Antibiotic use for acute respiratory tract infections (ARTI) in primary care; what factors affect prescribing and why is it important? A narrative review

Background Antimicrobial resistance is an emerging global threat to health and is associated with increased consumption of antibiotics. Seventy-four per cent of antibiotic prescribing takes place in primary care. Much of this is for inappropriate treatment of acute respiratory tract infections. Aims To review the published literature pertaining to antibiotic prescribing in order to identify and understand the factors that affect primary care providers’ prescribing decisions. Methods Six online databases were searched for relevant paper using agreed criteria. One hundred ninety-five papers were retrieved, and 139 were included in this review. Results Primary care providers are highly influenced to prescribe by patient expectation for antibiotics, clinical uncertainty and workload induced time pressures. Strategies proven to reduce such inappropriate prescribing include appropriately aimed multifaceted educational interventions for primary care providers, mass media educational campaigns aimed at healthcare professionals and the public, use of good communication skills in the consultation, use of delayed prescriptions especially when accompanied by written information, point of care testing and, probably, longer less pressurised consultations. Delayed prescriptions also facilitate focused personalised patient education. Conclusion There is an emerging consensus in the literature regarding strategies proven to reduce antibiotic consumption for acute respiratory tract infections. The widespread adoption of these strategies in primary care is imperative.


Introduction
Antimicrobial or antibiotic resistance (AMR) is an increasingly serious threat to global public health [1]. Consequently, there is an emerging risk that standard antibiotic treatments no longer work making infections harder or impossible to control [2]. Increasing consumption of antibiotics is associated with the development of antibiotic resistance at individual, community, country and regional levels [3][4][5][6][7]. It is estimated that 25,000 humans in the EU die annually as a result of infections caused by resistant bacteria, at a societal cost of approximately €1.5 billion annually [8].
The rate of use of antibiotics in Ireland for quarter 4 of 2016 was 23 defined daily doses (DDD) per 1000 inhabitants per day, and this is up from 20 DDD per 1000 inhabitants per day in 2009 [9]. In comparison to other EU countries, antibiotic use in Ireland is mid-range [9]. The Scientific Advisory Committee of the Irish National Disease Surveillance Centre (NDSC) was tasked in 2001 to produce a strategy document in response to the growing problem of antimicrobial resistance. This resulted in the BStrategy for the control of Antimicrobial Resistance in Ireland^(SARI) [10], which produced a number of national guidelines and advised the Irish Health Services Executive (HSE) on matters relating to the prevention and control of antimicrobial resistance and healthcare-associated infection. The work of SARI has more recently been taken over by the National Clinical Programme for the prevention of healthcare-associated infection (HCAI) and antimicrobial resistance (AMR) under the auspices of the HSE. This has led to HSE guidelines issued for antibiotic prescribing in primary care [11].
Over the last 30 years, no major new types of antibiotics have been developed [12]. This in combination with increasing AMR means that we are dealing with a finite and diminishing antibiotic resource. Therefore, prudent antibiotic stewardship programmes, aiming to ensure the judicious use of antimicrobials by preventing their unnecessary use, have been established [1,10,[13][14][15][16].
Acute respiratory tract infection (ARTI), which incorporates the term Bupper respiratory infection^(URTI), is the most common reason for antibiotic prescription in adults, and these prescriptions are often inappropriate [17]. The benefits of antibiotics are marginal for the management of most cases of ARTI [18][19][20][21][22][23][24][25], including sore throat [26,27]. With few exceptions [28], inappropriate prescribing of antibiotics for patients with mainly URTI is common [29][30][31][32][33]. It is estimated that 75% of overall antibiotic prescribing takes place in primary care [34]. Large variations in antibiotic prescribing for URTI exist and are difficult to explain [22,35]. Some potential explanations include the fact that many general practitioners (GPs) do not think that antibiotic prescribing in primary care is responsible for the development of antibiotic resistance [36][37][38][39][40] and, on average, acute cough can last from nine [41] to 18 days [42], while public expectation is for a duration of 7-9 days [42].
This paper reviews the literature on factors affecting antibiotic prescribing for ARTI in primary care. We consider specifically the effects of patient expectation and desire for antibiotics to treat respiratory symptoms, other patient characteristics, primary care provider (PCP) characteristics and the setting of the consultation. We also review the evidence behind current strategies employed to address this public health challenge.

