Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections in pregnancy in low- and middle-income countries: A systematic review.

BACKGROUND: Sexually transmitted and genital infections in pregnancy are associated with adverse pregnancy and birth outcomes. Point-of-care tests for these infections facilitate testing and treatment in a single antenatal clinic visit and may reduce the risk of adverse outcomes. Successful implementation and scale-up depends on understanding comparative effectiveness of such programmes and their comparative costs and cost effectiveness. This systematic review synthesises and appraises evidence from economic evaluations of point-of-care testing and treatment for sexually transmitted and genital infections among pregnant women in low- and middle-income countries.
METHODS: Medline, Embase and Web of Science databases were comprehensively searched using pre-determined criteria. Additional literature was identified by searching Google Scholar and the bibliographies of all included studies. Economic evaluations were eligible if they were set in low- and middle-income countries and assessed antenatal point-of-care testing and treatment for syphilis, chlamydia, gonorrhoea, trichomoniasis, and/or bacterial vaginosis. Studies were analysed using narrative synthesis. Methodological and reporting standards were assessed using two published checklists.
RESULTS: Sixteen economic evaluations were included in this review; ten based in Africa, three in Latin and South America and three were cross-continent comparisons. Fifteen studies assessed point-of-care testing and treatment for syphilis, while one evaluated chlamydia. Key drivers of cost and cost-effectiveness included disease prevalence; test, treatment, and staff costs; test sensitivity and specificity; and screening and treatment coverage. All studies met 75% or more of the criteria of the Drummond Checklist and 60% of the Consolidated Health Economics Evaluation Reporting Standards.
CONCLUSIONS: Generally, point-of-care testing and treatment was cost-effective compared to no screening, syndromic management, and laboratory-based testing. Future economic evaluations should consider other common infections, and their lifetime impact on mothers and babies. Complementary affordability and equity analyses would strengthen the case for greater investment in antenatal point-of-care testing and treatment for sexually transmitted and genital infections.

Introduction

Sexually transmitted and genital infections (henceforth, referred to as STIs) during pregnancy are associated with a number of adverse pregnancy and birth outcomes [1-9] and their burden is highest in low- and middle- income countries (LMICs) [10-12]. Among the most prevalent infections are the curable STIs: syphilis, gonorrhoea (NG), chlamydia (CT), trichomoniasis (TV) and bacterial vaginosis (BV) [13, 14]. Untreated STIs in pregnancy can be associated with miscarriage, pre-term birth, stillbirth, low birth weight and neonatal eye and respiratory infections [15, 16].

Evidence shows that early detection and treatment of HIV and syphilis during pregnancy reduces the risk of adverse pregnancy and birth outcomes [17-20]. The World Health Organization (WHO) currently recommends HIV and syphilis testing for all pregnant women attending antenatal clinics [21-23]. The recommendation is driven by the commitment to eliminate the mother-to-child-transmission of HIV and syphilis [24]. Effective screening programs in antenatal clinics therefore play a pertinent role in ensuring early detection and treatment of HIV and syphilis, which directly improve maternal and child health.

Interventions utilised for diagnosing and/or treating STIs are illustrated in Table 1. For more than two decades, the diagnosis of STIs in many LMIC settings has been based on the WHO-endorsed strategy of syndromic management i.e. clinical diagnosis with no laboratory confirmation [25]. Syndromic management is often inaccurate and misses asymptomatic infections that make up a significant proportion of STIs in women [26, 27]. Laboratory-based diagnosis is beyond the reach of many health services in LMICs due to technical requirements and costs and even where it is available, delays in testing and the provision of results often prevent the timely initiation of treatment [28-31].

Table 1

Common STI interventions to detect and treat sexually transmitted and genital infections in low- and middle- income countries.

Intervention	Definition
Syndromic management	Identification of signs and symptoms associated with STIs and commencing treatment to alleviate symptoms and treat the infection [32].
Laboratory-based testing	Diagnosing STIs by determining the etiological agents responsible for the current infection. Testing requires skilled personnel and controlled conditions specific to a laboratory setting. Results may not be available to the clinician until several days later, requiring patients to return for the results and treatment at a later date [33].
Point-of-care testing	Diagnosing STIs by determining the etiological agent responsible for the current infection at the time of the initial patient consultation. Specimen transport is minimised or not required. Minimal training is required to perform the test. Testing may be done onsite in-front of or near to the patient. Patients should ideally receive results and treatment prior to leaving the clinic [34].

Technological advancements and the drive to find diagnostic solutions suitable for use at point-of-care have led to the development of a number of accurate, portable, simple-to-use and low- cost tests that are reshaping the global landscape of STI diagnosis and management [35]. These include rapid, point-of-care tests for HIV and syphilis that have been adopted and scaled up in many antenatal clinic settings and more recently, molecular assays for the diagnosis of CT, NG, and TV [36-38]. These new tests hold considerable promise for LMICs [39-41].

As with any health technology, it is crucial to consider any clinical benefits of point-of-care testing and treatment for STIs in pregnancy along with the associated costs of implementation and scale-up. This is especially important for LMICs that need to prioritise investment across a range of diagnostic technologies, treatments, and diseases within a relatively small fiscal space [42]. For these countries, there is considerable interest in weighing up the potential savings associated with the rapid delivery of results, reduced loss-to-follow-up and reduction of facility costs versus the long-run benefits associated with a laboratory confirmed diagnosis [43]. A new landscape for diagnostics is emerging in LMICs [37, 41, 44] and understanding the resource implications [45] and equity impact [46-48] on the health system and patients is a priority for policymakers [49]. To date, there have been no systematic reviews of the economic evidence relating to point-of-care testing and treatment for STIs in pregnancy. Consolidation and appraisal of studies in this field is timely and necessary for formulating strategies to achieve Sustainable Development Goal (SDG) 3: ensuring good health and well-being [50]. SDG 3 includes ending preventable deaths of newborns and children under 5 years of age and ensuring universal access to sexual and reproductive healthcare services [50, 51]. This systematic review examines economic evaluations of point-of-care testing and treatment of the most burdensome, curable STIs in pregnancy in LMICs. The specific objectives of this review are to:

Identify and synthesise the evidence from economic evaluations of point-of-care testing and treatment for STIs in pregnancy in LMICs;

Compare and contrast the findings, including key drivers of costs and cost-effectiveness; and

Appraise methodological and reporting quality using the Drummond 10 point checklist [52] and the Consolidated Health Economics Evaluation Reporting Standards (CHEERS) checklist [53].

Materials and methods

The methods for this review follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (S1 File). The methodology summarised below adheres to the published systematic review protocol [54]. The review is registered in PROSPERO (CRD42018109072).

Literature search and study selection

A comprehensive literature search was conducted by two researchers (OPMS and NB) in MEDLINE, Embase, and Web of Science and completed in April 2020. The search terms, shown in Table 2, were developed with the help of medical librarians to ensure a sensitive search specific to the objectives of the systematic review. All database searches were identical. Keywords and MeSH terms framed the searches, while truncation was used to capture multiple terms and the Boolean operators “OR” and “AND” to combine sub-heading search terms. Thereafter, a simplified version of the search was conducted in Google Scholar and restricted to the first 100 citations. Finally, a hand search of the bibliographies of the articles selected for full text review was conducted.

Table 2

Search terms used to identify economic evaluations of point-of-care testing and treatment for STIs in pregnancy in LMIC.

Sub-heading search terms	Search terms
Economic Evaluations	Cost-Benefit Analysis/
	(cost effectiveness or cost benefit analysis or cost utility or cost analysis).mp.
Point-of-Care testing and treatment	Point-of-Care Testing/
	("point of care" or "rapid" or "bedside" or "near to patient" or "lateral flow" or "test*" or "screening").mp.
STIs	GONORRHEA/
	exp Syphilis/
	exp Trichomoniasis/
	exp Chlamydia/
	bacterial vaginosis.mp.
	exp "bacterial vaginosis"/
	(STI or STD or "sexual transmitted disease*" or "sexual* transmitted infection*").mp.
Pregnancy	("pregnancy" or "pregnant women" or "ANC" or "antenatal").mp.

Two researchers independently conducted the database searches, screened records, titles and abstracts using Microsoft Excel (version 365), read full texts, and hand searched references of the included articles (OPMS and NB). The full texts were independently assessed against the eligibility criteria. A third senior researcher resolved any disagreements (VW).

Studies were included in this review if they:

conducted a full economic evaluation (comparing the costs and consequences of two or more options and include cost effectiveness analyses (CEAs), cost utility analyses (CUAs), cost-benefit analyses, or cost consequence analyses [55]) or a partial economic evaluation (measuring program or disease costs without comparisons with alternative options or outcomes [55]) of a point-of-care testing and treatment intervention for syphilis, NG, CT, TV or BV;

focused on pregnant women;

took place in at least one LMIC, as defined by the World Bank [56]; and

were full papers published in a peer-reviewed journal (commentaries, conference abstracts, editorials, protocols, and review papers were excluded).

No publication date nor language filter was applied.

Data extraction and analysis

Data were extracted into a form developed, using Microsoft Excel (version 365), specifically for this review. Variables extracted were guided by the categories in the CHEERS checklist and included: study setting; type of STI; type of economic evaluation; time horizon; type of testing and treatment intervention and comparator; study perspective; types of costs; measures of effectiveness and cost-effectiveness; and sensitivity analysis results.

Data extraction highlighted the significant methodological heterogeneity between studies in terms of interventions, study design, cost categories, and health outcomes. Consequently, neither a meta-analysis nor a sub-group meta-analysis was able to be performed as originally envisaged [54]. The analysis for this review was limited to a descriptive summary and narrative synthesis. This entailed the tabulation of study characteristics and a comprehensive assessment of relevant themes [57]. In addition, percentage differences were calculated to demonstrate a unitless relative difference and describe how cost-effective an intervention is. The percentage difference is calculated by finding the absolute difference in cost-effectiveness ratios dividing them by the average of the two values and multiplying this by 100. Percentage differences alongside cost-effectiveness decision rules and key outcomes emphasize between study heterogeneity. No additional statistical analyses were performed.

Study appraisal

Two checklists were used to appraise the methodological quality and reporting standards of the studies included in this review. Methodological quality was assessed using the 10 point, 13-criteria Drummond checklist [52] and reporting quality was appraised using the CHEERS checklist [53]. Together the checklists ensure reporting transparency and consistency of appraisal of the studies included in this systematic review and optimised their comparability across common themes. Each item on the checklist was assigned ‘Yes’, ‘No’ or ‘Unclear’ and where the checklist item was not applicable, ‘N/A’, was used. All studies included in this systematic review were independently appraised by two researchers (OPMS and NB). A third, senior, researcher (VW) resolved any disagreements.

Results

Search results

A total of 532 studies were identified after the initial search of the electronic databases. Sixteen studies fulfilled the inclusion criteria. The selection strategy is illustrated by the PRISMA flow diagram in Fig 1.

Fig 1

PRISMA flow diagram of the study selection process for point-of-care testing and treatment for STIs in pregnancy in LMIC.

Tables 3 and 4 summarise the key results of this systematic review. All studies assessed antenatal point-of-care testing and treatment for syphilis [58-70], except for one that focused on CT [71]. Two studies that assessed syphilis also evaluated point-of-care testing and treatment for HIV in pregnant women [72, 73]. Most studies were conducted in Africa (n = 10) [58, 59, 61–64, 66, 70, 71, 73], while three were conducted in Latin and South America [67, 68, 72] and the remaining three involved cross-country comparisons [60, 65, 69]. Of the single country analyses, four countries were classified as low-income countries [63, 68, 70, 73] and seven as middle-income countries [59, 62, 64, 66, 67, 71, 72]. For the remaining multi-country studies, the majority of countries were middle-income countries [58, 60, 61, 65, 69].

Table 3

Summary characteristics of economic evaluations for point-of-care tests for STIs in pregnancy in LMIC.

