Background

Antimicrobial stewardship

Antimicrobial resistance (AMR) is a critical issue requiring effective antimicrobial stewardship (AMS) [1]. AMS can prevent antibiotic overuse, misuse, and abuse [1] and reduce drug resistance [2], costs, and hospital stays [3, 4]. This is particularly important for low- and middle-income countries (LMICs), who suffer a greater AMR [4] and infectious disease burden [5]. Well-developed stewardship measures applicable to LMICs are thus a research priority.

Point-of-care testing

Diagnostic tests to support appropriate treatment and prescribing are a key component of AMS programmes. However, lower-level healthcare settings in LMICs have limited laboratory and diagnostic capacity. Point-of-care testing (POCT), in which patient specimens are analysed outside of a clinical laboratory, at the site of patient care, by staff who have not been formally trained in laboratories, offers a means of reaching more patients with diagnostic services [6]. POCT represents a promising avenue for enhancing antimicrobial stewardship. For example, C-reactive protein (CRP) testing provides real-time assessment of likelihood of bacterial infection, reducing antibiotic prescribing in primary care [7,8,9,10]. In a review, 44% of patients who received CRP tests were prescribed antibiotics at initial consultations for upper respiratory tract infections (RTIs), compared to 63% of those untested [11]. More specific POCTs for particular infectious diseases like malaria can also guide appropriate antimicrobial therapy, thereby contributing to reduced disease burden and resistance, and enabling rapid diagnosis for a disease that previously only relied on clinical diagnosis. A malaria POCT test-and-treat strategy in Zambia reduced paediatric prevalence of malaria by 17% [12]. Moreover, POCTs can help reduce unnecessary use of drugs, staff, and equipment [13], lowering costs. For example, Helicobacter pylori screening reduces the number of patients referred for endoscopy [14].

Implementation in private settings

POCT in private retail settings such as pharmacies and drug shops is particularly promising for LMICs [15], where trained workforce and infrastructure are lacking [16], as they are independent of expensive, centralised laboratories [17]. POCTs are easy to perform, interpret, and transport [18]. Although rapid diagnostics are available in large public hospitals, with median availability of malarial diagnostics reaching 91.6% in 10 LMICs [19], these services are often overloaded [20]. In contrast, primary care tends to lack diagnostics, with only 19.1% median availability [19]. Adding additional diagnostic services in hospitals and primary care would burden the national budget. Hence, POCTs in private settings could make diagnostic services more accessible.

Moreover, pharmacies are an ideal checkpoint for antimicrobial use, as regulations around dispensing antimicrobials are poorly enforced in LMICs, with frequent over-the-counter non-prescribed antimicrobial sales [21]. By detecting or ruling out infection, POCTs can help providers recommend appropriate treatment [22,23,24,25,26,27]. Their use in pharmacies and drug shops can reduce unnecessary treatment and improve care-seeking behaviours [25,26,27,28], while still providing revenue from test sales.

Aim

In this study, we reviewed evidence for implementation of POCTs for infectious diseases in private retail settings in LMICs, to inform future studies and policy design.

Methods

Search strategy

This review was structured with reference to the Scale for the Assessment of Narrative Review Articles (SANRA) [29] and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [30]. We searched PubMed/Medline on 26/06/2023, using MeSH headings and synonyms for ‘infectious disease’, ‘rapid diagnostic testing’, and ‘pharmacy’ (full search terms Additional file 1). A manual search of references from other studies was also conducted to include relevant studies. No limit on date of publication was imposed on studies for inclusion.

Inclusion criteria were developed referencing population-intervention-comparison-outcomes (PICO):

  • Participants: pharmacies and private retailers in LMICs

  • Interventions: implementation of staff-performed POCT

  • Comparisons: N/A

  • Outcomes: feasibility and impact of implementation

Non-private or non-interventional studies (reporting only test accuracy, hypothetical or modelling studies, lacking actual implementation); or not on infectious diseases, were excluded.

Two independent investigators selected papers for full-text screening using Covidence, resolving conflicts by discussion. Investigators designed an abstract screening tool [31] (Additional file 2) and randomly selected 20 papers for standardisation of screening.

