Monthly Archives: June 2016

“Choices”

Antimicrobial Resistance, The Art of Microbiology,

2011-July-HI-2VitekCardsNEW

How many antibiotic susceptibilities do you release to the clinicians for any particular isolate?

At the extreme end of the scale, some microbiology laboratories think nothing of releasing 15-20 antibiotic susceptibilities for the one isolate. This sometimes happens in labs rich enough to have a Vitek or Phoenix where the system panels contain many different antibiotics, often customised for a particular institution.

However more is not always necessarily better…

The main reasons for not routinely releasing an antibiotic susceptibility on an isolate are as follows:

  • Unnecessarily broad spectrum for the clinical indication.
  • Low threshold to developing antibiotic resistance.
  • Relatively poor efficacy compared to other antibiotics for the clinical indication.
  • Logistically difficult to use.

The majority of mainstream diagnostic labs probably give a choice of 5-10 antibiotics per isolate. Even then that is still a lot of antibiotics for the clinicians to mull over. It also gives a lot of scope for non-standardised practice…

More laboratories, including my own, are looking more and more towards focused antibiotic reporting, in order to “guide” clinicians towards the optimal treatment regime.

For example we routinely release 2 antibiotics for a Staphylococcus aureus isolate from a general wound swab. (If MALDI-ToF succeeds in reliably differentiating between MSSA and MRSA, then one could potentially release flucloxacillin/cloxacillin alone, without doing formal susceptibility testing, and with a comment “Further susceptibilities available on request”)

We routinely release 4 antibiotics on urinary E.coli isolates. I am thinking of reducing this to 3.

For MSSA (Methicillin Susceptible Staphylococcus aureus) isolates from blood cultures and invasive sites we routinely release just the one antibiotic, flucloxacillin. The same principle applies with Streptococcus pneumoniae and penicillin from blood cultures. By doing this we make it crystal clear what the expected standard treatment is. Of course we have a few others ‘up our sleeve’ should they be required in the case of allergy, treatment failure etc.

And then of course it is always important to remember that we have the choice of releasing (or testing) zero antibiotics, where the clinical details are absent or they are not indicative of acute infection. We should probably use the zero option more, along with a comment saying “Antibiotic susceptibilities available on request”.

Occasionally we get calls from clinicians, a little bit upset or grumpy that we have not released more antibiotics. Being thick skinned, I don’t mind this at all. It gives me the opportunity to build relationships with the clinicians, and also to explain our thinking…

Michael

 

“Taking humans out of the equation”

Future of Microbiology,

IMG_4662

There are now several smartphone apps available for anti-microbial prescribing. Some of these can even be programmed with your institution’s local guidelines. Locally, we use the “Microguide app”

However as far as I am aware, the apps currently related to antibiotic prescribing do not intake/process patient information. (Please correct me if I am wrong on this.)

But that is the next step…

Two key questions related to good antibiotic stewardship are as follows:

  • Based on the clinical findings, does this patient actually need an antibiotic?
  • If so, which antibiotic would be best ?

Humans are generally good at writing antibiotic guidelines. However they are not so good at following them. That’s because lots of other things get in the way, such as emotion, past experience, fatigue, and the desire to do everything possible to help the patient in front of them.

It will only be another few years before apps designed to answer these questions will be available. They might be rudimentary to start off with, but it won’t be long before they become sophisticated and will be every bit as competent as physicians are at making most of the routine decisions. (Think about chess computers…)

I am sure these guidelines will require regulatory approval (FDA, etc.) and it will be interesting to see how these bodies approach such apps.

In 2050, I will be 77, and coming up to the (revised) retirement age. In 2050, clinical decision making apps in all branches of medicine will be commonplace, I am sure of it. Computers are (at present) unemotional, and therefore very well placed to perform this job of deciding who exactly needs an antibiotic, and if so, which one.

However I suspect such apps will be in everyday use long before 2050…

Michael

 

“The Blind Microbiologist”

Confessions of a Microbiologist, The Art of Microbiology, , ,

blindI have authorised a little flurry of urine cultures recently with ESBL producing organisms. A good proportion of these had no clinical details, and a significant proportion were from Rest Home residents.

It is the lack of clinical information which frustrates me the most..

These urine samples could have been taken for any one of the following reasons:

  • a)  The nurse at the Rest Home had decided to routinely dipstick the urine of all the residents.
  • b)  The urine was taken and sent to the lab because there was a “whiff” of urine from the patient.
  • c)  The patient is known to be MDRO colonised, so another urine was taken to ‘check’ if it is still there.
  • d)  The patient gets a 6 monthly urine sample sent to the lab because they are known to have a history of recurrent UTIs.
  • e)  The patient is new to the practice or Rest Home so an initial screening urine was taken off.
  • f)  The patient had symptoms or signs suggestive of an acute urine infection.

It is dangerous to assume that the majority of urines arriving at your lab (without clinical details) are specifically for reason F.

My bet is that this is not the case…

When we release antibiotic susceptibilities from the laboratory on a urine culture, we are essentially giving the green light for antibiotic use. An ‘endorsement’ of sorts, a licence to treat, regardless of the clinical indication. In essence we are reporting susceptibilities “blind”…

However reason F is the only reason listed above where I would actually want to routinely release susceptibilities. It could easily be argued that for the laboratory to release susceptibilities on an MDRO without there being any clinical details, simply represents poor antibiotic stewardship.

My laboratory has recently made the provision of clinical details mandatory for Infectious serology tests. If no clinical details, then the serum is simply stored until they are provided. Uptake by clinicians has been very good, and there have been virtually no complaints. What is there to complain about? The quality gains that we have made from this have been very significant.

I am now wondering whether a similar model should be applied to urine samples. Obviously specimen integrity (and ease of recollection) should be taken into account, but there are a couple of possible models as follows:

  • No clinical details, urine stored for maximum 24 hrs until provided. Otherwise recollect.
  • No clinical details, urine processed, but susceptibilities only released on provision of clinical details.

As microbiologists, I don’t think we should be scared of having these kinds of conversations. In fact, I believe we owe it to our patients…

Michael

p.s. Book sales going well, approximately 50 copies ordered so far. They are being posted within 24 hours of receipt of order but obviously will take a few days to arrive depending on location. Let me know if any problems in the ordering process. Thanks.