2.2. HIV testing for a changing epidemic
People’s knowledge of their own HIV status and that of their partners is essential to the success of the HIV response (1). HIV testing services provide a pathway to HIV prevention, treatment, care and other support services. HIV testing services refer to the full range of services that should be provided with HIV testing, including counselling (pre-test information and post-test counselling); linkage to appropriate HIV prevention, treatment, care and other clinical services; and coordination with laboratory services to support quality assurance and the delivery of accurate results (1).
The overarching goals of HIV testing services are to:
identify people living with
HIV by providing high-quality testing services for individuals, couples and families;
effectively link individuals and their families to
HIV treatment, care and support and to HIV prevention services, based on their status; and
support the scale-up of high-impact interventions to reduce
HIV transmission and HIV-related morbidity and mortality.
All HIV testing services should continue to be provided in accordance with WHO’s essential five Cs: consent, confidentiality, counselling, correct test results and connection or linkage to prevention, care and treatment.
HIV testing services should always be voluntary. Protecting and maintaining client confidentiality is important, especially when offering testing as part of partner services and when the pre-test information session includes questionnaires screening for risks, symptoms or indicator conditions. An enabling environment that removes barriers such as stigma, discrimination and criminalization and age-of-consent issues is important for increasing access to and uptake of HIV testing services, especially among those at high ongoing risk and key populations ().
Addressing critical enablers for HIV testing services programmes.
2.4. HIV testing service delivery approaches
WHO recommends differentiated HIV testing service delivery approaches in a people-centred approach to address the needs of a variety of population groups, contexts and epidemic settings (Chapter 7 provides further details on people-centred care and differentiated service for HIV treatment). It is essential for countries to provide a strategic mix of WHO-recommended differentiated HIV testing service delivery approaches (Box 2.1). These include facility-based testing, community-based testing, HIV self-testing, partner services including provider-assisted referral and social network–based approaches. Trained lay providers and peers can now often provide an HIV diagnosis within a single visit in a health facility or community setting using rapid diagnostic tests.
Programmes need to routinely review and use their data to select and optimize the implementation of HIV testing service delivery approaches to reach populations and geographical settings with the largest proportion of people living with HIV who do not know their status.
See Chapter 5 of the Consolidated guidelines on HIV testing services, 2019 (1), which describes in detail the service delivery approaches for HIV testing services.
2.4.1. Facility-based HIV testing services
Recommendationsa
High-HIV-burden settings
HIV testing should be offered to all populations and in services (for example, services for sexually transmitted infections, hepatitis, TB, children under five, immunization, malnutrition, antenatal care and all services for key populations) as an efficient and effective way to identify people with HIV.
Low-HIV-burden settings
HIV testing should be offered for:
Source: Consolidated guidelines on HIV testing services, 2019 (1).
- a
These recommendations were developed in 2007 and revised in 2019.
Facility-based HIV testing services encompass testing in a health facility or laboratory setting. Facility-based HIV testing services can be provided at stand-alone HIV testing services sites (often referred to as voluntary counselling and testing sites) or routinely offered at clinical sites (often referred to as provider-initiated testing and counselling).
Routine facility-based HIV testing services can be offered at a range of public and private health facilities. This aims to increase the coverage of HIV testing services, provide earlier diagnosis, normalize HIV testing, remove the need for personal motivation to seek HIV testing services and reduce missed opportunities for HIV testing services. HIV testing services have been successfully and effectively integrated in some clinical settings such as antenatal care and TB services. Significant opportunities exist for integrating HIV testing services into many clinical services. The decision where and in which facilities to routinely offer HIV testing services should be guided by the local epidemiology and HIV testing services coverage gaps.
2.4.2. Facility-based HIV testing for infants and children
Recommendations (2016)
In settings with a high burden of HIV infection, infants and children with unknown HIV status who are admitted for inpatient care or attending malnutrition clinics should be routinely tested for HIV
(strong recommendation, low-certainty evidence).
In settings with a high burden of HIV infection, infants and children with unknown HIV status should be offered HIV testing in outpatient or immunization clinics
(conditional recommendation, low-certainty evidence).
Good practice statement (2016)
In all settings, the biological children of a parent living with HIV (or who may have died from HIV) should be routinely offered HIV testing services and, if found to be either infected or at high risk of infection through breastfeeding, should be linked to services for treatment or prevention and offered a broader package of voluntary provider-assisted referral.
Source: Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection: recommendations for a public health approach – second edition (3).
Background
Access to infant diagnosis is mostly limited to infants born to mothers enrolled and retained in programmes to prevent the mother-to-child transmission of HIV. These women generally have very low vertical transmission rates, so most infants diagnosed will test negative. By contrast, mothers who receive inadequate or no interventions to prevent mother-to-child transmission will have much higher transmission rates and yet their infants are unlikely to be tested and identified as HIV-infected. This contributes to the large gap between coverage of and need for ART among children and persistently high HIV-related mortality among children. Previous WHO guidelines have emphasized the importance of case-finding and testing outside programmes to prevent mother-to-child transmission to identify children who did not benefit from such interventions, but for a variety of reasons facility-based1 testing and counselling for children has not been optimally implemented (4).
Rationale and supporting evidence
A systematic review was undertaken to compare the standard approach of testing infants and children in programmes to prevent mother-to-child transmission with testing in a range of clinical settings outside such programmes (5). The primary outcomes examined were yield of testing in terms of the HIV seropositivity rate and acceptability by caregivers. The objective was to provide additional evidence to reinforce and contextualize guidance on testing children for HIV.
No studies directly compared the yield of testing within programmes to prevent mother-to-child transmission with testing outside of these programmes. However, 24 studies were identified that reported on the yield of facility-based testing and counselling for children younger than five years in a variety of settings, including inpatient, outpatient, nutritional rehabilitation centres and immunization clinics. Twenty-two of the 24 studies were conducted in sub-Saharan Africa, and 18 of 22 were in settings with high HIV prevalence (>5%). There was one study from Asia and one from Oceania. One study provided data for both outpatient and immunization clinics (6), but the rest assessed yield in only one setting (16 inpatient studies, 2 outpatient studies, 3 nutrition centres and 2 immunization clinics).
One third of the studies were conducted during or after 2013, when WHO issued guidance on using lifelong ART for all pregnant and breastfeeding women. The yield of positive test results was very high in inpatient settings for children (22.5%, 95% confidence interval (CI) 16.0–29.0%) and high in nutrition centres (14.2%, 95% CI 2.3–26.1%). The rates were lower in immunization clinics (3.3%, 95% CI 0–6.9%) and outpatient settings (2.7%, 95% CI 0.3–5.2%). Positivity rates varied significantly by geographical region. Across 18 studies in eastern and southern Africa, the prevalence was 22.6% (95% CI 17.2–28.0%), whereas in four studies conducted in western and central Africa (where the population HIV prevalence is lower), the prevalence was less than half, at 9.7% (95% CI 2.2–17.2%). Asia and Oceania had too few studies to perform subgroup analysis.
Eight studies were identified that used a universal testing approach in inpatient settings for children; these were compared with eight studies that assessed symptoms to determine which children to test. Although symptom-based testing approaches showed a slightly higher yield of positive results (23.1%, 95% CI 14.9–31.3% versus 21.9%, 95% CI 12.4–31.4%) this difference was not statistically significant.
Data from countries with lower prevalence were limited, but one study from western Africa reported a positivity rate of 25% in nutritional centres (7), suggesting that if the coverage of maternal ARV drugs used to prevent mother-to-child transmission is poor, the yield of facility-based testing and counselling of children in selected settings may be high, even when the overall HIV prevalence in the country is low. Unpublished data from Ethiopia suggest that prevalence rates among children – even in inpatient settings – have declined significantly over the past 10 years, but testing of children of index clients has remained consistently high at greater than 5% (T. Tsague, UNICEF, personal communication, June 2015).
There are no reports on the cost–effectiveness of testing children for HIV in specific health-care settings for children. Integrating HIV services (including HIV testing) into other health programmes has been found to be generally cost effective, but the cost–effectiveness of facility-based testing and counselling for children (especially in immunization programmes and outpatient clinics, where the yield of positive results is likely to be lower) depends on factors such as maternal prevalence and the coverage of services to prevent mother-to-child transmission (8). In settings with high maternal HIV prevalence and low coverage of services to prevent mother-to-child transmission, testing infants and children is likely to be highly cost effective as a strategy to prevent HIV-associated mortality. Moreover, facility-based testing and counselling during infancy may identify infants who are exposed to HIV with detectable antibodies but are not yet infected, providing an opportunity to prevent transmission during breastfeeding.
Of the 24 studies assessed in the systematic review, 13 reported on the rates of caregiver acceptance of HIV testing children. Acceptance rates varied by the location of testing and by region, but the overall mean acceptance rate was high, at 92% (range 73–100%). Most caregivers were motivated to accept testing by a desire to know the child’s HIV status (78%). A small minority (5%) reported being influenced by other parents whose children had been tested. In a study in South Africa to inform about the acceptability and feasibility of routine HIV testing in immunization clinics, just over half of all eligible children and caregivers accepted HIV testing (9). The Guideline Development Group made a strong recommendation to provide routine HIV testing for infants and children admitted for inpatient care or attending malnutrition clinics, citing existing vast programme experience and testing yield along with high levels of feasibility and acceptability, despite the low-certainty evidence.
