Facilitators and barriers to point-of-care testing for sexually transmitted infections in low- and middle-income countries: a scoping review

Abstract

Background

Sexually transmitted infections (STIs) in low- and middle-income countries (LMICs) are predominantly managed by syndromic management. However, most STIs are asymptomatic. These untreated STIs cause individual morbidity, and lead to high STI prevalences.

There is increasing interest in the use of point-of-care tests (POCTs) for STIs in LMICs, which could facilitate same day testing and treatment. To best utilise these tests, we must understand the facilitators and barriers to their implementation. The aim of this review is to explore how point-of-care testing for STIs has been implemented into healthcare systems in LMIC and the facilitators and barriers to doing so.

Methods

A scoping review was conducted by searching MEDLINE, Embase, Emcare, CINAHL, Scopus, LILACS, the Cochrane Library, and ProQuest Dissertations and Theses for studies published between 1st January 1998 and 5th June 2020. Abstracts and full articles were screened independently by two reviewers. Studies were considered for inclusion if they assessed the acceptability, feasibility, facilitators, or barriers to implementation of point-of-care testing for chlamydia, gonorrhoea, trichomoniasis or syphilis in LMICs. Thematic analysis was used to analyse and present the facilitators and barriers to point-of-care STI testing.

Results

The literature search revealed 82 articles suitable for inclusion; 44 (53.7%) from sub-Saharan Africa; 21 (25.6%) from Latin American and the Caribbean; 10 (12.2%) from East Asia and the Pacific; 6 (7.3%) from South Asia; and one (1.2%) multi-regional study. Thematic analysis revealed seven overarching themes related to the implementation of POCTs in LMICs, namely (i) Ideal test characteristics, (ii) Client factors, (iii) Healthcare provision factors, (iv) Policy, infrastructure and health system factors, (v) Training, audit, and feedback, (vi) Reaching new testing environments, and (vii) Dual testing.

Conclusion

Implementation of POCTs in LMICs is complex, with many of the barriers due to wider health system weakness. In addition to pressing for broader structural change to facilitate basic healthcare delivery, these themes may also be used as a basis on which to develop future interventions. The literature was heavily skewed towards syphilis testing, and so more research needs to be conducted assessing chlamydia, gonorrhoea, and trichomoniasis testing, as well as home or self-testing.

Introduction

Sexually transmitted infections (STIs) cause significant morbidity worldwide [1]. Low-income countries have the highest prevalence of gonorrhoea, trichomoniasis, and syphilis [1], where contributing factors include poor access to healthcare, a lack of affordable diagnostics, and sociocultural barriers. In these settings, syndromic management is recommended, which is the provision of treatment to cover most organisms that could cause a set of signs and symptoms. However, it will therefore miss asymptomatic infections, which comprise the majority of STIs [2, 3]. Due to the limitations of syndromic management and a lack of laboratory capacity in resource-limited settings, there have been calls for the development and implementation of point-of-care tests (POCTs) for STIs as a tool to improve STI control [4, 5].

Notwithstanding the limited laboratory capacity in many low- and middle-income countries (LMICs), there are additional benefits to the provision of same day results. Laboratory testing away from the point-of-care requires multiple steps between sample collection and treatment of a positive case, with each step increasing the risk of pre-treatment loss-to-follow-up [6, 7]. Furthermore, time, cost, and infrastructural barriers may be particularly stark for patients accessing healthcare in LMIC [8]. Patients may not have the means to re-attend healthcare for their results and treatment, after their initial attendance. This was demonstrated amongst young people in Harare, Zimbabwe tested for chlamydia (CT) and gonorrhoea (NG) as part of a community-based testing strategy, where testing was laboratory-based, and results were available the following week. Of those with a positive test, 33.5% remained untreated despite attempts at active follow-up [3].

The 2021 World Health Organization (WHO) guidelines for the management of symptomatic STIs re-affirmed syndromic management as the standard of care where timely laboratory diagnosis is not available [9]. Importantly however, these guidelines introduced some guidance for the use of molecular assays for STIs with same day results, if these technologies were available. Prior to their implementation more widely, there is a need to understand the facilitators and barriers to the integration of POCTs into existing health systems, as well as the acceptability and feasibility of their use. Additionally, it is imperative to identify any gaps in the literature, to ensure resources are allocated appropriately to answer important questions in implementation research.

Multiple review papers have previously answered research questions relating to the use of POCTs for STIs in LMICs. This includes articles confirming the high sensitivities and specificities associated with syphilis rapid diagnostic tests (RDTs) [10, 11], and noting stockouts as an issue with regards to supply chain management of POCTs generally, including syphilis RDTs [12]. Systematic reviews have also demonstrated the positive impact of syphilis POCTs on syphilis screening rates in antenatal care [13, 14]. However, although some of these reviews have touched upon aspects of implementation, no review has primarily focussed on the facilitators and barriers to implementation. POCTs have the potential to significantly change the way STIs are managed in resource-limited settings. However, their implementation will only be successful if we understand how best to use them. The aim of this review is to explore how point-of-care testing for STIs has been implemented into healthcare systems in LMIC and the facilitators and barriers to doing so.

The specific objectives are to determine: (1) the facilitators and barriers to using POCTs for STIs in LMIC; (2) the acceptability and feasibility of using POCTs for STIs in LMIC; (3) How POCTs for STIs been incorporated into different models of care in LMIC; and (4) what gaps are present in the research knowledge base regarding the use of POCTs for STIs in LMIC.