Outcomes

A data extraction table was created using Google Sheets 2023. Two authors independently performed extraction of study characteristics, methods, and outcomes:

  • Uptake: proportion of treatment-seeking patients receiving POCT out of total population studied

  • Positivity: proportion of patients receiving a POCT who tested positive

  • Treatment provision: proportions of treatment-seeking patients and of patients not tested receiving intended treatment(s)

  • Adherence to POCT results: proportions of POCT-negative patients not receiving intended treatment(s) and of POCT-positive patients receiving intended treatment(s)

  • Referrals: proportion of patients referred elsewhere

  • Test accuracy: sensitivity, specificity, or positive predictive value of test (if available)

  • Safety and accuracy of performance of test: proportion of providers safely and correctly performing, interpreting and disposing of POCT (if available)

  • Recommended/median POCT retail price (if available)

  • Opinions of providers/customers on POCT

Proportions were expressed as percentages. Individual outcomes were chosen or calculated from cluster data. Additional characteristics extracted were training length/content, supervision, demand-generation activities, referral, and guidelines for those who tested positive or negative on the POCT.

Quality assessment

Article quality was checked against seven of ten relevant features published by the Consensus Working Group of the Joint Programming Initiative on Antimicrobial Resistance: randomized design; use of controls; multi-centre study design; sustainability of intervention (> 12 months); sample size calculation (where relevant); prospective design; and correction for confounding variables [32]. Funding was evaluated for conflicts of interest.

Results

Study selection

From our search strategy, 848 titles were identified, of which 63 studies were screened in full (Fig. 1). 21 studies were excluded based on pre-defined inclusion/exclusion criteria. 19 studies in high-income countries were excluded after review to focus on LMICs. 23 studies were included, including one study detailing the policy implications from another study included in this review [33, 34].

Fig. 1
figure 1

PRISMA flow diagram for studies analysed in this narrative review

Full size image

Study design and characteristics

Table 1 describes characteristics of the included studies. 19 studies were on malaria, two were on a mix of malaria and pneumonia, one was on paediatric fever management, and one was on respiratory tract infections of viral or bacterial aetiology. Study designs included nine randomised controlled studies, four controlled, non-randomised studies, five uncontrolled interventions, one interventional pre-post study, one cross-over interventional study and three retrospective analyses of RCTs. Hansen et al. (2017) was a cost-effectiveness analysis of Mboyne et al. (2015)’s malaria study [26, 34].

Table 1 Study information and characteristics
Full size table

Studies lasted between three [41, 52, 53] and 108 months [49]. Of the studies where the setting of the study was described, nine were in rural areas [23, 24, 26, 36,37,38, 43, 44, 46, 50, 51, 53], five in urban/suburban regions [33, 34, 41, 45, 49, 52], and two were in a mix of both [42, 47]. The number of outlets investigated per study ranged from two [41] to 317 [49].

Only eight studies described the consumable equipment they provided to the private retailers, like antiseptic pads [36], free gloves and sharps disposal [23, 38, 44, 47], or bins [39, 40, 43]. 20 studies implemented training for providers, covering study protocol, signs/symptoms, evaluation, and diagnostic criteria. Intervention arms also received training on POCT administration, interpretation, and disposal. Training lasted between 90 minutes [53] and six 2-day sessions [44].

For malaria, guidance for patients testing positive were provision of the appropriate medication, such as ACTs for malaria [23, 24, 26, 34,35,36,37, 41,42,43,44, 47, 52]. For other POCTs, patients satisfying disease criteria based on the test results were supposed to be prescribed amoxicillin for pneumonia [24, 37, 42], antibiotics for CRP levels greater than or equal to 100 mg/l [45], or zinc sulfate tablets for non-bloody diarrhoea [42]. For negative test results, clients were often referred to formal care [23, 36, 42, 47], especially if there was fever [34]. Sellers were to recommend stopping antimalarials [41] or against their purchase [24, 26, 34, 37, 43], or in cases of CRP less than 30 mg/l, to not dispense antibiotics [45]. Kwarteng et al. (2019) provided symptomatic treatment [43]. In Onwunduba et al. (2022), cases with CRP of intermediate levels between 30 mg/l to 100 mg/l, sellers were asked to use their professional judgment to decide on prescription of antibiotics [45].