Implementation considerations
Although the guidance for active case finding and facility-based testing and counselling among children has been in place since 2007, the uptake of this recommendation has been poor. Issues around the legal age of consent and provider discomfort with disclosure have contributed to this lack of uptake, especially for adolescents and older children. A recent study in six primary clinics in Zimbabwe identified several other factors, including a perceived lack of importance attached to testing older children and a sense that testing was not warranted if children were asymptomatic (10). Lack of time and reagents and discomfort with approaching male caregivers were also noted as reasons for not testing. A WHO survey of health workers, policy-makers and programme managers from 17 countries found that almost half the respondents thought that testing children in immunization clinics would be either easy or very easy to do, suggesting that this policy is highly feasible to implement. Experience from countries that have been trying to roll out facility-based testing and counselling of children highlights the importance of thorough linkage to care and services for children who are exposed or infected. Linkage to care may be easier for children in inpatient settings than for those in busy outpatient clinics. The negative impact of HIV testing on the uptake of other essential childhood interventions, such as immunization, has been cited as an argument against integrating testing in immunization clinics (11). A study in the United Republic of Tanzania showed that, although integration of HIV testing resulted in an increase in immunization rates in urban centres, there was a decrease in rural facilities, possibly reflecting higher levels of stigma in rural communities (12).
2.4.3. Community-based HIV testing services
Recommendations (2019)
High-HIV-burden settings
Community-based HIV testing services are recommended, with linkage to prevention, treatment and care services, in addition to routine facility-based testing, for all populations, particularly key populations
(strong recommendation, low-certainty evidence).
Low-HIV-burden settings
Community-based HIV testing services are recommended for key populations, with linkage to prevention, treatment and care services, in addition to routine facility-based testing
(strong recommendation, low-certainty evidence).
Source: Consolidated guidelines on HIV testing services, 2019 (1).
Community-based testing refers to HIV testing services offered in the community, outside a health facility. WHO recommended community-based HIV testing services in 2013 to expand testing, especially among key populations and their partners, young people, men and others who may be less likely to test in facilities.
Community-based HIV testing services can be delivered in many ways and in different settings and venues. These include HIV testing services at fixed locations in the community, including community-based voluntary counselling and testing sites, mobile outreach in hotspots and community sites such as parks, bars, clubs, cruising sites and saunas and at events, places of worship, workplaces and educational establishments, sometimes with the use of mobile vans. Community-based HIV testing services can also be delivered at peoples’ homes, usually referred to as home-based HIV testing services (1). Home-based HIV testing services can either be offered to eligible members in all households in a certain area (that is, door-to-door) or be more focused – for example, in the context of HIV partner services. Community-based HIV testing services can be conducted by trained lay providers and peers using rapid diagnostic tests and the test for triage strategy.
Community-based HIV testing services can be delivered either alone or in combination with testing and screening for other infections such as TB, viral hepatitis and sexually transmitted infections or as part of other community services such as maternal and child health, and contraception. Appropriate training and supervision of providers is needed when combining HIV testing services with other infections and services. Some HIV testing service approaches and models, when carefully designed and focused, can effectively reach priority populations such as men, key populations and the partners of people living with HIV.
2.4.4. HIV self-testing
Recommendations (2019)
HIV self-testing should be offered as an approach to HIV testing services(strong recommendation, moderate-certainty evidence).
Remarks
Providing
HIV self-testing service delivery and support options is desirable.
Communities need to be engaged in developing and adapting
HIV self-testing models.
HIV self-testing does not provide a definitive HIV-positive diagnosis. Individuals with a reactive test result must receive further testing from a trained tester using the national testing algorithm.
Sources: WHO recommends HIV self-testing – evidence update and considerations for success (13) and Consolidated guidelines on HIV testing services, 2019 (1).
HIV self-testing is a process in which people collect their own specimen (oral fluid or blood) using a simple rapid HIV test and then perform the test and interpret their results when and where they want. Box 2.2 summarizes key messages on HIV self-testing for providers, self-testers and the community.
HIV self-testing has emerged as an effective tool in expanding HIV testing services among people at risk of HIV who may not otherwise test and those at ongoing risk who need to test frequently (14,15). HIV self-testing is a convenient and confidential option for HIV testing. HIV self-testing is safe and accurate, and lay users can perform HIV self-testing and achieve performance comparable to that of trained health-care workers. Globally, use of HIV self-testing within differentiated national HIV testing services plans has expanded and is helping countries to reach national and global goals.
Box 2.2Key messages for providers, self-testers and communities
HIV self-testing is a test for triage and does not provide a definitive HIV-positive diagnosis. A reactive (positive) HIV self-testing result is not equivalent to an HIV-positive diagnosis. All reactive HIV self-testing results need to be followed by further testing by a trained provider to confirm HIV status, starting with the first test in the national testing algorithm.
Nonreactive HIV self-testing results should be considered HIV-negative, with no need for immediate further testing except for those starting PrEP. For people starting or already taking PrEP, HIV self-testing will not usually replace initial or subsequent quarterly facility visits and testing.
Those with invalid HIV self-testing results need to repeat the test using another HIV self-testing kit or to seek testing from a trained provider. Any person uncertain about their HIV self-testing result should be encouraged to seek testing from a trained provider.
HIV self-testing is not recommended for people living with HIV who are receiving ART, since false-negative HIV self-testing results can occur. Those who are HIV-positive but not receiving ART should be encouraged and supported to initiate ART.
Retesting following a negative self-test result is necessary only for those at ongoing risk, such as people from key populations and those reporting potential HIV exposure in the preceding 12 weeks.
HIV self-testing means testing yourself. HIV self-testing is for individuals who want to test and learn their HIV status on their own. Offering a HIV self-testing kit to a sexual partner, friend or adult family member and encouraging them to use the self-test can often be a good way to help them learn their HIV status. However, a person should never be coerced or forced to self-test. Coercive or mandatory use of an HIV self-testing kit should never be supported or encouraged and is not considered self-testing.
WHO does not recommend that parents or guardians use HIV self-testing kits to test their babies or children. HIV self-testing will not provide a correct result for children younger than 18 months old because the mother’s antibodies may still be present in the infant.
HIV self-testing service delivery models and support tools
HIV self-testing kits can be distributed through various channels, including those supported by public or donor funding or in the private sector as well as through public–private partnerships (). If possible, offering choice in HIV self-testing service delivery options and type of test kits (such as between kits using oral fluid or blood) can help to reach more people.
HIV self-testing service delivery models.
Many people can perform HIV self-testing correctly with minimal or no support; however, some may need and want support, and it should be made available. This can include video instructions, virtual support or in-person demonstration or training. Providing a range of support options is important, if feasible. Support tools and packages should be adapted to address the local context, population needs and community preferences as well as considering available resources. Programmes are encouraged to define a minimum support package to accompany HIV self-testing implementation. This package can be regularly reviewed and adjusted as programmes expand and scale up. Available resources and community preferences need to be considered.
2.4.5. HIV partner services
Recommendations (2019)
Provider-assisted referral should be offered to people with HIV as part of a comprehensive package of testing and care(strong recommendation, moderate-certainty evidence).
Social network–based approaches can be offered as an HIV testing approach for key populations as part of a comprehensive package of care and prevention(conditional recommendation, very-low-certainty evidence).
Good practice statement (2019)
In all settings, biological children with a parent living with HIV (or who may have died from HIV) should be routinely offered HIV testing services and, if found to be either infected or at high risk of infection through breastfeeding, should be linked to services for treatment or prevention and offered a broader package of voluntary provider-assisted referral.
Note: Partner services include partner notification, contact tracing, index testing and family-based index case testing for reaching the partners of people living with HIV. These guidelines define partner services as encompassing a range of partner services packages and approaches, including social network–based approaches.
Source: Consolidated guidelines on HIV testing services, 2019 (1).
Partner services offer voluntary HIV testing services to the sexual and/or drug-injecting partners of people living with HIV. This is an effective way of identifying additional people living with HIV. Partners who are diagnosed with HIV can be linked to treatment services, and those who are HIV-negative and at ongoing risk of acquiring HIV can be linked to effective HIV prevention. When partner services are offered, HIV testing services should also be offered to the biological children of the person living with HIV if the children’s HIV status is unknown.
HIV partner services can be delivered in many ways, including patient referral and provider-assisted referral (16). If feasible and acceptable to the client, voluntary provider-assisted referral (see Box 2.3) should be given priority, since it is more effective and provides the opportunity to offer comprehensive prevention interventions to partners who are HIV-negative but remain vulnerable to acquiring HIV.
Box 2.3Definition of provider-assisted referral
In provider-assisted referral (also called assisted partner notification or index testing), a trained provider asks people living with HIV about their sexual and/or drug-injecting partners and then, with the consent of the HIV-positive client, informs the partners of their potential exposure to HIV. The provider then offers voluntary HIV testing services to these partners. The provider can contact partner(s) by telephone or email or in person and offer them home-based HIV testing services or invite them to visit a facility to receive HIV testing services.
Globally, the adoption of HIV partner services policies and implementation is increasing. Despite progress, members of key populations and their partners do not often fully benefit from HIV partner services and provider-assisted referral (14,17). Although partner services are safe, feasible and effective among members of key populations and their partners (14,17), implementation remains limited. This is often attributed to policy and structural barriers, confidentiality concerns and the reluctance of members of key populations to identify their partners to providers because they fear stigma and discrimination (17,18).