Methods

The protocol for this scoping review has been previously published [15]. The scoping review was conducted according to Joanna Briggs Institute methodology [16]. MEDLINE, Embase and Emcare (Ovid SP), CINAHL (EBSCO), Scopus, LILACS and the Cochrane Library, including the Cochrane Central Register of Controlled Trials (Wiley) were searched for articles published in English between 1st January 1998 and 5th June 2020, with the search terms (variations on and synonyms of) “sexually transmitted infections”, “point-of-care testing”, and “low- and middle-income countries”. Sources of unpublished studies and grey literature searched included ProQuest Dissertations and Theses. The full search strategy is described in Additional file 1: Appendix S1.

Inclusion criteria

Studies were considered for inclusion if they included participants receiving or healthcare professionals providing point-of-care testing for chlamydia, gonorrhoea, trichomoniasis, and/or syphilis. Samples had to be tested at the site of sample collection. Additionally, health care and systems assessments not directly involving patients were also considered for inclusion. Human immunodeficiency virus (HIV) testing was not included in this review and so studies focusing on HIV testing alone were excluded.

Studies based in countries defined as low, lower-middle, and upper-middle economies by the World Bank, were considered for inclusion [17]. For the search strategy, the filter used for LMIC was a Cochrane filter based on the 2009 classification of countries by The World Bank [18]. This was to prevent exclusion of studies from countries that were low or middle income at the time of publication but that have since become high-income countries.

Only studies published from 1998 onwards were considered for inclusion. This was because the earliest RDTs for syphilis were developed around this time [19, 20]. Of note, rapid plasma reagin (RPR) testing was available prior to 1998, which can be used for on-site testing with provision of same day results. However, given the change in the diagnostic landscape following the development of syphilis RDTs, it was nevertheless felt appropriate to only consider studies featuring RPR after 1998.

This review was focussed on the implementation of POCTs for STIs and the facilitators and barriers to doing so. As a result, studies were only considered for inclusion if they, at least in part, assessed the acceptability, feasibility, facilitators, or barriers to implementation of point-of-care STI tests. This could be on a local, regional, national, or international level. The protocol defined facilitators as factors that promote implementation or adoption of POCTs, and barriers as factors that impede implementation or adoption of POCTs [15]. It was recognised that the definitions for feasibility and acceptability were likely to vary between studies. Feasibility was generally felt to refer to the ease and success with which POCTs were implemented logistically, whereas levels of uptake of POCTs by individuals was taken to be a surrogate for acceptability. How the POCT was implemented into a model of care was broadly split into: (i) full integration into existing services; (ii) a standalone project requiring patient consent for enrolment into the study prior to testing, but whereby treatment was prescribed based on POCT results; and (iii) a standalone project requiring patient consent for enrolment into the study prior to testing, and where POCT results required confirmatory testing prior to provision of treatment. This differentiation is necessary as interventions run by a dedicated team of study staff, not fully integrated into routine care, may not be wholly representative of how testing would work in a real-world situation.

Types of sources

This scoping review considered primary studies with quantitative, qualitative, or mixed-methods study designs. Post-hoc changes to the protocol were made for practical reasons, due to the high number of studies suitable for inclusion. After initial abstract screening, it was decided to specifically exclude review papers, pure modelling studies, and pure economic evaluations. Similarly, abstracts for which no full text was available were also excluded.

Study selection

Following the initial search, all identified records were collated and uploaded into EndNote X9 (Clarivate Analytics, PA, USA) and duplicates removed. The finalised list of records was then uploaded into Microsoft Excel, where titles and abstracts were independently screened by KM and RW for assessment against the inclusion criteria for the review. Each article was assigned a unique reference number, to ensure fidelity when merging spreadsheets to collate reviewer decisions and comments prior to discussion. Potentially relevant papers were retrieved in full, and the full text of selected citations were independently assessed in detail against the inclusion criteria by KM and RW. Any disagreements that arose between the two reviewers at each stage were resolved through discussion. Provision for further discussion with a third reviewer (JV) was included in the protocol, however this was not required.

The reference lists of the full text articles included in the review were screened for additional papers by KM. A list of articles was created based on article title alone, with duplicates and previously considered articles excluded. The abstracts and full texts of these articles were then assessed by KM and RW in the same manner as for the main study selection process.

Data extraction

Data was extracted from papers included in the scoping review using a data extraction tool developed and published with the protocol [15]. The protocol specified that the data extraction tool would be modified as necessary during data extraction. The only revisions to the data extraction tool post-publication of the protocol were the addition of the data points “STI screened” and “sample type and method of sample collection”. All data was independently extracted from papers by two reviewers, including descriptive characteristics, to enhance accountability and ensure the same articles were being compared. Both reviewers then discussed this data to decide what would be written in the final table.

Data analysis and presentation

Descriptive characteristics of research studies were presented in tabular form, including population, setting, POCT used, as well as data indicative of acceptability and feasibility.

Thematic analysis of the extracted data was used to analyse and present the facilitators and barriers to point-of-care STI testing [21]. NVivo 12 (QSR International) was used to assist with coding and theme development. Both KM and RW independently coded the notes made on the facilitators and barriers for the first ten studies. The two reviewers discussed the codes and developed a common set of initial codes. KM subsequently coded the remainder of the papers, adding new codes where necessary. An inductive approach was then used to develop themes, where were reviewed, named, and defined. Coding and themes were iteratively reviewed and refined, with regular thematic discussions between the two reviewers. Due to the nature of this project as a scoping review, coding continued past data saturation [22]. The final set of themes were presented as a narrative summary. A tree map was also produced using NVivo 12, with the size of themes and sub-themes proportional to the number of different sources coded. The Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist is included as Additional file 2: Appendix S2.

