CADTH Health Technology Review

Interventions to Influence the Use of Antibiotics for Acute Upper Respiratory Tract Infections

Rapid Review

Authors: Wade Thompson, Suzanne McCormack

Key Messages

• Thirteen systematic reviews (SRs) were identified. Two of these SRs addressed delayed antibiotic prescribing and 12 of these SRs investigated family medicine interventions. Delayed antibiotic prescribing reduced antibiotic use for upper respiratory tract infections (URTIs) compared to immediate prescribing and did not appear to impact patient satisfaction or re-consultation rates; however, there was less evidence on clinical outcomes, health care utilization, or antibiotic resistance. One systematic review concluded that there was no difference between delayed and immediate antibiotics for many clinical outcomes, but that immediate antibiotics may modestly improve symptoms for acute otitis media (AOM) and sore throat compared with delayed antibiotics.

• Various family medicine interventions were evaluated; however, there were generally few relevant primary studies in each of these reviews for most of the interventions. Rapid or point-of-care tests to guide the treatment of URTIs appeared to reduce antibiotic prescribing compared to control or usual care; however, there was relatively little evidence on clinical outcomes for these tests. One systematic review on procalcitonin-guided treatment suggested that this intervention led to no difference in the number of days with restricted activities or rates of treatment failure compared to control, while another systematic review found that C-reactive protein (CRP) point-of-care tests to guide antibiotic prescribing resulted in no difference in recovery or time to resolution of the symptoms; however, this was based on only 2 randomized controlled trials (RCTs). Evidence was mixed on different patient- or provider-directed interventions, such as education, training, and tools. In some of the primary studies from the eligible SRs, these interventions reduced antibiotic prescribing; while in others, they had no effect on antibiotic prescribing.

Context and Policy Issues

Acute upper respiratory tract infections (URTIs) result from viral or bacterial pathogens that infect the upper airways.¹ URTIs comprise a range of conditions and symptoms such as cough, sore throat, acute otitis media (AOM), pharyngitis, tonsillitis, rhinitis, common cold, influenza, and sinusitis.¹ While serious complications are rare, patients can develop pneumonia and associated adverse sequelae.¹ A substantial proportion of health care visits to primary care and other family medicine settings are due to suspected URTIs.²

Most URTIs are viral in origin, meaning they are unlikely to benefit from antibiotics and can often be effectively managed remotely with supportive treatments. Despite this knowledge, antibiotics are often inappropriately prescribed.^3,4 A descriptive analysis in Ontario estimated unnecessary antibiotic prescribing in primary care at rates of 48.4% for acute sinusitis and 52.6% for acute bronchitis.⁵ Further, more than 80% of unnecessary antibiotic prescriptions were attributed to respiratory tract conditions and infections.⁵ Overprescribing contributes to excessive use, unwanted side effects (e.g., allergic reactions, gastrointestinal symptoms, emergency department visits),⁶ and antibiotic resistance.⁷ Antibiotic resistance is a global public health threat and its reduction is a priority.⁸

Initiatives such as Choosing Wisely aim to reduce antibiotic overuse and advise against using antibiotics for URTIs that are viral in origin.⁷ However, while the risks of inappropriate and overuse are established, antibiotic prescribing is challenging and providers must balance the pressure of diagnostic uncertainty, time constraints, and patient expectations and preferences against the undesirable consequences of overprescribing.⁸ Thus, there is interest in management approaches that aid decision-making and reduce excessive prescribing and use.

Various interventions are available that address inappropriate antibiotic prescribing and the pressures of providing in-person care for patients with URTIs. These include delayed (post-dated) antibiotic prescriptions⁹ and different interventions to guide the use of antibiotics (e.g., clinical decision support tools, rapid point-of-care testing [POCT], viral prescription pads).¹⁰ The desired outcome is to reduce the overprescribing of antibiotics and antibiotic resistance, and potentially to improve clinical outcomes (e.g., reduce harms from antibiotics) and reduce health care utilization such as hospital and emergency department visits.

Changes to health care service delivery and organization due to the COVID-19 pandemic mean providers in primary care face difficulty in caring for patients with URTIs.⁷ Practical guidance and interventions continue to be required to assist clinicians in making care decisions for patients with upper respiratory tract symptoms. Resources such as the Using Antibiotics Wisely campaign's updated The Cold Standard toolkit provide resources for clinicians to support virtual and in-person management.¹¹ However, there continues to be an interest and need for evaluating contemporary evidence to identify which interventions might be most effective for supporting shared decision-making and optimizing the care of people with URTIs.

This report aims to summarize the evidence regarding the clinical effectiveness of delayed antibiotic prescription-filling interventions and other family medicine interventions to influence the use of antibiotics for suspected acute URTIs.

Research Questions

1. What is the clinical effectiveness of a delayed antibiotic prescription filling intervention to influence the use of antibiotics for suspected acute upper respiratory tract infections?

2. What is the clinical effectiveness of other family medicine interventions to influence the use of antibiotics for suspected acute upper respiratory tract infections?

Methods

Literature Search Methods

A limited literature search was conducted by an information specialist on key resources including MEDLINE, the Cochrane Database of Systematic Reviews, the international HTA database, the websites of Canadian and major international health technology agencies, as well as a focused internet search. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were URTIs and antibiotics. Search filters were applied to limit retrieval to health technology assessments, SRs, meta-analyses, or network meta-analyses. Where possible, retrieval was limited to the human population. The search was also limited to English-language documents published between January 1, 2016, and December 14, 2020.

Selection Criteria and Methods

One reviewer screened citations and selected studies. In the first level of screening, titles and abstracts were reviewed, and potentially relevant articles were retrieved and assessed for inclusion. The final selection of full-text articles was based on the inclusion criteria presented in Table 1.

Exclusion Criteria

Articles were excluded if they did not meet the selection criteria outlined in Table 1, they were duplicate publications, or they were published prior to 2017. SRs in which all relevant studies were captured in other more recent, or more comprehensive, SRs were excluded. Primary studies in SRs were not considered relevant for the report if they only included patients with lower respiratory tract infection ([LRTI], e.g., pneumonia) or a mixture of patients with LRTI and URTI and did not report outcomes for URTI separately. Further, primary studies in SRs were not considered relevant if they took place in settings other than family medicine/general practice (GP) settings or primary care clinics (e.g., studies were not considered relevant if they took place in the emergency department or a pediatric clinic).

Critical Appraisal of Individual Studies

The included SRs were critically appraised by one reviewer using A MeaSurement Tool to Assess systematic Reviews 2 (AMSTAR 2).¹² Summary scores were not calculated for the included studies; rather, the strengths and limitations of each included publication were described narratively.

Summary of Evidence

Quantity of Research Available

A total of 382 citations were identified in the literature search. Following the screening of titles and abstracts, 343 citations were excluded and 39 potentially relevant reports from the electronic search were retrieved for full-text review. Four potentially relevant publications were retrieved from the grey literature search for full-text review. Of these potentially relevant articles, 30 publications were excluded for various reasons and 13 publications met the inclusion criteria and were included in this report. These were all SRs. One SR was on delayed antibiotic prescribing,¹³ one SR was on delayed antibiotic prescribing and family medicine interventions,¹⁴ and 11 SRs were on family medicine interventions.^15-25 Some primary studies were included in more than one SR. The 2 SRs on delayed antibiotic prescribing reported on 4 of the same RCTs.^26-29 Two of the SRs on family medicine interventions reported on the same 3 primary studies^30-32 but reported outcomes differently from one another and 1 reported an additional outcome. Another 2 SRs reported on 1 of the same primary studies.³³ A table of primary study overlap is provided in Appendix 5.

Appendix 1 presents the PRISMA³⁴ flow chart of the study selection. Additional references of potential interest are provided in Appendix 6.

Summary of Study Characteristics

Additional details regarding the characteristics of included publications are provided in Appendix 2.

Study Design

All 13 eligible studies were SRs and all had broader inclusion criteria than the present report. Specifically, the SRs reported on a mixture of both LRTI and URTI, and/or a variety of clinical settings (e.g., emergency department, pediatric clinics, and primary care). One SR by Mortazhejri et al.¹⁴ conducted in 2020 (search date November 2016) examined studies on both delayed antibiotics (7 RCTs) and interventions aimed at patients and the public (6 studies comprising different interventions and settings) for URTIs. Only 4 of the studies on patient and public interventions were relevant for this report; they were all RCTs.

Another SR by Spurling et al.¹³ on delayed antibiotics for respiratory tract infections (RTIs) was a living SR initially published in 2017 (search date May 2017) and with a most recent search of November 2020. No additional studies had been identified since the 2017 search. This SR contained 11 RCTs in total, which broadly addressed RTIs in a variety of clinical settings. Five RCTs were relevant for this report, which addressed URTIs in a family medicine setting.

Eleven SRs addressed family medicine interventions. The SR by Carvalho et al.²¹ was published in 2020 (search date February 2020) and contained 22 studies in total, which covered a range of RTIs in different settings. Two studies (one pre-post and one cohort study) were relevant for this report. Cohen et al. published an SR²² in 2020 (search date June 2019), which contained 5 RCTs in different settings. Three of these RCTs were relevant to this report. Another 2020 SR (search date March 2019) by Fraser et al.²⁵ contained 38 primary studies on sore throat in a variety of settings and 3 RCTs were relevant for this report, as they were conducted in a family medicine setting specifically. Van Hecke et al. conducted an SR²⁴ in 2020 (search date January 2020), which contained 35 studies on a range of different POCTs in different outpatient settings. Two of these studies (1 RCT 1 one retrospective chart review) were relevant for this report, as they were on URTI in family medicine.