Author and Reference number	Study setting	Infection studied	Perspective	Comparators	Time Horizon	Cost components	Health outcomes	Efficiency measures
Partial Economic Evaluations
Shelley et al (2015) [64]	Zambia	Syphilis	Provider	Rapid Syphilis Test (RST) rolloutvs.RST pilot	One year or less	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Transport6. Supervision7. Quality Assurance and Control	Not applicable	Average cost per woman screened
Sweeney et al (2014) [63]	Tanzania	Syphilis	Provider	RSTvs.Rapid Plasma Reagin (RPR)	One year or less	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Training6. Information Education Communication (IEC) material	Not applicable	Average cost per woman screened
Obure et al (2017) [72]	Colombia	Syphilis and HIV	Provider	Dual HIV and Syphilis point-of-care test (dRDT)vs.HIV point-of-care test (hRDT) and RST	One year or less	1. Test2. Staff3. Treatment4. Supplies/ Consumables	Not applicable	Average cost per woman screened
Levin et al (2007) [60]	Bolivia & Mozambique	Syphilis	Provider	RSTvs.RPR	Not reported	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Training6. IEC material7. Lab8. Guidelines9. Promotion	Not applicable	Average cost per woman screened
Full Economic Evaluations
Bristow et al (2016) [73]	Malawi	Syphilis	Societal	dRDTvs.1. hRDT2. hRDT and RST3. hRDT and TPHA	Lifetime	1. Test2. Staff3. Treatment4. Patient out-of-pocket (OOP) expenses5. Cost of delivery and immediate post-natal costs	Adverse pregnancy outcomes	Total cost & DALYs averted
Owusu-Edusei et al (2011) [61]	Sub-Saharan Africa	Syphilis	Societal and Provider	Dual RST (dRST)vs.1. RST2. Onsite RPR3. Lab based RPR and TPHA4. No screening	Lifetime	1. Test2. Staff3. Treatment4. Patient OOP expenses5. Cost of delivery and immediate post-natal costs	Adverse pregnancy outcomes	Total cost & DALYs averted
Kuznik et al (2015) [65]	Latin America and Asia	Syphilis	Provider	RSTvs.No screening	Not reported	1. Test2. Staff3. Treatment	1. Neonatal death2. Still birth3. Congenital syphilis	incremental cost/DALY averted
Terris-prestholt et al (2015) [69]	Peru, Tanzania and Zambia	Syphilis	Provider	1. RPR2. RST3. dRST (and treat all positives)4. dRST (only treat if nTrp is positive)5. RST followed by RPR6. RPR followed by RST7. RST followed by dRST (and treat all positives)8. RST followed by dRST (only treat if nTrp is positive)9. MDAvs.No screening	Not reported	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Fixed clinic costs6. RPR equipment7. System costs	1. Neonatal death2. Still birth3. Congenital syphilis	cost/DALY averted
Kuznik et al (2013) [58]	Sub-Saharan Africa	Syphilis	Provider	RSTvs.No screening	Not reported	1. Test2. Staff3. Treatment	1. Neonatal death2. Still birth3. Congenital syphilis	average cost/DALY averted
Rydzak and Goldie (2008) [62]	South Africa	Syphilis	Not reported	1. RST2. RPR confirmed with TPHAvs.No screening	Lifetime	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Patient OOP expenses6. Cost of delivery and immediate post-natal costs	1. Neonatal death2. Still birth3. Congenital syphilis4. Low birthweight (LBW)5. Miscarriage	discounted costs saved per 1000 women
Schackman et al (2007) [68]	Haiti	Syphilis	Societal (CEA) and Provider (Scale up)	RSTvs.1. Syndromic management (in rural setting)2. RPR (in urban setting)	Not reported	1. Test2. Staff3. Treatment4. Patient OOP expenses	1. Neonatal death2. Still birth3. Congenital syphilis	Total incremental cost/DALY averted
Vickerman et al (2006) [70]	Tanzania	Syphilis	Not reported	RST (4 types of tests)vs.RPR	Not reported	1. Test2. Staff3. Treatment	Adverse birth outcomes	total cost/DALY saved
Blandford et al (2007) [59]	South Africa	Syphilis	Provider	1. Off-site RPR then TPHA;2. Onsite RPR3. RSTvs.No screening	One year or less	1. Test2. Staff3. Treatment4. Supplies/ Consumables	Congenital syphilis	total incremental cost/cases averted
Mallma et al (2016) [67]	Peru	Syphilis	Not reported	RSTvs.RPR	Not reported	1. Test2. Staff3. Treatment4. Supplies/ Consumables	Adverse birth outcomes	cost/DALY averted
Romoren et al (2007) [74]	Botswana	Chlamydia	Provider	1. syndromic management with Azithromycin treatment;2. point-of-care testing with Erythromycin treatment;3. point-of-care testing with Azithromycin treatmentvs.Syndromic management with Erythromycin treatment	One year or less	Point-of-care testing and treatment:1. Test2. Staff3. TreatmentSyndromic management:1. Staff2. Treatment3. Training4. Supervision	1. Neonatal death2. Still birth3. Congenital syphilis	incremental cost/cases cured
Larson et al (2014) [66]	Zambia	Syphilis	Provider	No screening programvs.1. 62% of antenatal care attendees tested, only 10% of positive cases were treated;2. 62% of antenatal care attendees tested, all positive cases were treated; and3. All antenatal care attendees tested and all positive cases were treated	One year or less	1. Test2. Staff3. Treatment4. Supplies/ Consumables5. Training	1. Neonatal death2. Still birth	total cost/DALY averted

CEA: cost-effectiveness analysis; DALY: disability adjusted life year; dRDT: dual HIV and syphilis rapid diagnostic test; hRDT: HIV rapid diagnostic test; IEC: Information Education Communication; MDA: mass drug administration; nTrp: Non-treponemal; RST: rapid syphilis test; RPR: rapid plasma regain testing; TPHA: Treponema pallidum Hemagglutination Assay.

* Quality Control/ Quality Assurance refers to reviewing the quality of all the factors required for effective testing and treatment for syphilis in pregnancy.

** Larson et al (2014) utilised country representative statistics, from a previously conducted evaluation study, to build their scenarios. Country statistics showed that 62% of antenatal clinic attendees were tested for syphilis while only 10% of the test positives were treated.

Table 4

Summary results extracted from the economic evaluations for point-of-care tests for STIs in pregnancy in LMIC.

Author	Results	Cost-effectiveness decision rule	Drivers of cost and cost-effectiveness	Key findings	Generalisability of results	percentage difference
Partial Economic Evaluations
Shelley et al (2015) [64]	Pilot period• Average unit cost/woman tested: USD 1.49• Average unit cost/woman treated: USD 14.12Rollout period• Average unit cost/woman tested: USD 4.84• Average unit cost/woman treated: USD 72.73	Average cost per woman tested and treated is lower than the baseline average cost per woman tested and treated (Pilot program)	1. Cost (RST test kit)2. Screening coverage3. Supply wastage	RST rollout had higher costs than RST pilot	Not stated	105.5% (test); 135% (treated)
Sweeney et al (2014) [63]	RPR• Average cost/woman tested: USD 2.32• Average cost/woman treated: USD 12.96RST• Average cost/woman tested: USD 1.92• Average cost/woman treated: USD 21.40	Average cost per woman tested and treated is lower than the baseline average cost per woman tested and treated (RPR)	1. Screening coverage2. Supply wastage3. Time taken to test	RST had higher costs than RPR	Not stated	18.9% (test); 49.1% (treat)
Obure et al (2017) [72]	RSTa• Average Unit Cost/woman tested: USD 10.26• Average Unit Cost/woman treated: USD 607.99.dRDT• Average Unit Cost/woman tested: USD 15.89• Average Unit Cost/woman treated: USD 1859.26	Average cost per woman tested and treated is lower than the baseline average cost per woman tested and treated (single hRDT and RST)	No sensitivity analysis	RST (and hRDT) had lower costs than dRDT	Not stated	43.1% (test); 101.4% (treat)
Levin et al (2007) [60]	RPR• Average Unit Cost/woman tested: USD 1.43 (Bolivia) and USD 0.91 (Mozambique)• Average Unit Cost/woman treated: USD 40.09 (Bolivia) and USD 12.25 (Mozambique)RST• Average Unit Cost/ woman tested: USD 1.91(Bolivia) and USD 1.05 (Mozambique)• Average Unit Cost/ woman treated: USD 40.77 (Bolivia) and USD 13.45 (Mozambique)	Average cost per woman tested and treated is lower than the baseline average cost per woman tested and treated (RPR)	No sensitivity analysis	Mozambique: RST had higher costs than RPRBolivia: RST had lower costs than RPR	Not stated	28.7% (for testing in Bolivia) and 1.7% (for treating in Bolivia);14.3% (for testing in Mozambique) and 9.3% (for treating in Mozambique)
Full Economic Evaluations
Bristow et al (2016) [73]b	hRDT only• Total cost: USD 21 875 298• Effectiveness: 110 875 DALYdRDT• Total cost: USD 21 479 390• Effectiveness: 108 693 DALYhRDT and RST• Total cost: USD 21 864 363• Effectiveness: 110 691 DALYhRDT and TPHA• Total cost: USD 21 893 483• Effectiveness: 110 697 DALY	Incremental costs and DALYs is lower than the baseline (RST & hRDT; HIV only; hRDT &TPHA)	1. Prevalence2. Screening coverage3. Risk of adverse outcome	dRDT had lower costs and was more effective than hRDT; hRDT; and RST; hRDT and TPHA in lab	Results generalisable in similar countries	1.9% (Total Cost); 2% (DALYs)
Owusu-Edusei et al (2011) [61]c	Dual RST• Total cost: USD 79 000• Effectiveness: 5 adverse pregnancy outcomes• DALYs: 42RST• Total cost: USD 76 000• Effectiveness: 2 adverse pregnancy outcomes• DALYs: 15Onsite RPR• Total cost: USD 84 000• Effectiveness: 11 adverse pregnancy outcomes• DALYs: 94Offsite (lab based) RPR and TPHA• Total cost: USD 86 000• Effectiveness: 13 adverse pregnancy outcomes• DALYs: 107No screening• Total cost: USD 106 000• Effectiveness: 39 adverse pregnancy outcomes• DALYs: 341	Cost-savings are greater than the baseline and over-treatment rates are lower than the baseline	1. Cost (RST test kit)2. Test sensitivity (performance)3. Cost associated with adverse pregnancy outcomes	dRST had lower costs but was less effective than RST;dRST had lower costs and was more effective onsite RPR;RPR and TPHA (lab); no screening	Not stated	33% (Total Cost); 183.1% (DALYs)
Kuznik et al (2015) [65]d	No screening program:Average ICER as not reported for either Asia or Latin AmericaRST:Asia: average ICER was USD 53/ DALY averted; andLatin America: average ICER was USD 60/ DALY averted	WHO threshold (ICER below country’s Gross Domestic Product (GDP) per capita	Prevalence	RST cost-effective compared to comparator	Not stated	No baseline figure provided to calculate percentage difference
Terris-prestholt et al (2015) [69]	No screening program:ICER: Reported as not applicableMost cost-effective outcomes:1. Peru: RST ICER was USD 53.69/ DALY averted;2. Tanzania: MDA ICER was USD 8.7/ DALY averted; and3. Zambia: MDA ICER was USD 5/ DALY averted	ICER is lower than the baseline scenario (no screening program)	1. Cost (fixed clinic costs)2. Screening coverage3. RST reactivity rate4. Antenatal clinic attendance	Peru: RST cost-effective compared to comparators;Tanzania: RST not as cost-effective as MDA;Zambia: RST not as cost-effective as MDA	Alludes to generalisability of results, but not definitively	121.7% (for Peru); 167.5% (for Tanzania); 170.6 (for Zambia)
Kuznik et al (2013) [58]d	No screening program:Average ICER was not reportedRSTAverage ICER was USD 11/ DALY	Threshold of an ICER less than Gross National Income (GNI) per capita	Prevalence	RST cost-effective compared to comparator	Not stated	No baseline ICER provided to calculate percentage difference
Rydzak and Goldie (2008) [62]	No screening program• Discounted costs saved/1000 women: USD 110 220RPR• Discounted costs saved/1000 women: USD 161 310RST (dominates)• Discounted costs saved/1000 women: USD 170 030	Cost-savings are greater than the baseline scenario (no screening program)	1. Cost (labour, RST test kit price and other supplies)2. Test sensitivity (performance)3. Cost associated with adverse pregnancy outcomes	RST cost-effective compared to comparators	Results not generalisable to other settings or scenarios	200%
Schackman et al (2007) [68]	Rural setting:RPR: ICER was USD 10.64/ DALYRST: ICER was USD 6.83/ DALYUrban setting:RST: ICER was USD9.95/DALY	Thresholds:1. ICER lower than USD50/DALY recommended for health interventions in resource- poor countries; and2. ICER lower than the GDP per capita of country where intervention is being implemented	1. Prevalence2. Test sensitivity (performance)	RST cost-effective compared to comparators	Results generalisable to HIV scale-up in resource poor settings	No baseline ICER provided to calculate percentage difference
Vickerman et al (2006) [70]	Test using serum:RPR• Cost/DALY: USD 12/DALYRST (Determine)• Cost/DALY: USD 17/DALYTest using whole blood:RPR• Cost/DALY: USD 16.8/DALYRST (Determine)• Cost/DALY: USD 14.1/DALY	Threshold cost-effectiveness ratio is lower than the baseline (RPR test)	Test sensitivity (performance)	RST cost-effective compared to comparators	Not stated	19.2% (using serum) and 32.5% (using whole blood)
Blandford et al (2007) [59]	Off-site RPR and confirmatory TPHA• Cases averted: 18• Cost/case averted: USD 82 per case.Onsite RPR• Cases averted: 16• Cost/case averted: not published (but states “dominated”RST• Cases averted: 27• Cost/case averted: USD104 per case.	ICER is lower than the baseline scenario (no screening program)	1. Prevalence2. Test sensitivity (performance)3. Relative distribution of active/latent syphilis	RST cost-effective compared to comparators	Not stated	33.8%
Mallma et al (2016) [67]	RST• Cost/DALY averted: USD 46/ DALY avertedRPR• Cost/DALY averted: USD 109/ DALY averted	WHO threshold (ICER is below the country’s GDP per capita)	1. Prevalence2. Cost (salaries)3. Test positivity rates4. Screening coverage5. Number of hours worked	RST cost-effective compared to comparator	Not stated	81.3%
Romoren et al (2007) [74]	Syndromic Management• Cases cured (erythromycin): 800• Cost/ cured case (erythromycin): USD 66• Cases cured (azithromycin): 1500• Cost/ cured case (azithromycin): USD 21point-of-care test (0.75 sensitivity- base case)• Cases cured (erythromycin): 3200• Cost/ cured case (erythromycin): USD 35• Cases cured (azithromycin): 3500• Cost/ cured case (azithromycin): USD 31	Willingness-to-pay threshold (cost-effectiveness ratio is lower than the willingness to pay threshold)	1. Prevalence2. Cost (point-of-care test)3. Test sensitivity (performance)4. Probability of partner notification	point-of-care test combined with Azithromycin treatment cost-effective compared to comparators	Results generalisable to other sub-Saharan countries	62.5% (using erythromycin) and 94.9% (using azithromycin)
Larson et al (2014) [66]	Evaluation study conditions (62% of antenatal clinic (ANC) attendees tested and 10% of positive cases treated):• USD628/ DALY averted62% of ANC attendees tested, and all positive cases treated:• USD66/ DALY avertedAll ANC attendees tested, and all positive cases treated:• USD60/ DALY averted	ICER is lower than the baseline scenario	1. Prevalence2. Cost (training)3. Number of antenatal clinic attendees tested and treated4. Staff time	RST (testing and treating all antenatal clinic attendees) cost-effective compared to comparators)	Results not generalisable to other settings or scenarios	165.1%