Results

The literature search performed on 5th June 2020 returned 1633 unduplicated titles and abstracts. Of these, 1414 were excluded based on title and abstract screening, leaving 219 records that were sought for retrieval. Despite assistance from the Brighton and Sussex NHS Library and Knowledge Service, the full texts for two articles were unable to be retrieved. A further 67 were abstract-only articles and were therefore excluded. There were therefore 150 full text reports assessed for eligibility, of which 73 were found eligible for inclusion. Reasons for exclusion are reported in the PRISMA diagram in Fig. 1. A further nine additional articles were identified through screening the references of the included articles, resulting in a total of 82 articles included in this review.

Fig. 1

PRISMA flow diagram

Data on the 82 articles is presented in Table 1, including study design, country, study population, and STI POCT used. The most common region was sub-Saharan Africa, where 44 (53.7%) studies took place, followed by 21 (25.6%) in Latin America and the Caribbean, 10 (12.2%) in East Asia and the Pacific, and 6 (7.3%) in South Asia. One (1.2%) multi-site study took place across three regions. All fingerprick and venous blood samples were healthcare worker (HCW)-collected, rather than self-testing.

Table 1 Table of studies included in review detailing information on study design, context, and measures of acceptability

The majority (85.4% = 70/82) of articles featured syphilis testing, whereas only 11 (13.4%) included trichomoniasis (TV) testing, 10 (12.2%) included CT testing, and 10 (12.2%) included NG testing.

Testing was provided in antenatal care only in 46 (56.1%) articles; 15 in urban/sub-urban regions; 14 in rural areas; and 17 covering both urban and rural areas. 14 (17.1%) papers focussed testing on key and vulnerable populations including sex workers and their clients, men who have sex with men (MSM), transgender women (TGW), migrant domestic workers, and homeless individuals. Other testing locations included STI/HIV services (7/82 = 8.5%), home testing (4/82 = 4.9%), primary healthcare (2/82 = 2.4%), and as part of demographic surveillance surveys (2/82 = 2.4%). One (1.2%) paper featured a survey of healthcare professionals across a range of testing locations [96]. A further six (7.3%) articles included testing for both pregnant women and another population [69,70,71,72,73,74].

Acceptability, feasibility and integration into care of point-of-care testing

Measures of acceptability for studies included in the review are reported in Table 1. Generally, high levels of uptake of point-of-care testing were noted. Importantly, in some of the studies with lower levels of uptake there was evidence of ‘overlap’ with feasibility whereby testing was not always offered due to stock or staffing issues [40], or where the requirement for venepuncture for confirmatory testing was noted as a bigger deterrent than the POCT [79]. Lower levels of uptake were also noted to cluster in studies featuring marginalised populations including MSM, TGW and sex workers [75, 79, 85, 89].

How feasibility was measured or could be interpreted varied significantly between studies (see Table 1 for full list of feasibility measures). Overall, POCTs were felt to be feasible and were implemented successfully.

Table 1 also describes whether POCTs were fully integrated into routine care, or whether they were standalone projects. Importantly, only one study investigated the use of an algorithm to allocate testing. Verwijs et al. developed an algorithm whereby they tested all participants for TV, but only tested for CT, NG, and syphilis, if they had a “positive risk score” [97]. Their algorithm had sensitivities ranging between 68.5 and 76.0% and specificities ranging between 97.4 and 100.0% for CT, NG, TV and syphilis, which was noted to be far superior to the WHO syndromic management algorithm.

Thematic analysis of facilitators and barriers to point-of-care testing for STIs in LMIC

Of the 82 studies included in the review, data on both facilitators and barriers, on facilitators alone, and on barriers alone, was able to be extracted from 53, 12, and 15 articles, respectively. Thematic analysis of the extracted facilitators and barriers revealed seven over-arching themes related to the provision of point-of-care testing for STIs in LMIC. These are (i) Ideal test characteristics, (ii) Client factors, (iii) Healthcare provision factors, (iv) Policy, infrastructure and health system factors, (v) Training, audit, and feedback, (vi) Reaching new testing environments, and (vii) Dual testing. Figure 2 demonstrates a tree map showing these themes and their sub-themes.

Fig. 2

Tree map showing thematic analysis of facilitators and barriers to point-of-care testing for STIs in LMIC (the area associated with individual themes and sub-themes is proportional to the number of different sources coded)

Ideal test characteristics

Accurate

In two studies assessing test attributes that affected willingness to test by clients, no potential for a false positive result had the largest impact [72, 81]. Additionally, health workers valued accurate tests, with concordance between POCTs and laboratory tests reinforcing their confidence in them [74]

Equipment-free

Requirements for an electrical supply, sufficient space, and ongoing maintenance needs are important barriers relevant to testing using the GeneXpert platform for CT, NG and TV [64, 67, 68, 99, 100]. The RPR test was also felt to be overly complex for on-site testing [41], with equipment or generator failure sometimes leading to an inability to complete screening [95].

Conversely, the lack of a need for an external power source or any form of laboratory infrastructure was often cited as one of the main advantages for simpler tests such as syphilis RDTs [47, 70, 92, 94].

Preferences were also noted for pre-prepared testing kits [58], that did not require any additional parts in order to conduct the test [62], as well as algorithms to simplify the whole testing process [47]

Rapid

A rapid turnaround time was noted from both clients and healthcare providers across a large number of studies as a major facilitator to point-of-care testing [36, 42, 47, 53, 59, 61, 62, 70, 72,73,74,75, 80, 81, 83, 88, 89, 94, 96, 104]. This was noted to reduce anxiety by reducing the wait time for clients [83]. Reduced analytic time will also increase potential testing capacity; in a study by Badman et al., having access to only a single four-module GeneXpert machine with an analytic time of ninety minutes, resulted in only 56% of pregnant women being enrolled due to limited analytic capacity [65].

