Antibiotic prescribing in primary care, adherence to guidelines and unnecessary prescribing - an Irish perspective

Antibiotic resistance is a major concern globally. Antibiotic consumption increases the likelihood for an individual to develop bacterial resistance [1]. The majority of antibiotic prescribing takes place in primary care and general practitioners (GPs) have been encouraged to prescribe antibiotics more rationally and to only give antibiotics when necessary [2].

Wide variation in antibiotic prescribing practices has been shown to exist in Europe demonstrating more than a threefold difference in antibiotic prescribing rates between countries, without any logical reasoning to explain the variation [3].

In the Republic of Ireland (ROI), there has been increasing levels of resistance and antibiotic use [4]. Quinolone resistance rates in E.coli increased from 5% in 2002 to 23% in 2008. Although ROI has seen small reductions in antibiotic use since 2009, high seasonal variation is still apparent which, in other countries, is also associated with high antibiotic consumption (France, Greece, Portugal and Italy) [3]. The reductions in antibiotic use seen in ROI have been associated with the stabilisation of rates of penicillin non-susceptible Streptococcus Pneumoniae (PNSP). The proportion of PNSP increased in ROI from 10.3% in 2004 to 23.1% in 2008 and decreased in 2010 to 18.2% [5]. The reductions in antibiotic use in ROI is low in comparison to that of Belgium and France where sustained campaigns have been associated with reductions of 36% and 26.5% respectively [6, 7].

Non-clinical factors such as patient pressure have been shown to influence prescribing even in the absence of a clinical indication for an antibiotic [8, 9]. The majority of ROI citizens must pay a fee to visit their GP and this has shown as a possible influence on the GP’s decision to prescribe an antibiotic during the consultation [10].

The purpose of this study was to describe specific diagnoses for which systemic antibiotics are prescribed in primary care by GPs in the ROI, to assess adherence to recent national guidelines and gain more information on antibiotics that are prescribed unnecessarily.

Data were collected from 171 GPs nationally who recorded information on 16,899 consultations. The mean (±SD) number of consultations recorded per GP was 98.82 ± 5.37. The majority of GPs who participated completed a demographics questionnaire (84.80%, 145). All GPs were from combined GMS and private practices and situated in various settings: urban (37.24%, 54), rural (24.83%, 36) and mixed (37.93%, 55)

The majority of attendees were females (58.80%, 9,936) and held a GMS card (53.83%, 9,096).

An antibiotic was prescribed in one-fifth of consultations recorded (20.16%, 3,407). A quarter of these (25.74%, 877) were prescribed to a child between the ages of 1 and 14 years. A small proportion had received an antibiotic for the same presenting complaint in the previous two weeks prior to the consultation (8.22%, 280).

Reasons for an antibiotic prescription

Symptoms/diagnoses associated with the respiratory system accounted for the majority of antibiotic prescriptions (64.72%, 2,205), followed by skin (10.21%, 348) and urinary tract disorders (8.63%, 294). Overall, 22.63% of consultations (3,824) recorded either a diagnosis or symptoms of the respiratory system. The majority of these consultations received an antibiotic prescription (57.66%, 2,205). Children aged 4–14 years had the highest consultation rate where a respiratory symptom/diagnosis was recorded (33.94%, 1,298). Patients under the age of 65 were twice as likely to consult with respiratory symptoms/diagnoses than younger patients (OR 2.08, 95% CI 1.87-2.23).

Children aged from 0–14 years had the lowest percentage rate of antibiotic prescribing when presenting with respiratory symptoms (52.25%, 767). The highest percentage was seen for patients aged 15–64 (62.23%, 1,104) (Figure 1). In consultations with a respiratory diagnosis, 66.08% (1,420) resulted in an antibiotic prescription, in comparison with 48.22% (787) of those with respiratory symptoms. High prescribing rates (>85%) were seen in patients with throat, ear and sinus infections recorded. There was also considerable antibiotic prescribing for conditions not included in the guidelines such as upper respiratory tract infection (URTI) (33.10%, 187), cough (35.82%, 173) and sore throat (53.11%, 128) (Figure 2).

Necessary antibiotics

Most of the antibiotics that were prescribed were deemed to be necessary for the condition they were being prescribed for (71.76%, 2,445) and almost a quarter of these prescriptions were for co-amoxiclav (24.09%, 589). There were 315 (9.25%) prescriptions that were thought not to be necessary and prescribers were unsure of 602 (17.67%) prescriptions; amoxicillin was the most common antibiotic choice in these instances (36.19%, 114 and 29.73%, 179 respectively) (Figure 3). Amoxicillin was twice as likely to be prescribed when the antibiotic was thought to be unnecessary (OR 2.86, 95% CI 2.22-3.69). URTI was the most common reason for an unnecessary prescription (17.46%, 55) followed by cough (12.06%, 38) and flu-like symptoms (9.52%, 30).