ANC: Antenatal clinic; CEA: cost-effectiveness analysis; DALY: Disability-adjusted-life-years; GDP: Gross domestic product; GNI: Gross national income; dRDT: dual HIV and syphilis rapid diagnostic test; hRDT: HIV rapid diagnostic test; ICER: Incremental cost effectiveness ratio; IEC: Information Education Communication; MDA: mass drug administration; nTrp: Non-treponemal; RST: rapid syphilis test; RPR: rapid plasma regain testing; WHO: World Health Organization.

a. Only syphilis was included in the analysis, results for HIV were excluded.

b. Did not publish an ICER (or Cost/DALY) as the difference between each comparator would be smaller than 1, and therefore not significant.

c. Did not publish an ICER (or Cost/DALY), as it may not have yielded a meaningful outcome.

d. Reported average ICERS.

Economic evaluations

Type of economic evaluation

Economic evaluations were categorised as either partial or full evaluations Table 3 indicates that four studies were partial economic evaluations [60, 63, 64, 72]. The remaining 12 studies were full economic evaluations, which included two cost-consequence analyses [61, 73] and 10 CEAs [58, 59, 62, 65–71], seven of which were CUAs [58, 65–70].

Perspectives, costs, and outcomes

Nine studies conducted an economic evaluation using trial data [59, 60, 63, 64, 66–69, 72], while seven studies utilised existing literature [58, 61, 62, 65, 70, 71, 73]. Ten studies took a provider perspective [58–60, 63–66, 69, 71, 72], three took a societal perspective [61, 68, 73] and three studies did not specify the perspective taken [62, 67, 70]. The most common types of costs measured were direct and indirect medical costs related to tests and/or treatment, staff time, and other consumables such as medical supplies. Patient costs were rarely captured (n = 4) due to a provider perspective being taken by most studies.

All studies in this review compared a point-of-care testing and treatment intervention against one or more comparators as shown by Fig 2. Four studies compared laboratory-based testing with a point-of-care testing and treatment strategy [60, 63, 67, 70]. Six studies compared point-of-care testing and treatment against multiple comparators including laboratory testing, another point-of-care test, mass drug administration (MDA), syndromic management and/or no screening program [59, 61, 62, 68, 69, 73].

Fig 2

Economic evaluation comparators for point-of-care testing and treatment for STIs in pregnancy in LMIC.

MDA: Mass Drug Administration; POC: point-of-care.

Most cost-effectiveness studies found point-of-care testing and treatment more cost-effective compared to laboratory-based testing [62, 67, 68, 70], no screening [58, 65, 66] or syndromic management [71] as shown in Table 4. For the full economic evaluations (n = 12), a variety of effects were measured. Most studies measured effectiveness by estimating the number of cases of adverse outcomes averted, such as miscarriage, stillbirth, neonatal death and congenital syphilis [58, 61, 65–71, 73].

Three-quarters of the full economic evaluations (n = 9) calculated incremental cost effectiveness ratios (ICERs), typically defined as the incremental cost per disability adjusted life years (DALY) averted or the incremental cost per case averted/cured [35, 37–40, 42, 43, 45, 46]. The ICERs reported varied considerably due to the heterogeneity of methods used to derive cost-effectiveness; therefore, making comparability difficult. The percentage differences emphasise the heterogeneity between studies included in this review. The greatest percentage difference between comparators was 170.6% (Cost/DALY of 63.1 and 5), while the smallest was 19.2% (Cost/DALY of 17 and 11.8).

Nine studies indicated a time horizon; the time horizon was typically restricted to 12 months or less [59, 63, 64, 66, 71, 72]. Only three studies modelled lifetime costs and effects of point-of-care testing and treatment for STIs in pregnancy on mothers and babies [61, 62, 73]. These studies did not explicitly indicate whether cost-effectiveness or outcomes were sustained over time.

Key drivers of economic evaluations

A sensitivity analysis was performed in 14 studies (87.5%) by varying key assumptions and recording the impact this had on the findings of the evaluation. Ten studies conducted a univariate sensitivity analysis and of those studies, six conducted a multivariate sensitivity analysis [59, 61, 63, 64, 68, 73] and three conducted a probabilistic sensitivity analysis [58, 69, 71] in addition to the univariate sensitivity analysis. The main drivers for costs and cost effectiveness are illustrated in Table 4 and included: STI prevalence [58, 59, 65–69, 71, 73], cost of test, cost of treatment, cost of training or salaries and wages [58, 61, 62, 64, 66–69, 71], test sensitivity and specificity [58, 59, 61, 62, 68, 70, 71], and treatment and/or screening coverage [63, 64, 66, 69, 73].

Quality appraisal

All studies met 75% or more of the methodology criteria of the Drummond Checklist and 60% for the CHEERS checklist (Fig 3). Fig 3 illustrates the shortcomings with respect to reporting criteria in the reviewed papers. The appraisal based on the Drummond checklist (Table 5), indicated that a key gap was the reporting of all relevant costs with few studies taking account of training, quality control and quality assurance. The appraisal using the CHEERS checklist (Table 6) highlighted that only half the studies (n = 8) clearly outlined the time horizon for the analysis [59, 61–64, 66, 71, 72], most taking a one year or shorter time horizon [59, 63, 64, 66, 71, 72]. Half the studies (n = 8) utilised analytical methods to deal with missing values and uncertainty [58, 61, 63, 65, 67, 68, 70, 71]. Only a third of the studies (n = 6) explained variations in the data by different population characteristics [60, 64, 67, 70–72]. The most common sub-group analyses were by type or location of health facility (hospital or health centre and urban or rural).

Fig 3

Assessment of methodological and reporting quality of economic evaluations of point-of-care testing and treatment for STIs in pregnancy in LMIC (%).

Table 5

Drummond 10-point checklist.

		References
		[65]	[69]	[58]	[62]	[68]	[70]	[59]	[67]	[71]	[66]	[64]	[63]	[72]	[60]	[73]	[61]
1	Well defined research question stated	1	1	1	1	1	1	1	0.5	1	1	0.5	1	1	1	1	1
2	Comprehensive description of competing alternatives	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
3	Evidence of program effectiveness included	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
4	All relevant cost and consequences for each alternative identified	0.5	1	0.5	0	0.5	0	0.5	0	0.5	1	1	1	0.5	1	0.5	0.5
5	Costs and consequences measured accurately and appropriately	0.5	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
6	Costs and consequences valued credibly	1	1	1	0.5	1	1	0.5	1	1	1	1	1	1	1	1	1
7	Costs and consequences adjusted for differential timing	1	1	1	1	1	0	0	1	0	1	1	0.5	0	1	0.5	0.5
8	Incremental analysis of costs and consequences performed	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
9	Allowance made for uncertainty in cost and consequence estimates	0.5	0.5	0.5	0.5	0.5	1	0.5	1	1	0.5	1	0.5	0.5	0.5	0.5	0.5
10	Presentation/Discussion included all concerns raised in the results	1	1	1	0.5	1	1	1	1	1	1	1	1	1	1	1	1
Score		8.5	9.5	9	7.5	9	8	7.5	8.5	8.5	9.5	9.5	9	8	9.5	8.5	8.5

In this Table 1 denotes that the checklist item is clearly included in the study; 0, that the checklist item is not included in the study; and 0.5, that although the item is present, it is not clear.

Table 6

Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist.