Demand-generation activities were paper-based like roadside posters, media-based like newspapers or radio, and/or verbal promotions during healthcare consultations or durbars held by traditional community leaders [23].

Methods of assessing outcomes

Table 2 summarises study methodologies.

Table 2 Studies’ methods for assessing outcomes
Full size table

Studies collected data on POCT implementation and/or treatment decisions by forms/questionnaires filled out by drug vendors [23, 34, 35, 41, 46, 50, 53], direct observation of the vendors or mystery clients [24, 36,37,38, 42,43,44,45, 47, 52]. Patient-based methods included sales data, administrative/patient records [26, 38] or provider/household surveys [24, 36,37,38, 44, 50, 52]. Simmalavong et al. (2017) used epidemiological data [49]. Qualitative methods included focus group discussions [39, 40, 43] or interviews with providers [51, 52].

To assess accuracy of POCT administration, studies used mystery clients [23, 36, 45, 47], direct observation [24, 37, 42, 44], microscopy of blood slides [18, 34, 43], or checking POCTs [43]. Two studies reported manufacturer’s specificity and sensitivity [23, 35]. Three assessed sensitivity and specificity directly by comparing malaria POCT results to blood microscopy [26, 34, 41]. Cohen et al. (2012) randomly checked unused tests [38].

Few studies reported how they determined retail price if authors recommended a specific price. Three malaria studies referenced previous willingness-to-pay studies [26, 34, 44].

Testing and treatment outcomes

In general, studies showed that implementation of POCT could lead to feasibly high uptake levels and adherence to treatment guidelines (Table 3).

Table 3 Outcomes of testing and treatment
Full size table

This was dependent upon factors like adequately designed training, demand generation, linkage to care, support for providers, and appropriate financial remuneration. For example, in one longitudinal study of malaria POCTs, uptake increased when monthly check-ins, financial incentives, or more intensive support was implemented in subsidized schemes of POCTs, from 3.0% to 6.4% (monthly check-ins), 2.7% to 11.9% (financial incentives), and 5.4% to 13.0% (intensive support) [36]. Moreover, six studies reported adherence of treatment outcome to positive and negative test results above 90% [23, 24, 26, 34, 44, 47]. These factors will be discussed individually below.

Training for providers

POCT implementation requires comprehensive training before implementation and our review finds that it should cover topics including POCTs’ importance and benefits, its administration, interpretation, waste disposal, and counselling after results [40, 54, 55]. Firstly, emphasising the need, value, and accuracy of POCTs may improve uptake. In a Kenyan malaria study that only taught epidemiology and POCT procedure/management but did not emphasise the need for POCTs, uptake was as low as 30.4% [47]. Conversely, when providers were educated on POCT’s value and felt aligned with professionals through training, higher provider uptake was observed [36, 51, 56, 57], reaching 97% in a study where drug shops were trained in an integrated community case management style where providers were educated on the need and how malaria POCTs worked [42]. Instilling belief in the need for POCTs could address other factors limiting POCT uptake – for example, some providers only used POCTs for specific patient profiles or disobeyed guidelines, believing NGOs had eradicated local malaria [51] or endemicity was declining [52]. Alternatively, some providers relied on clinical judgement if they perceived shortcomings of POCTs [35]. One study found poor adherence to negative malaria POCT results, as 20–41% of malaria-negative mystery clients were told by providers that they were positive and suggested that this was due to providers’ “mistrust of POCT results” despite their competence in POCT performance [47]. This is despite high sensitivity of malaria POCTs (91.7%–100% [23, 34]), which could be highlighted in training to instil confidence.