In 2019, Burstein et al. conducted an SR²⁰ (search date January 2016), which contained 34 studies addressing antibiotic use in a broad range of conditions and settings. One controlled before-after study on URTIs was relevant from this SR. Another 2019 SR (search date April 2018) was conducted by Health Information and Quality Authority¹⁹ and contained 13 primary studies on a range of RTIs. Five of those studies (1 before-after study, 1 cross-sectional study, 1 observational study, 2 RCTs) were relevant for this report, as they were specifically on URTI.

In 2018, Deniz et al. conducted an SR¹⁶ (search date June 2017) containing a total of 7 studies in either primary care or secondary care settings. Three observational longitudinal studies conducted in primary care were relevant for this report. Kochling et al. also conducted an SR in 2018¹⁷ (search date August 2016), which contained 17 studies on any RTI, and 3 RCTs from this SR were relevant for this report (specifically investigating URTIs). Lane et al. conducted an SR in 2018¹⁵ (search date April 2016), which included 3 primary studies (1 RCT, 1 cohort study with a historical control, and 1 retrospective cohort study). McDonagh et al. conducted an SR in 2018²³ (search date January 2018), which was an update to a 2016 Agency for Healthcare Research and Quality comparative effectiveness report.³⁵ This SR contained 96 studies on any RTI in a mixture of settings. The results were not reported in a way that allowed for determination of the relevance of primary studies.

One SR from 2017 (search date February 2017) was conducted by Schuetz et al.¹⁸ This SR contained 32 studies, in total, on different RTIs in a mixture of settings. Two of those RCTs were relevant for this report; the outcome data from this SR on URTI specifically was reported in a separate publication by Odermatt et al.³⁶

As described, all SRs had broader inclusion criteria than the present review. Only the characteristics and results of the subset of relevant studies will be described further in this report.

Country of Origin

The SRs were conducted in Canada,¹⁴ Australia,¹³ Portugal,²¹ France,²² the UK,^15,24,25 the US,^20,23 Ireland,¹⁹ the Netherlands,¹⁶ Germany,¹⁷ and Switzerland.¹⁷

Patient Population

For the primary studies in the Mortazhejri et al. SR¹⁴ on delayed antibiotics, the patient population was those patients in any age group with URTI seeking treatment in a general/family practice setting. Two of the eligible RCTs were conducted in children (n = 315 and n = 81), 1 in adults (n = 149), and 4 were conducted in both children and adults (n = 129, n = 716, n = 556, and n = 398). In the Spurling et al. SR,¹³ among 5 eligible primary studies, 1 RCT was conducted in children (n = 712) with sore throat, 1 (n = 398) was in adults with uncomplicated RTI, 1 (n = 191) was in adults and children with cough, 1 was among children with AOM (n = 315), and 1 (n = 129) was among adults and children with the common cold.

With respect to family medicine interventions, relevant studies from SRs were those conducted in family medicine, GP, or the primary care clinic setting. The 2 relevant primary studies in the Carvalho et al. SR²¹ were conducted in the general population (children and adults). One study was on patients with sinusitis (comprising 438 visits) and another on those with sinusitis or pharyngitis (comprising 240 visits).

In the Cohen et al. SR,²² the 3 relevant RCTs were all conducted among patients with acute sore throat or a clinical diagnosis of pharyngitis or tonsilitis. One RCT was conducted in those greater than or equal to 3 years of age (mean age 31, n = 424), 1 in adults (age not reported, n = 533), and 1 in patients aged 14 to 60 years (mean age 32, n = 557). The relevant primary studies in the SR by Fraser et al.²⁵ were the same 3 from Cohen et al.; however, these SRs reported outcomes differently.

The 4 relevant RCTs from the Mortazhejri et al. SR¹⁴ included 2 RCTs (n = 558 and n = 499) among children, 1 among patients with no age reported (n = 914), and 1 among adults (n = 2923) — this SR focused on URTI in general, but the authors did not specify the type of RTI in the primary studies.

The SR by Van Hecke et al.²⁴ included 2 relevant reports, which were both conducted among children with acute sore throat (n = 1,307 and n = 176).

The population in the 1 relevant report from the Burstein et al. SR was one with URTI but was not further described (although the authors noted that all studies in their SR targeted adults).²⁰

In the Health Information and Quality Authority SR, 1 study was conducted in patients of all ages with either acute sinusitis, tonsilitis, or otitis media (n = 367 GPs, patient characteristics was not described), 1 was conducted in adults with acute cough or sore throat (n = 120), 1 was conducted among 4,264 adults with RTI in general (authors reported URTI results separately, but the number of patients with URTI was not reported), another was conducted among 258 adults with either LRTI or rhinosinusitis (reported separately, but the number with rhinosinusitis was unclear), and 1 study among 560 patients with acute rhinosinusitis did not report the age of patients.¹⁹

All 3 relevant studies in the Deniz et al. SR¹⁶ were conducted among children with AOM. For the 3 relevant studies in the Kochling et al. SR, 1 was conducted among patients aged 14 to 60 with acute pharyngitis (n = 557); 1 was among patients older than or 21 years (n = 914) with either runny nose, blocked nose, cough, or sore throat; and 1 was among patients older than 16 years with acute cough (n=1,651).¹⁷

In the SR by Lane et al.,¹⁵ 1 primary study was conducted among 22 medical residents and examined prescribing for URTI and bronchitis (results were reported separately for each) but did not provide patient characteristics; another study looked at 28,301 patient encounters (for 69 physicians) on febrile respiratory illness, but patient characteristics were not provided; and a third study was conducted among patients with influenza-like illness (in 27 GP practices), but patient characteristics were not provided.

The primary studies in the McDonagh et al. SR²³ were not reported individually based on the type of RTI and setting — the authors noted that among the 96 studies in their SR, 21 studies were among patients with AOM, 32 studies were on sore throat, 7 studies were on rhinitis, 22 studies were on sinusitis, and 16 studies were on cough and common cold.

In the SR by Schuetz et al., 1 relevant RCT was conducted among people with upper or lower RTIs (n = 458, mean age 40 years), while the other relevant RCT was also conducted among people with upper or lower RTIs (n = 571, mean age 40 years).¹⁸ The patients with URTI (n = 644) from both studies were reported in a separate publication — an individual patient meta-analysis by Odermatt et al.³⁶

Interventions and Comparators

The delayed antibiotic interventions in the Mortazhejri et al. SR¹⁴ included a mixture of different approaches, including delayed prescriptions, patient-led delayed prescriptions (prescription given to patient at time of initial visit and patients given instructions to fill the prescription after a given time), post-dated prescriptions, delayed collection (a prescription was not provided at the time of visit but kept at office reception or the pharmacy for the patient to fill if symptoms were not improved after 2 to 7 days [depending on study]), and delayed re-contact. The comparator was not described for each study, but for the SR patient, intervention, comparator, and outcome (PICO) question, the comparator was noted to be usual care or no intervention. In the Spurling et al. SR, the intervention was delayed antibiotic use (use or advice to use antibiotics more than 48 hours after initial consultation) and the comparators were immediate antibiotic use and no antibiotic use.¹³

The family medicine interventions varied. In the Carvalho et al. SR, the intervention was e-health tools.²¹ The intervention in 1 relevant primary study was a best practice alert, while in the other relevant study the intervention was a multi-component intervention for providers comprised of didactic teaching, guideline use, and a clinical decision support system (CDSS) integrated into the electronic medical record. The comparator in both studies was the pre-intervention period.

In the Cohen et al. SR, the intervention of interest was rapid tests alone or in combination with scoring systems and the comparator was management based on clinical grounds with or without scoring systems.²² In the Fraser et al. SR, the intervention was rapid POCTs for group A Streptococcus (Strep A) with or without clinical scoring systems compared to antibiotic prescribing based on clinical judgment and/or clinical scoring tools alone.²⁵

In the Mortazhejri et al. SR,¹⁴ the intervention of interest was patient-centred interventions, which varied depending on the study. Interventions in the relevant primary studies included patient education sessions, pamphlets, websites, and pamphlets plus videos. The comparator was no intervention or usual care.

The SR by Van Hecke et al. investigated diagnostic POCTs for various diseases.²⁴ In the relevant studies, the intervention was POCTs for sore throat (one study was OSOM Strep A Test, the other was latex agglutination antigen detection method). The comparator in 1 study was usual care, while the comparator in the other study was the pre-intervention period.

In the 1 relevant study in the Burstein et al. SR, the intervention was a mixture of brochures and a media campaign aimed at patients, plus small group education sessions and algorithms for providers implemented in 1 community.²⁰ The comparator was communities not receiving the intervention.

The intervention of interest in the Health Information and Quality Authority SR was CRP POCTs either with or without communication, training, and education, while the comparator was standard care alone.¹⁹ In the Deniz et al. SR, the intervention of interest was guidelines on AOM, while the comparator was a pre-guideline introduction.¹⁶

The Kochling et al. SR studied various primary care interventions to reduce antibiotic prescriptions.¹⁷ The interventions of interest in the relevant studies were a rapid antigen Strep A test in one study, a multi-component intervention in another study (communication training for physicians, handouts for patients, waiting room poster), and patient education on causes of URTI in another study. The comparator was usual care in 2 studies, while the comparator was education on influenza vaccination in the other study.

