Is there a role for empiric scabies treatment in itchy patients without clear clinical signs of scabies? 

I have seen many astute clinicians miss scabies or consider it after months of treating for AD with biologics and other systemics (including myself). Though this is very rare my personal rule of thumb is to treat with oral ivermectin if at any point patient is not responding the way I expect to treatment. 

So we see scabies more often.

So I will treat a patient and family empirically if either they or I have a concern - with topicals.

1. Case of girl at University of Michigan who has been seen by derm and heme onc - and halfway through her chemo for leukemia cutis her rash was getting worse and unbearable - scabies was found on biopsy done for ruling out drug rash and was also the cause of her abnormal blood counts.
2. We have had cases where we have treated empirically in kids with no rash but new onset itching. We have had two cases where later the family would call and let us know their son was doing great and their cousins were diagnosed - they had gotten it on a family trip and the second it was from staying at a friend's house after school.

Now for cases that are where you favor delusions of parasitosis - we will treat empirically one round and do a biopsy of one lesion - with the agreement in the chart if both of those fail to find an infestation the patient will be willing to consider other options.

In Southern Delaware, we find scabies commonly. 

Our worst problems were on two occasions when patients were treated by primary care and dermatology physicians for 1 year with steroids, with a diagnosis of atopic dermatitis. In both cases, those patients were eventually hospitalized at my institution for other reasons, and found to have Norwegian Scabies, resulting in the infestation of LARGE numbers of nurses, physical therapists, and other workers (more than 30 people) by the time the diagnosis was made.

As an inpatient ID doc, if a patient has a pruritic rash and any family member with the same, he is isolated and treated, regardless of skin scraping results, regardless of presence or absence of classic burrows. Rarely we might treat empirically even without a symptomatic close contact. 

The treatment is a single dose of topical permethrin, and in the morning isolation can be stopped. And this allays a lot of understandable worries from the nursing staff, especially given our past experience.

Unfortunately, we do not have any diagnostic test capable of ruling OUT scabies with certainty - scrapings and biopsies lack enough sensitivity, hence the occasional need for empiric treatment. Like all things, this depends upon your local prevalence of the disease and whether or not you've been "burnt" before as we have.

Rarely, if ever unless a patient has exposure to a partner/family member with documented scabies or resides in a nursing home with an outbreak. Way too many "scabies" are diagnosed/treated by FPs, PAs, and NPs as part of a shotgun regimen.

Never. Unless you want to lose your compass and begin down a path of constant guessing. Not a good place to find yourself as a dermatologist.