Reliable in adverse conditions

In LMIC, testing may need to be conducted in challenging conditions including field sites with high temperatures, dusty conditions, and inadequate lighting [59, 77, 95]. As a result, tests must be robust and useable in these settings, returning low numbers of invalid tests [77]. Similarly, the need for appropriate storage conditions was a potential barrier to the use of POCTs [23, 26, 92, 94]. As such, test kits that are more resilient to fluctuating, extreme environmental conditions may be better suited to use in LMIC.

User-friendly and easy to interpret

User-friendliness of POCTs was a commonly noted facilitator to their use [24, 35, 42, 50, 59, 62, 64, 70, 74, 80, 82, 88, 92, 96, 97, 101,102,103,104]. Specifically, clear instructions [104], ease of interpretation [70, 102, 104] and minimal training requirement [47, 59, 64, 65, 70, 89, 92, 104] were described. Additionally, the similarity to other rapid tests, such as those for HIV or malaria, was felt to aid uptake and implementation by HCWs [35, 47, 57, 74].

Challenges to conducting testing including difficulties with extracting blood for fingerprick testing [42, 50, 51, 74], results interpretation [24, 27, 30, 42, 71, 80, 90], visual difficulties of healthcare professionals affecting reading of results [59, 61], and errors as a result of mistakes with timing and volume of buffer [74]. Importantly, the more complex the test, the more stringent the training requirements will be, which is a potential barrier to its use [27, 68, 71, 100].

Client factors

Cost of clinic attendance, testing and treatment

Clients noted multiple barriers to reaching clinics and to undertaking testing and treatment. These included the actual cost of testing and treatment for the client or their partner, as well as transport costs to attend healthcare facilities [43, 50, 69, 72, 81]. Young et al. noted that sometimes pregnant women would not attend referral services, which were often further away, due to additional travel costs [50].

As well as the direct financial costs associated with accessing healthcare, indirect financial and non-financial costs may also be present. Clients often noted a “lack of time” for testing, particularly where additional or long waiting times were present [32, 40, 66, 67, 85,86,87]. This may be a particular barrier for clients who have competing priorities such as the need to go to work [87]. This is further exacerbated when travelling long distances is required to access healthcare [31, 57, 65, 85, 102].

Knowledge and awareness of risk and testing

Clients’ awareness of their risk of infection and the importance of testing, as well as their trust in the testing of process, were important factors governing individuals’ willingness to be tested. Low perception of risk was often noted as a reason for test refusal [38, 47, 87, 91], as was a lack of interest or awareness in the need for testing [32, 60, 67, 85, 86, 91, 92].

Furthermore, misconceptions existed surrounding both the consequences of infections and the testing process itself. These included concerns about clandestine HIV testing [79], mistaking rapid syphilis testing for HIV testing [47], beliefs that testing for syphilis involved a urine sample or vaginal examination [57], and limited knowledge of the consequences of syphilis, particularly for babies [47, 57]. Additionally, worries about positive results acted as barriers to testing uptake [38, 40, 87, 92]. However, three of the four studies where this was noted involved dual HIV/syphilis testing and so the fear and stigma surrounding HIV may have had an impact.

Conversely, where clients were engaged in the testing process, there was often noted to be a willingness to wait for their results, and an appreciation of same day results [30, 47, 53, 57, 63, 73, 88, 97, 103]. This positive health-seeking behaviour was also demonstrated in a study by Gadoth et al., where 58 out of 65 women who screened positive for an STI returned for at least one follow-up test of cure [64].

With poor knowledge noted to be an important barrier to testing uptake, satisfactory pre-test counselling is therefore an important facilitator to testing, with provider counselling noted as a reason for satisfaction with some testing services [28, 53]. Conversely, inadequate patient counselling may be detrimental to testing uptake and some studies noted concerns in this regard [50, 60]. For example, Fleming et al. noted that mothers did not always feel fully informed about the tests and were wary of asking staff for more information [43]. Additionally, only 17.5% of pregnant women in a study in dispensaries in rural Kenya had syphilis explained to them during the consultation [49]. Baker et al. also noted that in Uganda, expectant mothers were noted to have poor knowledge for the reasons behind syphilis screening and would avoid testing if there were long waiting times [44]. Furthermore, Nkamba et al. noted that “some providers and clinic administrators stated that it is difficult to provide information to women due to high rates of illiteracy”. They note that suggestions to tackle this included use of images or drawings to supplement counselling [57].

General uptake of and trust in health services

Uptake of testing within healthcare environments is also determined by general uptake of health services within communities. This can work both positively and negatively. For example, high levels of access to antenatal care (ANC) by pregnant women are often noted, even in rural environments where accessibility may be difficult [29, 45]. However, delayed attendance at ANC was documented regularly in studies across sub-Saharan Africa [29, 44, 50, 57]. Late diagnosis and treatment may affect the efficacy of treatment in preventing congenital syphilis.

Multiple factors will affect general uptake of health services. However, trust was noted to be an important factor in some studies, including general distrust of the medical system [57], and trust in the organisation providing testing [80]. This may be particularly important for key populations and marginalised groups.

Stigma and the influence of partners and peers

Stigma surrounding STI testing was noted to be an important barrier. Nkamba et al. reported that women had fears of being stigmatised by other community members and their partners [57]. As a result, some people did not want to be seen at a health facility. This may affect an individual’s perception of the whole process. When asked about a rapid syphilis testing service set up in Mongolian ANC services, 38.6% of respondents either agreed or strongly agreed that it was felt to be “stressful and less confidential” [53].