To address some of the challenges in scaling up HIV partner services among key populations, especially the challenges of confidentiality, in 2019 WHO recommended social network–based HIV testing approaches as an approach for reaching the sexual or drug-injecting partners and social contacts of the members of key populations. These approaches also can expand the scope of testing to the HIV-negative partners and social contacts of members of key populations, thus making testing services more acceptable and normalizing their use (16). Social network-based approaches are safe, acceptable and feasible and may identify additional people living with HIV.
All partner services should be voluntary. Whenever partner services and social network–based approaches are offered, the client should be informed of their benefits and cautions and assured that their decisions about contacting partners and other people from their social networks are voluntary and not pressured. Although provider-assisted referral should be encouraged, clients should have the opportunity to choose from all available options for partner services, including social network–based approaches, or to decline it altogether. They can choose different methods for different partners. For example, they may prefer to inform their primary partners themselves (patient referral), but they may not be comfortable informing other partners and instead choose provider-assisted referral.
2.4.6. Strategic planning for HIV testing services
Despite increases in the number of HIV tests conducted every year, 19% of the people living with HIV are unaware of their status and, in many settings, HIV testing services are not sufficiently focused (19). In many settings with high treatment and coverage rates, poorly targeted HIV testing services continue to miss people living with HIV who are at greatest risk and do not know their status. These include key populations globally and, in settings with a high burden of HIV infection in southern Africa, men, adolescents and young people (15–24 years old). These realities require new focus and new approaches to reach people with undiagnosed HIV early in their infection. WHO’s 2019 HIV testing services guidelines responds to this changing face of the HIV epidemic (1). They support the development and scale-up of evidence-informed HIV testing services approaches in facilities and in community settings for those who need HIV testing, prevention and treatment services.
Countries need to adopt a strategic mix of service delivery approaches to achieve equitable access to HIV testing services, based on the local context, the nature of the epidemic, priority populations and gaps and available resources. The national HIV testing services plan should facilitate diagnosing as many people living with HIV as early as possible and give priority to reaching the population groups with higher HIV risk in which the gap in knowledge of HIV status is greatest. Once diagnosed, HIV testing services should support effective linkage to appropriate post-test services. People living with HIV who learn their status without adequate support may not link to care or may be lost to follow-up.
Focused HIV testing services approaches are needed, especially in resource-limited settings (1). HIV testing services can be optimized by giving priority to new innovative approaches such as HIV self-testing and partner services, including provider-assisted referral (index testing) and social network–based HIV testing services approaches and focusing them on specific health services, priority populations and geographical settings. Optimized facility-based HIV testing services remain an important approach, especially in settings with a high burden of HIV infection. In addition, community-based testing can effectively reach key populations and other priority groups with a range of delivery models such as through community-based fixed sites or mobile outreach in hotspots and community sites such as parks, bars, clubs, cruising sites and saunas and at events, places of worship, workplaces and educational establishments or through home-based HIV testing services. Test for triage can be used in community-based HIV testing services using trained lay providers or peers and a single rapid diagnostic test and then referring and linking all people with a reactive test result to appropriate HIV prevention, care and treatment services (see also Chapter 7 on linkage from HIV testing to enrolment in care).
For certain population groups, retesting is recommended (see Box 2.4). The primary goal of retesting should be to enable those who have previously tested HIV-negative to stay HIV-negative and to identify those who have become HIV-positive as early as possible so that they can start treatment. Retesting among people who are HIV-negative or of unknown status has two key purposes: (1) monitoring the effectiveness of HIV prevention interventions and (2) identifying and treating new HIV infections as early as possible when prevention efforts fail. Globally, most people who have an HIV-negative test will not need retesting (1).
In low-HIV-burden settings, retesting all pregnant and postpartum women during pregnancy is not advised. In high-HIV-burden settings, retesting is advised for all pregnant women with an unknown or HIV-negative status during late pregnancy (third trimester). Catch-up testing is needed if the first test or retest is missed or delayed. High-HIV-burden countries could consider an additional retest in the postpartum period for specific districts or regions with high HIV burden or incidence, women from key populations or who have a partner with HIV who does not have suppressed viral loads.
See Chapter 7 of the Consolidated guidelines on HIV testing services, 2019 (1), which describes in detail the strategic planning considerations for effective and efficient HIV testing services.
Box 2.4Suggested optimal retesting frequency for various population groups
All settings
Only specific groups of people in high-HIV-burden settings or individuals with HIV-related risks need post-test counselling messages encouraging retesting at the appropriate intervals. WHO guidance recommends annual retesting for:
all sexually active individuals in high-
HIV-burden settings; and
people who have ongoing
HIV-related risks in all settings, including:
- –
key populations, defined as men who have sex with men, people in prison or closed settings, people who inject drugs, sex workers and transgender people;
- –
country- or epidemic-specific risk groups such as men and adolescent girls and young women in east and southern Africa; and
- –
people with a known HIV-positive partner.
Retesting in special groups: in certain situations, individuals who have been tested for
HIV in the past can be retested. These include:
- –
individuals presenting with a diagnosis or receiving treatment for sexually transmitted infections or viral hepatitis;
- –
individuals with a confirmed or presumptive TB diagnosis;
- –
outpatients presenting with clinical conditions or symptoms indicative of HIV; and
- –
individuals with recent HIV risk exposure.
More frequent retesting, that is, every 3–6 months, may be warranted based on individual risk factors and as part of broader HIV prevention interventions, such as individuals taking PrEP who require quarterly HIV testing or key populations who present to services with a sexually transmitted infection.
Retesting for pregnant and postpartum women
High-HIV-burden settings
Retest all pregnant women with unknown or HIV-negative status in late pregnancy, at the third-trimester visit. If either the first test or retest is missed or delayed, catch-up testing is needed.
An additional retest for women of unknown or HIV-negative status in the postpartum period can be considered. Countries could consider an additional postpartum test in specific districts or provinces with high HIV burden or incidence and among women from a key population or who have partners with HIV who do not have suppressed viral loads.
Low-HIV-burden settings
Retest pregnant women with unknown or HIV-negative status who are in serodiscordant relationships, whose partner does not have suppressed viral loads on ART or have other known ongoing HIV risk in late pregnancy – at a third-trimester visit. If either the first test or retest is missed or delayed, catch-up testing is needed.
An additional retest for women of unknown or HIV-negative status in the postpartum period can be considered among women from key populations or who have partners with HIV who do not have suppressed viral loads. Countries could also consider an additional postpartum test in specific districts or provinces.
Source: Consolidated guidelines on HIV testing services, 2019 (1).
2.4.7. Implementation considerations for priority populations
Because of shifts in HIV epidemiology (20,21), all settings must focus efforts increasingly on priority populations that remain underserved by existing approaches (Box 2.5). Priority populations are those that: (1) are most affected by HIV and have high ongoing HIV risk; (2) are critical to achieving and sustaining low HIV incidence; and/or (3) have specific individual or structural HIV-related vulnerabilities (16). Although key populations and their partners are a priority in all settings, other populations may be a priority based on country context, setting or local epidemiology. This often includes men, adolescents and young people, pregnant women, infants and children, serodiscordant couples, sexual and drug-injecting partners of people living with HIV as well as migrants, refugees, displaced populations and other vulnerable groups. See Chapter 6 of the Consolidated guidelines on HIV testing services, 2019 (1), which describes in detail the key considerations for implementing HIV testing services for priority populations.
Box 2.5Recommendations and implementation considerations for priority populations
Infants and children (3,22–25)
Recommendations
The addition of nucleic acid testing (NAT) at birth to existing early infant diagnosis testing approaches can be considered to identify HIV infection among HIV-exposed infants(conditional recommendation, low-certainty evidence).
In settings with a high burden of HIV infection, infants and children with unknown HIV status who are admitted for inpatient care or attending malnutrition clinics should be routinely tested for HIV(strong recommendation, low-certainty evidence).
In settings with a high burden of HIV infection, infants and children with unknown HIV status should be offered HIV testing in outpatient or immunization clinics(conditional recommendation, low-certainty evidence).
Point-of-care nucleic acid testing should be used to diagnose HIV among infants and children younger than 18 months of age(strong recommendation, high-certainty evidence).
Rapid diagnostic tests for HIV serology can be used to assess HIV exposure among infants younger than four months of age. HIV-exposure status among infants and children 4–18 months of age should therefore be ascertained by HIV serological testing the mother(conditional recommendation, low-certainty evidence).
Rapid diagnostic tests for HIV serology can be used to diagnose HIV infection among children older than 18 months following the national testing strategy(strong recommendation, moderate-certainty evidence).
An indeterminate range of viral copy equivalents should be used to improve the accuracy of all nucleic acid–based early infant diagnosis assays(strong recommendation, moderate-certainty evidence).
Good practice statements
National regulatory agencies are encouraged not to delay the adoption of point-of-care early infant diagnosis by conducting further evaluations but instead to adopt a rapid and streamlined registration and national approval process for immediate implementation.
In all settings, biological children with a parent living with HIV (or who may have died of HIV) should be routinely offered HIV testing services and, if found to be either infected or at high risk of infection through breastfeeding, should be linked to services for treatment or prevention and offered a broader package of voluntary provider-assisted referral.
Key populations (1,26)
Recommendations
HIV testing services should be routinely offered to all key populations both in the community and in facility-based settings.