The intervention of interest in the Lane et al. SR was the use of real-time epidemiological data to improve prescribing by reducing diagnostic uncertainty.¹⁵ The comparator was a pre-intervention period in one study, a historical control in another study, and was described as “control” in one RCT.

The SR by McDonagh et al. investigated various interventions to reduce inappropriate antibiotic use compared to standard care (no strategy of improving antibiotic use); however, no relevant reports could be identified from this SR because of the way the results were reported.²³

Finally, the intervention in the Schuetz et al. SR was the use of procalcitonin measurements to guide antibiotic decisions and the comparator was standard care.¹⁸

Outcomes

In the Mortazhejri et al. SR,¹⁴ the outcomes of interest were prescription or use of antibiotics, patient satisfaction with the consultation, and reconsultation for the same illness. In the Spurling et al. SR, the outcomes of interest were clinical outcomes (e.g., malaise, fever, pain), antibiotic use, patient satisfaction, adverse effects of antibiotics, complications of the disease, and reconsultation.¹³

For family medicine interventions, antibiotic use was reported as an outcome in the SRs by Carvalho et al.,²¹ Cohen et al.,²² Fraser et al.,²⁵ Mortazhejri et al.,¹⁴ Van Hecke et al.,²⁴ Burstein et al.,²⁰ Health Information and Quality Authority,¹⁹ Deniz et al.,¹⁶ Kochling et al.,¹⁷ Lane et al.,¹⁵ and Schuetz et al.¹⁸ (outcome data from Odermatt et al.³⁶). Guideline adherence was an outcome in the Carvalho et al. SR,²¹ while inappropriate antibiotic prescription rate was an outcome in the Fraser et al. SR.²⁵ Cohen et al. reported the outcome of recovery, defined as substantial improvement or complete recovery at day 7.²² Health Information and Quality Authority examined time to resolution of symptoms as an outcome.¹⁹ The rate of reconsultation was reported by Cohen et al.,²² Mortazhejri et al.,¹⁴ Van Hecke et al.,²⁴ and Health Information and Quality Authority ¹⁹ Patient satisfaction was reported by Mortazhejri et al.¹⁴ and Health Information and Quality Authority ¹⁹ Cohen et al. reported on the rate of complications attributed to the index infection.²² As aforenoted, the Schuetz et al.¹⁸ SR reported on outcomes for patients with both LRTI and URTI — Odermatt et al. was a separate publication based on this SR, which reported on patients with URTI, only, and reported the number of days with restricted activity and the treatment failure rate at 28 days.³⁶

Summary of Critical Appraisal

Additional details regarding the strengths and limitations of included publications are provided in Appendix 3.

Systematic Reviews

Delayed Antibiotics

Overall, the reviews contributing data on delayed antibiotics had few sources of potential bias. The main weaknesses that were identified related to under-reporting the rationale for study inclusion criteria, lack of context provided for the impact of risk of bias and heterogeneity on the review results, and non-critical weaknesses with the search strategies. Details are available in Table 4 and the following summary.

The 2 eligible systematic reviews^13,14 reported their research questions and inclusion criteria in detail following the PICO format. They included a reference to a protocol published a priori. Both reviews explicitly listed eligible study designs and exclusions but did not provide a rationale for the criteria utilized.

Both reviews employed comprehensive literature searches with some limitations. One review¹⁴ did not conduct the search within 24 months of publication of the review (the search was conducted in November 2016 and the review was published in 2020), meaning that studies published since 2016 were not captured in this review., Further, the authors did not search reference lists or include a grey literature search.¹⁴ The other SR is a living SR and had an up-to-date search but did not outline a formal grey literature search.¹³

Both reviews performed study selection and data extraction in duplicate and provided a formal list of excluded studies (with reasons for exclusion) and described the included studies in detail.

Both reviews used appropriate tools for assessing the risk of bias — the Effective Practice and Organisation of Care (EPOC) risk-of bias assessment guide adapted by study design¹⁴ and the Cochrane risk-of-bias tool,¹³ respectively. One review did not assess the potential impact of risk of bias on the pooled results¹⁴; the other review planned sensitivity analysis based on risk-of-bias assessments.¹³ Both reviews accounted for the impact of the risk of bias in interpreting and discussing the results of the review and documented primary study funding sources.

Criteria and statistical approaches for pooling were appropriate in both reviews, but 1review lacked details about the specific analytical approach.¹³ Both reviews formally assessed and investigated sources of heterogeneity; however, the scope of planned and executed analyses investigating factors contributing to heterogeneity was narrow in both cases. Both reviews planned to assess publication bias using funnel plots, but this was not possible because of the number of included studies. One review further considered publication bias¹³ within the Grading of Recommendations Assessment, Development and Evaluation (GRADE) assessments, and unpublished data were sought and used. Both reviews detailed funding sources and declared no conflicts of interest.

Family Medicine Interventions

The reviews contributing data on family medicine interventions had varying degrees of bias. Several critical and non-critical weaknesses were identified with select reviews, including under-reporting of details of and rationale for study inclusion criteria or study characteristics, incomplete efforts to identify studies through alternative search approaches, lack of duplicate selection or extraction, lack of disclosure of excluded studies, inadequate assessment of the risk of bias, inappropriate methods for quantitative synthesis, lack of consideration of the risk of bias and publication bias in quantitative synthesis, and lack of consideration of the risk of bias and heterogeneity in the interpretation of review findings. Details are available in Table 5 and the following summary.

The 12 reviews reporting on family medicine interventions mostly included sufficient detail about scope and inclusion criteria, including all of the PICO components. Some reviews were either missing sufficient detail about main inclusion criteria (mainly comparators where relevant)^17,20,21 or were lacking detail (e.g., time frame for follow-up) or clarity in their presentation.¹⁵

Registered protocols were available for most of the included reviews, although not all described deviations from the protocol (if applicable).^{14-16,18,21-25} Some reviews either stated that the review was not pre-registered or did not reference a protocol or a priori methods, making it challenging to assess whether analyses were pre-planned.^17,19,20

Most reviews listed the eligible study designs, including specific trial or study design features, for inclusion; however, only one review explicitly included a rationale for included and excluded study designs.²⁴

Most reviews included a comprehensive literature search strategy; however, almost all were missing at least one desired component, such as searching trial registries, scanning reference lists, consulting topic experts, or including a formal grey literature search strategy. The search in 1 review was not conducted within the desired time frame of 24 months.¹⁴

Except for 2 reviews, all others conducted study selection in duplicate. The 2 reviews noted conducted initial screening using one reviewer, with selections checked by a second, potentially increasing the likelihood of missing relevant studies.^15,23 Apart from 4 reviews, most conducted data extraction in duplicate. Two^23,25 used a single review for data extraction, with results checked by another, potentially increasing the likelihood of recording errors. Methods of data extraction, including the number of reviewers involved, were unclear for 2 reviews.^20,24 Most reviews did not include a comprehensive list of excluded studies with reasons for exclusion, making it challenging to assess the accuracy and completeness of study selection. In most cases, review authors noted the number of excluded studies and aggregate reasons for exclusion in text or the PRISMA diagram of the review.

Most reviews described the characteristics of the included studies in adequate detail. However, specific information about interventions and comparators, and time frame of follow-up for various outcomes, were often missing. Several studies had more serious omissions. One did not adequately describe the study design, comparators, details of the interventions, and follow-up time frame.²⁰ The Odermatt et al. publication³⁶ reported outcomes from the Schuetz et al.¹⁸ SR — the characteristics of the relevant studies were well-described in the original Schuetz et al. publication; however, the studies were not described in detail in the Odermatt et al. publication (although the characteristics of the patients were described).

A satisfactory tool or approach for assessing the risk of bias was employed in most cases. One review used a scale that did not assess all relevant domains, notably failing to consider allocation concealment and selective outcome reporting.²¹ One review used a set of predefined criteria in place of a formal tool, which addressed most potential sources of bias except for bias resulting from multiple measurements and bias due to methods used to ascertain exposures and outcomes.²³ Another SR extended the Cochrane risk-of-bias tool for RCTs to non-randomized studies.²⁴ One review did not formally assess the risk of bias.²⁰ Five reviews disclosed primary study funding sources,^{14,18,19,22,25} while the others failed to report this information. The Odermatt et al. publication (reporting outcomes from the Schuetz et al. SR) did not include detail on study quality, although this was formally assessed in the original Schuetz et al. SR.¹⁸

Of the 6 reviews that conducted meta-analysis, most used appropriate methods for combining results statistically, including a formal assessment of the appropriateness of pooling based on study similarity and a description of the type of meta-analytic model.^{14,18,19,22,25} One review pooled randomized and non-randomized studies without providing a rationale.²⁴ Of these reviews, one formally planned to assess the potential impact of risk of bias on the pooled analyses through sensitivity analysis²² and 1 planned sensitivity analysis based on the main quality indicators.¹⁸ The other reviews did not formally assess the potential impact of risk of bias on the pooled results. In one SR¹³ examining clinical outcomes for delayed prescribing versus immediate prescribing, the sample size and event rates for clinical symptoms were relatively low, making it difficult to draw firm conclusions on these outcomes. The difference in pain between the 2 groups in some of the primary studies may have been clinically relevant but were not statistically significant. Most reviews did not formally assess publication bias. Two reviews either formally assessed publication bias or provided a rationale for why it was not feasible, sought unpublished data, and considered publication bias in their GRADE assessments of the certainty of the evidence.^18,19 Some reviews pursued unpublished studies or data, or considered publication bias informally as part of the GRADE assessments.^14,22,25

In interpreting and discussing the results of the review, 6 reviews accounted for risk of bias by directly discussing the potential impact of methodological considerations on the results, incorporating risk of bias considerations in assessments of the certainty of the findings (e.g., GRADE) or in establishing a hierarchy of the importance of findings.^{14,17-19,22,25} The remaining reviews either summarized the risk of bias results without putting them in context with the results or failed to discuss the risk of bias in interpreting the results.^{15,16,20,21,24} A satisfactory explanation for and discussion of the heterogeneity observed in the results was presented by some reviews, including a description of the relevant sources of heterogeneity and their potential implications or factoring heterogeneity into assessments of the certainty of the findings (e.g., GRADE).^{14,17-19,22,25} Other reviews failed to discuss heterogeneity, mentioned it briefly without discussing implications, or limited the discussion of heterogeneity to select possible sources when other factors may have impacted the findings.^{15,16,20,21,24} All included reviews disclosed any conflicts of interest, or lack thereof, and relevant funding sources.