Peers could have a direct positive influence or negative influence on testing uptake. Sabido et al. noted that knowledge of people who had already been tested was one of the primary motivators for testing, alongside perceived risk of infection [80]. However, Hall et al. reported that when friends did not accept testing, this sometimes led to an unwillingness to test [87].

Treatment of partners is a key component of STI control to prevent reinfection of the index patient and break the cycle of transmission. Partner resistance to testing and treatment [40, 64], and also a lack of provision for partner notification within clinical settings [28] may be detrimental to STI control at both individual and community levels. Furthermore, barriers to same day treatment for positive results may include a preference by healthcare providers to ensure treatment of both the index case and their partner together at a separate appointment, potentially negating one of the key benefits of point-of-care testing [57].

Individual preferences for method of testing

Across different studies, individuals had different preferences for the mode of testing. Generally, there was a preference for fingerprick testing over venepuncture [53, 69, 70, 72, 76, 80], with some specifically mentioning that was because less blood was taken [29, 38, 47]. Other noted benefits included that visibility of the results increased clients trust in them [24, 50] and perceived increased confidentiality through rapid testing [83].

However, barriers to fingerprick testing included pain [38, 51, 76, 79, 80, 82, 87, 92], general fear of fingerprick testing [86], worry about infection from the fingerprick [76, 87] and concerns that testing would affect their sex work [76].

Regarding sampling for CT, NG and TV, a preference for vaginal swabs over urine samples was noted in two studies [68, 101], as well as a preference for self-collection of swabs [68]

Context specific factors affecting testing uptake

Multiple studies reported factors associated with uptake of point-of-care testing for STIs, predominantly syphilis [34, 53, 56, 75, 76, 79, 87, 101]. Higher education level [34, 53, 101] and increased age [53, 76, 101] were noted to be associated with higher levels of testing uptake in three studies each. Other factors that were reported as being associated with higher or lower levels of uptake appeared to be much more context-specific, often reported in only one study. This included risk factors such as presence of STI symptoms [79], increased number of sexual partners [87] and different locations for sex work [75, 76, 79].

Healthcare provision factors

Effect on workload, waiting times and clinic flow

Point-of-care testing appeared to have different effects on different existing healthcare systems. Crucially, the ability to test and treat at the same visit meant that less visits were required with reduced time to treatment [24, 28, 36, 38, 42, 47, 53, 61, 73, 74]. Some studies also noted that introduction of testing actually reduced clinic waiting time [24, 50, 73] and allowed more patients to be tested per day [38, 47].

However, other studies reported increased consultation times [35, 71], increased waiting times [53, 65, 97], and disruption to clinic flow [32, 74, 99]. The analytic time was often an issue causing delays [50, 74], with the RPR noted to be particularly time-consuming [23, 42]. Badman et al. found that integration of CT/NG/TV testing into routine practice increased waiting times by 2 h [65]. As discussed above, for clients with pressures on their time, this may have an adverse effect on collection of results and treatment completion. In that study, 11 out of 67 patients with positive results had to leave prior to receiving treatment. The introduction of POCTs also lead to more complex consultations requiring multiple procedures [35, 68, 71, 96] and was often felt to generally increase workload for staff [24, 28, 32, 44,45,46, 50, 53, 61, 62, 71, 78, 99].

Staff knowledge and motivation

Multiple studies noted healthcare provider enthusiasm and satisfaction through offering a same day testing and treatment service, or a desire for point-of-care testing where it was not yet available [24, 50, 53, 57, 60, 74, 93, 99]. It was obviously important for HCWs to have knowledge of the benefits of testing and treatment [28, 57], with benefits derived for both HCWs and clients through learning about new infections [63]. Job satisfaction is also likely to improve if POCTs can provide tangible benefits to care [50, 96]. Importantly, clients valued caring staff [80].

As a result, important barriers to implementation included poor knowledge and awareness amongst HCWs of the burden of STIs, particularly syphilis in pregnancy [23, 26, 44, 45, 57], as well as poor attitudes, low motivation or a lack of commitment to screening [23, 45, 50, 60, 92].

Poor documentation and record keeping

Inadequate record keeping was a commonly reported theme, with test results and treatment not consistently recorded [25, 26, 28, 33, 43, 49, 60, 71]. HCWs specifically complained that too much documentation was required [28, 35] and that there was a lack of integration of client records [25]. Where there is no single source to monitor outcomes, such as a central ANC register with all relevant data, monitoring of outcomes becomes more difficult and there is also likely to be duplication of data if recording takes place across multiple registers or records.

Confidence in point-of-care testing

Multiple studies noted reduced confidence in point-of-care testing, which has the potential to hinder implementation [55, 61, 69, 70, 76, 80, 82, 83, 85]. For example, although most HCWs interviewed by Marks et al. reported confidence in the point-of-care results, a number of individuals noted that their confidence in the test was reduced as a result of differing results between point-of-care and laboratory testing, as well as previous experiences of discordant results with malaria RDTs [74].

Strategies to address worries about change

Although task shifting was noted by some to be a perceived benefit of POCTs [96], some healthcare providers worried about change to existing structures and processes. Laboratory workers worried that POCTs would compromise their authority [32, 38, 61]. However, identification of workplace champions and the explanation the key roles of different providers, such as laboratory workers overseeing quality control, helped to manage these worries and facilitate implementation [61].

Policy, infrastructure and health system factors

Supply chains, stockouts and procurement

Stockouts were a significant issue in the delivery of testing services across a range of settings [23, 24, 26,27,28,29, 32, 33, 35, 38, 40, 41, 43,44,45, 49, 56, 57, 60, 92, 96]. Additionally, there were also accessibility issues with syphilis tests and treatment never available in some clinics [48, 93, 96]. For example, Vani et al. report that syphilis testing in Madhya Pradesh was not offered to clients attending facilities below district level [26].