Overall, this study has identified the huge burden of respiratory symptoms/conditions in general practice in Ireland. Nearly a quarter of all the consultations recorded were due to respiratory illnesses (22.63%). This is higher than the UK where only approximately 15% of GP consultations are due to these ailments [14]. As with other countries, the most common cause for an antibiotic prescription is due to symptoms or conditions relating to the respiratory system. Overall, children (1–14 years) received the highest proportion of the antibiotic prescriptions and this is probably due to the high number of consultations involving the respiratory system in this age group. Surprisingly, patients (15–64 years) presenting to the GP have the highest rate of antibiotic use per respiratory consultation, with over 60% of these consultations receiving an antibiotic prescription. Other studies have also shown that children visit their GP more often with respiratory conditions but in proportion do not receive more antibiotics than adults [14, 15]. Research has also shown that antibiotics are not justified to reduce the risk of serious complications of these conditions but may be warranted for those over the age of 65 where the number needed to treat to prevent pneumonia post chest infection is much lower than younger age groups (39 vs 119) [16]. Therefore it would be expected to see higher rates of antibiotic use in the elderly cohort. Possible explanations for the higher rates of antibiotic use seen in younger patients (15–64 years) may be due to structure of the health care system in ROI. The elderly, the very young (<5 years) and the most socio-economically disadvantaged are more likely to be GMS card holders due the eligibility criteria. The average cost for a GP consultation for a private patient in ROI is €51 [17]. GPs have been shown to over-estimate the patient’s expectation for a prescription and this influences the decision to prescribe [18]. This may be more evident in ROI as GPs may be more sensitive to patients’ expectations if the patient is paying for the consultation [10]. It may be particularly difficult for the GP not to give a prescription under these circumstances; they may provide a prescription to reduce the need for re-consultation if required. Diagnostic uncertainty and fear of negligence have also been quoted as drivers of antibiotic overprescribing [19].

There is considerable variation in antibiotic prescribing between the respiratory clinical entities with a substantial proportion of prescribing occurring for respiratory symptoms. GPs may have had an inclination to label symptoms as a diagnosis to gain endorsement of their prescribing; diagnostic labelling has been shown to be associated with high levels of antibiotic prescribing [20].

Antibiotics are not recommended for symptoms such as cough, sore throat but a high level of prescribing was recorded in these conditions. This is the reason why the majority of deviations from guidelines in this study were due to the diagnosis/reason for consultation. A study in Canada showed that 5% of URTI consultations received an antibiotic; this is in contrast to 33.10% in this study [21]. The majority of consultations that diagnosed an ear infection received an antibiotic (89.49%); USA and the UK have similar rates while the Netherlands has much lower rates of 56% [22, 23]. In a recent study in the USA, 30% of children diagnosed with a viral respiratory infection were prescribed an antibiotic [24].

In comparison to other countries, the level of agreement to guidelines is low. However, these guidelines have been recently developed and should be disseminated widely to have any impact. In the Netherlands, about three-quarters of antibiotics prescribed are first-line choice [25]. A third of patients who had been prescribed an antibiotic in the previous two weeks (i.e. for the same illness) had been prescribed co-amoxiclav which is recommended as second-line (broad-spectrum antibiotic); this could suggest that these cases were not treatment failures but rather viral illnesses.

The majority of tonsillitis presentations are not being treated with narrow-spectrum antibiotics as considered appropriate. This has a major impact on increasing resistance. The usage of amoxicillin for otitis media (first line) is lower than expected at 42.42%, other countries in Europe have rates of 82% with co-amoxiclav only used in 7% of cases [23]. Studies have described how a reduction in antibiotic usage carries a parallel decrease in resistance [26, 27]. The high incidence of co-amoxiclav has been shown in other studies in ROI, with the consumption increasing by 80% from 2000–2005 while the use of amoxicillin decreased by 9% in the same time period [28, 29]. In the absence of guidance from public health bodies, it could be argued that third-party representatives from the pharmaceutical industry may have contributed to the high usage of second line and third line agents [30].

In contrast to the high rates of co-amoxiclav, amoxicillin was more likely to be prescribed than other antibiotics when the prescription was for deferred/delayed use and also when the GP did not think that the antibiotic was necessary for the condition being treated. It could be argued that this study provides possible evidence that GPs prescribe antibiotics such as amoxicillin as ‘impure placebos’ or ‘pseudoplacebos’, these agents consist of biologically active agents that have specific efficacy for some conditions but used as a placebo for another condition i.e. antibiotics for probable viral infections such as URTI and cough. A survey in Denmark found that 70% of GPs had prescribed antibiotics as a placebo intervention in the previous year [31]. The most common reason quoted for this was to ‘follow the wish of the patient and avoid conflict’ [31]. Prescribers also report generally positive attitudes about the use of medication to promote the placebo effect [32].

This study clearly indicates that guidelines need to be disseminated at a local level with a focus to reduce antibiotic prescribing for minor respiratory conditions and reduce broad-spectrum antibiotic prescribing whenever necessary. Prescribing a course of antibiotics increases the patient’s risk of developing bacterial resistance for up to a year [1]; GPs should be mindful of this issue when prescribing antibiotics that are not necessary. It is particularly important to increase public awareness regarding appropriate antibiotic use to ease patient pressure on GPs to prescribe.

This study provides evidence that antibiotics are probably not being used as prudently in the community in ROI as they may be other countries. Children present to their GP with respiratory symptoms more often than adults, but higher rates of antibiotic prescribing occur in adults with similar symptoms. This study also shows that amoxicillin may be being used for its placebo effect rather than as specific treatment for bacterial infection.