Section/Item (including item number) on CHEERS checklist			References
			[65]	[69]	[58]	[62]	[68]	[70]	[59]	[67]	[71]	[66]	[64]	[63]	[72]	[60]	[73]	[61]
Title and abstract	1	Title	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	2	Abstract	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Introduction	3	Background	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Methods	4	Target pop. & sub-groups	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	5	Setting and location	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	6	Study perspective	Y	Y	Y	Y	Y	N	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
	7	Comparators	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	8	Time horizon	Y	N	N	Y	N	N	Y	N	Y	Y	Y	Y	Y	N	N	Y
	9	Discount rate	Y	Y	Y	Y	Y	N	N	Y	N	Y	Y	Y	N	Y	Y	Y
	10	Health outcomes	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	N	N	Y	Y
	11a	Effectiveness measures: single-study estimates	NA	Y	NA	NA	NA	Y	NA	N	NA	NA	NA	NA	NA	NA	NA	NA
	11b	Effectiveness measures: synthesis-based estimates	Y	NA	Y	Y	Y	NA	Y	N	Y	Y	NA	NA	NA	NA	Y	Y
	12	preference based outcome measurement/ valuation	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	NA	NA	NA	NA	Y	Y
	13a	Resource/Cost estimates: single study-based	Y	NA	NA	NA	NA	NA	Y	Y	NA	Y	Y	Y	Y	Y	NA	NA
	13b	Resource/Cost estimates: model-based	NA	Y	Y	Y	Y	Y	NA	NA	Y	NA	NA	NA	NA	NA	Y	Y
	14	Currency, price data and conversion	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	15	Model choice	Y	N	Y	Y	Y	N	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
	16	Assumptions	N	N	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y
	17	Analytical methods	Y	N	Y	N	Y	Y	N	Y	Y	N	N	Y	N	N	N	Y
Results	18	Study parameters	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
	19	Incremental costs and outcomes	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	NA	NA	NA	NA	Y	Y
	20a	Characterising uncertainty: single study-based	Y	Y	Y	Y	Y	Y	Y	Y	NA	Y	Y	Y	N	N	Y	Y
	20b	Characterising uncertainty: model-based	NA	NA	NA	NA	NA	NA	NA	NA	Y	NA	NA	NA	N	N	NA	NA
	21	Characterising heterogeneity	N	N	N	N	N	Y	N	Y	Y	N	Y	N	Y	Y	N	N
Discussion	22	Summary key findings	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Other	23	Funding source	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N
	24	Conflicts of interest	Y	Y	Y	N	Y	Y	N	Y	Y	Y	Y	Y	Y	N	Y	N

Y: Yes (included in the article); N: No (Not included in the article); NA: Not applicable (not applicable to the type of study).

Discussion

This review identified 16 economic evaluations set in LMIC on point-of-care testing and treatment for STIs in pregnancy and presented a synthesis of the evidence. All but one study focused on syphilis and most were set in African countries. The majority of studies in this review suggest that point-of-care testing and treatment for syphilis and CT in pregnancy can be cost-effective in LMIC settings when compared to no screening programs, laboratory-based testing and/or syndromic management. These studies also indicate that point-of-care testing and treatment for STIs is most cost-effective where access to alternative testing mechanisms is limited, including laboratory testing facilities. Further, there was considerable variation in the types of costs and outcomes utilised by economic evaluations, as well as in the time horizons and sample sizes, which proved to be a challenge in comparing the costs and cost effectiveness of interventions across the different settings. The decision rules for cost-effectiveness and percentage differences in Table 4 demonstrate the between study variation. The decision rules indicate that cost-effectiveness is dependent upon the decision rule utilised. Further, percentage difference highlights the difference between the largest and smallest cost-effectiveness measures per study. The largest percentage difference between comparators was 170% [69] while the smallest 19% [70], which demonstrates the large variability with respect to measures of cost-effectiveness. Key drivers of cost and cost-effectiveness were STI prevalence and costs (including the cost of tests, treatment, training costs, and salaries or wages.

Our review also highlighted some gaps in the evidence, in terms of methodology and scope. First, the evidence is limited both in geographic and infection scope. Most studies were conducted on the African continent, with only one study conducted in Asia [65] and none in the Pacific. Further, point-of-care testing and treatment for syphilis in pregnancy dominated the literature. Only one economic evaluation focussed on CT and there were no economic evaluations of other prevalent, curable STIs such as NG, TV or BV. The lack of studies from Asia-Pacific countries is of concern given the high STI prevalence, particularly for common infections including syphilis, NG, TV, CT, reported among pregnant women in the region [68, 72, 75] as well as the widespread integration of point-of-care testing and treatment for syphilis in antenatal clinics [22]. Although existing economic evaluations provide guidance to cost-effectiveness, context matters. In particular, results from economic evaluations cannot always be generalised because local disease epidemiology, social, cultural and financial barriers to implementation vary between settings [76]. These could lead to differences in the uptake of testing and treatment between settings, and therefore, also differences in the cost of scale up.

In the last decade, the number of studies based in LMICs evaluating the utility of point-of-care tests for the detection and treatment of STIs in pregnancy has increased [38, 77], which has influenced the number of economic evaluations conducted on these types of tests. Yet with few cheap, accurate and easy to use point-of-care tests for CT, NG, TV and BV available [78], there have not been many effectiveness trials conducted in LMIC. As such, the data available to conduct robust economic evaluations is limited. Therefore, there is a need to expand the scope of effectiveness studies and accompanying economic evaluations of point-of-care testing for high-burden STIs in pregnancy [38, 79, 80].

Second, most studies included in this review took the provider perspective. Only three studies took a societal perspective. The paucity of studies conducted from the societal perspective potentially excludes and/or underestimates the direct and indirect costs to women such as reducing patient waiting times and reducing the number of clinic visits required to receive results, which in turn is expected to reduce the considerable direct and indirect costs incurred by many women seeking antenatal care in LMICs [81, 82]. Consequently, these direct and indirect costs can be, in many settings, barriers to participation as costs are often borne to the patient as OOP expenses and can lead to catastrophic health expenditure, which might lead to a drop in uptake/participation.

Third, most economic evaluations reviewed had short time horizons of 12 months or less. Short time horizons do not consider the lifetime implications of effective point-of-care testing and treatment of STIs in pregnancy. Economic evaluations in this field should ideally incorporate the long-term costs and benefits to both mother and baby from adopting point-of-care testing in order to capture said costs and benefits that occur later in life.

Our review indicated that all but two studies conducted a sensitivity analysis, which demonstrate the variables that drive costs and cost-effectiveness in each study. The most common drivers of cost and cost-effectiveness were STI prevalence and costs (which included the cost of tests, treatment, training costs, and salaries or wages). The sensitivity analyses of the studies included in this review did not explore the potential variation in loss-to-follow-up. These findings are consistent with other systematic reviews that have indicated that HIV prevalence and the cost of the HIV test are key drivers of cost-effectiveness of HIV screening in pregnancy [83] and key populations, such as sex workers [84].

We also found that most studies applied a narrow definition to the costs of point-of-care testing including only the cost of test kits, treatment costs and staff salaries. Costs associated with other core activities such as quality control, quality assurance and procurement are not negligible, and thus these studies may have underestimated the true cost of POC testing [85]. Underestimating costs leads to inaccurate projected costs for implementation and scale up, which consequently means that insufficient budget will be allocated to the implementation and scale up of the intervention. Insufficient funding, will inevitably mean that scale up will not go as planned, may be suspended, or may not run successfully [45]. Standardising the methodology, or introducing guidelines, for measuring costs may mitigate the likelihood that crucial costs are excluded from analyses. In addition, this will aid the comparison of costs across studies, which was particularly difficult in this review.

Finally, increased attention should also be paid to the broader issues of affordability and equity. Despite recent recommendations by the International Society for Pharmacoeconomics and Outcomes Research to conduct budget impact analyses alongside economic evaluations [45], no studies in this review did so although two did mention its importance [58, 62]. Budget impact analyses explore economic and financial consequences highlighting the affordability of program implementation. They specifically factor in short-run costs to cater for government budget allocations [45, 86] and are especially relevant in settings where health care resources are highly constrained [49]. There have been ongoing calls for incorporating equity within economic evaluations of health interventions [46, 48]. This is particularly relevant for point-of-care testing and treatment, which can reduce inequalities in healthcare by diagnosing and treating STIs in a single visit to a health facility. This is important in populations with limited access to laboratory facilities [87]. Future studies in this field should explore ways to expand cost-effectiveness analysis to address health equity concerns including distributional cost-effectiveness analysis which quantifies the distribution of costs and effects by equity-relevant variables such as socioeconomic status, location, ethnicity, sex and severity of illness [88]. Alternatively, non-health benefits such as financial risk protection could be measured using extended cost-effectiveness analysis [47, 89].

Economic evaluations generate evidence to optimally allocate resources and provide the foundation for the delivery of healthcare interventions [55]. This review has highlighted that point-of-care testing for STIs in pregnancy has proven to be cost effective in African settings. The results may be generalisable to similar settings, however consideration should be taken with the shortcomings revealed by this review, including the heterogenous nature of economic evaluations, the perspective taken and the short run time horizons. Suggesting that evidence should be interpreted with caution, and an economic evaluation conducted prior to implementing programmes to fill these gaps and provide better information to aid uptake, scale up and indicate priority settings in resource constrained contexts. Going forward, extended cost-effectiveness analyses and budget impact analyses provide evidence for the implementation of a cost-effective intervention, which is equitable, affordable, and sustainable. Extended cost-effectiveness analyses highlight the distributional effects within a given populations as a result of the implementation of an intervention; while budget impact analyses determine the investment required by the government to implement an intervention; and demonstrate affordability given available resources and budget constraints.

This review has some limitations, with respect to the synthesis of key results, the significant variability in economic methods, particularly regarding the types of costs and outcomes measured, meant that the planned meta-analysis was not possible [54]. Further, few studies utilised a societal perspective to determine costs, which may have understated the true value associated with point-of-care testing and treatment for STIs in pregnancy. However, the descriptive and narrative synthesis provides valuable insight into these limitations and gaps and suggests direction for future research. Additionally, most studies took a narrow view of costs and measured them along a short-run time horizon, which highlights a short-coming of the studies included in this review. This systematic review did not include grey literature, which can also be an important source of evidence but less likely to be subject to independent scientific peer-review, where additional bias is introduced. Finally, a further enhancement of economic evaluations would be the inclusion of equity and affordability analyses.

Conclusion

Our review indicates that point-of-care testing and treatment for syphilis in pregnancy is cost-effective in LMICs compared to laboratory-based testing, syndromic management and where no testing programs have been implemented. The review also revealed that key drivers of cost and cost-effectiveness are STI prevalence and costs (including the cost of tests, treatment, training costs, and salaries or wages. It also highlights important gaps in the published literature that requires urgent attention by researchers. These gaps include broadening the scope of economic evaluations to include common curable STIs, such as CT, NG, TV and BV; the geographical representation of studies; the costs borne to pregnant women and their families as a result of accessing testing and treatment for STIs; the life time costs and effects of testing and treatment on mothers and their babies and finally; the budget and equity implications of implementing point-of-care testing and treatment versus other screening practices.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist.

(DOCX)

Click here for additional data file.

(XLSX)

Click here for additional data file.

3 Nov 2020

PONE-D-20-31634

Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections (STIs) in pregnancy in low- and middle-income countries (LMICs): A systematic review

PLOS ONE

Dear Dr. Saweri,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

This is a well-written manuscript. I concur with the reviewers' comments.

Please submit your revised manuscript by Dec 18 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Remco PH Peters, MD, PhD, DLSHTM

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please amend either the title on the online submission form (via Edit Submission) or the title in the manuscript so that they are identical.

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: GENERAL/MAJOR COMMENTS

This systematic review of economic evaluations of point of care STI tests is well-designed, well-conducted and generally well-described.

However, the critical interpretation might be deepened – both in (A) the descriptive analysis in Table 3 and (B) the Discussion.

(A) Failing the option to do a formal statistical meta-analysis (due to heterogeneity in outcome measures and …), the crux of the findings are in the descriptive Table 3. In this table, however, between Results and outcome/conclusion the relation is not always clear, nor are the comparative outcome definitions under Results always clear. Please see Specific comments below.