Different contexts will need different stock management and supply chain solutions. Garcia et al. reported that central procurement of dual HIV/syphilis tests was an improvement on the previous system of clinics buying locally, which resulted in variation in test kits, stockouts, and poor-quality control [61]. Nnko et al. reported that introduction of POCTs for syphilis led to improved supply chains of test kits and penicillin [47].

Complex funding pathways

Supply chain management and funding pathways are intimately linked. Therefore, complex funding and ordering pathways may have detrimental effects on stock management [23, 26, 44]. In LMIC, there are often separate, vertical funding structures for different disease programmes [23, 44, 50]. This may result in very uneven funding for different programmes, based on funder preference rather than need. This may also result in “borrowing” of supplies from other programmes with insufficient resources [23]. Other concerns regarding funding include high initial investment costs [64], source of long-term funding [43, 53] and proof of cost-effectiveness [41].

Engagement with stakeholders including authorities, policymakers and international partners

Introduction of point-of-care testing at scale is a large undertaking, and requires significant collaboration with relevant stakeholders, with studies noting this as a key to success [31, 38, 61, 75, 92]. Frequent communication and genuine engagement with authorities and policymakers was stated to be crucial, to help instigate policy change [38, 61]. Furthermore, engagement with local stakeholders is also essential to ensure support and buy-in from the community [75].

Unfortunately, some studies noted a lack of support or mentorship from international partners, who were perceived to be more focussed on other conditions such as prevention of mother to child transmission of HIV [44, 50].

Integrating into and learning from existing processes

The ability to integrate effectively into existing clinic processes was an important facilitator [37, 50, 53, 68, 97]. This requires the intervention to be flexible and adaptable to different working conditions [50]. Additionally, it is preferable to learn from and improve existing, functioning systems, rather than try and implement an entirely new system through a top-down approach [25]. Otherwise, the consequences may be disruption of existing efficient work processes. Although De Schacht et al. reported an overall increase in coverage of syphilis testing following implementation of POCTs, a significant decline was seen in some health facilities with previously high screening rates [46].

Guidelines as facilitators to testing

The inclusion of point-of-care testing into guidelines and protocols was noted to facilitate their use [27, 35, 57]. Conversely, lack of inclusion in guidelines or policy documents, or lack of guidelines entirely were barriers [26, 28, 43, 57].

Weak, fragmented health systems

One of the fundamental issues to implementing a new intervention into healthcare systems in LMIC, is that they are often weak, highly fragmented, and have significant infrastructural barriers [27, 50, 57]. This may include water shortages, power cuts and load shedding [57, 67]. This, in conjunction with high patient volumes [50, 57, 99] may in turn may result in poor working conditions for staff [50], exacerbating high staff turnover, staff shortages, and absenteeism [26, 27, 38, 40, 43, 49, 50, 57, 92, 99]. Implementation of a new testing strategy into existing pathways may therefore require additional resources over and above the testing equipment, for successful integration [51]. For wider scale implementation, one must also consider uptake across both public and private facilities, and across different levels of care, where incentives for uptake of testing may differ [32, 34, 60].

Training, audit, and feedback

Provision of high-quality training to all relevant staff

High-quality training was an important facilitator to testing [27, 50, 53, 58, 60, 61, 80]. Difficulties were therefore reported when training was felt to be insufficient [24, 26, 28, 35, 43, 45, 57, 96], not uniform [50, 92], or where training was not provided to all relevant staff [23, 35]. Watson-Jones et al. noted that often only one HCW was trained at each site, which was problematic for leave and illness cover, or if they were transferred before a replacement was trained [23].

Particular aspects of the testing process that HCWs had difficulty with were communicating positive results [40, 92], the added complexities of dual testing [27], and understanding treatment regimens and storage conditions [92]. Additionally, some staff were reported to not follow available guidelines [35, 44].

Supervision with refresher or remedial training

Ongoing supervision with focussed refresher or remedial training was identified as important in addressing poor testing practices and maintaining high standards [23, 27, 40, 42, 49, 50]. Garcia et al. noted that daily monitoring and supervision in the first months following introduction of point-of-care syphilis testing into ANC provided the necessary support to incorporate testing into their already busy schedules [61]. Supervision frequency was able to be gradually reduced as HCWs adapted to the new schedule. Similarly, Young et al. noted that HCWs appreciated observation and remedial training to ensure maintenance of skills [50].

Quality assurance and control, and accountability

Several studies noted a lack of quality assurance and control. In some settings, this was difficult due to a wide variation in the type of test kits used for screening [26]. However, in other situations there were simply few provisions in place to provide quality assurance and control, and triggers for action, such as fluctuating monthly prevalence rates, were not acted upon [23]. Barriers to the provision of quality assurance and control were similar to barriers to testing itself, including a “lack of trained staff, dedicated time, transport and reporting infrastructure” [24]. In one study where no quality control programme was in place, health workers suggested monthly quality control checks with positive and negative controls, with internal checks provided by laboratory staff [28].

Ongoing monitoring and evaluation of testing programmes may be hampered if STI screening is not a priority within health ministries [26]. This is unfortunate as real-time data is helpful in motivating staff, addressing issues early, and enables accountability [32].

Reaching new testing environments

Outreach services

Point-of-care testing can be an important enabler of testing for marginalised and at-risk populations, for whom access to clinical services is difficult. Mobile testing has been shown to reach higher risk individuals than clinic-based testing [84]. However, testing in outreach scenarios may have different facilitators and barriers to their implementation, that must be considered. For example, studies assessing syphilis and HIV testing in MSM have noted the importance of testing availability in alternative venues such as bars and clubs, and public places like parks [85], as well as environments that were “relaxed and gay-friendly” [83].