Secondly, training must include clear guidelines for negative tests. Studies with the worst adherence to negative test results – 48.4% in Ikwuobe et al. (2013) and 57.7% in Cohen et al. (2012), which also had the poorest uptake (16%) – had (sometimes deliberately) vague guidelines [38, 41]. The former lacked guidelines on handling negative results and suspension of antimalarials only after a pharmacist/patient discussion without study authors [41]. In the latter, staff were to proceed as “[they] would normally”, only being told “how POCTs work and how to use them” [38]. Hutchinson et al. (2017) reported vendors’ “anxiety around the management of a negative mPOCT [malaria POCT] result” as negative results might reveal “diagnostic uncertainty” regarding the illness, negatively impacting vendors’ reputation [40]. Hence, managing negative test results well is important for provider uptake and to reassure customers. This is also important from the providers’ perspective to maintain drug sales, which will be discussed below.

Guidelines for negative tests appear to be particularly important for diseases in which concerns about under prescription are dominant. Of studies with large discrepancies between adherence to positive versus negative test results, studies tended to report higher adherence to positive results than negative (Table 4) [35, 38, 40, 41, 53]. This could reflect a reluctance to under-prescribe medication and miss a diagnosis in malaria. In contrast, in other studies of notifiable diseases like HIV and HCV, low positive adherence may indicate stigma: for example, an HCV study in a high-income country (excluded from this review) reported 100% adherence to negative test results but only 28.2% adherence to positive test results [58]. However, data on this in the studies included in this review is limited, as the single study on CRP POCTs only reported adherence to negative CRP test results (30.4%) as they only collected data through mystery clients [45].

Table 4 Studies where positive and negative adherence differ
Full size table

Thirdly, providers should consider the format in which training is implemented. Certification of a formal training course could assure customers of providers’ credibility and instil self-confidence in providers, particularly in LMICs [26, 34, 39, 40]. For example, Klepser et al. mandated a “Collaborative Institutional Training Initiative program” and POCT certificate course [60,61,62]. This would capitalize upon the increased legitimacy in the eyes of customers already conferred upon vendors using POCTs. For example, customers were surveyed on their opinion of the outlets administering POCTs. In Uganda, POCTs gave legitimacy to vendors, who were perceived to have unclear credentials [33, 39] or purely profit-driven [40]. This change was attributed to the involvement of external project supervisors/government and new technology [39], with sharps and blood testing shifting vendors into the “category of an endorsed professional” [34, 40]. Such outward-projecting improvements in retailers’ image also benefited vendors: for example, in Uganda, vendors reported outward-directed benefits, as government partnership conferred legitimacy, status, and confidence about safety from authorities’ raids [39]. They attributed this to POCT technology marking them as endorsed professionals, particularly visibly drawing blood using recognisably medicalized objects like gloves, needles, and packaged lancets [39]. Hence, formalization of training and certification in POCT could reinforce these benefits and improve uptake amongst customers by legitimizing shops. Studies performed after the COVID-19 pandemic, in which pharmacists took on a big role in testing, might show further changes in this direction of customers’ attitudes toward pharmacist-performed POCT, as has been shown for high-income countries [63,64,65,66]. However, none of the studies performed after the pandemic directly assessed changes in these attitudes.

Some customers worried sellers were “unskilled” in the practice of POCTs and risked HIV infection [51] or injury, or overcharged [40]. In contrast to these concerns, six studies reported that POCTs were accurately performed, with performance of measures of safety/test administration/waste disposal above 90% [23, 24, 34, 36, 38, 43] and vendors felt more confident about making medical decisions by reducing guesswork [40, 51, 67]. However, a few studies like Poyer et al. (2018) reported more inconsistent levels of accuracy, between 14.7% and 100% attainment of safety outcomes [47], while Soniran et al., (2022) reported 66.7% attainment of safety outcomes compared to 40% in control arms [50]. Some commonly missed steps before the procedure included not checking expiry dates [38], explaining the test, or testing away from other clients [47]. Steps missed during the test were drawing the right amount of blood [38, 50], using antiseptic [36], wearing gloves [47], checking the time after adding the buffer [50]. Afterwards, providers failed to immediately dispose of lancets in sharps bins [47, 50], or waited < 15 min before reading the result [47]. We recommend that a checklist be provided, highlighting these commonly missed steps to ensure that the proper procedures are followed.