Community-based HIV testing, with linkage to prevention, treatment and care, should be offered, in addition to routinely offering testing in facilities, for key populations in all settings
(strong recommendation, low-certainty evidence).
Social network–based approaches can be offered as an HIV testing approach for key populations as part of a comprehensive package of care and prevention(conditional recommendation, very-low-certainty evidence).
Adolescents (26)
Recommendations
HIV testing services, with linkages to prevention, treatment and care, are recommended for adolescents from key populations(strong recommendation, very-low-certainty evidence).
Adolescents should be counselled about the potential benefits and risks of disclosing their HIV-positive status and empowered and supported to determine whether, when, how and to whom to disclose(conditional recommendation, very-low-certainty evidence).
Settings with a high burden of HIV infection
In settings with a high burden of HIV infection, HIV testing services, with linkage to prevention, treatment and care, are recommended for all adolescents(strong recommendation, very-low-certainty evidence).
Settings with a low burden of HIV infection
HIV testing services, with linkage to prevention, treatment and care, should be accessible to adolescents in low and concentrated epidemicsa(conditional recommendation, very-low-certainty evidence).
Good practice statement
Governments should revisit age-of-consent policies, considering the need to uphold adolescents’ rights to make choices about their own health and well-being(with consideration for different levels of maturity and understanding).
Pregnant women, couples and partners (1,26,27)
Recommendations
All pregnant women should be tested for HIV, syphilis and hepatitis B surface antigen (HBsAg)a
at least once and as early as possible(syphilis: strong recommendation, moderate-certainty evidence; HBsAga: strong recommendation, low-certainty evidence).
Dual HIV and syphilis rapid diagnostic tests can be the first test in HIV testing strategies and algorithms in antenatal care.
Provider-assisted referral should be offered to all people with HIV as part of a voluntary comprehensive package of testing and care(strong recommendation, moderate-certainty evidence).
Couples and partners should be offered voluntary HIV testing services with support for mutual disclosure(strong recommendation, low-certainty evidence).
Women who disclose any form of violence by an intimate partner (or other family member) or sexual assault by any perpetrator should be offered immediate support. Health-care providers should, as a minimum, offer first-line support when women disclose violence. If health-care providers are unable to provide first-line support, they should ensure that someone else (within their health-care setting or another that is easily accessible) is immediately available to do so
(strong recommendation, indirect evidence).
Health-care providers should ask about exposure to intimate partner violence when assessing conditions that may be caused or complicated by intimate partner violence, to improve diagnosis and identification and subsequent care(strong recommendation, indirect evidence).
Good practice statements
Mandatory or coercive testing is never warranted. In consultation with the client, the provider should assess the risk of harm, the most appropriate approach for couple and partner testing, including more supportive options such as provider assistance, and situations that make couple or partner testing inadvisable.
- a
Now referred to as settings with a low burden of HIV infection.
- a
Particularly in settings with a ≥2% HBsAg seroprevalence in the general population.
2.6. Strategies to make HIV testing services accessible
Several WHO-recommended health programming practices can improve the accessibility and efficiency of HIV testing services in clinical and community settings, such as integrating HIV testing services with other testing and health services, decentralizing HIV testing services to primary health care facilities and outside the health facilities in the community and task sharing of HIV testing service responsibilities to increase the role of trained lay providers.
Integration
Integration is the co-location and sharing of services and resources across different health areas and includes offering HIV testing, prevention, treatment and care services alongside other relevant health services. WHO recommends integrating HIV services, including HIV testing services, with a range of other relevant clinical services, such as those for TB, viral hepatitis, sexually transmitted infections, maternal and child health, sexual and reproductive health, primary health care, key population programmes such as harm-reduction programmes for people who inject drugs and, in priority countries, voluntary medical male circumcision programmes. The primary purpose of such integration is to make HIV testing services more convenient for people attending health facilities for other reasons and to increase the uptake of HIV testing. Integration is appropriate in all epidemic settings and is especially important where the HIV prevalence is high and should be designed according to the focus populations and context.
Decentralization
Decentralization of HIV testing services refers to providing HIV testing services at peripheral health facilities such as primary health care facilities and outside health facilities in the community. Decentralization of HIV testing services may be appropriate in both high-prevalence and low-prevalence settings. Providing HIV testing in places closer to people’s homes may reduce transport costs and the waiting times experienced in central hospitals and thereby increase uptake. For example, community-based HIV testing services may be more attractive for men, young people and key populations, who are otherwise less likely to test in facilities (3). Close collaboration between community programmes conducting HIV testing and nearby health facilities and health-care providers is likely to improve rates of early enrolment in care. Linkage for ART and care services should be provided as quickly as possible, ideally in all decentralized sites and programmes.
Decentralization of HIV testing services may not always be appropriate or acceptable to potential users. In some settings, centralized HIV services may provide greater anonymity than neighbourhood services for key populations or others who fear stigma and discrimination. In some low-prevalence settings, decentralizing HIV testing services may be inefficient and costly. Context, needs, service gaps and overall costs and benefits should be weighed to determine the extent and manner of decentralizing HIV testing services.
Task sharing
Recommendation (2015)
Lay providers who are trained and supervised to use rapid diagnostic tests can independently conduct safe and effective HIV testing services(strong recommendation, moderate-certainty evidence).
Source: Consolidated guidelines on HIV testing services, 2015 (26).
Many countries continue to face shortages of trained health workers. Task sharing – the rational redistribution of tasks from types of health-care providers with longer training to types with shorter training – is a pragmatic response to health workforce shortages. Task sharing seeks to increase the effectiveness and efficiency of the available personnel and thus enable the existing workforce to provide HIV testing services to more people.
Trained lay providers and peer workers can support task sharing. A lay provider is defined as any person who performs functions related to health-care delivery and has been trained to deliver specific services but has received no formal professional or paraprofessional certificate or tertiary education degree. Lay providers can be trained to deliver all testing services, including pre-test information, performing HIV rapid diagnostic tests, interpreting test results and reporting HIV status, offering post-test counselling and supporting linkage to prevention, treatment and care services. Peers can be trained to function as lay providers.
WHO recommends that trained and supervised lay providers be able to provide HIV testing services, both in the community and in health facilities. A test-for-triage approach using single rapid diagnostic tests in the community or HIV self-testing with linkage to further testing at facilities can support the role of trained lay providers and community-based HIV testing services.
For further details, see Chapter 7, which includes a strong recommendation on task sharing of specimen collection and point-of-care testing with non-laboratory personnel when professional staffing capacity is limited.
2.7. Maintaining the accuracy and reliability of HIV diagnosis
WHO guidance on HIV diagnosis and testing strategies
Western blotting
Western blotting and line immunoassays should not be used in national HIV testing strategies and algorithms(strong recommendation, low-certainty evidence).
HIV testing strategy and algorithm
WHO recommends that all HIV testing algorithms achieve at least 99% positive predictive value and use a combination of tests with ≥99% sensitivity and ≥98% specificity.
The first test in an HIV testing strategy and algorithm should have the highest sensitivity, followed by a second and third test of the highest specificity.
Countries should consider moving to a three-test strategy as HIV positivity within national HIV testing service programmes falls below 5% – meaning all people presenting for HIV testing services should have three consecutive reactive test results in order to receive an HIV-positive diagnosis.
Dual HIV/syphilis rapid diagnostic tests can be the first test in HIV testing strategies and algorithms in antenatal care.
WHO suggests using a testing strategy for HIV diagnosis that is suitable for HIV diagnosis during surveillance and routinely returning HIV test results to participants.
Retesting prior to ART initiation
All people newly diagnosed with HIV should be retested to verify their HIV status prior to starting ART, using the same testing strategy and algorithm as the original diagnosis.
Retesting among people living with HIV who already know their status, who are on treatment, is not recommended as it can provide incorrect results if the person with HIV is on ART.
Source: Consolidated guidelines on HIV testing services, 2019 (1)
Providing a correct HIV diagnosis, as quickly as possible, is critical to all HIV testing services and national programmes. Accurate diagnosis enables newly identified people living with HIV to start ART sooner, which has immediate benefits for their health and, through provider-assisted referral (index testing), for the health of their partners and the community (1). To achieve accurate results for children older than 18 months, adolescents and adults, WHO recommends that countries use an HIV testing strategy or algorithm that combines rapid diagnostic tests and enzyme immunoassays that, when used together, achieve a positive predictive value of at least 99% (). The positive predictive value indicates the probability that an HIV-positive diagnosis is correct.
WHO encourages all countries to use three consecutive reactive tests to provide an HIV-positive diagnosis. Because of declines in HIV prevalence among those untreated (treatment-adjusted prevalence) and decreasing HIV positivity in HIV testing services programmes, countries currently using two consecutive reactive tests to provide an HIV-positive diagnosis are advised to move toward using three reactive tests as their treatment-adjusted prevalence and national HIV positivity in HIV testing services programmes fall below 5%. Countries with a low burden of HIV infection, with national HIV prevalence below 5%, are reminded to continue to use three consecutive reactive tests to provide an HIV-positive diagnosis.
WHO standard testing strategy for HIV-1 diagnosis (among people ≥18 months of age). A1:Assay 1 (first test); A2: Assay 2 (second test); A3: Assay 3 (third test). Assay (test) are HIV rapid diagnostic tests (RDTs) or enzyme immunoassays (EIAs). (more...)