This must be balanced however with the difficulty of finding discrete locations within alternative venues [85], concerns about professionalism and quality of testing in non-traditional settings [83] and the need for linkage to care for those with positive results, if treatment is unable to be offered on-site [76, 77, 85].

Self and home testing

The availability of user-friendly POCTs allows for the possibility of self- and home-testing [83, 101, 102]. This may allow for increased privacy and confidentiality [101]. Importantly, in a study by Lippman et al., out of 910 eligible women, only one declined to participate because of feeling uncomfortable with self-collection and testing. Important barriers to address include worries about errors when self-testing [101] and ensuring clear instructions to facilitate self-testing and reading results [101, 102].

Testing in remote and rural regions

In remote and rural regions, barriers for transport of tests and consumables include large distances, cost and availability of transportation and significant geographical barriers [57, 92]. For example, in regions of the Brazilian Amazon, where air and fluvial transport were the only viable transport options, there was no access to some clinical bases during the dry season [92]. Although point-of-care testing may alleviate some of these barriers, in comparison to laboratory-based testing, rigorous planning is required when delivering testing programmes in remote regions to prevent insufficient consumables in the field preventing testing [95].

Point-of-care testing enables decentralisation of care [40, 50, 74, 92]. In addition to preventing sample loss when sending to a laboratory for testing [41, 42], it also presents an opportunity for testing in areas where previously no testing may have been available, at least without significant travel time [25, 74]. It has also been suggested that point-of-care testing may be particularly beneficial in low-volume testing sites [70], where transport of samples to a central laboratory may not be feasible for either structural or financial reasons.

Dual testing

Dual testing saving time and human resources

Across all papers where dual testing was examined, usually with a combined HIV/syphilis RDT, dual testing was felt to be beneficial in terms of saving both time and human resources [27, 28, 61, 73], as well as reducing the number of blood draws for patients [28, 51, 69, 70, 72, 81]. Barriers to their use identified by Maddox et al. included the need for dual tests to provide cost-savings over and above two single-pathogen tests, as well as the additional procedural steps required to conduct the dual test [27].

The potential benefits of dual treponemal and non-treponemal syphilis testing were also stated across a range of studies, to prevent either the need for confirmatory laboratory testing, or overtreatment [46, 70, 75, 79, 80, 89].

Positive synergy of testing for multiple conditions

In addition to increased efficiency at the point of care, dual tests may allow STI tests to benefit from structures in place for HIV testing. For example, linking syphilis testing with HIV testing was felt to provide benefits both in terms of an improved perceived value of syphilis testing given its integration with the routine, well-funded HIV test, as well as preventing stockouts because of stronger HIV supply chains [27].

A positive synergistic effect was also found when implementing integrated packages of separate tests and interventions [46, 61]. For example, Dassah et al. found that pregnant women screened for HIV, or who received intermittent preventative therapy for malaria, were more likely to be screened for syphilis [34].

Conversely, following the introduction of syphilis POCTs into ANC in Zambia, there was a significant increase in ANC attendance, HIV testing, and the number of women receiving antiretroviral prophylaxis [38]. It was suggested this could be due to increased awareness of HIV and syphilis in pregnancy as well as “greater efficiency of integrated services” [38].

Discussion

The aim of this scoping review was to assess how point-of-care testing for STIs has been implemented into healthcare systems in LMIC and the facilitators and barriers to doing so. This review has revealed a broad, diverse evidence base featuring different study methodologies, different POCTs and target populations, and different approaches to testing. The use of thematic analysis has allowed us to infer seven key themes from this data, that may be helpful in the design of future interventions and the implementation of POCTs into existing clinical services.

Ideal test characteristics generally conformed with the WHO “ASSURED” criteria, a set of target characteristics for POCTs for STIs, whereby they should aim to be affordable, sensitive, specific, user-friendly, rapid, robust, equipment-free, and deliverable in resource-limited settings [105]. Ongoing diagnostics research is necessary to develop POCTs for CT and NG that meet more, if not all these criteria. For example, the binx io CT/NG assay is now available on the market and has an analytic time of only thirty minutes [106], however current cost is likely to be prohibitive to a widescale rollout in LMICs. Importantly, the need for equipment was a barrier to the use of POCTs, and so accurate RDTs are still the most suitable type of test.

The overarching themes “client factors”, “healthcare provision factors”, and “policy, infrastructure and health system factors” mirror the “patient-level”, “facility-level”, and “health system” factors noted to influence the implementation of rapid syphilis and HIV testing in ANC by Swartzendruber et al. [107]. Stockouts, quality assurance, and preference for fingerprick testing over venepuncture were noted by both reviews.

Despite the range of studies included in this review, two sub-themes within “client factors”, namely “individual preferences for method of testing” and “context specific factors affecting testing uptake”, emphasise the need for ongoing research, particularly qualitative studies, at a local level to develop nuanced strategies, suitable for the populations served, and to target sub-populations with low uptake. Context-specific client factors associated with uptake were often variable, only noted in single studies, and did not coalesce to form themes across different environments. As such, this makes extrapolation to other settings difficult. Increased age was noted to be associated with increased syphilis testing uptake in multiple studies, which aligns with the literature, with youth reported to lack knowledge about STIs and available services, as well as experience additional barriers related to acceptability and accessibility of services [108]. Regarding methods of testing, a preference for vaginal swabs over a urine sample was noted in two included studies [68, 101]. This concurs with a previous systematic review, which found a slightly higher preference for vaginal swabs over urine samples [109]. However, again it is of utmost importance to assess the needs of the local population, and to offer multiple options where possible.