Summary of Findings

Delayed Antibiotic Prescriptions

Antibiotic Use

In a meta-analysis of 6 RCTs involving 1,788 patients, Mortazhejri et al. reported that patients in the delayed antibiotic group were less likely to use antibiotics compared to those in the immediate prescription group (odds ratio [OR] = 0.09; 95% confidence interval [CI], 0.03 to 0.23).¹⁴ The authors conducted subgroup analysis comparing delayed patient-led or post-dated prescriptions (i.e., patients received at index consultation) to immediate prescription, and delayed collection (prescription not provided at time of visit but kept at office reception or pharmacy for patient to fill if symptoms had not improved after 2 to 7 days [depending on the study]) to immediate prescriptions, and found both approaches had similar effects in reducing antibiotic use compared to immediate prescriptions. The authors concluded that delayed antibiotic prescriptions significantly reduced the use of antibiotics for URTIs. Spurling et al. also investigated various delayed prescription strategies.¹³ Four of the RCTs in the Spurling et al. SR were also included in the meta-analysis from Mortazhejri et al.¹⁴ Relevant RCTs from the Spurling et al. SR all demonstrated significantly reduced antibiotic use for both delayed prescriptions received at the time of consultation and delayed collection compared to immediate antibiotics.¹³ In 2 studies, the odds of antibiotic use were lower for delayed prescriptions at time of visit compared to immediate prescription; the OR for antibiotic use was 0.20 (95% CI, 0.09 to 0.44) in one study and 0.04 (95% CI, 0.01 to 0.13) in another study. In another 4 studies, the odds of antibiotic use were lower for delayed collection versus immediate antibiotics; the 4 relevant studies had ORs of 0.03 (95% CI, 0.01 to 0.09), 0.00 (95% CI, 0.00 to 0.07), 0.00 (95% CI, 0.00 to 0.02), and 0.05 (95% CI, 0.02 to 0.08). In this SR, 1 relevant RCT reported that delayed prescriptions at time of visit increased antibiotic use compared to no antibiotic prescription (OR = 3.47; 95% CI, 1.34 to 9.01), while delayed collection increased antibiotic use in 2 studies (OR = 2.14; 95% CI, 0.81 to 5.66, in 1 study; OR = 3.18; 95% CI, 1.85 to 5.46, in the other study). The authors concluded that delayed prescriptions markedly reduced the use of antibiotics compared to immediate antibiotics and the least antibiotic use was in the no-antibiotic group.¹³

Patient Satisfaction

Four RCTs from the Mortazhejri et al. SR reported on patient satisfaction.¹⁴ Compared to patients receiving immediate antibiotics, those receiving delayed antibiotics were more satisfied in 1 study (OR = 1.47; 95% CI 0.32, to 6.85), less satisfied in 2 studies (OR = 0.61, 95% CI, 0.25 to 1.49; and OR = 0.32; 95% CI, 0.16 to 0.65), and there was no difference in 1 study (P = 0.14, no further detail provided). Spurling et al. also compared the number of patients satisfied with delayed antibiotics versus immediate antibiotics.¹³ Among relevant RCTs, the ORs comparing delayed versus immediate antibiotics crossed 1 in 4 studies, suggesting no difference in patient satisfaction between delayed antibiotics or immediate antibiotics. In one study, the odds of satisfaction were decreased in the delayed group compared to the immediate antibiotic group (OR = 0.32; 95% CI, 0.16 to 0.65). The authors of this SR¹³ concluded that delayed prescriptions did not result in different satisfaction levels compared to immediate antibiotics and that satisfaction was high in both groups.

Reconsultation

Four RCTs in the Mortazhejri et al. SR reported on the intention to re-consult and 1 RCT reported on the actual reconsultation rate.¹⁴ There was no difference in the proportion of patients intending to re-consult in the delayed antibiotic group versus the immediate antibiotic group in 2 studies (73% versus 65% in 1 study [OR = 1.50; 95% CI, 0.71 to 3.17] and 69% versus 86% in the other study [OR = 0.62; 95% CI, 0.19 to 2.06]). In 2 studies, the proportion intending to re-consult was lower in the delayed antibiotic group (57% versus 79% in 1study [OR = 0.35; 95% CI, 0.22 to 0.55] and 63% versus 83% [OR = 0.35; 95% CI, 0.20 to 0.62] in the other study). In 1 RCT, there was no difference in the actual reconsultation rate between groups at one month (P = 0.56, no other data provided). While the findings were mixed, these results suggest that delayed antibiotics appear to have either no effect on, or slightly reduce, the intention to re-consult. The authors of the SR did not comment on or make an overall conclusion about the effect of delayed antibiotics on reconsultation rates.

Clinical Symptoms

The Spurling et al. SR reported on pain, malaise, and fever.¹³ Two RCTs in this SR reported the number of participants with pain on days 3 to 6 in the delayed antibiotic group versus the immediate antibiotic group. In 1 study, there was a numerically greater proportion of participants with pain in the delayed group (13/61 [21%]) compared to the immediate antibiotic group (9/58 [16%]) (OR = 1.47; 95% CI, 0.58 to 3.77); however, this was not statistically significant. In another primary study, the proportion of participants with pain was numerically higher in the delayed antibiotic group than in the immediate antibiotic group (28/111 [25%] versus 15/101 [15%]), but this difference was not statistically significant (OR = 1.93; 95% CI, 0.96 to 3.88). One study in this SR found that a higher proportion of patients experienced pain on day 3 in the delayed antibiotic group (45/150 [30%]) compared to the immediate antibiotic group (19/135 [14%]) (OR = 2.62; 95% CI, 1.44 to 4.76). Finally, 1 RCT in the Spurling et al. SR found no difference in the proportion of patients with fever on days 3 to 6 in the delayed antibiotic group (5/67 [7%]) compared to the immediate antibiotics group (6/62 [10%]) (OR = 0.75; 95% CI, 0.22 to 2.60).¹³ The authors of the SR concluded that, for many clinical outcomes, there was no significant difference between delayed and immediate antibiotics. However, they noted that symptoms for AOM and sore throat may have been modestly improved by immediate antibiotics compared with delayed antibiotics.

Adverse Effects of Antibiotics

One RCT in the Spurling et al. SR found no difference in the proportion of patients experiencing vomiting in the delayed antibiotic group compared to the control group (15/179 [8%] versus 18/215 [8%], OR = 1.00; 95% CI, 0.49 to 2.05).¹³ Three RCTs in the Spurling et al. SR reported on rates of diarrhea in the delayed antibiotic group versus immediate — in 2 studies, there was no difference in the proportion of patients experiencing diarrhea in the delayed antibiotic group compared to the immediate group (11/67 versus 12/62 [OR = 0.82; 95% CI, 0.33 to 2.02] in 1 study and 23/179 versus 23/215 [OR = 1.23; 95% CI, 0.67 to 2.28] in another study). One study reported a lower rate of diarrhea in the delayed antibiotic group compared to the immediate antibiotic group (14/150 [9%] versus 25/135 [19%], OR= 0.45; 95% CI, 0.22 to 0.91). Two studies found no difference in the rates of rash for delayed antibiotics versus immediate — the rates in one study were 11/180 (6%) versus 14/215 (7%) (OR = 0.93; 95% CI, 0.41 to 2.1) and the rates in other study were 8/150 (5%) versus 6/135 (4%) (OR = 1.21; 95% CI, 0.41 to 3.58). The authors of the SR did not comment on this specifically; however, these results suggest that there appears to be no difference in the rates of adverse effects of antibiotics for delayed antibiotics compared to immediate antibiotics, with one study suggesting possible lower rates of diarrhea for the delayed antibiotic strategy.

Family Medicine Interventions

Antibiotic Use and Prescribing

One pre-post study in the Carvalho et al. SR examined antibiotic use, finding that a best practice alert reduced antibiotic prescriptions at sinusitis visits in the year following its introduction compared to the year before (86% before introduction versus 62% after, P < 0.01).²¹ The authors of this primary study concluded that clinical reminders in the electronic medical record could be an effective tool to reduce inappropriate antibiotic use.

