Identifying treatment thresholds

The primary focus of diagnostic decision-making for participants was concerned with identifying major indications for antibiotic treatment. These were judged to include the nature and severity of illness based on presentation of symptoms and signs, in the context of the patient’s medical history. A majority of participants adopted a risk stratification approach in undertaking clinical assessment.

It’s a combination of things… For example, for an upper respiratory tract infection, tonsillitis, pharyngitis, you know, there’s a Centor guidance. So that’s where you have a checklist of things. Does this person have cervical lymphadenopathy? Do they have a fever? Do they have like absence of a cough, you know. Do they have exudate on their tonsil? So, then if you have a score of 3 or more then they have antibiotics. (Int 1, GP).

Risk stratification approaches included additional patient factors such as patient age and the presence of comorbidities including COPD, asthma, diabetes, cancer or a history of pneumonia. Whereas, many followed risk assessment protocols based explicitly on local or national clinical guidelines, some participants stressed the importance of clinical judgement in making safety-driven decisions.

You don’t want to miss something very serious. So, that’s where your clinical judgement and decision-making skills play a major role. And experience, obviously, because these are things I deal with every day. (Int 14, Nurse)

Threshold-guided decision-making spanned the continuum from ‘I am prescribing’ to ‘I am not prescribing’. Diagnostic uncertainty was part and parcel of the threshold-guided decision-making: prescribers pointed to the difference between more and less obvious cases, characterised by equivocal, ambiguous and non-convincing evidence. One participant contrasted several hypothetical scenarios:

‘…a patient with COPD, bronchiectasis, I may have a lower threshold for treating than a very fit and well 20-year-old, even if that 20-year-old had a productive cough with green sputum, their chest is clear, I’m not likely to give them antibiotics. Well they’re not feverish, whereas if they’re an 80 something with a history of COPD then I’d have a lower threshold for starting antibiotics because they’re likely to have less reserve and more likely to have complications from an infection’ (Int 20, GP).

Confidence in prescribing

Appropriate prescribing and not just a reduction in antibiotics emerged as a priority for participants, who reflected on their own performance from different perspectives. In general, participants reported a high level of confidence in prescribing but also noted occasional limitations:

I feel confident but that doesn’t mean necessarily that I think I’m making the right decision in every case. Sometimes when I’m making perhaps the wrong decision, I’m making that maybe because of patient pressure or because of my unwillingness to tolerate risk. (Int 22, GP).

Many participants acknowledged changes towards less prescribing over the last few years:

I prescribe less because I guess we’re more aware now of drug resistance than we were 5 years ago. It’s much more talked about and we’re seeing it more. But also, I’m now more confident in having that difficult discussion with the patient. (Int 5, GP)

Risks of prescribing and non-prescribing

Seven participants explicitly identified safety as a priority in infection management. All participants demonstrated vigilance to risks arising both from prescribing antibiotics and not prescribing. The fear was expressed of ‘missing something’ that could cause deterioration and consequently, participants admitted ‘being cautious’ and favoured prescribing antibiotics. At the same time, the common concern was also the avoidance of prescribing unnecessarily. Among the risks of prescribing, several side effects were reported, most commonly, gastrointestinal upsets, nausea, Clostridium difficile infection and thrush but also allergic, anaphylactic reactions, antibiotic resistance and less common side effects such as liver problems (failure). Participants also observed long-term adverse consequences of inappropriate prescribing:

I think, certainly for children, I think if you prescribe antibiotics and they don’t need them and then they have a rash because they’ve got a virus and then a penicillin allergy on their notes for the rest of their lives… I think another consequence is that if you prescribe inappropriately, it’s very difficult for another healthcare professional, down the line, to explain to that patient, you’re almost saying the other person was wrong. (Int 15, Nurse)

Risks of non-prescribing generated a shorter list with sepsis being the most concerning consequence.

Sepsis…that’s one thing I do worry about. If I see someone who’s got a high temperature and a high heart rate… then I think about those factors and I think actually if this was in my clinical judgement – if I left this for 2 days, then I think they would be crossing that line (Int 26, GP).

Three prescribers who acknowledged the tendency to overprescribe, did so, in one case, because they assessed antibiotics’ benefits to exceed harms and in two cases because of potential litigation following a missed serious bacterial infection:

Because medico-legally you’re much more likely to be brought up on missing something and not prescribing antibiotics than giving antibiotics when it wasn’t necessary… if there’s any uncertainty about prescribing antibiotics I would always err on the side of giving them because the risk, however small of missing an infection that then gets worse would be enough for me to give antibiotics. (Int 19, GP)

Facing antimicrobial resistance

Participants shared concern for the global rise in antimicrobial resistance. At the same time, they acknowledged lacking in-depth microbiological knowledge: “we talk more about not prescribing and prescribing correctly than resistance itself’ (Int 9, Pharmacist). Meanwhile, they had to deal with the consequences of the antimicrobial resistance in their daily practice:

I’ve had a few patients that have had MRSA [Methicillin-resistant Staphylococcus aureus]. I’ve had a few people who have had PVL [Panton-Valentine leukocidin form of MRSA] infections, skin infections with multiple resistance… So we can sometimes struggle to find an antibiotic that’s oral, that’s then suitable. I’ve got a Type 1 diabetic, young lady, who has very poorly controlled diabetes and recurrent boils and abscesses on her back. And we did a swab of that and yes, there was only one oral antibiotic that was sensitive - everything else was resistant. (Int 23, GP)

antimicrobial resistance was most commonly encountered in older women with urinary tract infections:

‘I think sometimes you do see, for example, in the UTI breakdown, some people have quite resistant UTIs and that becomes difficult’. (Int 15, Nurse)

‘I’ve been a GP for about 10 years and you’ve already seen that certain antibiotics just aren’t working anymore, and we need to change the way that we’re doing things and you know we used to give trimethoprim locally first line for UTIs. Resistant in the majority of cases. So, we’re giving nitrofurantoin’. (Int 10, GP)

There was mention of difficulties in conveying information about resistance to patients—discussing it in the encounters and emphasising community impact may have been less efficient than focussing on individual risks. There was also a worry that primary care is running out of antibiotics despite the strategies of second-line and third-line antibiotics:

They (patients) literally cannot have any, they’ve got an E. coli infection that’s not sensitive to amoxicillin or nitrofurantoin or trimethoprim or even cefalexin or the cipro. It’s just like literally multiply-resistant. And there’s some quite virulent, my understanding is it is strains of bacteria where antibiotics will not work. And then you kind of get to the hard-core ones. (Int 11, GP)

In such cases of failure of several course of antibiotics, referral to secondary care, possibly for intravenous therapy were reported as the only options. Other times, where the resistant organism could be tackled in primary care, the last resort was a longer course or long-term prophylactic antibiotics. More investigations and consultations with microbiologists about unresolved infections appeared to precede these decisions.

Managing patient expectations

Participants identified patient pressure as a factor in their decision-making but they shared the view that patients differ in terms of their expectations regarding antibiotics. On the one hand, increased knowledge of the appropriate indications for antibiotic therapy (not for viruses) and understanding of antimicrobial resistance from public health and media campaigns was noted. On the other hand, patient pressure in a form of implicit expectations or explicit demands remained frequent: readily prescribed in the past, antibiotics had a profile of immediate cure in large parts of patient population:

… so many people have been mis-prescribed antibiotics in the past that I think they just won’t believe you that they don’t need them (Int 25).

A GP summarised this ambivalence:

There’s a reasonable cohort now who come in and say they don’t want them (antibiotics). They’ve read, they’re educated, they know that they’re contributing potentially to resistance and they don’t want to risk the side effects. But there’s also a large cohort still who come in and say, “My cough’s gone to my chest, I need antibiotics. So, it’s trying to often you know, get through those barriers and explain to them that their chest is clear.” (Int 10, GP)

Eliciting expectations, educating patients and delayed prescription were the key strategies for managing patient expectations. Explaining assessment results and positive language were deemed important for the success of the consultation; several participants preferred the time-saving mode of giving out written information about the expected length of illness (for example, about the duration of sinusitis with and without antibiotics) and about side effects of antibiotics. Elicitation of expectations included asking patients: “What were you hoping for when you came in today?” (Int 26, GP). Delayed prescriptions were used by all but three of the participants interviewed. This was considered as a form of partnership, of shared decision-making between the clinician and the patient:

… that helps patients because at least psychologically they have got an antibiotic, but they know they can’t use it straightaway. (Int 25, GP)

Communicating risks

As above, participants demonstrated that the commitment to reduced prescribing was dependent on patient understanding of the need for antibiotics. This meant that at times building and maintaining relationships were prioritised and led to prescribing decisions, as an interviewee reported:

Much of my job is trying to build a rapport with someone and build a rapport so that we can have a conversation that’s therapeutic. If someone has come in adamant that they want antibiotics there is some conversation to be had there. Why did you get this idea from? What is it that you believed this would do? And what is your previous experience? Now, if they’re not willing to go into that today, I may actually give them a short course of antibiotics with the understanding that we have another conversation. This is a way of building some trust (Int 29, GP).

The participants differed in terms of how they dealt with risk in encounters with patients: some were liberal prescribers who tended to avoid complaints and patient frustration, others preferred having difficult conversations on non-antibiotics’ course of actions. Among liberal prescribers, there was the notion of offering antibiotics in order to be safe. In the case of non-prescribing, prescribers sometimes delved into lengthy explanations in order to secure patient adherence:

When I’m explaining that there’s no sign of bacterial infection and we don’t want to give you antibiotics if we don’t need to. Most people go, ‘Oh yes, yes, no, of course not.’ But some people might say, ‘Oh, well, you know.’ Then I will go into the reasons why, you know. ‘Well actually you might get side effects, you know, it can make you, give you diarrhoea, it can give you thrush. And things can become resistant to it and it won’t be helpful for you in the future. (Int 30, GP)

Advice on possible warning signs: ‘safety netting’ emerged as a dominant risk reduction strategy:

I will give them (patients) an awful lot of safety-netting, and tell them what, ‘If this doesn’t get better, this is when you come back.’ You know, or ‘These are the signs of you getting worse,’ or what they do if they are getting worse. (Int 16, GP)