Crucially, many of the barriers identified across multiple themes were symptoms of wider health system weakness, including high workload, infrastructural barriers, poor documentation, supply chain issues, complex funding pathways, and inadequate quality assurance. Importantly, POCTs are a useful tool that can ameliorate some problems. However, they are not a “fix-all” and globally we must continue to invest in the fundamentals of healthcare provision, to strive to achieve universal health coverage (UHC), as per the sustainable development goals [110]. Reid et al. reported that health workforce density ranked as the most important element in determining UHC in LMIC [111]. Without investment in developing the capacity and accessibility of the health workforce, POCTs will be unable to fulfil their potential. Furthermore, similarly to the review by Kuupiel et al., stockouts were a significant issue for the provision of point-of-care testing [12]. Robust funding, procurement, and supply chain processes are essential to ensure an uninterrupted supply of consumables for both testing and treatment, to ensure populations receive appropriate care.

This scoping review has revealed some important gaps in the literature, where further primary research is warranted. The evidence base supporting the use of syphilis RDTs is much stronger than for CT, NG, and TV, which is mirrored by the recommendations for the use of syphilis RDTs in ANC by the WHO [112, 113]. As such, more research is required to trial POCTs for CT, NG, and TV, in different settings and with different implementation strategies. Importantly, syphilis and dual HIV/syphilis RDTs have been shown to be cost-effective in settings with varying prevalences [114, 115]. Unfortunately, point-of-care testing for CT and NG is still relatively expensive. In addition to research on cost-effectiveness of CT/NG point-of-care testing, consideration should be made to the development of algorithms to allocate tests to individuals based on their risk of an STI, such as done by Verwijs et al. [97], which may allow conservation of resources. However, a risk prediction tool may be inappropriate if prevalence is high even amongst individuals without risk factors. For example, non-selective testing was felt to be more appropriate for young people in Zimbabwe by Kranzer et al., as risk prediction tools developed were insufficiently sensitive for CT/NG [116].

Additionally, self-testing in LMIC was only trialled in two studies assessing the use of the OSOM Trichomonas Rapid Test [101, 102]. However, the scoping review by Ong et al. notes syphilis self-testing as a potential strategy to expand screening in key populations [117]. Since the initial search, studies in China and Zimbabwe have explored the role of syphilis self-testing in MSM [118, 119], both of which noted the potential of self-testing to complement facility-based testing and facilitate testing of key populations.

This review has several strengths. It was conducted according to Joanna Briggs Institute methodology, with two independent reviewers reviewing papers for study selection, and extracting data. The scale of the review is also advantageous in that it reduced the influence of individual studies, and the plentiful data allowed for the development of rich themes. However, it is not without limitations. Although beneficial in some respects, the sheer number of relevant papers and coding of the data means that some of the nuances in the data may have been lost. There were also some instances where a single research project may have led to more than one publication, such as a clinical trial and a qualitative component, both of which were included because of the broad inclusion criteria [80, 82]. This may add additional weight to certain projects. However, the thematic nature of analysis, and the large number of included studies will have hopefully ameliorated this.

Regarding the search strategy, the approach was broad, and reference lists were also searched for suitable publications. As such, the risk of missing key papers was low. However, as the search was limited to English language articles only, potentially relevant studies may have been missed, which could have introduced bias into our review. Furthermore, the search was performed on 5th June 2020, and so more recent relevant studies will have been published since, of which some are known [118,119,120]. However, given the high number of included studies, and that saturation was reached, an updated search was not felt to be warranted.

This review has focussed on curable STIs, which excludes viral STIs including HIV, human papillomavirus, herpes simplex virus, hepatitis B, and hepatitis C. The different transmission dynamics, the inability to institute a same day testing and treatment strategy, and different public health considerations, mean it would have been inappropriate to include all these infections in the review. Bacterial vaginosis (BV) was also excluded as it is not an STI, and symptomatic BV would be adequately treated by syndromic management. Screening and treating asymptomatic BV, particularly in pregnancy, is controversial and currently not recommended [121, 122].

Further limitations include that there was no cohesive, objective measure of feasibility, and test uptake was an imperfect surrogate for acceptability. It was not often possible to disentangle acceptability and feasibility, for example if test uptake was not possible due to a stockout. Furthermore, if testing required consent for a research study, potentially with additional questionnaires or follow-up, this may not truly represent uptake in normal circumstances.

Conclusions

Overall, this review contributes to the current evidence base by offering a cohesive synthesis of a range of studies. Regarding future research and practice, the themes and sub-themes could be used as a basis for a checklist or toolkit when implementing or evaluating the use of POCTs for STIs in LMICs. Crucially, this review emphasises the nuances of testing in different environments, including different workloads, different infrastructure, and different populations. These must be considered when designing interventions, ideally with the meaningful engagement of the clients, healthcare workers, and communities involved. It also demonstrates the critical state of healthcare infrastructure in some LMICs, and the stark contrast with health systems in high income countries. This review will hopefully draw further attention to these severe limiting factors to the provision of routine healthcare in LMICs. Importantly, POCTs are not a panacea and will not be suitable for all environments. In addition to continuing to strive for more equitable access to healthcare globally, we must continue to evaluate the suitability for implementation of POCTs into health systems on their merits and limitations. In particular, more research is required for CT, NG, and TV testing, and self-testing, especially regarding their implementation and cost-effectiveness. As discussed, POCTs have the potential to expand access to testing to new populations, whether by reaching marginalised populations with poor access to formal healthcare, or by being accessible to those living in remote or rural locations. However, centralised, laboratory testing will always be a key component of any national or international testing strategy, and laboratorians must continue to be supported.