(B) The discussion dwells on methods of economic evaluations and to what extend the existing, reviewed studies met or not the optimum standards. Apart from this partly academic matter, what is the more practical conclusion from the presented evidence? Concretely, if a health minister (in country X) considers changing its policy from syndromic treatment to point-of-care testing (for, say, syphilis), what could s/he learn from this review? Should a new (local?) study be carried out, and if so what would be the preferred methodology and design? Or is the evidence in favor of POC testing already clear and uncontested, in all or some (which=?) settings?

SPECIFIC COMMENTS

RESULTS: Table 3:

Shelley et al: What is meant with ‘Quality control measures valuable to scale-up of RST’? How does the ‘quality control’ relate to the Results? Please expand the Results such that this becomes clear. Also – if space allows – comment on how rural and urban settings differed, e.g. which setting had higher syphilis prevalence, or what was the driver of the comparative cost and screening and treatment volumes?

Bristow et al: What was the criterion for ‘best option’? I suppose, cost per DALY averted? If so, please add ‘cost per DALY averted’ to the results (possibly replacing cost and DALY). Or if you keep the outcome DALY: is this DALYs lost or DALYs averted?

Adding an outcome per infection or DALY averted should help to make the study summary more comparable to other studies higher and lower within Table 3.

Owusu-Edusei et al: Same comment as for Bristow, on cost versus cost per DALY, and the definition if ‘more effective’, and of DALY (lost or averted).

Terris-Prestholt et al.: The ‘No screening program ; ICER reported as not applicable’ is redundant, as this is the definition of the counterfactual/comparator, ICERs are expressed against this counterfactual/comparator.

Kuznik et al: ICER is for what? Per infection averted? Please write that out. And clarify, what does an ICER mean for the ‘No screening program’, what was the comparator or counterfactual scenario here?

• Rydzak: To interpret the ‘RST dominates’, please write out what is the comparator scenario for each scenario (No screening; RPR, and RST), for example:

No screening compared to …

• RPR compared to No screening

• RST compared to ?RPR? or ?No screening?

Schackman et al. Please clarify how Rural and Urban differed, e.g. in prevalence? Without that info, the differential results are meaningless to the reader.

Vickerman et al: From the 4 outcomes (cost/DALY) given, RPR has lower cost/DALY for Serum but RST has lower cost/DALY for Whole Blood. Therefore the conclusion ‘test using serum/whole blood: Cost-effective outcome = RST’ is not obvious from the results – the results suggest RST as more cost-effective for whole blood but not for serum. Please clarify; for example, add into the Results cell, the corresponding cost per DALY for Serum + Whole Blood combined.

Blandford: As for Kuznik et al., above, please clarify the respective comparators for each of the 3 testing options. Also clarify ‘Not published but states ‘Dominiated’.

To facilitates readers’ matching between Conclusion and Results, I’d rephrase the Conclusion as ‘Off-site RPR and confirmatory TPHA’. And remove ‘(averts more cases)’ as that’s obvious from, and just a repeat of, the Results.

Romoren et al: From the results, between the 4 options, the most cost effective looks to be Syndromic Management with azithromycin, at USD 21. In spite of that result, the conclusion states POC with azithromycin, despite it costing USD31 per cured case. Please reconciliate and clarify.

Larson et al.: Where does the 62% come from, what algorithm or policy was this the result of?

At reading the conclusion, I infer back that ‘RST’ must correspond to test and treat all, i.e. the third of the 3 options. Please write this out in the Results. And perhaps shorten the corresponding scenario names in column Intervention versus Comparator.

DISCUSSION

The Discussion mentions many wise and true things, but they are not too well structured, leaving the reader without a clear understanding of the overall current evidence or its policy implications.

We would recommend structuring the discussion of limitations between:

(1) Limitations that affect the core outcomes of studies analyzed (in Table 3), i..e. that bias those comparisons, notably:

• Societal perspective, which may cause under-stating the value POC compared to laboratory-based policies, if the savings from POC are for a large part on the patient/client side.

• Types of cost: if the variability across studies is important, then please show it – adding a column with ‘Cost components included’ to Table 3. (To make space in Table 3, you might remove the column Intervention versus Comparator’ which gets repeated in the next-right column in any case.?)

(2) More general limitations or qualifications such as:

• Most studies costed only test kits, treatment and staff, but not QAQC or procurement: if this applies across all scenarios compared within a study, that comparison is not biased.

• Affordability and equity: True but somewhat beyond the scope of this review?

• Short time horizons: A true and important limitation – but not one that biases the comparison across scenarios within a study: Per treatment or maternal infections cured, the longer-term effects should be similar regardless of through which screening algorithm that treatment or cure was produced, not?

CONCLUSION:

The start and end of the conclusion as written are more an Introduction. I suggest to narrow and focus the conclusion on the part in the middle, from ‘Our review indicates…’ to ‘… other screening practices’.

Regarding the lack of geographical representation of the studies:

Why is that a problem? Is there reason to think that the alternative screening algorithms would compare differently in different settings (e.g. in Latin America or Asia-Pacific)? Why, what setting-specific determinant (e.g. STI prevalence??) would be the driver of varying cost-effectiveness rankings? If you believe this is truly important, then make it more concrete, beyond a description of geographic representation, in the 3rd paragraph of the Discussion.

Reviewer #2: Please see attached comments.

This manuscript provides the results of a systematic review of economic evaluations of antenatal point-of-care (PoC) testing and treatment for sexually transmitted infections (STIs) in low- and middle-income countries (LMICs). The authors identified 16 economic evaluations in the peer reviewed literature that met the inclusion criteria. Because of heterogeneity, authors were only able to provide a summary of the literature, and concluded that PoC testing and treatment during pregnancy is likely to be cost-effective compared to no testing, syndromic management, and lab-based testing. This is an important topic likely to be of interest to the readers of PLOS ONE. Most countries continue to provide sub-optimal antenatal STI care through the syndromic approach, missing infections and overtreating pregnant women. A major barrier to expanding etiological testing is cost and the uncertainty around the cost-effectiveness of testing. I suggest accepting this manuscript with some corrections and clarifications outlined below.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Adriane Wynn

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: PLOSONE_CEA POC testing_27Oct2020.docx

Click here for additional data file.

22 Jan 2021

Responses to Reviewers

Reviewer #1 comments:

Failing the option to do a formal statistical meta-analysis (due to heterogeneity in outcome measures and …), the crux of the findings are in the descriptive Table 3. In this table, however, between Results and outcome/conclusion the relation is not always clear, nor are the comparative outcome definitions under Results always clear. Please see Specific comments below.

RESPONSE: Thank you for your comment. The results and outcomes presented in Table 3 are results published by each study included in the review. Table 3 in the manuscript has been revised to clarify the relationship between the results and outcome/conclusion. The table now have 5 columns that present cost components, health outcomes (where applicable), efficiency measures, the drivers of cost effectiveness, and key findings of each economic evaluation. Each column has been systematically completed to improve coherence and comparability across the studies included in this review (see pages 9-17, line #190).

The discussion dwells on methods of economic evaluations and to what extend the existing, reviewed studies met or not the optimum standards. Apart from this partly academic matter, what is the more practical conclusion from the presented evidence? Concretely, if a health minister (in country X) considers changing its policy from syndromic treatment to point-of-care testing (for, say, syphilis), what could s/he learn from this review? Should a new (local?) study be carried out, and if so what would be the preferred methodology and design? Or is the evidence in favor of POC testing already clear and uncontested, in all or some (which=?) settings?

RESPONSE: Thank you for your comment, we have edited the discussion to explain a practical recommendation for implementation (see page 27-28, lines #354-369).

Specific Comments:

RESULTS: Table 3:

1) Shelley et al: What is meant with ‘Quality control measures valuable to scale-up of RST’? How does the ‘quality control’ relate to the Results? Please expand the Results such that this becomes clear. Also– if space allows – comment on how rural and urban settings differed, e.g. which setting had higher syphilis prevalence, or what was the driver of the comparative cost and screening and treatment volumes?

RESPONSE: The term ‘quality control’ is listed in column 8 of Table 3 and is now explained in the footnotes of Table 3 (denoted by *).

2) Bristow et al: What was the criterion for ‘best option’? I suppose, cost per DALY averted? If so, please add ‘cost per DALY averted’ to the results (possibly replacing cost and DALY). Or if you keep the outcome DALY: is this DALYs lost or DALYs averted?

Adding an outcome per infection or DALY averted should help to make the study summary more comparable to other studies higher and lower within Table 3.

RESPONSE: This has now been made clearer through the inclusion of columns 8 and 9 of Table 3.

3) Owusu-Edusei et al: Same comment as for Bristow, on cost versus cost per DALY, and the definition if ‘more effective’, and of DALY (lost or averted).

RESPONSE: This has now been made clearer through the inclusion of columns 8 and 9 of Table 3.

4) Terris-Prestholt et al.: The ‘No screening program ; ICER reported as not applicable’ is redundant, as this is the definition of the counterfactual/comparator, ICERs are ex pressed against this counterfactual/comparator.

RESPONSE: The redundant statement has now been removed.

5) Kuznik et al: ICER is for what? Per infection averted? Please write that out. And clarify, what does an ICER mean for the ‘No screening program’, what was the comparator or counterfactual scenario here?

RESPONSE: This has now been clarified in Table 3 with a new column called ‘efficiency measures’ (column 9, row 10).

6) Rydzak: To interpret the ‘RST dominates’, please write out what is the comparator scenario for each scenario (No screening; RPR, and RST), for example: No screening compared to … RPR compared to No screening RST compared to ?RPR? or ?No screening?

RESPONSE: Table 3 has been revised and comparators for all full economic evaluations are detailed in column 5 of Table 3, while efficiency measures are now illustrated in column 10 of Table 3.

7) Schackman et al. Please clarify how Rural and Urban differed, e.g. in prevalence? Without that info, the differential results are meaningless to the reader.

RESPONSE: This has now been clarified in column 11, row 12 of Table 3.

8) Vickerman et al: From the 4 outcomes (cost/DALY) given, RPR has lower cost/DALY for Serum but RST has lower cost/DALY for Whole Blood. Therefore, the conclusion ‘test using serum/whole blood: Cost-effective outcome = RST’ is not obvious from the results – the results suggest RST as more cost-effective for whole blood but not for serum. Please clarify; for example, add into the Results cell, the corresponding cost per DALY for Serum + Whole Blood combined.

RESPONSE: The key result for this study has now been clarified in column 11, row 13 of Table 3.

9) Blandford: As for Kuznik et al., above, please clarify the respective comparators for each of the 3 testing options. Also clarify ‘Not published but states ‘Dominiated’.

To facilitates readers’ matching between Conclusion and Results, I’d rephrase the Conclusion as ‘Off-site RPR and confirmatory TPHA’. And remove ‘(averts more cases)’ as that’s obvious from, and just a repeat of, the Results.

RESPONSE: Comparators for this study have now been added in column 5 and key study findings have been amended as requested and included in column 11.

10) Romoren et al: From the results, between the 4 options, the most cost effective looks to be Syndromic Management with azithromycin, at USD 21. In spite of that result, the conclusion states POC with azithromycin, despite it costing USD31 per cured case. Please reconciliate and clarify.

RESPONSE: This has now been clarified in column 11, row 17of Table 3. The results and the conclusion are now consistent.

11) Larson et al.: Where does the 62% come from, what algorithm or policy was this the result of?

At reading the conclusion, I infer back that ‘RST’ must correspond to test and treat all, i.e. the third of the 3 options. Please write this out in the Results. And perhaps shorten the corresponding scenario names in column Intervention versus Comparator.

RESPONSE: The source of the 62% has now been added Footnote (denoted **) of Table 3). The link between the results and comparators have also been made clearer. The scenario names have also been made clearer. We have listed an explanation in the footnotes of Table 3 (denoted with **).

Discussion:

The Discussion mentions many wise and true things, but they are not too well structured, leaving the reader without a clear understanding of the overall current evidence or its policy implications.