Demand generation and community sensitization

Community sensitization can help potential customers recognise the importance of POCT and appropriate treatment [24, 54, 57, 68]. Patients were more likely to consent to receiving a POCT if they had used one before or were aware of its availability [44]. For example, two studies in this review with the highest adherence to test results implemented a POCT program that integrated with a community awareness campaign [24, 37]. Conversely, lack of familiarity contributed poor uptake in some studies which lacked any demand generation [27, 69,70,71], as evidenced in studies like Cohen et al. (2012), and Ikwuobe et al. (2013). Community sensitization could improve customers’ acceptance of test outcomes and reduce pressure on providers to meet customers’ expectations, improving provider adherence too [47, 51]. Furthermore, some providers did not comply with national malaria guidelines as they could not ensure patients finished the full antimalarial course [52] – community sensitization could teach patients to receive treatment properly.

The content of sensitization should emphasize the benefits of testing at private retailers without diminishing other healthcare sectors: for example, a qualitative study suggested that sensitization efforts should focus on emphasising the need for testing regardless of the location (public or private), and leverage the trust in drug shops, which could increase uptake at private retailers without diminishing the work of public healthcare [59]. In the reviewed studies, over 80% of customers were satisfied with POCTs [34, 72,73,74], citing the convenience of testing locations [51, 74], as they would not have to travel to hospitals [40].

Programs need to consider the most effective methods of community sensitization, which may vary depending on the level of economic development or degree of trust in medical technology in the area. For example, two studies communicated the need for POCTs through community leaders, which may be better trusted by communities in LMICs [23, 50]. None of the included studies surveyed the participants on where they had heard of the POCT service, a question that could be included in future questionnaires to evaluate the most effective form of advertising.

Linkage to care

Several studies identified formal linkage to support as important for success. Barriers included resistance from healthcare workers to the POCT program [34] and failure of patients to honour referrals [43]. For example, health-workers were concerned about the same issues that were reported by private sellers as being beneficial to private retailers. They felt untrained vendors encroached on professional boundaries, deeming them untrustworthy, and did not acknowledge the paperwork vendors used to refer customers [39]. Health-workers also worried about decreased malaria testing at their facility [39]. However, some health-workers recognised improvements like customers’ shorter travel times [40].

Suggestions to bridge public and private sectors include capacity-building programs for pharmacists with health workers to address mistrust of health workers [43] or partnership with government agencies and professional bodies [47]. Community sensitization on accreditation of pharmacists and the formalisation of training could address health workers’ preconceptions about unprofessionalism [39]. Secure communication platforms should be established to protect data security. Two studies further proposed integrating POCT programs into nationwide malaria surveillance systems [44, 49].

Support for providers

Adequate provider supervision and support is important immediately after POCT implementation and in the long run. One study found price subsidies for providers to buy POCTs were most effective when accompanied by monthly intensive support as it led to a larger increase in uptake (from 5.4% to 13%) than just check-in visits (from 3% to 6.4%) [36]. Another study commented that prolonged support is necessary because provider behaviours, particularly for malaria, are “driven by ingrained behaviours” and thus difficult to change quickly [23]. This was echoed in focus group discussions with providers, as although some providers felt POCTs were easy to use and optimised workflow by reducing guesswork [51], others felt their workload had increased. However, adequate support and minimizing the Hawthorne effect need to be balanced [34]. The Hawthorne Effect can occur when practitioners modify their behaviour knowing that they are being monitored during interventional studies [34]. A possible support schedule could provide intensive support initially and decrease intensity over time [75, 76].

In LMICs, as well as individual provider support, other forms of support that should be considered in POCT implementation programs include addressing systems-level barriers [51, 77]. Providers relied on research teams for waste disposal or struggled to collaborate with public health facilities [51, 77]. Hence, this study recommends that a systematic approach to integrating private retailers into waste management systems and infrastructure be considered for safe POCT disposal.