Cohen et al. examined the number of participants provided with an antibiotic prescription for sore throat with rapid test-guided therapy with or without clinical scoring systems compared to treating on clinical grounds with or without a scoring system.²² All RCTs reported lower prescription rates for rapid test-guided approaches — 1 RCT reported rates of 86/213 versus 124/211 (risk difference = –18%; 95% CI, –28 to –9), 1 RCT reported 167/367 versus 263/385 (risk difference = –23%; 95% CI, –30 to –16), 1 RCT reported 98/223 versus 133/208 (risk difference = –20%; 95% CI, –0.29 to –0.11), and another reported RCT 63/196 versus 155/274 (risk difference = –24%; 95% CI, –33 to –16). These authors also meta-analyzed 2 RCTs comparing antibiotic dispensing rates for rapid tests with or without scoring systems to those for treating on clinical grounds with or without scoring systems and reported lower dispensing rates in the rapid test group (156/445) compared to the clinical grounds group (197/455) (risk difference= –7%, 95% CI, –17 to –2). The authors concluded that rapid testing to guide antibiotic treatment for sore throat probably reduces antibiotic prescription rates but may have little or no impact on dispensing (although there were statistically fewer antibiotics dispensed in the rapid testing arm).

Fraser et al.²⁵ reported on the same 3 RCTs as Cohen et al.²²; however, Fraser et al. reported results by individual study arm in each RCT. In 1 RCT, the number of participants receiving antibiotics was lower for those receiving rapid Strep A test-guided therapy plus clinical criteria (Centor) compared to clinical criteria alone (123/281 [44%] versus 168/262 [64%]). In another RCT, the authors reported the number of participants provided with an antibiotic prescription was lower when therapy was guided by a rapid Strep A test plus clinical score (18% received immediate prescriptions plus 23% received delayed) compared to a clinical score alone (16% received immediate prescriptions plus 41% received delayed). Finally,1 RCT reported the number of participants who received antibiotics with use of a rapid Strep A test plus decision rule (39/102, 38%) versus the rapid test alone (32/120, 27%) versus the decision rule alone (94/170, 55%) versus usual care (82/141, 58%). The authors in this SR concluded that antibiotic prescription rates or use was higher in control arms compared to those incorporating POCTs.

Mortazhejri et al.¹⁴ examined antibiotic prescribing for various interventions directed at patients. In 1 RCT, there was no difference in the number of antibiotics prescribed per patient for an education pamphlet and video compared to control (mean difference ‒0.3, P = 0.23). In another RCT, there was no difference in antibiotic prescriptions for a pamphlet versus control (OR = 1.20; 95% CI, 0.84 to 1.72). Another RCT found reduced odds of receiving an antibiotic prescription at index consultation for booklets on RTIs aimed at parents of children compared to control (OR = 0.29; 95% CI, 0.14 to 0.60). Finally, 1 RCT reported no difference in antibiotic prescriptions between an interactive website and control (relative risk [RR] = 1.02; 95% CI, 0.82 to 1.43). The authors concluded that the effect of interventions varied depending on the type of material provided.

Van Hecke et al. examined antibiotic prescribing for POCTs to treat sore throat versus control.²⁴ One RCT reported antibiotic-prescribing rates of 261/581 for POCT versus 364/726 for usual care (RR = 0.90; 95% CI, 0.80 to 1.01), suggesting no difference in antibiotic-prescribing rates for these strategies. In a pre-post study described in this SR, the use of a POCT reduced antibiotic prescription rates post-implementation (34/68 versus 65/108 pre-implementation, RR = 0.48; 95% CI, 0.33 to 0.69). The authors concluded that POCT did not reduce antibiotic prescribing in RCTs but did show an effect in non-randomized studies.

In the Burstein et al. SR, 1 controlled before-after study reported that a multi-faceted intervention led to a 16% reduction in antibiotic prescriptions for URTIs compared to a 2% reduction in the control group (P < 0.06).²⁰ The authors concluded that the intervention led to a modest reduction in antibiotic prescriptions for URTIs.

Health Information and Quality Authority examined antibiotic prescribing at index consultations for the use of CRP POCTs compared to usual care.¹⁹ In a meta-analysis of 2 RCTs, the authors reported antibiotic prescription rates of were lower for the CRP group compared to the usual care group (80/260 versus 109/250, RR = 0.72; 95% CI, 0.58 to 0.90). In a meta-analysis of 2 non-randomized studies, the authors reported lower antibiotic prescription rates for CRP compared to usual care (183/335 versus 287/345, RR = 0.65; 95% CI, 0.59 to 0.73). The authors concluded that the use of CRP POCT to guide the management of patients with RTIs leads to reduced antibiotic prescribing at index consultation.

The Deniz et al. SR reported on 2 pre-post studies comparing antibiotic prescription rates before and after the introduction of a guideline on AOM.¹⁶ In one study, antibiotic prescription rates fell by 12% and in another study there was no change. The authors of this SR concluded that the introduction of guidelines seems to have a modest impact on antibiotic prescription rates, at best.

Kochling et al. examined different primary care interventions.¹⁷ In 1 RCT, the OR for the prescription of an antibiotic dropped to 0.58 (95% CI, 0.43 to 0.78) 6 weeks after a multi-component intervention was introduced and increased to 1.52 (95% CI, 1.19 to 1.95) in the usual care group. At 12 months, the OR was 0.72 (95% CI, 0.54 to 0.97) for the intervention group and 1.31 (95% CI, 1.01 to 1.71) in usual care group. The authors stated that these ORs corresponded to a 60% relative reduction in antibiotic prescription rates at 6 weeks and a persistent 40% relative reduction at 12 months for the intervention compared to usual care. In another RCT, a rapid antigen test for Strep A reduced antibiotic prescribing compared to usual care (123/281 versus 168/262, rate difference = ‒20%). Another RCT in this SR reported no difference in the rate of antibiotic prescriptions for patients receiving education on causes of URTI compared to those receiving control education (94/457 versus 81/457, rate difference = 3%).

Lane et al. examined the effect of locally relevant epidemiological data provided to prescribers on antibiotic-prescribing rates.¹⁵ In 1 cohort study with a historical control, the antibiotic-prescribing rates fell to 9% from 26% (P = 0.01) in the 3 years following the introduction of a surveillance program compared to the control period. In another study, antibiotic-prescribing rates were lower when prescribers received epidemiological data during a pandemic compared to a non-pandemic period (antibiotics prescribed at 3054/7789 visits versus 9741/20512, OR = 0.72; 95% CI, 0.68 to 0.77). One RCT found no difference in antibiotic-prescribing rates when GPs were proved with a heat map and decision support tool compared to control (41% prescription rate for intervention versus 41% for control, P = 0.90). The authors concluded that surveillance systems could support antibiotic decision-making but that few interventions had been evaluated and existing studies were methodologically weak.

Odermatt et al.³⁶ reported outcomes of an individual patient meta-analysis based on the Schuetz et al. SR.¹⁸ This publication included individuals with URTI from 2 RCTs in the Schuetz et al. SR, which examined procalcitonin to guide antibiotic prescribing compared to control. These authors found that procalcitonin-guided interventions reduced antibiotic prescription rates compared to control (59/332 versus 159/312, OR = 0.2; 95% CI, 0.1 to 0.3). The authors also found that the mean number of days exposed to antibiotics was lower in the procalcitonin group versus control (1.2 days [standard deviation 2.8] versus 3.7 days [4.0]). The authors concluded that procalcitonin-guided therapy reduced antibiotic exposure in URTI patients without compromising outcomes (which are reported in the sections Days With Restricted Activity and Treatment Failure at 28 Days that follow).

Guideline Adherence

One study in the Carvalho et al. SR examined antibiotic-prescribing guideline adherence among providers receiving a multi-component intervention versus a historical control. The authors found that guideline adherence was 91% following the intervention compared to 77% in the control group.²¹

Recovery

One RCT in the Health Information and Quality Authority SR reported that no difference was found in recovery at 7 days for the use of CRP POCT to guide antibiotic prescribing compared to usual care (27/118 versus 31/125, RR = 1.03; 95% CI, 0.89 to 1.18).¹⁹

Time to Resolution of Symptoms

Two RCTs in the Health Information and Quality Authority SR examined the median time to resolution of symptoms for the use of CRP POCT to guide antibiotic prescribing compared to usual care. Both studies reported no difference in the time to resolution of symptoms.¹⁹

Reconsultation

The Cohen et al. SR meta-analyzed 2 RCTs that compared reconsultation for the use of rapid test-guided therapy with or without scoring systems to those for treatment on clinical grounds with or without scoring systems.²² There was no difference in reconsultation rates between the 2 groups (59/571 versus 51/590, OR 1.12 = 95% CI, 0.57 to 2.21).

The Mortazhejri et al. SR¹⁴ examined various patient-directed interventions and identified 2 RCTs that examined reconsultation rate. One study found no difference in reconsultation rates over 1 year for patients receiving an online program compared to control (19% versus 19%, OR = 0.93; 95% CI, 0.73 to 1.16). Another study found that the intention to re-consult was lower for patients receiving an interactive booklet compared to control (55% versus 76%, OR = 0.34; 95% CI, 0.20 to 0.57).