Definitions
Where not otherwise specified, the term primary care provider (PCP) refers to all healthcare professionals dealing with the public in the primary care setting including GPs and non-medical professionals such as nurse practitioners, practice nurses, maternal child health nurses and pharmacists. Where specific studies mention particular types of PCP, this is indicated.

Search strategy
While this article is not intended to be a systematic review, a comprehensive search of the literature was performed through Cochrane Library, Embase, PubMed Central, Scopus, Medline and CINAHL, looking at English language journals from 1997 to date. Original studies were included. Review papers such as editorials, opinion pieces, studies from secondary care, case reports, articles written prior to 1997 and studies involving lower respiratory tract infections only were excluded. The search terms used were BRespiratory tract infection^or Bupper respiratory tract infection^, Bantibiotic^or Bantibacterial agents^, Bpatient expectations^or Bpatient attitudes^, antibiotic* and prescri*, and Bupper respiratory infection^or Bupper respiratory tract infection^or Bupper respiratory infection (URI) in children^or Bupper respiratory infection (URI) in adults or adolescents^. Duplicates were excluded during this process, and bibliographies were screened by two of the authors (JOD ROC) for further relevant papers. The search strategy used is outlined in detail in Table 1.

Results
The effect that different factors play in the PCP's decision to prescribe antibiotics to treat acute respiratory infections may be categorised as follows.

Time constraints
GPs and other PCPs working in highly pressurised clinical environments managing high patient volumes are more likely to prescribe antibiotics for ARTI [29,38,[43][44][45]. The level of antibiotic prescriptions issued increased in line with numbers of patients seen per day, resulting in shorter consultations [29,45]. Suggested reasons for this excess antibiotic prescribing were lack of time in the consultation to discuss management alternatives and to inform the patients about the poor efficacy of antibiotics [38,43,44].

Primary care providers' perceptions of patients' expectation for antibiotics
GPs and other primary care doctors are more likely to prescribe antibiotics to patients who expect them or whom they believe expect them [36,40,41,43,44,[46][47][48][49][50][51][52][53][54][55][56]. This experience is replicated with other non-medical PCPs [44]. Patient expectation has been described as an all-encompassing term that is affected by factors such as limited time in the consultation, diagnostic uncertainty and poor doctor patient communication [36]. High prescribers were concerned about patient satisfaction and were unaware that they differed from their peers [56].

Primary care providers' personal factors
A Canadian analysis prescribing for patients aged 66 years or older with non-bacterial ARTIs showed that primary care physicians who were in mid or late career or who were seeing high patient volumes, or who were trained outside of Canada or the USA were more likely to prescribe antibiotics [29]. However, the physician rationale for prescribing was not studied. A systematic review of studies from both ambulatory care and hospital settings concluded that inadequate knowledge and misconceptions of prescribing are prevalent among physicians from the USA, the UK and Peru, with pocket-sized guidelines seen as an important source of information [40]. Such misconceptions included prescribing antibiotics for purulent nasal discharge and thinking that occasional use of narrow spectrum antibiotics had a negligible effect on AMR [40]. A single cross-sectional study looking at 5937 ARTI visits to 102 primary care physicians in Canada found no association between empathy or burnout and antibiotic prescribing for ARIs in primary care [57].
Based on this, we can deduce that doctors' professional training, their career stage and time pressures can be seen as "Respiratory tract infection" or "upper respiratory tract infection" or "antibiotic" or "antibacterial agents" or "patient expectations" or "patient attitudes" Database Results based on inclusion criteria Scopus 49 Search string 2 ("upper respiratory tract infection" OR urti) AND (patient AND expectation) AND (antibiotic* AND prescri*))

Database
Results based on inclusion criteria Scopus 20 Search string 3 "patient expectations" and "antibiotics" or "antibacterial agents" and "upper respiratory infection" or "upper respiratory tract infection" or "upper respiratory infection (URI) in children" or "upper respiratory infection (URI) in adults or adolescents" Database Search string 5 "patient expectations" and "antibiotics" or "antibacterial agents" and "upper respiratory infection" or "upper respiratory tract infection" or "upper respiratory infection (URI) in children" or "upper respiratory infection (URI) in adults or adolescents". Database EBSCO 22 Search string 6 "patient expectation" AND "upper respiratory infection" AND "antibiotics" AND ("upper respiratory tract infection" OR urti) patient AND expectation) AND TITLE-ABS-KEY (antibiotic* AND prescri*) Database important factors affecting their decision to prescribe antibiotic for ARTI (Table 2).

