Those that know me, will be aware that I am a passionate believer in having clinical rationale and clinical context as pre-requisite for all microbiology samples that are processed by the laboratory.
Exactly the same principles should apply for antibiotic prescriptions…
There are various reasons why a clinician might prescribe an antibiotic:
- Fear: That if an antibiotic is not prescribed, then any suspected infection might increase in severity or even be potentially life threatening.
- Pressure: Pressure to prescribe an antibiotic from the patient or their relatives.
- Action: To be seen to be doing something positive for the patient.
- Justification: That prescribing an antibiotic justifies the cost or time of the consultation.
- Loyalty: If an antibiotic is not prescribed then the patient may go elsewhere to get what they want.
- Bacterial infection: That there is a genuine suspicion of a bacterial infection where the evidence shows that antibiotics are indicated in that particular clinical scenario.
It would be foolish to think that the real prescribing reason is always the last one on this list.
Humans are subject to all sorts of internal biases, and external pressures. We are in essence, fundamentally flawed. It is therefore wrong to expect us to always prescribe antibiotics for the right reasons as opposed to the “not so right” ones.
In other words we need tight controls on our decision making behaviours.
At the moment a clinician can write a prescription for most antibiotics without including any indication as to why the antibiotic was prescribed.
This needs to change.
It is my belief that all antibiotic prescriptions should have the clinical indication for prescribing included on the prescription form as a pre-requisite for dispensing, in both community and hospital settings.
Otherwise the pharmacist is essentially dispensing blindly.
Once such a system is in place, then specific criteria can start to be applied for certain infections in order for an antibiotic prescription to be valid/approved.
Then we can start getting some real controls in place for the purposes of antibiotic stewardship.
The days of clinicians being able to request laboratory tests and prescribe antibiotics without providing a clinical rationale are numbered.
A wound swab arrives into the microbiology laboratory..
…Because there are no clinical details we don’t know whether a Gram stain might be worthwhile.
Because there are no clinical details, we don’t know whether to add anaerobic culture or set up a yeast agar for this swab.
Because there are no clinical details, we don’t know whether this Pseudomonas aeruginosa isolate on the plate is potentially significant…
Microbiology laboratories all over the world (including our own) receive hundreds of wound swabs every day. Often there are no clinical details or only a very cursory acknowledgement to this requirement, such as “?infection”, or “discharging wound”.
You know the story well…
So often we are left to process and report the wound swab with no idea of what is actually wrong with the patient, and what exactly we should be doing with the swab.
In my opinion clinical details involve two main elements, which overlap to a certain extent:
- Contextualisation- What type of infection are we dealing with here? Taking the example of skin infections; Is it an impetiginous lesion, is it a boil, is it an area of cellulitis, is it a post-surgical wound, is it a burn, etc, etc? By providing clinical context we can start to work out what we are looking for and ascertain the importance of potential pathogens on the agar plates. Is there anything we should know about the patient? Are they immunocompromised? Do they have antibiotic allergies? Are they failing treatment? Is there anything unusual about the clinical presentation? Was it caused by a bite? Was there overseas travel?
- Justification- Taking the example of skin infections again, most patients with impetigo & boils, along with many other types of uncomplicated skin infections, do not need a wound swab sent to the lab. For most skin infections, swabbing should be the exception, not the rule. So what was it about this particular patient that prompted the clinician to send a wound swab? Was it because the patient failed first line treatment? Was it because they had an associated fever? Was it because the infection was getting rapidly worse?
Here are a few examples of hypothetical clinical details that both contextualise and justify the swab.
- “Impetiginous lesions on face getting worse despite topical anti-septic treatment. Past history of MRSA colonisation.”
- “Area of spreading cellulitis L lower leg. Patient diabetic.”
- “Carbuncle L buttock, patient has an anaphylactic allergy to penicillin.”
- “Cat scratch to R hand. patient now has fever and increasing erythema. Commenced on amoxycillin-clavulanate.”
It’s not rocket science, nor does it have to be a story.
If the clinician is required to contextualise and justify the laboratory request, then the advantages are two-fold. Not only can the scientist paint a picture in their head of what is going on with the patient, and process accordingly, but also the clinician may start to think twice about why they are sending a particular sample to the laboratory in the first place…
When I worked in a microbiology laboratory in Glasgow we hardly ever reported antibiotic susceptibilities on Staphylococcus aureus when cultured from sputum. Now I am working in New Zealand, and we almost always do…
So who is right and who is wrong?
The problem is that Staphylococcus aureus in the sputum can mean several different things:
- It may simply represent oropharyngeal contamination of the sample.
- On the flipside it may signify a severe necrotising pneumonia in an immunocompromised or post-influenza patient.
- And thirdly, we know that Staphylococcus aureus can colonise or occasionally cause “low level” infection in the architectually damaged lung (e.g. cystic fibrosis, COPD, bronchiectasis)
So in summary, it can mean absolutely nothing, or it could signify a life threatening illness…
The clinical context and sample quality are clearly key here to working out what is going on. However, in actual practice, the sputum sample often arrives into the laboratory without any clinical details, so we are processing blind.
“Just do your stuff, and give us the result…”
So how should we manage this problem from a laboratory point of view?
Here are a few potential solutions:
- Reject sputum samples for culture where the Gram stain shows lots of squamous epithelial cells representative of oropharyngeal contamination. (A lot of labs have now adopted this approach, including my own)
- Report susceptibilities routinely on Staphylococcus aureus from hospitalised (& cystic fibrosis) patients only.
- Add a comment saying that close clinical correlation is required in the interpretation of this result & susceptibilities will be available on request only.
or all of the above…
The clinical context is always important for the laboratory to issue a correct report. However, for sputum samples growing Staphylococcus aureus, it is absolutely critical.
Or one could be even stricter, and just say, “no clinical details, no test”…