One RCT from the Van Hecke et al. SR found that reconsultation events were lower for POCT-guided treatment of sore throat compared to usual care (RR = 4.70; 95% CI, 2.94 to 7.51). No further data were provided, so the direction of RR and definition of outcome were unclear).²⁴

The Health Information and Quality Authority SR found 2 RCTs on the reconsultation rate for the CRP-guided treatment of URTIs compared to usual care.¹⁹ In 1 study, the reconsultation rate was 33/129 in the CRP group and 23/129 in the usual care group (RR = 1.43; 95% CI, 0.89 to 2.30). In another study, the reconsultation rate was 165/894 in the CRP group and 149/812 in the usual care group (RR = 1.01; 95% CI, 0.82 to 1.23).

Patient Satisfaction

In the Mortazhejri et al. SR, 1 RCT reported no difference in satisfaction for patients receiving an interactive booklet compared to control (90% versus 94%) (OR = 0.64; 95% CI, 0.33 to 1.22).¹⁴

In the Health Information and Quality Authority SR, 1 RCT found that the proportion of patients satisfied with care was higher for those receiving CRP POCT-guided antibiotic therapy compared to usual care (90/118 [76%] versus 79/125 [63%]).¹⁹ In another RCT, there was no difference in patient satisfaction for CRP-POCT testing versus usual care (RR = 1.01; 95% CI, 0.86 to 1.16 [event rates not reported]).

Complication Attributed to Index Infection

In the Cohen et al. SR, the authors examined the number of participants with a complication attributed to the index infection for the use of rapid tests with or without scoring systems to guide the treatment of sore throat compared to treatment on clinical grounds with or without scoring systems.²² In 1 RCT, the rate of complications was 2/213 in the rapid testing group and 0/211 in the clinical grounds group; in the other RCT, there were no events in either group.

Days With Restricted Activity

Schuetz et al. (data from Odermatt et al.³⁶) compared procalcitonin-guided antibiotic therapy to usual care and found no difference in the median number of days with restricted activity between the 2 groups (4.2 days [IQR 5 to 14] versus 4.3 days [IQR 5 to 13]) in a meta-analysis of 2 RCTs.¹⁸ The authors concluded there was no difference in the number of days with restricted activities between the groups.

Treatment Failure at 28 Days

Schuetz et al. (data from Odermatt et al.³⁶) also examined treatment failure at 28 days for procalcitonin-guided antibiotic therapy compared to usual care.¹⁸ There was no difference in the rate of treatment failure between the 2 groups (110/132 versus 106/132) (OR = 1.0; 95% CI, 0.7 to 1.4) based on a meta-analysis of 2 RCTs. The authors concluded there was no difference in the rate of treatment failure.

Limitations

Delayed Antibiotics

Several primary studies in the SRs on delayed antibiotics examined antibiotic use and patient satisfaction; however, there were fewer data on clinical outcomes such as symptoms, health care utilization, quality of life, and hospitalizations, as well as on antibiotic resistance and appropriateness of prescribing. Further, studies focused on rates of prescribing or use without information on whether antibiotics were over-or underprescribed.

The eligible SRs included studies with a mixture of patient populations (adults, children); however, it was unclear whether there were any differences in outcomes based on the population (Moratzhejri et al. did not investigate this¹⁴ and Spurling et al. noted no difference in outcomes for children compared to adults, but it was not possible to determine whether this was true for URTI in primary care specifically).¹³ The RCTs in the 2 SRs were conducted in Western countries; however, only one trial was conducted in Canada. The study populations were not described in detail in the SRs, making it challenging to understand the generalizability of the patient populations to the Canadian context. Further, antibiotic-prescribing practices (e.g., types of antibiotics prescribed, baseline use of antibiotics)¹⁷ and health care system differences between countries make the generalizability of these studies to the Canadian context unclear.

Family Medicine Interventions

The SRs of family medicine interventions reported on antibiotic use and prescribing, and 4 SRs reported on reconsultation rates, but like the studies on delayed antibiotic interventions, there were less data on other clinical outcomes. Similar to the studies on delayed antibiotic interventions, family medicine interventions focused on rates of prescribing or use without information on whether antibiotics were over- or underprescribed.

There was little evidence on clinical outcomes, health care utilization, and antibiotic resistance. One study in one SR reported on recovery and time to resolution of symptoms, 1 SR reported on complications attributed to index infection,²² while 1 SR reported on days with restricted activity.¹⁸ In some of the SRs, the interventions of interest were quite broad (e.g., “primary care interventions” in the Kochling et al. SR¹⁷), making it difficult to draw conclusions about the effect of the interventions across the review. Few of the SRs reported on the same types of interventions, further challenging the ability to draw conclusions on the body of evidence. There was more relevant evidence on different diagnostic tests (e.g., rapid Strep A tests, CRP) compared to patient and provider-directed interventions (e.g., education, tools); however, there were generally few relevant studies available for each individual intervention. The patient or provider-directed interventions also differed from one another in the included components, making it challenging to draw conclusions across studies. Another limitation was that some of the SRs included relatively few primary studies with poor methodological quality. For example, in the Lane et al. SR,⁵ only 3 primary studies were found, which the authors concluded were methodologically weak. In many of the SRs, few primary studies were relevant for this report, making it challenging to draw conclusions about the evidence. Further, only 1 primary study in the 11 SRs on family medicine interventions was conducted in Canada. Given possible differences in primary care systems and antibiotic-prescribing practices in other contexts, it is unclear whether the results of the eligible SRs are generalizable to the Canadian primary care context.

Conclusions and Implications for Decision- or Policy-Making

This report identified 1 SR¹³ examining delayed antibiotics for URTIs, 1 SR examining both delayed antibiotics and patient-directed interventions to reduce unnecessary antibiotic use in URTIs,¹⁴ and 11 SRs^15-25 examining various family medicine interventions to influence the use of antibiotics for URTIs.

Delayed Antibiotics

Delayed antibiotic prescriptions appeared to significantly reduce the use of URTIs compared to immediate antibiotic prescriptions in the 2 SRs^13,14 examining this — this applied to both delayed prescriptions provided at the time of visit and delayed collection (prescription not provided at time of visit but kept at office reception or pharmacy for patient to fill if symptoms not improved after, for example, 2 days). Based on the individual studies from the SRs, delayed antibiotics generally did not appear to show any difference in patient satisfaction compared to immediate antibiotics, while results were mixed on the reconsultation rate. The intent to re-consult was lower in the delayed antibiotic group compared to immediate antibiotics (in 2 RCTs from one SR¹⁴), there was no difference in intent to re-consult in 2 RCTs from the same SR,¹⁴ and there was no difference in the actual reconsultation rate (in 1 RCT from one SR).¹⁴ The effect of delayed antibiotic prescribing on clinical symptoms of URTIs was less clear — 1 SR¹³ concluded that symptoms of AOM and sore throat might have been modestly improved by immediate antibiotics compared to delayed antibiotics; however, this was based on 1 RCT examining malaise and fever, and 2 RCTs examining pain in URTI.

Family Medicine Interventions

Various family medicine interventions were identified and the effects appeared to differ based on the individual interventions. Different rapid tests or POCTs were examined in eligible SRs,^18,19,22 which all appeared to reduce antibiotic prescribing. A well-conducted SR on rapid testing for group A strep (with or without scoring systems) to guide the treatment of sore throat suggested that this intervention reduced antibiotic-prescribing rates compared to treatment on clinical grounds, but rapid testing led to a smaller reduction in antibiotic dispensing.²² Another well-conducted SR suggested that CRP POCT-guided treatment reduced antibiotic prescribing for URTIs at index consultations compared to usual care.¹⁹ One SR further reported that procalcitonin-guided antibiotic therapy in URTIs reduced antibiotic exposure compared to control. There were less data on other outcomes, in particular clinical outcomes.¹⁸ The use of CRP POCTs to guide antibiotic prescribing in URTIs led to no difference in recovery or time to resolution of symptoms in 1 SR; however, this was based on only 2 RCTs.¹⁹ In 1 SR,²² there was no difference in reconsultation rates for rapid Strep A tests (2 studies), while in another SR there was no difference in reconsultation rates for CRP POCT (2 studies) compared to usual care.¹⁹ One RCT in 1 SR found that a POCT for sore throat reduced re-consultations compared to usual care.²⁴ The use of procalcitonin testing to guide therapy led to no difference in the number of days with restricted activities or rates of treatment failure compared to control based on 1 meta-analysis of 2 RCTs in 1 SR.^18,36

There was more heterogeneity in findings for different patient or provider-directed interventions, such as education or tools. For example, 1 SR found no difference in antibiotic prescribing for 3 different patient-directed interventions (compared to either no intervention or usual care) but found that a pamphlet on RTIs aimed at parents of children reduced antibiotic prescribing compared to usual care in 1 study.¹⁴ Two SRs each identified 1 primary study (i.e., 2 primary studies in total) that found that multi-component interventions involving provider interventions (e.g., education, training, decision support tools), as well as patient-directed interventions (e.g., handouts, posters), may reduce antibiotic prescribing compared to control and usual care. However, the only relevant study in 1 SR was a controlled before-after study²⁰ and the authors of the other SR were unsure whether the reduction in antibiotic prescribing was clinically meaningful for 1 of the primary studies.¹⁷ One SR found few studies examining the effect of infection surveillance systems to support antibiotic decision-making, noting that available evidence was methodologically weak. Finally, 2 studies on the effect of guidelines for the treatment of AOM had mixed results and the authors of this SR suggested that this intervention may have a modest impact on antibiotic prescribing, at best.¹⁶ Results on reconsultation were mixed for patient-directed interventions — in 1 SR, there was no difference in the reconsultation rate in 1 study and a reduction in intention to re-consult in another.¹⁴

Implications

There is increasing awareness of the overuse of antibiotics. Strategies are needed to mitigate potentially inappropriate antibiotic use; for example, the use of antibiotics to treat URTIs that are viral in origin.⁷ Strategies aimed at influencing antibiotic use should reduce such inappropriate antibiotic use while ideally either improving or not compromising clinical outcomes or health care utilization. In this report, most of the available evidence pertained to how interventions impacted overall antibiotic prescribing and use. There was evidence that delayed antibiotics reduced antibiotic prescribing and use compared to immediate antibiotics, while rapid test/POCT-guided therapy was shown to reduce antibiotic prescribing compared to usual care for URTIs. Thus, these strategies appear to be useful to implement in order to reduce antibiotic-prescribing rates for URTI. Delayed antibiotic prescriptions also do not appear to have a negative effect on patient satisfaction or reconsultation rates. However, it was less clear whether delayed antibiotics or rapid/POCTs had an impact on clinical outcomes, health care utilization, and antibiotic resistance, as there was little or no available evidence on these outcomes.