Patient expectation
Measurement of patient expectation for antibiotic treatment for ARTI varies from 74% [58,59] to 10% [41], with many measurements in between [60][61][62][63][64]. The study showing 10% patient expectation for antibiotics was conducted in China among patients presenting with ARTI symptoms and found that concern about illness severity and obtaining symptomatic treatment were the main reasons for consulting with ARTI rather than obtaining antibiotics [41]. The two studies showing the highest patient expectation rates were both studies of parents' attitudes to antibiotic prescribing for their children who were not sick at the time, and were based in Greece and Palestine. Patient satisfaction varies with antibiotic prescription policies for ARTI and patients were less satisfied in practices with low antibiotic prescribing rates, and a cautious approach to antibiotic prescribing may require a trade-off in terms of patient satisfaction [41,65]. Patients were also less satisfied when they expected but were not prescribed antibiotics [48, 66,67]. However, receiving information and reassurance from the HCP was also associated with high patient satisfaction (Table 3) [64,68]. Doctors may overestimate the pressure to prescribe antibiotics for acute cough [68][69][70] or other acute respiratory illnesses [50], often prescribing antibiotics for patients who did not request them [71]. There is mounting evidence that patients' expectations for antibiotics for ARTI have lessened in recent years, especially where the consultation is more patient centred [41,67,[72][73][74][75][76]. This illustrates the importance of a patient-centred consultation with good communication skills employed by the PCP.

Patient socioeconomic background
Patients with lower education level or who come from more deprived socioeconomic backgrounds who are likely to have less knowledge and understanding of the concept of antimicrobial resistance are more liable to be prescribed antibiotics for ARTI [38,47,[77][78][79]. However, studies from Ireland, China and Malaysia have shown that patients paying a consultation fee are also more likely to receive antibiotics for

Educational interventions for patients
Mass media interventions such as national TV advertising campaigns in Belgium and France [82] and repeated mass media campaigns in France and England [83,84] have been shown to reduce antibiotic prescribing for ARTI. However, these strategies work best when targeting both healthcare professionals and the public in mass media campaigns [85,86]. Single educational interventions of leaflets or leaflet/videotape mailed to patients had little effect on reducing antibiotic prescribing rates for elderly [87] or paediatric [88] populations. Also, passively leaving this literature in the waiting room was found to have no effect [89].

Educational interventions for GPs and other PCPs
Multifaceted educational interventions in general practice including visits by peer academics, regional 1-day seminars, internet-based training in communication skills and C-reactive protein (CRP) testing, all aiming to reduce antibiotic prescription rates for ARTI and to reduce the use of broad-spectrum antibiotics, have been shown to be effective (90)(91)(92). GPs may need further guidance on how to answer the concerns of patients without interpreting these questions as a demand for antibiotics, as well as educating the patient about antimicrobial resistance and supporting a good patient-practitioner relationship [93]. This educational process is hindered by the fact that guidelines issued for GPs vary considerably regarding categorisation of evidence and recommendations [94], taking little account of local antimicrobial resistance patterns in their recommendations [95].

Educational interventions for PCPs and public
Antibiotic use for adults diagnosed with ARTI can be reduced using a combination of PCP and patient educational interventions [85,86,96,97]. One such campaign based in Britain, Antibiotic Guardian, increased commitment to tackling AMR in both PCPs and members of the public, increased selfreported knowledge and changed self-reported behaviour particularly among people with prior AMR awareness [98]. In paediatric practice, a systematic review concluded that educational interventions targeting clinicians and parents of affected children are more effective than those for either group alone, and the most effective strategies address patient-clinician communication [99].