WHO also recommends that countries move away from using western blotting and line immunoassays in their national testing strategies or algorithms in favour of simpler and less costly rapid diagnostic tests and/or enzyme immunoassays to support the scale-up of HIV testing, prevention and treatment. Using a rapid diagnostic test, a trained lay provider can establish a HIV diagnosis within a single visit in a health facility or community setting (1).
WHO continues to recommend that all programmes retest people diagnosed with HIV prior to initiating lifelong treatment. This retesting to verify an HIV-positive diagnosis is intended to catch human errors such as mislabelling of test results or other random errors related to the test device, lot or testing site (1). Retesting is common among people living with HIV who already know their status, including those receiving treatment. Motivations for retesting vary including doubts about the accuracy of a previous test, feeling sick or healthy or wanting to check on or come to terms with an HIV-positive diagnosis. Such retesting is not recommended and can provide incorrect results if the person living with HIV is receiving ART. For some people who know their HIV status but have not initiated or discontinued treatment, retesting is an important opportunity to initiate or re-engage in care and build trust and gain familiarity with health-care workers and the process of linking to care.
In all settings, dual HIV and syphilis rapid diagnostic tests can be offered as the first test in antenatal care to increase testing and treatment coverage. See for WHO-recommended testing strategy for dual detection of HIV and syphilis infection in antenatal care settings.
It is important not to use the rapid dual HIV and syphilis test for:
See Chapter 8 of the Consolidated guidelines on HIV testing services, 2019 (1) for further considerations for selecting diagnostics for HIV diagnosis and use of dual HIV and syphilis rapid diagnostic tests in antenatal care.
In addition, quality assurance implemented through quality management systems is essential for any testing service, ranging from HIV testing conducted in laboratories and health facilities to community-based settings, including rapid diagnostic tests performed by lay providers. Detailed guidance on quality systems is provided in Chapter 9 of the WHO consolidated guidelines on HIV testing services, 2019 (1) and other relevant publications (28,29).
WHO-recommended testing strategy for dual detection of HIV and syphilis infection in antenatal care settings. A1:Assay 1, A2: Assay 2, A3: Assay 3, TP: Treponema pallidum (syphilis). A1 (Assay 1) is a dual HIV/syphilis rapid diagnostic test (RDT).
2.8. HIV diagnosis among infants and children
Background
Because mortality in the first year of life is very high among untreated infants living with HIV, early HIV testing, prompt return of results and rapid initiation of treatment are essential (30,31). HIV infection among infants can only be definitively confirmed with virological testing using NAT technologies. This is because maternal HIV antibody transmitted across the placenta may persist in the child up to 18 months of age, preventing the use of serological testing to diagnose HIV infection (32,33). Access to early infant diagnosis has improved significantly in recent years, but only about 60% of all HIV-exposed infants were tested by the second month of age in 2020 (19). For infants who are tested, delays in obtaining results and further losses in the testing-to-treatment cascade still occur (34) so that only 30% (35) of perinatally infected infants are effectively linked to services and start ART in a timely manner. Innovative approaches such as using assays at the point of care and adding a NAT at or around birth (0–2 days) can improve rapid identification and treatment initiation among infants (4,36,37).
Comprehensive HIV testing approach for infants and children.
Although early infant diagnosis is critical to minimize early mortality, other opportunities for testing are also essential to capture infants and children living with HIV who are infected postnatally or who were missed by infant diagnosis services (). Ensuring timely diagnosis of HIV infection for children requires a mix of interventions provided at different times at different care points. For children older than 18 months of age, serological testing is used in the same manner as in adults following the nationally validated testing algorithm. Since children poorly utilize voluntary counselling and testing services, facility-based testing is essential to improve the identification of children living with HIV, especially those who are born to mothers who have not received interventions for preventing mother-to-child transmission (3). Determining the exposure status of all infants and children who present to key entry points of health-care facilities, including malnutrition, TB and inpatient wards, and subsequently testing those identified as HIV exposed will support increased case-finding – each entry point has been observed to be a high-yield setting (5). Implementing index-case testing (also called family-based testing) has led to remarkable improvement in identifying children missed by infant diagnosis programmes (38), but more efforts are needed to widely scale up this intervention. Finally, targeted testing with the support of tools to increase the efficiency of testing at facilities at which low prevalence might not merit routine approaches is increasingly considered to manage competing priorities and limited resources at the facility (39).
2.8.1. Timing of virological testing
Recommendation (2016)
The addition of nucleic acid testing (NAT) at birth to existing early infant diagnosis testing approaches can be considered to identify HIV infection in HIV-exposed infants(conditional recommendation, low-certainty evidence).
Source: Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection: recommendations for a public health approach – second edition (3).
Background
Infants who have HIV detectable by NAT at birth are likely infected prenatally, will progress to disease rapidly and, in the absence of treatment, experience high mortality in the first few months of life (40,41). Infants infected at or around delivery may not have virus detectable by NAT for several days to weeks. The ability of NAT to detect virus in the blood may be affected by ARV drugs taken by the mother or infant for postnatal prophylaxis, resulting in false-negative results. This includes drugs present in the breast-milk as a result of maternal ART during breastfeeding. In addition, since HIV prevalence in the population decreases as a result of effective interventions to prevent mother-to-child transmission, the proportion of false-positive NAT results increases, underscoring the need to effectively confirm those identified as positive (42,43).
Finally, the ongoing risk of acquiring HIV during breastfeeding can delay the final determination of HIV status beyond 18 months. For all these reasons, identifying the optimal timing and frequency of infant testing is very challenging. Existing testing approaches have attempted to enhance programmatic simplicity and maximize the uptake of testing by aligning the timing of testing with the childhood immunization schedule. However, given the recent decline in the prices of assays and the expansion of infant diagnosis programmes, alternative testing approaches can be considered that maximize uptake, retention and timely treatment initiation while responding to changing epidemic and transmission dynamics.
The complexity of infant diagnosis is now growing because of significant scale-up of “treat all” (including pregnant and breastfeeding women), implementation of enhanced postnatal prophylaxis, reduced mother-to-child transmission rates and the increased relative contribution of postnatal transmission. Infant diagnosis can no longer be considered primarily a one-test process, since it now requires additional testing over the duration of exposure. Accordingly, several additional key considerations will be necessary to strengthen the infant testing cascade through the entire exposure period. This includes ensuring that ART initiation is not delayed for the infants found to be living with HIV.
Rationale and supporting evidence
The optimal timing of virological testing to diagnose HIV infection among infants is determined by four factors: (1) when infection occurs (prenatally, intrapartum or postpartum during breastfeeding); (2) the sensitivity and specificity and predictive values of the assay being used; (3) the mortality risk by age and (4) retention in the testing and treatment cascade (4). Relevant evidence that informed this recommendation includes survival curves, available data on the testing-to-treatment cascade and a recent diagnostic accuracy review on the performance of NAT at birth (0–2 days) and at four to six weeks of age in the context of exposure to ARV drugs (44).
Although concerns have been raised about the potential delay of HIV detection as a result of ARV drug exposure (42,43), no direct evidence currently confirms that the performance of NAT on dried blood spots at four to six weeks of age is lower in the context of ARV drug exposure (pooled sensitivity and specificity were 100% and 99.03% [95% CI 98.19–99.88%]). However, the quality of the available evidence is low, and more data on the performance of virological testing is urgently needed, especially in the context of maternal ARV drug exposure and enhanced (prolonged and multidrug) infant postnatal prophylaxis. Given the available evidence, the ability to detect both prenatal and intrapartum infections and to remain aligned with the provision of routine maternal and child health services such as scheduled immunization visits and co-trimoxazole prophylaxis, the age of four to six weeks remains the critical point at which to provide virological testing, as recommended in existing testing strategies (16).
The accuracy of diagnostic tests was reviewed in 2015 (44) to consider adding NAT at birth to detect perinatal HIV infection. Two studies were identified with overall sensitivity of 67.8% [95% CI 60.9–74.8%] and specificity of 99.73% [95% CI 99.4–100%], reflecting the difficulty of detecting intrapartum infections. Because of relatively poor sensitivity emerging from the currently available evidence, a single NAT at birth is likely to miss many infections and should only be considered as an additional opportunity for testing rather as substituting for the existing approach of testing at four to six weeks of age.
Overall, the empirical evidence was insufficient to recommend universally including NAT at or around birth (0–2 days) as a way to improve infant and programme outcomes. Nevertheless, this approach has potential benefits since it provides an additional opportunity for testing and enables earlier identification of infected infants in the context of poor scale-up of infant diagnosis. Linking birth testing to prompt ART initiation and care has the potential to reduce the early mortality and morbidity observed among infants who are infected prenatally and for whom the disease progresses more rapidly. From a programmatic perspective, there are potential advantages (but a lack of experience) with adding NAT at birth (0–2 days) and uncertainty around the clinical benefits and potential difficulties of treatment from birth as well as the potential complexity and cost of adding an additional test at a new service delivery point. There are also challenges associated with starting treatment among newborns and preterm infants given the available ARV drugs for this age group (see subsection 4.3).
In 2015, focus group discussions (45) with 105 women living with HIV from Kenya, Namibia and Nigeria suggested that earlier infant testing could be acceptable since mothers are motivated to avoid disease progression among infants. However, there were also concerns about the potential lack of understanding about the need to retest infants with negative NAT results and the associated loss to follow-up as well as potential emotional overload for women immediately after giving birth and the challenge of preserving confidentiality in the presence of family, partners and others in labour wards. Overall, women in the focus groups showed some reluctance to accept routine virological testing at birth and more favoured having a range of options from which to choose.