Other systemic factors relating to procurement included lacking weighing scales to calculate antimalarial dosage [43], or being undeliverable due to flooded roadways [44]. Simmavalong et al. (2017) described private facilities’ lack of control over POCT distribution left them vulnerable to “trickle down shortages” in their government-led program [49]. In one study, some shops disobeyed malaria guidelines as they lacked POCTs (3/65) or antimalarials (1) despite free central provision [52]. When designing POCT programmes in LMICs, comprehensive material and logistical support in addition to the POCTs themselves needs to be included.

Financial remuneration and pricing

POCT pricing should be carefully evaluated through willingness-to-pay studies, as it impacts provider income and patient uptake. Factors like manufacturing subsidy, distribution, cost of training, equipment, or supervision should be balanced against the price customers are willing to pay. Especially in LMICs, where the “combined cost of mPOCT [malaria POCT] and ACT is a barrier to rural folks” [43], a subsidy may be necessary to facilitate patient uptake.

Seven studies offered POCT for free to customers [23, 24, 35, 37, 41, 43, 46], while seven supplied POCT for free but sold at a subsidized price [26, 34, 38, 40, 44, 46, 47, 50], and the remaining studies did not explicitly state the price at which they were sold [35, 43, 49, 51, 52]. In Hutchinson et al. (2015), providers were given tests for free and sold at providers’ chosen price [39]. Retail prices for customers ranged between 0.18 USD (Myanmar [36]) to 2 USD (Uganda [40]).

These different financial schemes have been evaluated to mixed results – in one malaria study, there was no difference in uptake between districts that were or were not subsidised for POCTs [44]. Similarly, providing another financial incentive in addition to price subsidies did not significantly increase uptake compared to price subsidy with intensive support [36]. In contrast, a third study reported malaria POCT subsidies positively impacted uptake independent of a prior offer of an antimalarial subsidy, suggesting that subsidies influence the “next immediate action” (deciding whether to test) [46]. A cost-effectiveness analysis of malaria POCT in Myanmar also found that price subsidy coupled with information, education and communication provider-targeted strategy was the most cost-effective [78]. Possible cost-saving measures included reducing supervision, having shops cover some training fees or gloves.

Providers’ attitudes towards renumeration may be affected by disease endemicity and negative test management. For illnesses like malaria where negative tests could impact drug sales directly, there was “tension between the motivation of the shop owner to make a profitable drug sale” and the lack of income after a negative test [46]. This was echoed in Gwagwalada, Nigeria, considered meso-endemic for malaria, where reduced antimalarial purchase may generate significant income loss [41]. A pharmacist “expressed concerns about loss of sales”, reasoning it would be difficult to restrict profit-guided pharmacies’ antimalarial sales without alternative income [41].

In studies with high malaria positivity, providers felt the programme benefited from them financially, as they enjoyed increased drugs and POCT sales [24, 39]. Even negative tests offered opportunity to sell more drugs, changing from antimalarials to other drugs like antipyretics [23, 26], and greater polypharmacy and median spending by those testing negative [40]. However, drugs like paracetamol have a “lower profit margin”, so whether overall economic benefit occurs may be affected by the proportion of positive/negative tests in high/low endemic areas [37]. These further highlight the importance of providing clear guidelines for negative test results, as described earlier.

Quality assessment

Study quality varied greatly (Table 5): only 11 studies had a randomised research design and a control group, of which seven corrected for confounding factors. Although 17 studies were multi-centre, sample sizes varied drastically: some were extremely small with only 21 focus groups [39], while others reached 15,517 patients across 59 drug shops [34]. An epidemiological study pooled 2,301,676 tests across 317 pharmacies [49]. Three studies did not disclose funding sources [36, 41, 51] (Additional file 3).

Table 5 Quality assessment of the included studies
Full size table

Moreover, differences in methods of measurement between studies hampered direct comparison – one study measured uptake as proportion of households remembering receiving POCT [36] while others only included consenting clients [23, 34]. In one study, there was significant heterogeneity in willingness to provide testing amongst 92 drug shops [38]. A standardised way of quantifying POCT uptake should be used across studies to allow for comparison – for example, recording the proportion of febrile clients or clients suspected of a particular disease by the pharmacist.

