Antibiotic prescribing in primary care

Primary care is responsible for around 80% of all antibiotic prescribing in the UK’s National Health Service,1 with rates likely to be similar worldwide. Two studies in The BMJ add to the growing evidence base informing policy on antimicrobial stewardship, which has helped primary care clinicians reduce prescribing by 13% in the past five years1 without increasing serious complications, including sepsis.2 But the studies also highlight the daily challenge of ensuring that patients who are unlikely to benefit are not treated, whereas those who require antibiotics receive the right class, at the right time, at the right dose, and for the right duration. This task is made considerably more difficult by the absence of real-time microbiology in primary care.

Both studies use invaluable routine NHS data. The first of the studies, by Gharbi and colleagues (doi:10.1136/bmj.l525), explores the effect on bloodstream infections of different primary care antibiotic prescribing strategies for urinary tract infections (UTIs) in older people.3 This study is one of a growing number of studies investigating the relation between prescribing in primary care and serious infections,4 and it is timely because rates of bloodstream infection (and mortality) are increasing, particularly in this age group.1 Moreover, the UK government has committed to halving bloodstream infections by 2021,5 and Gharbi and colleagues provide evidence that delays in treatment of UTIs could be causing harm among older people.3

The team analysed 312 896 uncomplicated lower UTI episodes among 157 264 patients aged 65 years and older between 2007 and 2015. The authors showed that antibiotics were prescribed to 87% of patients on the same day (termed immediate prescribing). The remainder either had no record of an antibiotic prescription within seven days (7%) or had a prescription issued within seven days, but not on the day of diagnosis (termed deferred prescribing, in 6%). The authors then showed that bloodstream infections and mortality rates were significantly higher in the groups with no and with deferred prescriptions, compared with immediate prescriptions.

The relation might not be causal, however, and the implications are likely to be more nuanced than primary care doctors risking the health of older adults to meet targets for antimicrobial stewardship. Firstly, evidence shows doctors are cautious when managing infections in vulnerable groups.67 Secondly, deferred prescribing in this study is probably different from the delayed prescribing used in primary care. Most clinicians issue a prescription on the day of presentation, with verbal advice to delay treatment, rather than waiting for a patient to return or issuing a post-dated prescription. The group given immediate antibiotics in the study by Gharbi and colleagues likely contained some patients managed in this way. Thirdly, a deferred prescription could indicate diagnostic uncertainty. Perhaps a urine sample was sent (something the authors acknowledge they were unable to measure) and the laboratory result prompted treatment a few days later. Fourthly, “no prescription” could be a marker for same day admission, possibly with a bloodstream infection. Finally, a significant proportion of bloodstream infections in older people are not caused by urinary tract bacteria,8 and therefore are not preventable by UTI treatment.

It is also worth remembering that although bloodstream infections associated with the urinary tract are an important and increasing problem, the annual incidence of Escherichia coli related bloodstream infection (the most common) is 37/100 000,9 and only half of affected patients present first to primary care. Thus, a practice of 10 000 patients will see one or two patients each year with this bloodstream infection, compared to around 1800 UTI episodes10 managed in the same period and age group.

What are the implications for practice? Prompt treatment should be offered to older patients, men (who are at higher risk than women), and those living in areas of greater socioeconomic deprivation who are at the highest risk of bloodstream infections. Further research is needed to establish whether treatment should be initiated with a broad or a narrow spectrum antibiotic and to identify those in whom delaying treatment (while awaiting investigation) is safe.

The second of the studies, by Pouwels and colleagues (doi:10.1136/bmj.l440), builds on good evidence that “short” antibiotic courses are as effective as “long” courses for most infections treated in primary care.1112 The authors used data from 2013 to 2015 to show convincingly that adhering to the latest national guidance on length of treatment for common infections would result in 14 fewer days of antibiotic use for every 10 prescriptions issued, or to around 65 million fewer antibiotic days each year for the UK.12 This is double the number of antibiotic days that would be saved if the UK were to hit its 2016 target of halving unnecessary antibiotic prescriptions by 2020 (assuming a seven day current average course length).13

Prescribers cannot be held responsible for what they were doing before new guidelines were issued, but we can familiarise ourselves with new guidance from the National Institute for Health and Care Excellence on managing common infections and optimise practice from here on.14 Both clinicians and patients may need convincing to abandon longer courses of antibiotics, and future campaigns by Public Health England to “Keep Antibiotics Working” could usefully emphasise that when antibiotics are needed, shorter courses are sufficient to kill bacteria and less harmful than longer courses. Also, that some symptoms should be expected to persist beyond the end of the course, in some cases for up to four weeks.1516