I agree with the above two posters and would add my little rule of thumb. Given any case presentation in the stem of the question followed by the phrase: "What's the next best step," my brain does the following.
I look at the answer choices, be they diagnostic tests, x-rays, meds, etc. If, as said above, the patient is stable and needs no interventions (A,B,C's) then I consider which test has the most "bang for the buck" given the case. I'll try to cook up an example (it probably won't be perfect).
An 78 y/o man presents to the ED from nursing home for confusion. He's oriented to place but not to person or time. He has difficulty concentrating to answer your questions, seeming to nod off intermittently. He feels weak. Vitals: HR: 110 RR: 20 BP: 136/80 Temp: 96.3 SpO2: 98% on RA. PmHx includes: CAD, HTN, DM, and BPH. He is able to respond to deny chest pain or shortness of breath. He has not had cough or headache. He denies surgeries. He is unsure of his medications and whether or not he took them today. What's the next best step in evaluation?
A) CBC
B) Chest CT
C) D-Dimer
D) BNP
E ) Blood Cultures
F ) Consult hospital ethics committee
G) UA
H) Urine Tox
I suppose it's easy to breakdown an example once you've created it, but for the sake of argument, if this sort of question appeared here would be my approach. Taking the history into account I'm thinking infectious process. Looking at each order I wonder which is the most bang for the buck. In doing so I ask myself 1) what's a likely diagnosis and a test that will logically hone the field or give you a helpful result at all given the circumstances, 2) which test will change my management depending on its result, 3) are there easier/cheaper tests than others?
If I order a CBC and it comes back normal, do I send this guy home or change what I was going to do? No. If it comes back with an elevated or depressed WBC count will it change anything? No. He's got enough infectious symptomatolology (sic) to support a treatment as is.
Along those lines blood cultures aren't incorrect to order in a febrile patient, however if you order them "next" it doesn't help your management (unless there's a 3-day waiting period to buy guns and administer antibiotics) so it's not the NEXT best step.
Chest CT may show a pneumonia, P.E. or pleural effusion, but it's expensive and not going to be a logical first step in an elderly, febrile patient. If you were to order the D-Dimer, it will not change your management. Say it comes back high. Does this guy more likely have a P.E. or an occult infection. One could argue that he is tachycardic and confused which "could" be a P.E., but of course in the face of numerous other symptoms o' infection this test, even if positive, will not have a good doctor pumping this guy full of heparin. Similarly the BNP is not a first step. It tries to hammer down the CHF diagnosis for which this guy has risk factors but no findings suggestive of.
It's best to never involve the hospital ethics committee. While they debate, your patient dies. I can think of one practice question that actually set up a scenario such that the ethics committee was the right call. I think it was in COMBANK and I had to give them credit. It was epic. It probably won't happen on the real thing--certainly not on the USMLE.
If I order a UA and it comes back positive for infection (frequent in the experienced, aged population) then it will change what I do. I'll have a probable answer. The stem will give you clues to hang your hat on. In this gentleman, his diabetes and BPH and nursing-home-ness predispose to complicated UTI. A simple, cheap urinalysis (esp w/ culture) goes a long way in this case.
For sake of argument, place Fingerstick Blood Glucose in place of urinalysis and it becomes the best answer. A quick, cheap test that, if negative in an altered mental status patient w/ history of diabetes and poor med control, rules out a dangerous needs-to-be-treated diagnosis in 10 seconds.
I hope all that was coherent and helpful. I am on my way to bed. Anyone can rephrase it more succinctly, I don't mind.