We would recommend structuring the discussion of limitations between:

1) Limitations that affect the core outcomes of studies analyzed (in Table 3), i..e. that bias those comparisons, notably:

• Societal perspective, which may cause under-stating the value POC compared to laboratory-based policies, if the savings from POC are for a large part on the patient/client side.

• Types of cost: if the variability across studies is important, then please show it – adding a column with ‘Cost components included’ to Table 3. (To make space in Table 3, you might remove the column Intervention versus Comparator’ which gets repeated in the next-right column in any case.?)

2) More general limitations or qualifications such as:

• Most studies costed only test kits, treatment and staff, but not QAQC or procurement: if this applies across all scenarios compared within a study, that comparison is not biased.

• Affordability and equity: True but somewhat beyond the scope of this review?

• Short time horizons: A true and important limitation – but not one that biases the comparison across scenarios within a study: Per treatment or maternal infections cured, the longer-term effects should be similar regardless of through which screening algorithm that treatment or cure was produced, not?

RESPONSE: Thank you for your comment, the limitations have now been grouped as recommended (see page 28, lines #370-383).

Conclusion:

The start and end of the conclusion as written are more an Introduction. I suggest to narrow and focus the conclusion on the part in the middle, from ‘Our review indicates…’ to ‘… other screening practices’.

RESPONSE: Thank you for your comment, we have made changes to the Conclusion as recommended by the reviewer.

Regarding the lack of geographical representation of the studies:

Why is that a problem? Is there reason to think that the alternative screening algorithms would compare differently in different settings (e.g. in Latin America or Asia-Pacific)? Why, what setting-specific determinant (e.g. STI prevalence??) would be the driver of varying cost-effectiveness rankings? If you believe this is truly important, then make it more concrete, beyond a description of geographic representation, in the 3rd paragraph of the Discussion.

RESPONSE: A sentence has been added to the discussion emphasising the importance of conducting economic evaluations in a range of settings (see page 25, lines #292-295).

Reviewer #2:

Presentation, tables and figures

• Readers should be able to understand the study objective in each of the figure and table titles. They should stand alone.

RESPONSE: Thank you for your suggestion, we have renamed all the tables and the figures in this manuscript to reflect the study objective in each.

References

• Reference 8 is old. There are recent papers on the global burden of curable STIs

• Your only reference for adverse outcomes is for syphilis. Add in evidence for other STIs too.

• I think reference 18 should actually be 19.

RESPONSE: Thank you for your feedback; we have included recent references for chlamydia, gonorrhoea, trichomonas, and bacterial infection: REFS #1-REF#9 (see page 3, line #59). We have replaced reference #8 with a newer study: REF#14 (see page 3, line #62). We have also replaced reference #18 with two newer references: REF#26 and REF#27 (see page 3, line #75).

Abstract, Introduction, Methods, Results and Discussion

Abstract

Line 24: “tests” and “facilitates” don’t agree (should be testing and facilitates or tests and facilitate).

RESPONSE: This sentence has now been edited (see page 1, line #24).

Line 26: You mention “efficiency” often, but do you really mean effectiveness? What are your efficiency measures?

RESPONSE: Efficiency measures are now listed for all studies in column 10 of Table 3.

Line 37: should be “assessed”.

RESPONSE: This correction has been made (see page 1, line #38).

Introduction

Paragraph 2 covers many ideas including: effectiveness of early detection and treatment of HIV and syphilis; interventions for managing STIs; syndromic management for CT/NG/TV; and lab-based diagnosis out of reach. I suggest breaking this paragraph up to have one paragraph on HIV & syphilis testing guidelines/protocols (e.g. types of testing conducted), and effectiveness; and another paragraph dedicated to the infections managed by syndromic management, problems with syndromic management and barriers to testing.

RESPONSE: Thank you for your suggestion, we have amended the manuscript to reflect this suggestion (see page 3-4, lines #64-70 and lines #71-78).

Paragraph 3 could benefit from the more description of point-of-care testing, such as desirable attributes (accuracy, minimal training, rapid, low cost, portability), etc.

RESPONSE: Thank you for your suggestion. Desirable attributes for both serological and molecular point-of-care tests are described on page4 lines #83-85.

Minor

Line 60: When referring to the bacteria/protozoa for the first time, it would be good to spell out (e.g. Chlamydia trachomatis)

RESPONSE: Indeed, when referring to the causative organism (bacteria/protozoa) it is convention to list the name of the bacteria in full. However, for this manuscript, we are referring to, and identifying, the infection/disease (chlamydia) rather than the bacteria (chlamydia trachomatis), therefore we have listed the infection in the manuscript.

Line 85: All countries need to prioritize investment, but LMIC may have tighter budget constraints.

RESPONSE: This edit has been made on page 5 (see lines #94- #95).

Line 89: Is there a citation for the new landscape that is emerging? What do you mean by this?

RESPONSE: “…new landscape emerging” refers to the recently developed and effective technology for testing and treatment at point-of-care. This has been clarified on page 5 (line 98) by the inclusion of supporting references for Chlamydia, Gonorrhoea, and HPV (REFS #37, #41, #44).

Lines 100-105: Are you not also comparing and contrasting effectiveness measures?

RESPONSE: Thank you for your feedback, the purpose of this review is to systematically review the evidence on economic evaluations of point-of-care testing to detect STIs in pregnancy. We are comparing and contrasting the primary outcome of these studies, which include costs and cost-effectiveness, not effectiveness measures per se and clarified this (see page 6, lines #111-112).

Methods

With respect to the literature review, I have several questions:

1) “Partial/ full economic evaluation” should be defined in the methods.

RESPONSE: Brief definitions of partial and full economic evaluations have been included in the methods section (see page 7, lines #139-144).

2) What software was used for the screening (e.g. Covidence)?

RESPONSE:

We have revised the manuscript to include the software package (Microsoft Excel (version 365)) used for screening records, titles and abstracts (see page 7 line #135).

3) I have found that a lot of economic analysis has been conducted in the grey literature (e.g. by USAID/Health Policy Project, World Bank), why did you decide to exclude grey lit?

RESPONSE:

We excluded grey literature as it is may not always follow recommended guidelines for conducting economic evaluations and/or is not often peer reviewed. By excluding grey literature, we appreciate that we may have excluded some relevant evidence and acknowledge this as a limitation (see page 28, lines #379-381).

4) Does your search account for different spellings of diseases? (e.g. gonorrhea/ gonorrhoea/gonococcal)? What about prenatal (in addition to antenatal)?

RESPONSE:

The search terms were finalised in consultations with a medical librarian. The terms and syntax used (Table 2) account for different spelling. For example, “GONORRHEA/” takes the American spelling of Gonorrhoea (i.e gonorrhea) into consideration and different derivations of “Gonorrhoea”. Prenatal is the parent term for antenatal, however upon closer inspection during the development of the search strategy, the exclusion of prenatal did not influence the outcome of the search.

5) Was the third reviewer/resolver more senior or have more expertise in the subject area?

RESPONSE:

Yes, the third reviewer is more senior, and is a more experienced economist specialising in economic evaluations in low- and middle- income countries (LMIC) and has published 3 other systematic reviews in high impact journals including Lancet Infectious Diseases and Health Economics.

With respect to the data abstraction process, I have several questions:

6) Did you abstract data associated with your quality appraisals (e.g. perspective, exchange rate)? Do you have a table that could be published in the supplemental documents?

RESPONSE: We have not abstracted the data for the quality appraisals. However, in response to the comments from reviewer 1 we amended Table 3, which now presents relevant details extracted from the studies included in this review and somewhat overlaps with the checklist items in the quality appraisal. We have also amended the manuscript to reflect this (see page 7, lines #152-156).

7) What software was used for abstraction (e.g. Qualtrics/excel)?

RESPONSE: Microsoft Excel (version 365) was used for abstraction. The manuscript has been revised to make this clearer (see page 7, line #152).

Results

With respect to Table 3, I have several questions/comments:

1) I think it’s important to include the type of cost assessed in each study. As you note later, many studies did not include capital costs. Capital costs associated with POC testing often make this strategy unaffordable (compared to lab testing) because an assay must be placed at each clinical site.

RESPONSE: Types of costs (Cost components) are now detailed in column #7 of Table 3.

2) It would be helpful to also add a column for effectiveness outcomes. If there’s no room, at least define what health outcomes are reflected in the DALYs (e.g. PID, preterm birth, conjunctivitis, etc.)

RESPONSE: In Table 3, the column entitled ‘efficiency measures’ incudes the primary measure of effectiveness used in the economic evaluation (see column #9).

With respect to the results narrative, I have several questions/comments:

1) Please include more discussion on the details and quality of the outcomes included in the CEA/CUAs and separate out by syphilis/HIV and CT. Did authors use results from RCTs or meta-analyses? Were the effectiveness measures derived from an appropriate population – or are they all upper income populations? Was the intervention appropriate/similar? Do authors assume treated and cured infections have the same probability of an adverse outcome as never being infected?

RESPONSE: we have amended the text to highlight the number of studies conducted using trial data and those conducted using existing literature (See page 18, lines #200-201).

2) In the quality assessment, do you consider how costs were collected (e.g. micro-costing / bottom up costing)?

RESPONSE: Thank you for your comment, we did not include this in our analysis. However, we know that only four studies (all of the partial economic evaluations) clearly stated that they utilised an ingredients-based approach, three of which also indicated utilising a step-down approach to costing. All of the full economic evaluations (n=12) did not specify a method for how costs were collected, we could only establish that these studies took an ingredients based approach, however, could not establish if another approach was also utilised.

Line 226: Does “treatment coverage” include loss to follow-up? I would think loss to follow-up would be important when comparing PoC with lab-based testing.

RESPONSE: The key drivers presented in this sub-section are those that are identified by the studies included in this review, we have only collated them to compare/contrast them. Treatment coverage refers to the number of antenatal clinic attendees treated for an STI. However, we cannot extrapolate whether treatment coverage includes loss to follow-up (unless specifically stated in each study).

Discussion

1) The first paragraph isn’t really about your study. I would include this content in the introduction and use the first paragraph to summarize your most important findings.

RESPONSE: Thank you for your suggestion, we have amended the first paragraph of the discussion as suggested (See page 25, lines #267- 282)

2) Discuss drivers of costs/cost-effectiveness

RESPONSE: Thank you for your comment, we have revised the manuscript to include a paragraph on the drivers of cost and cost-effectiveness (See page 26, lines #318- 325).

3) Discuss the outcomes that were included. The effectiveness of CT/NG/TV infection testing and treatment in terms of preventing adverse health outcomes (besides PID) is very limited. This may be another important reason why countries are hesitant to introduce expensive testing.

RESPONSE: Thank you for pointing this out, indeed there is limited evidence demonstrating the effectiveness of testing and treating CT/TV/NG in pregnancy to reduce adverse pregnancy and birth outcomes such as miscarriage, still births, neonatal deaths etc. However, as detection of CT/TV/NG has improved over the last decade, there has been a growing number of random control trials investigating the effectiveness of detecting CT/TV/NG in pregnancy to reduce adverse pregnancy and birth outcomes, enabling the expansion of evidence and providing a definitive conclusion whether testing and treating STIs in pregnancy reduces adverse pregnancy and birth outcomes (See page 26, lines #296-303).

4) In the conclusion you mention a gap in terms of measuring the “life time costs and effects of testing and treatment on mothers and their babies,” but you don’t mention this in the discussion. What’s available in terms of lifetime costs of adverse outcomes? Do we need more research here?

RESPONSE: In the discussion we mention that most economic evaluations are based on a relatively short time-horizon and overlook these long-term costs and effects (see page 26 lines #313- #317).

5) Are the results similar to economic analyses in high income countries (e.g. Ong and Ditkowsky in Australia and the US)?