Delayed prescriptions
Awell-documented strategy for reducing antibiotic prescriptions for ARTI is the use of delayed prescriptions. These are valid prescriptions issued at the time of the consultation. The PCP usually negotiates with the patient that they are not to be used immediately but only if the patient feels that their symptoms deteriorate or do not improve as expected [93]. There is substantial evidence that the use of delayed prescriptions has been associated with reduced antibiotic use [48, [100][101][102][103][104][105][106][107][108]. The DESCARTE study has been looking at the symptomatic outcome of acute sore throat in a random sample of 2876 adults according to antibiotic prescription strategy in routine care. It concludes that in the routine care of adults with sore throat, a delayed antibiotic strategy confers similar symptomatic benefits to immediate antibiotics [109]. Another paper from the same study also concluded that a small advantage in terms of reduced re-consultation for a 10-day course of penicillin could not be ruled out, but the effect is likely to be small [110]. However, a prospective observational cohort study of 14 primary care networks in 13 countries found the strategy to be unhelpful in reducing antibiotic consumption [111]. A qualitative study of GPs, trainee GPs and nurse prescribers found that issuing delayed prescriptions was not considered to be a helpful strategy for managing patients with self-limiting respiratory tract infections within primary care [112]. A Cochrane systematic review concluded that delayed prescriptions reduced patient satisfaction in some trials, which seems to have little advantage over avoiding them altogether where it is safe to do so [104]. In many cases, patients are happy to receive delayed prescriptions for antibiotics for ARTI [102][103][104][105][106][107][108][109][110][111][112][113][114]. A recent qualitative study of patients concluded that delayed prescribing is acceptable no matter how the delay is operationalised, but explanation of the rationale is needed by the PCP [114]. However, not all GPs issue delayed prescriptions [115] and not all patients may be content to receive them as they felt less enabled by consultations which resulted in delayed prescriptions [100]. In summary, delayed prescriptions may be a useful adjunct for PCPs in giving focused education to the patient about the expected natural history of their ARTI and what symptoms and signs to look out for that might indicate deterioration. Patient-focused education combined with the use of educational leaflets or booklets has been shown to reduce antibiotic consumption in children and adults [116][117][118][119]. The success of these strategies could depend on the level of communication skills of the PCP [116][117][118][119].

PCP-patient communication
Poor doctor-patient communication has been implicated in inappropriate antibiotic prescribing [36]. Significantly,      explanation of long-term negative sequela does not appear to be a sufficiently strong incentive for patients and, consequently, antibiotic resistance needs to be explained as a more immediate health issue [61]. The need for, and effect of, such communication can be impressive. In a US study of 98 patients visiting family medicine clinical sites, whereas more than half the respondents recognised that treatment for colds did not require antibiotics, 70% erroneously indicated that viruses require antibiotic treatment and 95% of patients reported satisfaction when advised by their physician that antibiotic treatment was not necessary, even if they initially thought they needed antibiotics [120]. Training of PCPs in communication skills has been shown to reduce antibiotic prescribing for ARTI [92]. The results are even better when combined with point of care testing (POCT) [121]. A systematic review of the effectiveness of primary care-based interventions to reduce antibiotic prescribing for children with RTIs and a study involving structured interviews of parents of acutely ill children attending an out of hours service both indicate that clinical and communication skills in the PCP where they take a good history, examine the patient appropriately and give a good explanation of the cause of the illness help to improve appropriate antibiotic prescribing in paediatric practice [72,89]. Also in paediatric consultations for ARTIs, parents receiving combined positive (e.g. measures to reduce fever and pain) and negative (e.g. recommendation against need for antibiotics) treatment recommendations were more likely to give the highest possible visit rating, which may reduce the risk of antibiotic prescribing [122]. When the doctors' explanation is backed up by use of an information booklet, this reduced the number of antibiotics children consumed [116][117][118] without affecting parent satisfaction or numbers of return visits [118]. Such a booklet has high acceptability for both parents and clinicians [117]. Reductions in antibiotic prescribing for adults with ARTI have also been shown by the use of an information booklet [119].
The substance of what is communicated is also important. One study suggests that within-consultation communication aimed at reducing antibiotic expectations would be more effective if it is acknowledged that viral illness can be severe, thus validating the patient's decision to attend (e.g. viral pneumonia) and that bacterial infections can be self-limiting and therefore may not need an antibiotic [123]. It also suggests that clearer explanations of the symptoms and signs of a child's illness that indicate when antibiotics are and are not warranted would help reduce misunderstandings, as would reducing antibiotic prescribing that is not supported by evidence [123]. However, such communication which gives the patient more influence over the decision whether or not to prescribe an antibiotic may result in a variable outcome depending on the factors imposed by the healthcare system. A nine-country qualitative study described clinicians' accounts of the non-clinical factors that shape antibiotic prescribing decisions for lower respiratory tract infection [124]. It showed that PCPs in specific primary care networks in Europe report that their prescribing decisions are influenced by factors imposed by the healthcare system, including direct patient access to antibiotics, systems to reduce patient expectations for antibiotics and lack of consistent treatment guidelines [124].
A systematic review has shown that misunderstandings have occurred because parents' expressions of concern or requests for additional information were sometimes perceived as a challenge to the clinicians' diagnosis or treatment decision and may be an important contribution to the unnecessary and unwanted prescribing of antibiotics [125]. Two systematic reviews concluded that interventions that aim to facilitate shared decision-making such as enhanced communication skills and patient information leaflets reduce antibiotic prescribing in primary care in the short term [126,127]. Effects on longer-term rates of prescribing are uncertain, and it is unclear how any sustained reduction in antibiotic prescribing affects hospital admission, pneumonia and death [126].