Model-based analysis (46) undertaken in 2015 supported optimizing six-week testing before adding NAT at birth. In addition, it suggested that, under the ideal scenario of full uptake and retention (100% of HIV-exposed infants being tested and retained in the testing-to-treatment cascade), a two-NAT strategy, with the first test at birth and the second test after six weeks of age, improves survival compared with a single test at six weeks. Any testing programme, whether at birth or six weeks, must have a mechanism to return test results promptly and link infants living with HIV to care and ART. Based on programmatic, clinical and cost data from South Africa over the lifetime of HIV-exposed infants, the modelling found that a programme of birth and six-week testing could be very cost-effective in settings similar to South Africa. The model confirmed that false-positive results may be common (about 30 positive results of 100 may be false-positive), even with relatively high assay specificity (98.0–99.0%), especially if the risk of mother-to-child transmission is low (less than 2% at six weeks). Confirmatory testing is critical to minimize toxicity, stigma and costs for uninfected infants with false-positive results.
Given the risks, benefits, possible acceptability and potential cost–effectiveness, the 2016 WHO consolidated HIV guidelines (3) recommended that adding NAT at or around birth (0–2 days) can be considered where feasible but only in parallel with efforts to strengthen and expand existing infant testing approaches. Existing recommendations that infants with an initial positive virological test result should start ART without delay remain important. Nevertheless, a second specimen should be collected to confirm the initial positive virological test result. Immediate initiation of ART saves lives and should not be delayed pending the results of the confirmatory test.
Implementation considerations
As infant diagnosis programmes are further scaled up, every effort should be made to improve the uptake of NAT, strengthen retention along the testing-to-treatment cascade, ensure confirmation of NAT positive results with a second sample and ensure that infants who test negative by NAT are retained in care until a final diagnosis is made. If adding NAT at birth is being considered, effective linkage to maternal HIV screening at the time of delivery should be ensured and the following steps should be taken:
collection of data on the performance and feasibility of birth testing during implementation;
improvement of uptake and retention in the testing-to-treatment cascade;
active tracking of infants with negative NAT at birth to ensure that they return at six weeks for retesting and co-trimoxazole initiation; and
retesting of infants who test positive at birth with a second specimen as soon as possible, with
ART being started immediately after the first positive test and stopped if the second specimen tests negative.
Several countries have already started implementing NAT at birth. Several implementation considerations can be summarized from these experiences (24).
Countries that are considering birth testing should critically review current performance and opportunities for strengthening their six-week and overall infant diagnosis programme and consider key indicators (such as PENTA1 immunization visit coverage and attended delivery rate), so that the potential gains provided by birth testing can be investigated more fully. For example, in settings where the attended delivery rate is much lower than PENTA1 immunization visit coverage (six weeks), the added value of birth testing as a means of expanding infant diagnosis may be limited.
Pilot projects are a good way to start obtaining national experience on this innovative testing approach, but fully measuring the impact requires that programmes collect data on the feasibility and impact of birth testing and linkage to
ART initiation.
Targeted approaches that provide birth testing only for high-risk infants are expected to have a higher yield than routine birth testing. This approach may be potentially less resource intensive and present a lower burden for health-care workers.
Active tracking of infants with negative NAT results at birth is critical to ensure that they return at six weeks to be retested and start co-trimoxazole; establishing unique patient identifiers or other innovative mechanisms (such as bar codes) to track babies can be considered.
It is crucial that the turnaround time for reporting test results to health facilities and caregivers be rapid to optimize the benefit from NAT at birth. Same-day point-of-care assays should be used whenever possible.
Birth testing is acceptable to mothers, but challenges arise from the increased human resources needed, the difficulty of collecting blood samples from newborns, the need to ensure sample collection outside standard working hours and deliver results, linkage to
ART and the nature of the infant diagnosis system as a whole (stock-outs, referral mechanisms and delayed results).
The key to effective implementation is to ensure that newborns who have been identified as
HIV-infected are linked to treatment immediately and that age-appropriate formulations are available to start them on treatment.
Good leadership and coordination are needed to oversee service provision, support supervision, mentorship and the quality improvement cycle.
Ensuring accurate interpretation of the nine-month test
The 2016 WHO consolidated HIV guidelines (3) recommend that rapid diagnostic tests be used to assess HIV exposure among infants younger than four months, and HIV exposure among infants 4–18 months old should be ascertained by testing the mother. When the mother cannot be tested, current guidelines emphasize the importance of not considering a negative rapid diagnostic test result from an infant 4–18 months old as a definitive test of exposure. Box 2.6 highlights implementation issues.
Based on the 2016 WHO consolidated HIV guidelines (3), rapid diagnostic tests are serological assays that can also be used to exclude established infection among healthy, HIV-exposed infants nine months and older. However, changes in transmission dynamics and in policy and practice have complicated using rapid diagnostic tests for determining infection status. Substantial drug exposure for infants with implementation of the “treat all” policy for mothers and enhanced postnatal prophylaxis of HIV-exposed infants may have resulted in viral load reduction and delayed antibody development among infants living with HIV. Finally, the occurrence of maternal infection in late pregnancy or during the postnatal period may have caused a lack of passive HIV antibody transfer to the HIV-exposed infant. These factors increasingly jeopardize rapid diagnostic test accuracy at nine months of age as a means of correctly ruling out established infection among HIV-exposed infants. These concerns are supported by findings from Kenya and Uganda (47,48), where 15–40% of children younger than two years and identified as HIV-infected had a positive NAT but negative rapid diagnostic test.
Box 2.6Using rapid diagnostic tests: implementation considerations
Priority should continue to be given to testing mothers at all entry points to determine exposure status for infants and children younger than 18 months.
If the mother is absent or unable to be tested, the infant should have a rapid diagnostic test, but negative results for infants older than four months should not definitively exclude exposure, and follow-up testing is required.
If the mother is absent or unable to be tested and the infant presents with signs and symptoms of
HIV infection, perform a NAT.
Perform a NAT following any positive rapid diagnostic test for the mother or the infant and perform a confirmatory NAT following any positive NAT result.
Rapid diagnostic testing at nine months was initially recommended in the 2010 WHO recommendations on the diagnosis of HIV infection in infants and children (49), with the goal of targeting NAT for the HIV-exposed infants most likely to be infected (such as those with a positive rapid diagnostic test) as a cost-saving measure. However, because of declining mother-to-child transmission rates, increasing availability and lower costs of NAT, changing transmission and drug exposure dynamics and the fact that rapid diagnostic tests are less effective at determining the need for NAT testing, such a targeted approach may be less compelling. Further, the added programmatic complexity and potential for inappropriately interpreting test results have additional unintended consequences.
Given the challenges and data outlined above, replacing the rapid diagnostic test at nine months with NAT can now be considered to minimize the challenges of interpretation and simplifying the infant testing algorithm.
2.8.2. Infant diagnosis algorithm
Infant diagnosis throughout the exposure period is critical to identify all possible infants and children living with HIV who need treatment. Several interventions, including recommendations on when to test, where to test and with what to test aim to improve case-finding and rapid linkage to treatment.
Several key considerations underscore the new simplified algorithm ():
Finally, continuing infant retention in care remains critical until the end of the exposure period. More effort should be given to establishing a final diagnosis at 18 months of age or three months after breastfeeding ends, whichever occurs later. Although the coverage of the traditional six-week infant test is increasing and earlier time points are increasingly considered, the changing dynamics of transmission and increased drug exposure mean that increased efforts are needed to maintain follow-up throughout the entire exposure period. The aim is to ensure that all HIV-infected infants, including those infected in the postnatal period, are identified and receive treatment.
Simplified infant diagnosis algorithm. a Based on 2016 WHO Consolidated ARV Guidelines (3), addition of NAT at birth to the existing testing algorithm can be considered. b Point-of-care NAT can be used to diagnose HIV infection as well as to confirm positive (more...)
Confirmatory testing of positive test results
A cost–effectiveness analysis undertaken to assess the value of confirmatory testing in different scenarios highlighted that confirmatory testing is indeed cost-effective (50). Without confirmatory testing, this analysis showed that, in settings with mother-to-child transmission rates similar to those of South Africa, more than 10% of the infants initiating ART may in fact not be HIV-infected. Confirmatory testing of positive test results using a new sample, in accordance with WHO guidelines, may avoid this, although this policy is not consistently implemented (Box 2.7).
Programmes must ensure that all HIV-exposed infants are retained in care and tested appropriately throughout the entire exposure period, and all infants with a positive result should receive a confirmatory test. Further, those with repeatedly indeterminate test results should be actively tracked, retained and retested and their status should be resolved.
Box 2.7Giving priority to confirmatory testing of positive and indeterminate tests
Declining mother-to-child transmission rates globally have led to concerns about false-positive and indeterminate tests.
People with indeterminate results need immediate repeat testing and should be managed according to the standard operating procedures ().
People with repeated indeterminate results need a multidisciplinary team of health-care providers to support retention, tracking and status resolution.
ART programmes need to give priority to confirmatory testing of all positive test results using a new sample.
Clinical monitoring and further testing based on the national infant testing schedule need to be done until a definitive
HIV status is established.
