Point-of-care tests: the key to reducing antibiotic prescribing for respiratory tract infections in primary care?

KEYWORDS:

• Antimicrobial resistance
• infection
• point-of-care test
• primary care
• respiratory

1. Introduction

Across the international health-care community, multiple stakeholders are developing and evaluating point-of-care tests (POCTs) as a potential solution to antimicrobial resistance (AMR) through reduction of antibiotic prescribing. There is a good reason for this focus: AMR was declared one of the top-10 threats to global health by the World Health Organisation in 2019.

The majority of UK NHS antibiotics are prescribed in primary care [1], a pattern found worldwide [2,3]. At least 20% of these prescriptions are considered unnecessary [4]. Consequently, there is a critical need to safely reduce antibiotic prescribing. This is especially true for acute respiratory tract infections (RTIs), the problem most commonly managed by health services internationally [5] and the infections associated with most inappropriate prescribing [6,7]. Most RTIs are self-limiting; yet, antibiotics are prescribed in up to 67% of UK RTI consultations [8], with prescribing often attributed to clinical uncertainty regarding microbiological diagnosis and disease prognosis, leading to defensive (‘just-in case’) care [9]. It is a challenging situation: clinical presentation alone is not helpful for differentiating between viral and bacterial RTIs, and standard microbiological testing in laboratories is too slow for initial decision-making. In essence, clinical microbiology and infection diagnostics are not currently fit for the purpose in primary care.

2. Are point-of-care tests the solution?

The 2020 Wellcome Trust AMR report [10] and the UK commission ‘Review on Antimicrobial Resistance’ placed great emphasis on POCTs that distinguish viral from bacterial infection as a solution to AMR [11], with Chair Jim O’Neill recommending that ‘no antibiotic should be prescribed without a test.’ POCTs have potential to be ‘game-changing’ stewardship tools, supporting precision prescribing by bringing the laboratory to the consulting room and overcoming the prohibitive time-scale for standard microbiological testing.

Broadly speaking, there are two types of POCT for diagnosing infections. First, ‘microbiological’ POCTs, which test for the presence of specific pathogens. Second, ‘biomarker’ POCTs, which measure and attempt to distinguish host viral from bacterial immune responses, often via a blood sample.

Within Europe, there is marked variation in the use of POCTs in primary care for RTIs (ranging from 0% to >65% of consultations), with C-reactive protein (CRP) and group A streptococcal (strep A) tests most commonly used [12]. The COVID-19 pandemic has boosted public awareness of POCTs in testing for respiratory viruses and increased acceptability of using rapid diagnostic tests. Could this set the stage for the widespread uptake of POCTs in primary care?

3. Evidence for POCTs in primary care

Whilst this sounds promising, most POCTs have not been adequately assessed in primary care. The POCT market is booming and there have been encouraging results, although this has predominantly focused on secondary care. The primary care population is different, with potentially less severe infections and thus differing immune responses and microbiology. POCTs must be evaluated in their own right in primary care.

There are multiple questions to answer. Fundamentally, we need evidence of clinical effectiveness (such as reduction in antibiotic prescribing or improved patient outcomes), long-term cost effectiveness and safety in practice in primary care before widespread roll-out is even considered. Validity, including sensitivity, specificity, positive predictive value, and negative predictive value, should be considered in this specific clinical setting. This includes patients with underlying co-morbidities, such as those with COPD and asthma, who are commonly prescribed antibiotics for RTIs. Additionally, how POCTs might influence clinician/patient behavior and antibiotic prescribing in primary care should be explored, given the antibiotic prescribing decision is being made in the context of a conversation between clinician and patient. Whilst we focus here on the role of POCTs in reducing antibiotic prescribing, POCTs should also be evaluated regarding their ability to predict prognosis and help identify patients who will subsequently deteriorate.

Use of POCTs in primary care may have benefits, such as reducing antibiotic prescribing and AMR (on a personal and societal level) and reducing individual’s exposure to other antibiotic harms. POCTs may increase clinician and patient diagnostic confidence, potentially reducing repeat consultations during the same illness, and they may increase detection of bacterial infections requiring antibiotics, thus improving clinical outcomes. Both these factors may have a positive impact on POCT cost effectiveness. Further research is needed to understand where POCTs may best fit in the diagnostic pathway, and this is likely to be practice/population dependent. Might POCTs have a role as a triage tool, helping identify those requiring clinician appointments?

However, these benefits may not actually exist in practice and, indeed, there may be negatives to using POCTs. Crucially, the availability of POCTs may medicalize minor illness, increasing health-seeking behavior and demand for future testing for similar illnesses. These medicalization effects need to be taken into consideration when assessing long-term cost effectiveness. Furthermore, we must ensure that POCTs do not miss bacterial infections, worsening clinical outcomes. Additionally, the day-to-day time and cost requirements for already overstretched primary care systems may be prohibitive.