RESPONSE: Cost-effectiveness studies for STIs, including that of Ong and Ditkowsky, conducted in high income countries tend to not focus on testing at point-of-care unless conducted in a remote community. Thus making it difficult to compare results of studies based in high-income countries.

Conclusion

1) I disagree that your review “highlights the role that economic evidence plays in moving the global health agenda beyond the utility of antenatal point-of-care testing and treatment for STIs towards incorporating resource allocation, which is a major determinant of implementation and scale-up.” I think your review identifies and summarizes the literature, but more research is needed to move the agenda forward.

RESPONSE: We have toned down this statement to these concerns (see pages 28-29).

2) Again, mention cost/cost-effectiveness drivers.

RESPONSE: We have revised the manuscript to include the ‘drivers of cost and cost effectiveness’ in the conclusion (see page 29, lines #388-391).

3) Mention life time costs in the discussion if you are going to include in the conclusion

RESPONSE: Please see previous response. These costs are now mentioned in the discussion (see page 26 lines #313- #317).

Submitted filename: SaweriOPM_Manuscript_Responses to editor.docx

Click here for additional data file.

26 Feb 2021

PONE-D-20-31634R1

Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections in pregnancy in low- and middle-income countries: A systematic review

PLOS ONE

Dear Dr. Saweri,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Apr 12 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Remco PH Peters, MD, PhD, DLSHTM

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: GENERAL/MAJOR COMMENTS

The authors have made revisions in response to each comment, which has strengthened the Methods, Results and Discussion – but I wonder if some of the changes are adequate, and some may actually introduce confusion or misinterpretation.

Table 3 is more readable now. Still, this being the key results, it merits further improvement – and an attempt to restore some meaningful content dropped since the original submission.

SPECIFIC COMMENTS

RESULTS, Table 3:

Drivers of cost and cost-effectiveness: This is a useful addition, in principle – but does not really yet add insight in practice, because:

• The drivers mentioned are much broader than those that determine the ranking between study arms, i.e. RST versus comparators, shown in the left-hand column.

• The drivers mentioned pertain mostly to the absolute cost per outcome (e.g. $$ per case averted or per DALY) but that column got dropped since the original submission.

• The cost drivers mentioned are named too broad and vague, for example: ‘Cost’ does this mean procurement cost i.e., test price?

I would suggest to be more concrete and specific in listing cost drivers -- focusing on the specific factors that explain the comparison between RST and comparators – and explain the direction of effect – which study arm/comparator had higher or lower value for each cost driver. Or optionally, slit the column into 2:

• Drivers of ‘absolute’ cost per outcome;

• Drivers of the difference in cost-per-outcome between RST and comparators.

You could make space by removing the column Infection studied, replacing that with one start or footnote indicating the one study where this was not syphilis but chlamydia. And some other of the left-hand columns might be merged.

For example, in Shelley et al., which of the 5 factors explained that RST rollout was cheaper than pilot? Which of the two used which type of blood collection? Which of the two arms had longer or shorter lifetime, and how did that influence the cost?

Sweeney et al.

• If the efficiency measure was total cost per health facility, did each health facility in both study arms test and treat equally many persons?

• What is meant with Time taken to test, is that the personnel time per test? Was that time longer for RST than for RPR?

Levin: In the column cost drivers, could you state the factors that explained the contrasting cost ranking between Mozambique and Bolivia? That does not seem to require the study having included a sensitivity analysis (?)

Owusu-Edusei et al: Was cost associated with adverse pregnancy outcome really a cost driver? Isn’t that relevant only if the incidence of such outcomes differed between the study arms i.e. comparators – which is unlikely: any undetected and untreated infection will on average incur the same adverse outcomes, regardless of the study arm /comparator that failed it? This is an example of my general comment above, that the many items mentioned in this column appear mostly meaningless. I suggest you limit the contents of this column to factors that influence the comparison between study arms / comparators. Any factors that influence the overall cost are not relevant, since your table (or any other Result in the paper) do not deal with the absolute overall cost or cost-effectiveness measures – they all just center on relative comparison of RST against comparators.

& Same comment for Rydzak.

Similarly, for Blandford, did the relative distribution of active/latent syphilis differ between the study arms i.e., comparators? If not, it is not relevant in this table that focuses on comparing test policies.

Kuznik 2015 & 2013, and Vickerman: Rephrase ‘All comparators’ as ‘Comparator’ since these studies each had only 1 comparator.

Terris-Prestholt et al.: As for Levin above, please focus and limit the cost drivers mentioned on those that explain the difference in ranking (RST above or below comparators) between Peru, and Tanzania+Zambia..

Kuznik 2013: By definition, the Discount rate will always influence the absolute cost measure, but it will do so equally for all study arms and comparators, and all studies not just Kuznik – so this is a meaningless redundant item to mention here.

Rydzak, Shackman, & Romoren:: Rephrase ‘All comparators’ as ‘Both comparator’ since these studies each had 2 comparators.

Rydzak: 1000 women is women tested, or treated?

Romoren et al: Under point 3. Remove the redundant words ‘Point-of-care testing and Azithromycin treatment vs.’

In the right-most cell of this row, clarify that PoC is RST, and replace ‘and’ by ‘with’ to clarify that this (combo of test type + treatment type) is 1 study arm not 2.

Larson: The added footnote (answering my earlier comment: Where does the 62% come from, what algorithm or policy was this the result of?) is helpful. But I’m still left with the question, why only 10% of positive cases were treated, and which 10% was this, was this random or by any characteristic or policy criteria?

& Cosmetic/formatting: Make Table 3 shorter and more readable by adapting the column widths according to their average content.

DISCUSSION

Page 36: Please explain the ‘Largest vs smallest percentage difference, 170% and 19% - between what? For which outcome is this, and which 4 of the studies listed in Table 3 provided the extremes?

Page 37: ‘In particular, the fundamentals of an economy and disease epidemiology..’: This reads as poor-language generality, I don’t understand what specifically the author want to say. Please rephrase into something more concrete, or remove.

‘This has, in turn, led to a comparatively new and emerging field…’: I think the discussion could do without this perhaps exaggerated statement.

‘Economic evaluations… should ideally… incorporate the long-term costs and benefits…’: I am not convinced this is necessarily neede -- unless the longer-term costs and/or benefits differ between the intervention arms being compared? Which does not seem to be the case for POC STI tests. Please also see my related specific comments on various rows/study summaries in Table 3.

Page 40: ‘… the incorporation of extended cost-effectiveness analyses and budget impact analyses into cost-effective analyses…: This reads like triple circular.

Additionally, most studies took a narrow view… short-run time horizon, although this does not influence the comparability between studies, it does highlight a short/coming of the studies included in this review. Consider to shorten and simplify – the time horizon was discussed already earlier in the Discussion.

‘…therefore grey literature…’: These last 10 last words are redundant in the sentence.

‘Finally, a further enhancement … equity’: This duplicates the mentioning of equity in line 403 above?

Reviewer #2: The manuscript is improved and my comments have been addressed. My suggested minor revision is related to the percentage difference mentioned in the results and discussion. Please mention the percentage difference (and how it was calculated) as an outcome of interest in the methods.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Eline Korenromp

Reviewer #2: Yes: Adriane Wynn

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Peer review SAWERI Econ Point-of-care tests STI pregnancy LMIC_PLOSONE_25Jan2021rev1_ek1.docx

Click here for additional data file.

12 Apr 2021

Editor-in-Chief

Plos one

Manuscript Title: Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections in pregnancy in low- and middle-income countries: A systematic review

Authors: Olga PM Saweri, Neha Batura, Rabiah al Adawiyah, Louise Causer, William Pomat, Andrew J Vallely and Virginia Wiseman

On behalf of my co-authors, I would like to thank the editor and reviewers for providing a detailed and extensive second review of our manuscript. We greatly appreciate the time and effort put into the feedback we have received. We have revised the manuscript and our responses to the reviewers’ feedback are provided below.

Best,

Olga Saweri

RESULTS, Table 3:

1. Drivers of cost and cost-effectiveness: This is a useful addition, in principle – but does not really yet add insight in practice, because:

• The drivers mentioned are much broader than those that determine the ranking between study arms, i.e. RST versus comparators, shown in the left-hand column.

• The drivers mentioned pertain mostly to the absolute cost per outcome (e.g. $$ per case averted or per DALY) but that column got dropped since the original submission.

• The cost drivers mentioned are named too broad and vague, for example: ‘Cost’ does this mean procurement cost i.e., test price?

RESPONSE:

Thank you for your feedback. The results presented in Tables 3 and 4 illustrate the results published in each included study and our interpretation of those results. The drivers of cost and cost-effectiveness, which were previously included in Table 3 are now presented in Table 4; these drivers are the results of sensitivity analyses conducted by each study included in this review (see Column D in Table 4 on pages 16-22). The revisions made to Column D in Table 4 (pages 16-22) reflect the results published by each study included in this review and are less ambiguous. In addition, we specifically list the drivers of cost and cost-effectiveness in the manuscript text and cross-reference the results presented in Table 4 (see page 24, lines 248-252).

2. I would suggest to be more concrete and specific in listing cost drivers -- focusing on the specific factors that explain the comparison between RST and comparators – and explain the direction of effect – which study arm/comparator had higher or lower value for each cost driver. Or optionally, slit the column into 2:

• Drivers of ‘absolute’ cost per outcome;

• Drivers of the difference in cost-per-outcome between RST and comparators.

RESPONSE:

Thank you for this suggestion. We have revised the manuscript to reflect this wherever possible. However, this was challenging. Few studies explicitly differentiated between drivers of absolute cost per outcome and those the explain the comparison between study arms or comparators. Most studies do not explicitly state that the results of the sensitivity analysis were drivers of cost-effectiveness, cost, effectiveness or between arms (see Table 4, column D, pages 18-24). Where the text is not explicit, we have assumed that the results of the sensitivity analysis are drivers of cost-effectiveness for full economic evaluations (n=12) or drivers of cost for partial economic evaluations (n=4).

3. You could make space by removing the column Infection studied, replacing that with one start or footnote indicating the one study where this was not syphilis but chlamydia. And some other of the left-hand columns might be merged.

RESPONSE:

We have amended Table 3 and included an additional table (Table 4) to improve the spaces in each column heading to improve readability (see Table 3 on pages 10-15 and Table 4 on pages 16-22).

4. For example, in Shelley et al., which of the 5 factors explained that RST rollout was cheaper than pilot? Which of the two used which type of blood collection? Which of the two arms had longer or shorter lifetime, and how did that influence the cost?

RESPONSE:

In this paper, the sensitivity analysis parameters were screening coverage, test kit price, supply wastage, project life, and type of blood collection. Blood collection here refers to the difference in collecting blood via a finger prick and via venipuncture. The former is utilized for RST while the latter for RPR. The results of the sensitivity analysis indicate that cost per person screened is most sensitive to changes in screening coverage, however changes in RST test kit and supply wastage also impact cost per person. These results are now reflected in table 4 column D (pp18-24).

In response to the query about the time horizon, the costs of both the pilot and the rollout were collected over the same amount of time, just two years apart. Further, the sensitivity analysis measured the effect of the project life, whereby costs in the rollout period were more sensitive to a change in the project life compared to the pilot period.

5. Sweeney et al.

• If the efficiency measure was total cost per health facility, did each health facility in both study arms test and treat equally many persons?

• What is meant with Time taken to test, is that the personnel time per test? Was that time longer for RST than for RPR?

RESPONSE

No, in this study, health facilities did not test and treat equal numbers of women in the study arms. In the RPR arm, over a 9-month period between 2007 and 2008, 838 antenatal clinic (ANC) attendees were tested for syphilis at 6 of the 9 selected health facilities. This is 17.8% of ANC attendees. Further 27% of women tested positive for syphilis and about 66% of those who tested positive were treated. The results state that no confirmatory tests were completed, therefore the true prevalence is unknown.

In the RST arm, over a similar time frame (in 2009-2010) at the 9 selected health facilities, 9372 ANC attendees were tested (87% of total ANC attendees); 10% of antenatal clinic attendees tested positive, of which 92% were treated.