Finally, point-of-care infant testing is being implemented in several countries and settings (see subsection 2.8.3). Previously there was limited evidence on how to conduct confirmatory testing of positive point-of-care infant test results, but since the 2016 WHO consolidated HIV guidelines were published (3), several studies have been published on its performance. Two point-of-care infant technologies are included on the WHO list of prequalified in vitro diagnostic products (51). The results from both laboratory and field studies have shown performance comparable to that of laboratory-based technologies (52). Further, two patient impact studies have been published that highlight the significantly improved patient outcomes when using point-of-care early infant diagnosis technologies (53,54). Based on this updated evidence, point-of-care infant testing can be used to confirm positive test results.
2.8.3. Technologies to use for infant testing
Recommendation (2021)
Point-of-care nucleic acid testing should be used to diagnose HIV among infants and children younger than 18 months of age(strong recommendation, high-certainty evidence).
Source: Updated recommendations on HIV prevention, infant diagnosis, antiretroviral initiation and monitoring (22).
Background
Although significant recent investments in improving the diagnostic networks, centralized laboratories and sample collection networks have been made in most settings with a high burden of HIV infection, clear improvements in access to infant testing and treatment initiation of infants have not increased at the same rate. Substantial challenges and barriers remain. First, in 2019, only 60% of infants received an HIV NAT within the first two months of age (19). Further, only 53% of children younger than 15 years living with HIV were receiving ART in 2019. The mortality of untreated, perinatally infected infants peaks at two to three months of age, with about 35% dying by 12 months of age and 52% by 24 months of age (40,55). A recent systematic review of laboratory-based, standard-of-care infant testing found that the mean turnaround time from sample collection to the results received at the clinic was 4.5 days (34). The time between the results received at the clinic to receipt by the caregiver was 4 days. The mean age at infant testing was 74 days; however, the mean age at treatment initiation was 214 days (seven months). In addition, in a subset of studies, 15% of the infants living with HIV had died between infant testing and ART initiation.
HIV NAT for infant diagnosis that can provide results on the same day of sample collection, similar to those used for older children and adult HIV testing, are now available on the market and have been approved by regulatory authorities (51). Several of the device-based technologies available are multi-disease nucleic acid–based technologies that can be shared across diseases for other molecular assays. Additional device-free tests are being developed. In 2016, WHO conditionally recommended the use of point-of-care technologies for infant diagnosis (3). This was based on low-certainty evidence from two diagnostic accuracy studies available at the time. Subsequent studies, including patient impact and clinical studies, have been completed and the guidance presented here was updated in early 2021 (22).
Rationale and supporting evidence
Summary of review findings
A systematic review (56) of the clinical impact of using same-day point-of-care infant diagnosis compared with laboratory-based technologies identified seven studies (53,54,57–61) of more than 37 000 infants across 15 countries in sub-Saharan Africa. The studies included two randomized controlled trials and several large, well-characterized cohort studies. The studies directly evaluated similar outcomes, and only those that provided true point-of-care, same-day testing and results were included. The data and results were consistent across studies. Most studies had a low risk of bias for critical outcomes (including the time to receive results), except for retention in care and mortality outcomes, with the risk of bias noted to be serious given the limited number of studies and small sample sizes. The overall certainty of the evidence in this review was rated as high.
Median time from sample collection to delivering the result to the caregiver
Same-day point-of-care testing significantly reduced the time to deliver the result to caregivers (high-certainty evidence). Across all seven studies, the median time from sample collection to results received by the infants’ caregivers was 0 days (95% CI 0–0 days) for point-of-care testing, regardless of the test used, the age of the infant or the type of health-care facility. Same-day results were returned 97% of the time when tested by point-of-care testing versus 0% for standard of care. For standard of care, the median time from sample collection to the caregiver receiving the result was 35 days (range 8–125 days, 95% CI 35–37 days). Five of seven studies had a median time to the caregiver receiving the result of more than 30 days. Six studies reported the median time from sample collection to initiating ART among infants testing positive for HIV was 0 days (95% CI 0–1) when tested using point-of-care testing (2–6,8). When tested using point-of-care testing, 51% of infants living with HIV initiated ART on the same day as sample collection versus 0% when tested by the standard of care. For the standard of care, the median time from sample collection to treatment initiation was 40 days (range 6–127 days, 95% CI 34–43 days). The evidence was of high certainty overall.
Proportion of infants living with HIV initiating treatment within 60 days
The overall proportion of infants living with HIV initiating treatment within 60 days was 90% when tested at the point of care versus 54% when testing using the standard of care. The odds ratio of initiating treatment within 60 days was 7.9 (95% CI 5.4–11.5). The evidence overall was of high certainty.
Retention in care and mortality
Two studies provided follow-up data for infants living with HIV after diagnosis and initiating treatment (53,59). The first study, from Mozambique, found that infants tested using point-of-care testing were significantly more likely to be retained in care after 90 days of follow-up compared with those receiving standard-of-care testing (adjusted rate ratio 1.40) (53). The second study, from Zambia, found high mortality rates in both arms but no statistically significant difference in mortality or rates of viral suppression at 12 months of age; however, the sample size was small: only 20 of 81 infants living with HIV remained alive and in care at 12 months from both groups (59). Overall, the evidence for these outcomes was of very low certainty.
The systematic review had several limitations. First, all studies were from sub-Saharan Africa, although this is consistent with the fact that more than 90% of HIV vertical transmission is in the WHO African Region. Although most studies had a low risk of bias for retention in care and mortality outcomes, the risk of bias and imprecision were noted to be serious given the limited number of studies and small sample sizes. The hub-and-spoke and near point-of-care concepts could not be analysed with the data available. In some studies, the hub-and-spoke results were provided within the same-day point-of-care arm and thus excluded because of inability to differentiate same-day versus near point-of-care testing. Although data suggest that same-day testing improves the return of results and treatment initiation, additional studies comparing same-day point-of-care with near point-of-care (less than seven days) and the standard of care (laboratory-based testing) testing would provide a more reliable basis for assessing this outcome.
Costs and cost–effectiveness
A synthesis of available cost–effectiveness models was developed using four cost–effectiveness studies and two overarching modelling approaches (62). Johns Hopkins’ model focused on sub-Saharan Africa and Zambia (63,64); and the Cost–effectiveness of Preventing AIDS Complications (CEPAC)-Paediatric model focused on Zimbabwe (65,66). All studies reported that point-of-care testing was more cost-effective than the standard of care defined in each study. Health benefits were described in terms of life-years saved, additional people initiating ART and deaths averted. In most scenarios, integrating or sharing platforms across diseases (Xpert® TB testing or HIV viral load testing) resulted in point-of-care testing being cost-saving compared with the standard of care. In Zambia, point-of-care testing cost US$ 752 less than the standard of care per additional person initiating ART when the devices were shared across TB and HIV programmes.
Affordability
Current point-of-care infant diagnosis tests cost US$ 15–25 per test, with instruments costing about US$ 15 000.
Access to same-day point-of-care testing in four countries with a high burden of HIV (Malawi, Mozambique, Uganda and Zambia) is currently estimated to already be 30–50%. The estimated incremental cost to support access to 70%, 80% or 90% of HIV-exposed infants with point-of-care technologies would be US$ 60, US$ 109 and US$ 194, respectively (67). These costs, for both point-of-care and laboratory-based testing, could be amortized across (but were not calculated within) other programmes, such as TB programmes, that may also use the devices. The remaining proportion of HIV-exposed infants would require access to infant testing through referral to laboratory-based devices.
Implementing point-of-care testing in these four countries would result in considerably more infants living with HIV initiating ART. With 70%, 80% or 90% point-of-care implementation for infant testing, 149 000, 162 000 or 175 000 infants living with HIV, respectively, would initiate treatment versus just 110 000 if the current rates of point-of-care testing were maintained. This would result in a cost of between US$ 325 and US$ 622 per additional person initiating ART.
Ethically, concerns about costs should not be a barrier to adoption. If the clinical and public health evidence in its favour is as conclusive as it seems, then the global health community must work with national governments and local authorities to supply point-of-care testing for infants. Paths forward would include appealing to international agencies and directly to the companies that build these diagnostics to lower their costs as much as possible.
Values and preferences
In a study from Kenya (74 interviews and six focus group discussions) and Zimbabwe (85 interviews and eight focus group discussions) of community members and elders, data were collected before point-of-care testing was introduced and after it had been in use for at least three months (68). Reduced time to receive test results lowered caregiver anxiety about the child’s HIV status and enabled families to start treatment earlier. Some considered printed point-of-care results more trustworthy than conventional handwritten results, believing that this reduced the chance of human error; a few distrusted HIV results that were generated too quickly. Caregivers were supportive of receiving point-of-care infant testing; however, additional collaboration with community groups is needed to increase acceptance and demand.
In addition, an online survey was undertaken among 43 people living with HIV to determine their values and preferences for using point-of-care testing for infant diagnosis compared with laboratory-based testing (22). Most people living with HIV (72%) thought that collecting the sample, testing and providing the result within one hour would be acceptable. Half (51%) the respondents thought that knowing the HIV status as soon as possible would be worthwhile, and 41% thought that the benefit of same-day testing and results was that treatment could start immediately. The majority (81%) thought that testing, diagnosing and starting treatment for an infant on the same day was acceptable. Most people living with HIV (74%) thought that nurses would be able to test an infant for HIV and provide the test results at the same visit. Most respondents (72%) felt confident that health-care workers could do this.
