So I'm sitting here on one of the nicest overnight shifts I've had in longest time (hence the reason I'm able to post this lol), and I've had some funky UA results come back on me tonight... ie positive nitirites but all else negative, and are you treating negative UA's but they are symptomatic??? Trying to see what other experienced ER docs use as their guidelines for interpreting and treating the UA's.
Depends on the patient. Also depends on the test. Was it just the dipstick or was it sent for microscopy? If it's nitrites positive, microscopy was done and shows <5 or so WBCs/hpf, then it's pretty negative. Though the NPV is still not 100%, but close, probably around 95% (depending on your population). If microscopy was not done (some labs will only do it if you order it separately or if the urine dipstick is positive) then the NPV drops further. Assuming a sensitivity of 95% and specificity of 40% based on brief search of literature and my experience (which will vary a bit based on population and lab) gives you a negative likelihood ration of 0.13. At that point you have a test that should significantly lower your pretest probability. Now the question is what to do with that.
If it's a woman with typical UTI symptoms, who has had a UTI before, who says her current symptoms are typical of her previous UTI, I'd say the pretest probability is probably around 80%*. That means even with a great negative test with a NLR of 0.13, the post test probability is still 34%. So I would treat that. At that point it makes you wonder if you even needed to send the UA. I would not send it typically, unless I have reason to send a culture as well (previous resistant UTIs, recurrent UTI, etc).
If it's a man who has burning on urination, has not had a UTI before, and it's unclear why he would have a UTI, then the pre test probability in my guesstimate is at most 20% of him having a UTI. That means the post test probability is 3%. I'd probably not treat that. Instead I would send a culture, as well as a urine probe for gonorrhea and chlamydia and have him follow up for test results.
Obviously I am not busting out likelihood ratio calculators on every shift, but it helps to think through the reasoning behind your practice this way when you get a chance.
Also this is one of those things where it helps to keep track of patients you saw and follow them up to see what percent of your patients that you treated or did not treat turned out to have a positive or negative culture. It definitely helps you feel more comfortable with your practice and helps interpret lab tests in your patient population more intelligently.
*Assigning a pretest probability is the hardest part of EBM, and requires a fair amount of voodoo so if you disagree, plug in your own numbers on this calculator:
Sensitivity and Specificity, Likelihood Ratio Calculators - GetTheDiagnosis.org