Diagnostic uncertainty
Diagnostic uncertainty and fear of complications among the attending physicians is a common cause for prescribing antibiotics for ARTI [36,51,54]. A systematic review suggested that interventions which reduce uncertainty about appropriate ARTI management in primary care are likely to be effective in promoting prudent antibiotic use while remaining attractive to GPs and feasible in practice [74]. Point of care testing, discussed below, is also useful.

Point of care testing
Point of care testing (POCT) including CRP and procalcitonin has been shown to reduce antibiotic prescribing in primary care [121,127,134]. Their use is acceptable to patients [134][135][136] although clinicians have expressed concerns about using them in the consultation [136].

Out of hours prescribing of antibiotics
The suggestion that OOH antibiotic prescribing quality is worse than in daily practice does not seem founded as the higher OOH prescribing rates could be explained by a different population of presenting patients [137]. In Norway, antibiotic prescribing for ARTIs in OOH services is at the same level as in normal working hours, but with a higher prescription rate of penicillin V (PcV), which was close to the national goal of 80% proportion of PcV for ARTIs [45]. A suggested explanation was that doctors working in transparent out of hours units are more adherent to guidelines than doctors working in regular general practice are. Antibiotic prescribing increased during busy sessions [45].

Influence of daycare providers
Some daycare providers encourage parents of children with infections to consult general practice and seek antibiotics [138]. Parents' perceptions of daycare providers' requirements may override their own beliefs of when it is appropriate to consult and seek antibiotic treatment [138]. This is a potentially harmful attitude, which should be challenged.

Direct patient access to antibiotics (non-prescription use)
A systematic review published in 2011 looking at 35 community surveys from five continents showed that nonprescription antibiotic use occurred worldwide and accounted for 19-100% of antimicrobial use outside of northern Europe and North America [139]. Safety issues associated with non-prescription use included adverse drug reactions and masking of underlying infectious processes. Antimicrobial-resistant bacteria are common in communities with frequent non-prescription use, which has been speculated to play an important role in selecting and maintaining these high levels of community antimicrobial resistance. In middle-income to high-income countries with reliable access to healthcare practitioners, antimicrobials should be restricted to prescription-only status [139].

Strengths and limitations
This review employed a comprehensive search strategy, encompassed over 20 years of publications and reviewed 139 relevant papers. The key findings from individual papers were analysed by the authors and integrated in a logical manner that is of use to front-line clinicians dealing with patients presenting with ARTI in primary care. Limitations were that only English language studies were included and that other indications for antibiotic prescription in primary care such as urinary tract infection were not considered. However, the commonest indication for antibiotic prescription in primary care is for ARTI. Patient expectation is a very important reason for PCPs to prescribe antibiotics, and it can change with good communication skills and patient-centred approach to the consultation. The most recent literature was felt to be of most relevance to practicing clinicians.

Recommendations for further research
Future research should focus on developing and refining strategies that have been proven to be successful, such as multifaceted education of PCPs and the public, the use of delayed prescriptions and ensuring that consultations for ARTIs are not time pressurised. The use of less expensive resources such as non-medical practitioners as the first point of contact should be explored in the wider global context, especially in countries where there is direct access to antibiotics because of limited access to healthcare professionals. The role of other primary care resources such as community pharmacists and practice nurses should also be considered in the context of developing a more cohesive approach by all primary care health professionals who are consulted by people with ARTI. Coordination of policy led by the relevant postgraduate training bodies for these disciplines should be encouraged and developed in conjunction with the relevant state agencies. Studies should also focus on problem areas such as the attitudes of daycare providers and the difficulties of dealing with feepaying patients. Also, in areas where the literature is contradictory such as with delayed prescriptions, further study is warranted. Further development and testing of POCTs to help reduce clinical uncertainty is advised. The importance of cultural and resource factors needs to be evaluated further, especially in the setting of direct patient access to antibiotics.

Conclusion
Antimicrobial resistance is an important and growing health issue, and a considerable contributor is the overprescribing of antibiotics in primary care for ARTI.
Less pressurised consultations, training to enhance PCP communication skills, multifaceted education campaigns aimed at patients and PCPs, delayed prescriptions and point of care testing to reduce diagnostic uncertainty all help to reduce such inappropriate prescribing.  (11)