Discussion

In this review, we describe that many studies demonstrated POCT can improve diagnosis, referral, and treatment of infectious diseases [23, 24, 26, 34, 37, 44, 47]. Pharmacies are a potential point of intervention to manage infectious disease diagnosis and treatment/referral [48]. In Uganda, > 72% of care-seekers sought care for febrile children at drug shops [42, 64]. Pharmacies increase accessibility in terms of proximity to patients’ houses, cost, and opening hours compared to primary care [64]. In one study, 38% of patients presented outside normal clinic hours [62]. This may benefit populations in LMICs, whose inconsistent schedules, language barriers, or unreliable internet complicate appointment-scheduling [79]. Other studies echo the importance of pharmacies for non-emergency care-seeking [23, 80,81,82].

Comparison with literature

This review adds to a small body of literature on POCT in private retail settings [83]. No other review examined infectious disease testing, except Visser et al. (2016) on malaria POCTs [77]. Of the twelve papers studied, five were absent from our review: four were unpublished, while the published study compared three training intensities and sensitization and reported little effect on an overall poor uptake [84]. Although study authors reported uptake and adherence improved with longer training, frequent initial supervision, and low POCT prices, this was only based on three studies with lowest provider numbers, with multiple exceptions [77]. The authors also argued that POCT programs would not scale-up easily, but that less intense but more scalable programs had poorer outcomes [77]. This was more difficult to compare in our review considering our greater number and heterogeneity in studies. However, the authors cited similar concerns like waste management, provider and client expectations, training, and wider health-system integration [77].

Another review by Boyce et al. (2017) on malaria POCTs in sub-Saharan Africa across private, public and community healthcare included five retail studies also assessed in our review, and agreed that POCT providers had good adherence, execution, and sensitivity, but lacked specificity compared to community health workers and formal healthcare [85]. Their other concerns mirrored ours, adding that patients may purchase substandard non-ACT antimalarials, contraindicating the intent of POCTs [85]. Nevertheless, the authors’ comparison with other healthcare settings highlights the usefulness of retail settings [85].

Strengths and limitations

Our search strategy with two independent researchers means it is unlikely we missed many eligible studies within PubMed/Medline. However, five studies were neither randomised nor controlled, four were quasi-experimental (in which participants are assigned to control and intervention groups in a non-random manner), and another four were pilot studies that lacked a control arm for comparison. Moreover, under close supervision, retailers may behave differently than when unwatched – the Hawthorne Effect [34] – highlighting the usefulness of evaluation via mystery clients. Initial implementation schemes could consider a period of evaluation using this method to assess the quality of POCT performance. Moreover, data collection relying on self-reporting like household surveys/exit interviews or pharmacist questionnaires are subject to bias/recall issues. For example, Cohen et al. (2012) reported differences between POCT positivity reported by customers and providers (89% versus 60%), as customers may not admit to buying antimalarials after testing negative [38]. Recall bias may be prominent after long follow-up periods [28].

This review is limited to English studies on PubMed/Medline. High heterogeneity prevented formal meta-analysis. Most studies were on malaria, especially qualitative discussions, so outcomes are biased towards malaria studies. This review highlights the need for quality primary POCT research, particularly for non-malaria infectious diseases. For instance, there is little data on CRP, other than to differentiate viral and bacterial infections in primary healthcare [86].

Conclusion

Private retail providers are an important point-of-access for patients, facilitating prompt diagnosis and treatment. In LMICs, POCT interventions can improve treatment-seeking behaviour, reduce inappropriate antimicrobial use and resistance, and lessen the burden on public healthcare services. This review shows POCT is not only feasible in non-formal settings but also welcomed by shops and customers. Successful implementation in LMICs requires a comprehensive protocol, including community sensitization, training, reasonable pricing, infrastructure support for low-resourced pharmacies, and wider healthcare integration.