There was variability in the effectiveness of patient and provider-directed interventions to influence prescribing, such as patient education and provider training.^14,17,20 The utility of such interventions may depend on the context and type of material provided and the delivery method. Given the heterogeneity in findings and limited evidence on different patient and provider-directed interventions, policy-makers may wish to carefully examine the comparability of primary studies with their health care context when considering implementing such interventions. Similar to delayed antibiotic-prescribing interventions, there was limited evidence on clinical outcomes, antibiotic resistance, and health resource utilization with patient and provider-directed interventions.

Overall, syntheses of evidence specifically on URTI in primary care settings (family medicine, GP) will be helpful to provide more clear evidence on the effect of interventions to influence antibiotic prescribing for URTIs in this setting. Further, more data on clinical outcomes, antibiotic resistance, and health care utilization would help to improve the understanding of the effects of interventions aimed at influencing antibiotic prescribing for URTIs.

References

1.Thomas M, Bomar PA. Upper respiratory tract infection. Treasure Island (FL): StatPearls Publishing; 2020: https://www.ncbi.nlm.nih.gov/books/NBK532961/. Accessed 2020 Oct 28.

2.Finley C, Chan D, Garrison S, et al. What are the most common conditions in primary care? Systematic review. Can Fam Phys. 2018;64(11).

3.Hutchinson J, Jelinski S, Hefferton D, Desaulniers G, Parfrey P. Role of diagnostic labeling in antibiotic prescription. Can Fam Physician. 2001;47. Medline

4.Dolk FCK, Pouwels KB, Smith DRM, Robotham JV, Smieszek T. Antibiotics in primary care in England: which antibiotics are prescribed and for which conditions? J Antimicrob Chemother. 2021;73(suppl_2).

5.Schwartz K. Unnecessary antibiotic prescribing in a Canadian primary care setting: a descriptive analysis using routinely collected electronic medical record data. CMAJ Open. 2020:E360. Medline

6.Smith SM, Fahey T, Smucny J, Becker LA. Antibiotics for acute bronchitis. Cochrane Database Syst Rev. 2017;6:CD000245. Medline

7.Leis JA, Born KB, Theriault G, Ostrow O, Grill A, Johnston KB. Using antibiotics wisely for respiratory tract infection in the era of covid-19. BMJ. 2020;371. Medline

8.Handle with care: preserving antibiotics now and into the future. Ottawa (ON): Public Health Agency of Canada; 2019: https://www.canada.ca/content/dam/phac-aspc/documents/corporate/publications/chief-public-health-officer-reports-state-public-health-canada/preserving-antibiotics/Final_CPHO_Report_EN_June6_2019.pdf. Accessed 2021 Jan 21.

9.Choosing Wisely Canada. Delayed prescription. 2018; https://choosingwiselycanada.org/wp-content/uploads/2018/10/Delayed-Prescription-EN-.pdf. Accessed 2021 Jan 21.

10.Lee C, Jafari M, Brownbridge R, Phillips C, Vanstone JR. The viral prescription pad - a mixed methods study to determine the need for and utility of an educational tool for antimicrobial stewardship in primary health care. BMC Fam Pract. 2020;21(1):1-12. Medline

11.Choosing Wisely Canada. The Cold standard: second edition. 2020; https://choosingwiselycanada.org/perspective/the-cold-standard/. Accessed 2021 Jan 21.

12.Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ. 2017;358:j4008. Medline

13.Spurling GK, Del Mar CB, Dooley L, Foxlee R, Farley R. Delayed antibiotic prescriptions for respiratory infections. Cochrane Database Syst Rev. 2017;9:CD004417. Medline

14.Mortazhejri S, Hong PJ, Yu AM, et al. Systematic review of patient-oriented interventions to reduce unnecessary use of antibiotics for upper respiratory tract infections. Syst Rev. 2020;9(106). Medline

15.Lane I, Bryce A, Ingle SM, Hay AD. Does locally relevant, real-time infection epidemiological data improve clinician management and antimicrobial prescribing in primary care? A systematic review. Fam Pract. 2018;35(5):542-550. Medline

16.Deniz Y, van Uum RT, de Hoog MLA, Schilder AGM, Damoiseaux R, Venekamp RP. Impact of acute otitis media clinical practice guidelines on antibiotic and analgesic prescriptions: a systematic review. Arch Dis Child. 2018;103(6):597-602. Medline

17.Kochling A, Loffler C, Reinsch S, et al. Reduction of antibiotic prescriptions for acute respiratory tract infections in primary care: a systematic review. Implement Sci. 2018;13(1):47. Medline

18.Schuetz P, Wirz Y, Sager R, et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections. Cochrane Database Syst Rev. 2017;10:CD007498. Medline

19.Health Technology Assessment of C-reactive protein point-of-care testing to guide antibiotic prescribing for acute respiratory tract infections in primary care settings. Dublin (IE): Health Information and Quality Authority; 2019: https://www.hiqa.ie/sites/default/files/2019-05/HTA_C-reactive_Protein_Point_of_Care_Testing-FullReport.pdf. Accessed 2021 Jan 21.

20.Burstein VR, Trajano RP, Kravitz RL, Bell RA, Vora D, May LS. Communication interventions to promote the public's awareness of antibiotics: a systematic review. BMC Public Health. 2019;19(1):899. Medline

21.Carvalho É, Estrela M, Zapata-Cachafeiro M, Figueiras A, Roque F, Herdeiro M. E-health tools to improve antibiotic use and resistances: a systematic review. Antibiotics (Basel, Switzerland). 2020;9(8). Medline

22.Cohen JF, Pauchard JY, Hjelm N, Cohen R, Chalumeau M. Efficacy and safety of rapid tests to guide antibiotic prescriptions for sore throat. Cochrane Database Syst Rev. 2020;6:CD012431. Medline

23.McDonagh MS, Peterson K, Winthrop K, Cantor A, Lazur BH, Buckley DI. Interventions to reduce inappropriate prescribing of antibiotics for acute respiratory tract infections: summary and update of a systematic review. J Int Med Res. 2018;46(8):3337-3357. Medline

24.Van Hecke O, Raymond M, Lee JJ, et al. In-vitro diagnostic point-of-care tests in paediatric ambulatory care: a systematic review and meta-analysis. PLoS One. 2020;15(7):e0235605. Medline

25.Fraser H, Gallacher D, Achana F, et al. Rapid antigen detection and molecular tests for group A streptococcal infections for acute sore throat: systematic reviews and economic evaluation. Health Technol Asses. 2020;24(31):1-232. Medline

26.Arroll B, Kenealy T, Kerse N. Do delayed prescriptions reduce the use of antibiotics for the common cold? A single-blind controlled trial. J Fam Pract. 2002;51(4). Medline

27.Little P, Williamson I, Warner G, Gould C, Gantley M, Kinmonth A. Open randomised trial of prescribing strategies in managing sore throat. BMJ. 1997;314(7082). Medline

28.Little P, Gould C, Williamson I, Moore M, Warner G, Dunleavey J. Pragmatic randomised controlled trial of two prescribing strategies for childhood acute otitis media. BMJ. 2001;322. Medline

29.de la Poza Abad M, Mas Dalmau G, Moreno Bakedano M, et al. Prescription strategies in acute uncomplicated respiratory infections: a randomized clinical trial. JAMA Intern Med. 2016;176(1). Medline

30.Little P, Hobbs FDR, Moore M, et al. Clinical score and rapid antigen detection test to guide antibiotic use for sore throats: randomised controlled trial of PRISM (primary care streptococcal management). BMJ. 2013;347. Medline

31.Llor C, Mandrel J, Balagué-Corbella M, Gómez M, Cots JM. Impact on antibiotic prescription of rapid antigen detection testing in acute pharyngitis in adults: a randomised clinical trial. Br J Gen Pract. 2011;61(586):e244-e251. Medline

32.Worrall G, Hutchinson J, Sherman G, Griffiths J. Diagnosing streptococcal sore throat in adults: randomized controlled trial of in-office aids. Can Fam Physician. 2007;53(4). Medline

33.Llor C, Bjerrum L, Arana J, et al. C-reactive protein testing in patients with acute rhinosinusitis leads to a reduction in antibiotic use. Fam Pract. 2012;29(6). Medline

34.Liberates A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. J Clin Epidemiol. 2009;62(10):e1-e34. Medline

35.McDonagh M, Peterson K, Winthrop K, Cantor A, Holzhammer B, Buckley DI. Improving antibiotic prescribing for uncomplicated acute respiratory tract infections. Rockville (MD): Agency for Healthcare Research and Quality; 2016: https://effectivehealthcare.ahrq.gov/sites/default/files/pdf/antibiotics-respiratory-infection_research.pdf. Accessed 2021 Jan 21.