With respect to ‘the time taken to test’, there is a difference between the time taken to conduct an Rapid Plasma Reagin (RPR) test for syphilis compared to a point-of-care test, or rapid syphilis test (RST). An RST is ‘quicker’, the study does not specifically state the difference in time. The sensitivity analysis showed that both RST and RPR costs were sensitive to staff time.

6. Levin: In the column cost drivers, could you state the factors that explained the contrasting cost ranking between Mozambique and Bolivia? That does not seem to require the study having included a sensitivity analysis (?)

RESPONSE: It is not clear from Levin et al (2007), therefore we have not amended the manuscript. What is clear from Levin et al (2007) is that direct comparisons between total costs in each country would not lead to “meaningful” results as the ‘project periods and the numbers of women screened are not directly comparable’ (see page S50/S54).

With respect to the rank of costs, costs related to testing accounted for 50% to 84% of costs in both countries. Looking into the breakdown of costs, ‘tests and other supplies’ accounted for the greatest share of costs in both countries. In Bolivia, clinic staff salaries associated with testing accounted for the second largest share of costs followed by start-up costs. However, in Mozambique, lab supplies accounted for the second largest share of costs followed by clinic staff salaries.

7. Owusu-Edusei et al: Was cost associated with adverse pregnancy outcome really a cost driver? Isn’t that relevant only if the incidence of such outcomes differed between the study arms i.e. comparators – which is unlikely: any undetected and untreated infection will on average incur the same adverse outcomes, regardless of the study arm /comparator that failed it? This is an example of my general comment above, that the many items mentioned in this column appear mostly meaningless. I suggest you limit the contents of this column to factors that influence the comparison between study arms / comparators. Any factors that influence the overall cost are not relevant, since your table (or any other Result in the paper) do not deal with the absolute overall cost or cost-effectiveness measures – they all just center on relative comparison of RST against comparators.

RESPONSE:

Thank you for your comment. Yes, cost associated with adverse pregnancy outcome was a driver of cost, thus we have not made any amendments to the manuscript. Owusu-Edusei et al (2011) identifies drivers in the text, which we have included in this table. Table 2 (see page 1001/1003) illustrates that the incidence of adverse outcomes differs per study arm (total adverse pregnancy outcomes are highest when there is no screening program (n=39) and lowest when screening using an RST (n=2)).

Excerpt from Owusu-Edusei et al (2011) p1000/1003: “…we varied select variables and examined the total expected cost keeping all other variables constant. As expected, the relative test performance, test costs, and costs associated with adverse pregnancy outcomes were all influential.” The text continues to describe how all scenarios were cost-saving and the extent to which they were.

8. Same comment for Rydzak.

RESPONSE:

Thank you for your feedback, an excerpt from Rydzak and Goldie (2008) page 780/784 indicates which variables were drivers, including the ‘costs associated with congenital syphilis’, therefore we have not made this change to the manuscript. The excerpt: “variations in test performance, test costs, and costs associated with congenital syphilis had the greatest effect on results. The choice between the 2 single-visit strategies was influenced by their comparative test sensitivity, and the cost of the ICS [rapid syphilis test] test kit, labour and supplies. A screening strategy was considered the preferred strategy if it achieved greater benefits at lower cost than all other strategies.”

Further, table 3 (page 779/784) indicates the number of adverse pregnancy outcomes and the number of adverse outcomes averted under each comparator. The table demonstrates that adverse pregnancy outcomes differ per comparator.

9. Similarly, for Blandford, did the relative distribution of active/latent syphilis differ between the study arms i.e., comparators? If not, it is not relevant in this table that focuses on comparing test policies.

RESPONSE:

Thank you for your comment; Blandford et al (2007) illustrates on that ‘model effectiveness of antenatal screening approaches are most affected by test sensitivity’ (see page S64/S66). The article also states that prevalence of maternal syphilis and the distribution of active and latent syphilis indirectly impact cost-effectiveness through costs. Given this, we have not changed the drivers stated in Table 4 (see page 20).

10. Kuznik 2015 & 2013, and Vickerman: Rephrase ‘All comparators’ as ‘Comparator’ since these studies each had only 1 comparator.

RESPONSE:

Thank you for your suggestion. We have made relevant changes used consistent language across all studies. The key findings column has now been placed in Table 4 (see Column E, Table 4 on pages 16-22).

11. Terris-Prestholt et al.: As for Levin above, please focus and limit the cost drivers mentioned on those that explain the difference in ranking (RST above or below comparators) between Peru, and Tanzania+Zambia.

RESPONSE:

Thank you for your comment, Terris-Prestholt er al (2015) states on page S76/S80 that the ‘variability in the DALYs averted for this algorithm is due to the variability of RST reactivity rate (or positivity rate)’ in addition, the study also highlights that variability in total cost is driven primarily by ANC attendance, fixed clinic costs and screening coverage. These have all been included as drivers of cost and cost-effectiveness. We have not included ‘true prevalence’, as this did not affect the ranking (See Row 11, Column D in Table 4 on page 19)

12. Kuznik 2013: By definition, the Discount rate will always influence the absolute cost measure, but it will do so equally for all study arms and comparators, and all studies not just Kuznik – so this is a meaningless redundant item to mention here.

RESPONSE:

Thank you for your feedback, we have amended Table 4 to reflect this (see Column D, Table 4 on pages 16-22).

13. Rydzak, Shackman, & Romoren:: Rephrase ‘All comparators’ as ‘Both comparator’ since these studies each had 2 comparators.

RESPONSE:

Thank you for your suggestion. We have made relevant changes to Table 4 and used consistent language across all studies (see Column D, Table 4 on pages 16-22).

14. Rydzak: 1000 women is women tested, or treated?

RESPONSE:

Rydzak and Goldie et al (2008) states that their outcome measure is ‘discounted costs saved per 1000 women’, which includes both testing and treatment for syphilis.

15. Romoren et al: Under point 3. Remove the redundant words ‘Point-of-care testing and Azithromycin treatment vs.’ In the right-most cell of this row, clarify that PoC is RST, and replace ‘and’ by ‘with’ to clarify that this (combo of test type + treatment type) is 1 study arm not 2.

RESPONSE:

Thank you for your comment. We have amended the manuscript to reflect that ‘testing for chlamydia using a point-of-care test followed by treating those testing positives for chlamydia with azithromycin’ was the most cost-effective outcome (see Row 18, Column E, Table 3 on pages 14 and Row 18, Column D, Table 4 on page 21).

We would like to clarify that the point-of-care test here refers to a point-of-care test for chlamydia, and therefore is not a rapid syphilis test (RST). We have amended Table 3 and 4 to state ‘point-of-care’ test.

16. Larson: The added footnote (answering my earlier comment: Where does the 62% come from, what algorithm or policy was this the result of?) is helpful. But I’m still left with the question, why only 10% of positive cases were treated, and which 10% was this, was this random or by any characteristic or policy criteria?

RESPONSE:

We have edited the contents of table 3 to improve clarity (see Row 19, Column E of Table 3) and the footnote of Table 3 (see pages 14-15).

Larson et al (2014) built three scenarios using the results of an evaluation study exploring rollout of point-of-care testing and treatment for syphilis in antenatal clinics (ANC). The evaluation found that 62% of ANC attendees were tested for syphilis. All women testing positive should have been treated, but only 10% were. For clarity, the scenarios from this study are outlined below:

Scenario 1 (baseline): the results of the evaluation study found that 62% of attendees were tested for syphilis, while only 10% of those that tested positive were treated. Therefore scenario 1 emulated this: 62% of antenatal clinic attendees tested for syphilis and 10% of test positives were treated.

Scenario 2: explored the scenario of treating all antenatal clinic attendees testing positive yet keeping the number of antenatal clinic attendees tested for syphilis (62%) constant. Thus scenario 2 becomes testing 62% of antenatal clinic attendees and treating all antenatal clinic attendees that test positive for syphilis.

Scenario 3: which could be seen as the ideal scenario, explores the possibility of testing all antenatal clinic attendees and treating all of whom test positive for syphilis.

17. Cosmetic/formatting: Make Table 3 shorter and more readable by adapting the column widths according to their average content.

RESPONSE:

Thank you for your comment. We have edited and adjusted Table 3 to make it more readable (pages 10-15).

DISCUSSION

1. Page 36: Please explain the ‘Largest vs smallest percentage difference, 170% and 19% - between what? For which outcome is this, and which 4 of the studies listed in Table 3 provided the extremes?

RESPONSE:

The percentage differences have been calculated per study to express the measured differences between comparators of each study included in this review. We have included percentage differences in Column G of Table 4 (see pages 16-22) to illustrate the difference between the baseline figure and the most cost-effective comparator of each study included in this review. In addition to percentage differences the cost-effectiveness decision rule has been included in Column C of Table 4 to reflect how each study determined cost-effectiveness. Together these findings illustrate between-study variability (see lines 285-293, page 30). In addition, we have edited the methods section to highlight percentage differences briefly and their inclusion in the results (see lines 163-168, pages 7-8).

2. Page 37: ‘In particular, the fundamentals of an economy and disease epidemiology..’: This reads as poor-language generality, I don’t understand what specifically the author want to say. Please rephrase into something more concrete, or remove.

RESPONSE:

Thank you for your feedback, we have amended the manuscript (see lines 307-311, pages 30-31).

3. ‘This has, in turn, led to a comparatively new and emerging field…’: I think the discussion could do without this perhaps exaggerated statement.

RESPONSE:

Thank you for your comment, we have amended the manuscript (see lines 313-316, page 31).

4. ‘Economic evaluations… should ideally… incorporate the long-term costs and benefits…’: I am not convinced this is necessarily neede -- unless the longer-term costs and/or benefits differ between the intervention arms being compared? Which does not seem to be the case for POC STI tests. Please also see my related specific comments on various rows/study summaries in Table 3.

RESPONSE:

Thank you for your feedback, the effects of untreated sexually transmitted infections in pregnancy can extend after birth, both in women and in children. Estimates of long-term costs and benefits would help us understand whether the investment in such programs can have a positive impact on the health of mothers and babies later in life.

5. Page 40: ‘… the incorporation of extended cost-effectiveness analyses and budget impact analyses into cost-effective analyses…: This reads like triple circular.

RESPONSE:

Thank you for your suggestion, we have amended the manuscript to sound less ‘like [a] triple circular’ (see lines 379-382, page 33).

6. Additionally, most studies took a narrow view… short-run time horizon, although this does not influence the comparability between studies, it does highlight a short/coming of the studies included in this review. Consider to shorten and simplify – the time horizon was discussed already earlier in the Discussion.

RESPONSE:

We have shortened the sentence as suggested (see lines 394-395, page 33).

7. ‘…therefore grey literature…’: These last 10 last words are redundant in the sentence.

RESPONSE:

We have removed this sentence from the manuscript (see line 398, page 33).

8. ‘Finally, a further enhancement … equity’: This duplicates the mentioning of equity in line 403 above?

RESPONSE:

The previous round of reviewer comments suggested that the final paragraph of the discussion should be a ‘summary of the discussion’. Thus, the final sentence of this paragraph seems repetitious as it is a summary point of the discussion. We have amended the manuscript to sound less repetitious (see line 399-400, page 33).

Submitted filename: Saweri_ResponseToReviewers.docx

Click here for additional data file.

31 May 2021

Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections in pregnancy in low- and middle-income countries: A systematic review

PONE-D-20-31634R2

Dear Dr. Saweri,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Remco PH Peters, MD, PhD, DLSHTM

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

9 Jun 2021

PONE-D-20-31634R2

Economic evaluation of point-of-care testing and treatment for sexually transmitted and genital infections in pregnancy in low- and middle-income countries: A systematic review

Dear Dr. Saweri:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Prof Remco PH Peters

Academic Editor

PLOS ONE