Evidence for the use of POCTs can only be generated from well-designed, independent clinical studies. Yet, there is commercial pressure for POCT implementation. A balance must be struck between striving for innovation and the time required to generate robust evidence. We believe strict regulation of POCT use is vital, with the introduction of these diagnostic tests assessed with a rigor similar to that for new medications [13], as well as prevention of direct-to-consumer POCT advertising.

4. Microbiological point-of-care tests

The evidence gap for primary care use of microbiological POCTs is starting to be filled. In 2019, we conducted one of the first feasibility studies of a multiplex microbiological POCT in primary care, testing for 19 viruses and four atypical bacteria via a nose and throat swab. The results were encouraging, showing clinicians found the POCT acceptable, useful, and that it increased their diagnostic confidence and reduced predicted antibiotic benefit [14]. POCT use was limited by the time taken for test results and the absence of testing for typical respiratory bacteria (due to being commensally carried in the upper respiratory tract). We are now conducting a primary care randomized efficacy trial of the same POCT (https://fundingawards.nihr.ac.uk/award/NIHR131758). The trial includes a mixed methods investigation of microbial, behavioral, and antibiotic mechanisms influencing the primary and key secondary outcomes: same-day antibiotic prescribing for children and adults presenting to primary care with RTIs; and patient-reported symptom severity on days 2 to 4.

Microbiological POCTs have potential benefits beyond reducing antibiotic use. There is growing evidence that antiviral treatments may reduce symptom severity, particularly for patients in higher risk groups [15]. With novel antivirals in development, microbiological POCTs may optimize precision prescribing for viral, as well as bacterial, infections.

There are, however, some challenges with microbiological POCTs. The upper respiratory tract (throat/nose) is easily accessible for sampling, but the clinical significance of detected microbes is not understood, as bacteria and viruses causing common RTIs also harmlessly inhabit the upper respiratory tract (‘commensals’). The recent ‘strep A crisis’ in the UK illustrates this well: following reports of increased scarlet fever incidence and deaths from invasive group A streptococcal infection, ‘strep A’ POCTs were suggested as a means to identify children needing antibiotics. However, commensal strep A carriage is present in up to 15% of children [16].

Additionally, whilst obtaining microbiological samples for upper respiratory tract infection is relatively straightforward, it is not ethical or practical, beyond sputum which is not always available, to obtain samples for lower RTIs in primary care. Consequently, samples from the upper respiratory tract tend to be used, on the assumption that respiratory cilia move microbes from the lower to the upper respiratory tract. However, it is noteworthy that upper respiratory tract swabs taken from symptomatic people by trained clinicians have zero pathogen detection in up to 28% of cases [17,18].

5. Host-response ‘biomarker’ point-of-care tests

Clinically significant infection triggers a host inflammatory/immune response, which differs depending on whether the infection is bacterial or viral. ‘Biomarker’ POCTs measure host immune proteins as surrogate markers of infection etiology.

This approach has some advantages over ‘microbiological’ POCTs. First, it allows diagnosis when the infection site is not easily accessible, such as lower RTIs. Second, the results are not confounded by the detection of commensal bacteria or viruses. However, without the microbe detection afforded by microbiological POCTs, these biomarker POCTs do not allow targeted antibiotic/antiviral treatment.

Single biomarker POCTs using CRP have been evaluated in randomized controlled trials and shown to reduce antibiotic prescribing for adults with acute lower RTIs [19]. CRP POCTs have been recommended in the UK by NICE since 2015 for patients with suspected lower RTIs in primary care. However, the uptake has been low. Reasons may include questions around who funds such tests and because clinicians may be uncertain about how the test works (as an elevated CRP does not always mean the infection is bacterial) [20] [13].

The diagnostic accuracy of lone biomarkers, such as CRP, is unlikely to be high enough. CRP does not always correlate with bacterial load and can be significantly raised in viral infection [21]. Recently, novel combination biomarker POCTs have been developed, attempting to increase diagnostic accuracy. These have had encouraging results in secondary care but have not been evaluated thoroughly in primary care [22].

6. Conclusion

Industry is investing heavily in the development of POCTs and, whilst these tests have the potential to reduce antibiotic prescribing, thorough evaluation of their clinical effectiveness, cost effectiveness (both short- and long-term taking account of medicalization effects), and safety, specifically in the primary care setting, is required.

Novel combination biomarker POCTs are attractive, given their potentially increased diagnostic accuracy combined with usefulness for inaccessible infection sites and distinguishing pathogenic from commensal microbes. Microbiological and biomarker POCTs offer complementary information – might they be used in combination to identify, first, if the patient is mounting an immune response, and second, what the specific organism is, thus optimizing precision prescribing in primary care.

Declaration of interest

Emily Brown’s and Alastair Hay’s employer (the University of Bristol) has purchased the bioMereiux bioFire for the RAPID-TEST RCT (Award ID: NIHR131758, https://fundingawards.nihr.ac.uk/award/NIHR131758) under a standard procurement arrangement. They are not not receiving any direct or indirect financial or non-financial support from bioMereiux. Emily Brown is an NIHR Clinical Research Fellow (In Practice Fellowship). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewers Disclosure

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was not funded.