36.Odermatt J, Friedli N, Kutz A, et al. Effects of procalcitonin testing on antibiotic use and clinical outcomes in patients with upper respiratory tract infections. An individual patient data meta-analysis. Clin Chem Lab Med. 2017;56(1):170-177. Medline

37.Ginzburg R, Conway J, Waltermaurer E, Song W, Jellinek-Cohen S. Using clinical decision support within the electronic health record to reduce incorrect prescribing for acute sinusitis. J Patient Cent Res Rev. 2018;5(3). Medline

38.Rubin M, Bateman K, Alder S, Donnelly S, Stoddard G, Samore M. A multifaceted intervention to improve antimicrobial prescribing for upper respiratory tract infections in a small rural community. Clin Infect Dis. 2005;40(4). Medline

39.Hingorani R, Mahmood M, Alweis R. Improving antibiotic adherence in treatment of acute upper respiratory infections: a quality improvement process. J Community Hosp Intern Med Perspect. 2015;5(3). Medline

40.Dowell J, Pitkethly M, Bain J, Martin S. A randomised controlled trial of delayed antibiotic prescribing as a strategy for managing uncomplicated respiratory tract infection in primary care. Br J Gen Pract. 2001;51(464). Medline

41.Little P, Moore M, Kelly J, et al. Delayed antibiotic prescribing strategies for respiratory tract infections in primary care: pragmatic, factorial, randomised controlled trial. BMJ. 2014. Medline

42.Pshetizky Y, Naimer S, Shvartzman P. Acute otitis media--a brief explanation to parents and antibiotic use. Fam Pract. 2003;20(4). Medline

43.Worrall G, Kettle A, Graham W, Hutchinson J. Postdated versus usual delayed antibiotic prescriptions in primary care: reduction in antibiotic use for acute respiratory infections? Can Fam Physician. 2010;56(10). Medline

44.Altiner A, Brockmann S, Sielk M, Wilm S, Wegscheider K, Abholz H. Reducing antibiotic prescriptions for acute cough by motivating GPs to change their attitudes to communication and empowering patients: a cluster-randomized intervention study. J Antimicrob Chemother. 2007;60(3). Medline

45.Bjerrum L, Gahrn-Hansen B, Munck A. C-reactive protein measurement in general practice may lead to lower antibiotic prescribing for sinusitis. Br J Gen Pract. 2004;54(506). Medline

46.Briel M, Schuetz P, Mueller B, et al. Procalcitonin-guided antibiotic use vs a standard approach for acute respiratory tract infections in primary care. Arch Intern Med. 2021;168(18):2000-2007. Medline

47.Burkhardt O, Ewig S, Haagen U, et al. Procalcitonin guidance and reduction of antibiotic use in acute respiratory tract infection. Eur Respir J. 2010;36(3). Medline

48.Cals J, Schot M, de Jong S, Dinant G, Hopstaken R. Point-of-care C-reactive protein testing and antibiotic prescribing for respiratory tract infections: a randomized controlled trial. Ann Fam Med. 2010;8(2). Medline

49.Francis NA, Butler CC, Hood K, Simpson S, Wood F, Nuttall J. Effect of using an interactive booklet about childhood respiratory tract infections in primary care consultations on reconsulting and antibiotic prescribing: a cluster randomised controlled trial. BMJ. 2009;339. Medline

50.Hebert C, Beaumont J, Schwartz G, Robicsek A. The influence of context on antimicrobial prescribing for febrile respiratory illness: a cohort study. Ann Intern Med. 2012;157(3). Medline

51.Lee M, Pan D, Huang J, et al. Results from a patient-based health education intervention in reducing antibiotic use for acute upper respiratory tract infections in the private sector primary care setting in Singapore. Antimicrob Agents Chemother. 2017;61(5). Medline

52.Kavanagh K, O'Shea E, Halloran R, Cantillon P, Murphy A. A pilot study of the use of near-patient C-Reactive Protein testing in the treatment of adult respiratory tract infections in one Irish general practice. BMC Fam Pract. 2011;12. Medline

53.Little P, Stuart B, Francis N, et al. Effects of internet-based training on antibiotic prescribing rates for acute respiratory-tract infections: a multinational, cluster, randomised, factorial, controlled trial. Lancet. 2013;382(9899). Medline

54.Little P, Stuart B, Andreou P, et al. Primary care randomised controlled trial of a tailored interactive website for the self-management of respiratory infections (Internet Doctor). BMJ Open. 2016;6(4). Medline

55.Malecki M. Rapid strip tests as a decision-making tool about antibiotic treatment in children – a prospective study. Pediatr Pol. 2017;92(2):149-155.

56.Meier F, Howland J, Johnson J, Poisson R. Effects of a rapid antigen test for group A streptococcal pharyngitis on physician prescribing and antibiotic costs. Arch Intern Med. 1990;150(8). Medline

57.Rios L, Mallafre M, Gonzalez-Hidalgo R, Collado A, Carbonell J, Colomes L. Aplicación de una pauta terapéutica revisada de otitis media en una consulta pediátrica. Revista de Calidad Asistencial. 2001;18(1):5-8.

58.Shah N, Ridgway JP, Konchak C, Fahrenbach J, Brown EC, Robicsek A. 163. What's going around? A prospective cluster randomized trial to evaluate a novel, real-time, syndromic surveillance tool's effect on clinical decision making amongst primary care providers. Open Forum Infect. 2021;1(suppl_1).

59.Taylor J, Kwan-Gett T, McMahon EJ. Effectiveness of a parental educational intervention in reducing antibiotic use in children: a randomized controlled trial. Pediatr Infect Dis J. 2005;24(6):489-493. Medline

60.Temte J, Shult P, Kirk C, Amspaugh J. Effects of viral respiratory disease education and surveillance on antibiotic prescribing. Fam Med. 1999;31(2). Medline

61.Thompson P, Gilbert R, Long P, Saxena S, Sharland M, Wong I. Has UK guidance affected general practitioner antibiotic prescribing for otitis media in children? J Public Health (Oxf). 2008;30(4). Medline

62.Tyrstrup M, Beckman A, Mölstad S, et al. Reduction in antibiotic prescribing for respiratory tract infections in Swedish primary care- a retrospective study of electronic patient records. BMC Infect Dis. 2016;16(1):1-10. Medline

Appendix 1: Selection of Included Studies

Figure 1: Selection of Included Studies

Appendix 3: Critical Appraisal of Included Publications

Table 4: Strengths and Limitations of Systematic Reviews on Delayed Antibiotics Using AMSTAR 2¹²

•• = Yes; • = Partial Yes; X = No; AMSTAR = A MeaSurement Tool to Assess systematic Reviews; PICO = patient, intervention, comparison, outcome; RoB = risk of bias.

Appendix 5: Overlap Between Included Systematic Reviews

Table 8: Overlap in Relevant Primary Studies Between Included Systematic Reviews on Delayed Antibiotics

Appendix 6: Additional References

Relevant Systematic Reviews Published Between 2015 and 2017

1.Coxeter P, Del Mar B, McGregor L, Beller EM, Hoffmann TC. Interventions to facilitate shared decision making to address antibiotic use for acute respiratory infections in primary care. Cochrane Database Syst Rev. 2015;2015(11): CD010907. Medline

2.McDonagh M, Peterson K, Winthrop K, Cantor A, Holzhammer B, Buckley DI. Improving Antibiotic Prescribing for Uncomplicated Acute Respiratory Tract Infections. Rockville (MD): Agency for Healthcare Research and Quality; 2016: https://effectivehealthcare.ahrq.gov/products/antibiotics-respiratory-infection/research. Accessed 2021 Jan 21.

3.O'Sullivan JW, Harvey RT, Glasziou PP, McCullough A. Written information for patients (or parents of child patients) to reduce the use of antibiotics for acute upper respiratory tract infections in primary care. Cochrane Database Syst Rev. 2016;11:CD011360. Medline

4.Wallace E, Uijen MJ, Clyne B, Zarabzadeh A, Keogh C, Galvin R, et al. Impact analysis studies of clinical prediction rules relevant to primary care: a systematic review. BMJ Open. 2016;6(3):e009957. Medline

Systematic Reviews Excluded Because Study Population, Type of Infection, and Setting Were Unclear

5.Egilmezer E, Walker GJ, Bakthavathsalam P, et al. Systematic review of the impact of point-of-care testing for influenza on the outcomes of patients with acute respiratory tract infection. Rev Med Virol. 2018;28(5):e1995 Medline

6.Lee JJ, Verbakel JY, Goyder CR, et al. The clinical utility of point-of-care tests for influenza in ambulatory care: a systematic review and meta-analysis. Clin Infect Dis. 2019;69(1):24-33. Medline

7.Tonkin-Crine SK, Tan PS, van Hecke O, et al. Clinician-targeted interventions to influence antibiotic prescribing behaviour for acute respiratory infections in primary care: an overview of systematic reviews. Cochrane Database Syst Rev. 2017;9:CD012252. Medline