who presented in shock with an oral temperature of 98.1. I asked the nurse to check rectal temperature. She was a bit skeptical. The reading was 102.1 and she was surprised and a little disappointed at how unreliable the oral temperature can be.
The lesson is this: oral temperature is not sufficiently sensitive. It might work for screening, but slight abnormalities in high risk presentations may call for the gold standard: rectal temperature.
Take home points:
-Oral and even temporal temperature assessment can have low sensitivity
-Minor oral temperature elevation may call for rectal temperature assessment