Acceptability and feasibility
A study across eight African countries (Cameroon, Côte d’Ivoire, Eswatini, Kenya, Lesotho, Mozambique, Rwanda and Zimbabwe) comprised structured interviews with health-care workers providing infant testing services and semistructured interviews with national and regional laboratory managers or early infant diagnosis programme managers – before and after point-of-care infant testing was implemented (69). Health-care workers found point-of-care infant testing easy to use (74% said it was very simple to run the test) and were very satisfied with the rapid turnaround time and ability to initiate treatment for infants living with HIV sooner (93%). All health-care workers recommended that the country increase point-of-care infant testing, while 87% would want a device in their health-care facility. Laboratory managers also supported scaling up point-of-care testing, although they were cautious about the need for reliable infrastructure to operate platforms.
In addition, an online survey was provided to 51 health-care workers and 53 programme managers to determine the acceptability and feasibility of implementing point-of-care infant testing (22).
Survey of 51 health-care workers
Most (88%) felt comfortable running the test, delivering the result, counselling and starting treatment on the same day. Most health-care workers thought it would be acceptable (77%) and 65% prefer point-of-care infant testing, if available. The majority (88%) thought that the mother would accept same-day infant testing and in some cases positive diagnoses. Almost half (45%) of the health-care workers thought that implementing point-of-care infant testing would increase the workload in the clinic but that enough human resources were in place to implement it.
Survey of 53 programme managers
Most countries (72%) surveyed already have a policy for point-of-care infant testing; however, 85% indicated that most infant tests were done using standard-of-care laboratory-based testing. The majority (55%) thought point-of-care infant testing was preferable and feasible. More than half the programme managers (55%) did not think that the workload would increase if point-of-care infant testing was implemented either in the laboratory or in the clinic.
Diagnostic accuracy
A systematic review was prepared to provide summary estimates of the diagnostic accuracy of technologies capable of being used at the point of care. The performance overall was greater than 98% sensitivity and 99% specificity (70).
Feasibility
Several technologies are on the market and available for use at the point of care; two already have WHO prequalification (58). Many such devices have already been procured and are used for TB testing (Cepheid GeneXpert®) or infant diagnosis already (Abbott m-Pima™ and/or Cepheid GeneXpert®). Both tests use whole blood and do not require any additional equipment or expertise. The Abbott m-PIMA™ device can run about 6–8 tests per day, and the Cepheid GeneXpert® device can perform about 6–8 tests per module per day (71). Across 140 developing countries with a high burden of TB and HIV infection (Cepheid’s High Burden Developing Country programme), more than 11 600 devices have been delivered, comprising 52 000 modules. Nearly 12 million GeneXpert® TB cartridges were procured per year in 2017 and 2018; however, only 1.2 tests per module per day are currently being run. This leaves available capacity for expanding TB testing and considering HIV infant testing. Infant diagnosis should remain a priority when technologies are multi-purposed or shared across programmes.
Point-of-care technologies may not need to be procured for every health-care facility to reach most HIV-exposed infants. In most countries with a high burden of HIV infection, most HIV-exposed infants attend a small proportion of available health-care facilities. In an analysis from Malawi, Mozambique, Uganda and Zambia, 80% of HIV-exposed infants attended 32%, 33%, 12% and 10% of health-care facilities, respectively, indicating that modest procurement and focused placement of point-of-care technologies would affect many of the HIV-exposed infants (67). Further, 10% of health-care facilities in each country serve 49%, 46%, 75% and 80% of HIV-exposed infants, respectively, in these four countries.
Equity
Ethical and equity considerations were developed to guide the guideline discussions (72). Some of the conclusions found were the fair distribution of benefits and burdens at the population level (social justice), treating people as equally important (equity) and that infants should not be differentially disadvantaged relative to others in their communities when there is little to no risk of precluding the provision of other or ongoing health resources. If the rest of the community is not harmed from going without a specific resource by introducing point-of-care testing, then it is unclear what could possibly count against introducing it.
Rationale for the decision
The Guideline Development Group formulated a strong recommendation favouring point-of-care NAT to diagnose HIV among infants and children younger than 18 months of age. This was based on their judgement of the overwhelming benefits of the intervention, including, but not limited to:
more rapid testing and return of results to caregivers and clinicians;
increased retention in the testing-to-treatment cascade;
fewer health facility visits for caregivers to receive results and more reliability in the timing of results and possibly more likelihood for test documentation;
increased equity with adult
HIV testing – same-day testing and receiving the result;
increased access to
ART and faster initiation, which may reduce mortality; and
improved quality of services.
No major notable harm was identified. However, some considerations were noted around the general higher costs of testing (this was not viewed as a barrier to implementation), the more extensive network support required by health-care workers and the need for greater technical support and maintenance (service and maintenance, quality assurance and supply chain).
Implementation considerations
The Guideline Development Group highlighted several implementation considerations. First, point-of-care infant diagnosis technologies should be considered and used within the current infant diagnosis algorithm at any point when an NAT is required (24) (). Second, access to high-quality diagnostic testing should be continually expanded across HIV and other molecular testing needs, ideally combining laboratory-based and point-of-care technologies in an integrated diagnostic network. If point-of-care testing cannot be done, alternative options must be found, including ensuring rapid laboratory-based testing. Optimal placement of point-of-care technologies should be considered within the context of the overall health system, including other disease programmes and needs. This will create efficiency and support expansion and improved diagnostic services for HIV and other diseases.
Finally, ensuring adequate human resources, training (including technical, result interpretation, counselling and supply chain), service and maintenance and quality assurance should be carefully considered. Clear messaging, communication and literacy considerations should be implemented to support demand generation, scale-up, trust and utilization, including close collaboration with community groups. Maximizing the clinical impact of point-of-care testing requires ongoing strengthening of treatment and care services for neonates, infants and children, same-day linkage of infants to treatment and care, reliable procurement of appropriate formulations for children and supported supervision for health-care workers managing these young infants.
Research gaps
Although substantial evidence was available to review this question, further implementation research on quality assurance approaches could be considered to understand the sustainable delivery of point-of-care testing for infant diagnosis. Further, a potential dual-claim point-of-care test should be investigated that can be used across infants, children and adults, both for HIV diagnosis and viral load to streamline supply chain and create more efficient diagnostic systems.
In addition, tests are being developed that may be device-free and closer to a traditional rapid diagnostic test. These will likely support further decentralization and require no capital investment for health-care facilities, especially those with low volumes. Diagnostic accuracy and clinical impact studies for these tests would be beneficial.
2.8.4. Rapid diagnostic tests for HIV serology
Recommendation (2016 and 2018)
Rapid diagnostic tests for HIV serology can be used to assess HIV exposure among infants younger than four months of age. HIV-exposure status among infants and children four to 18 months of age should therefore be ascertained by undertaking HIV serological testing in the mother
(conditional recommendation, low-certainty evidence).
Rapid diagnostic tests for HIV serology can be used to diagnose HIV infection among children older than 18 months following the national testing strategy
(strong recommendation, moderate-certainty evidence).
Sources: Consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection: recommendations for a public health approach – second edition (3) and HIV diagnosis and ARV use in HIV-exposed infants: a programmatic update (24)
HIV antibody assays reliably detect HIV antibodies among children but cannot distinguish persisting maternal HIV antibody from antibodies produced by the child. A positive HIV antibody test among infants and children younger than 18 months of age therefore confirms exposure to HIV but cannot definitively diagnose infection. In contrast, the presence of HIV antibodies is a quick and reliable means of definitively diagnosing HIV infection among children older than 18 months because maternal HIV antibodies are usually no longer detectable. Children who start ART as early as three to six months of age are unlikely to develop antibody response to the virus and may falsely test HIV-negative using a serological assay. Antibody testing should therefore not be used to confirm or rule out infection among children who are already receiving ART (73–75).
Rapid diagnostic tests with performance comparable to that of traditional laboratory-based serological assays are commercially available. WHO guidelines recommend the use of HIV antibody testing with a minimum sensitivity of 99% and minimum specificity of 98% (3). These assays may be particularly appropriate for use in resource-limited settings since they can be performed in clinic or community settings with minimal infrastructure. However, some concerns exist about the performance of rapid diagnostic tests, especially their ability to determine exposure and effectively exclude HIV infection at different ages (49).
Assessing HIV exposure for infants and children younger than 18 months
A diagnostic test accuracy review was conducted to explore the performance of rapid diagnostic tests as serological assays to assess HIV exposure and HIV diagnosis at different points in time (76). The four studies identified showed that the diagnostic accuracy of current commercially available rapid diagnostic tests corresponded closely with the reference standard (enzyme-linked immunosorbent assay (ELISA)) among infants aged zero to three months, when maternal antibody is detectable, with an average sensitivity of 95.4% (95% CI 89.3–98.1%) and an average specificity of 99.7% (95% CI 92.2–100.0%). Rapid diagnostic test performance after four months was lower, with average sensitivity for identifying HIV exposure dropping to 51.9% (95% CI 40.9–62.8%), likely resulting from waning maternal antibodies.
Although rapid diagnostic tests have significant potential to increasing access to and uptake of HIV testing, including in rural and remote areas, the available evidence suggests a potentially high risk that these tests will not capture HIV-exposed infants older than four months of age. Testing mothers is the best way to ascertain exposure and should be given priority whenever possible. When mothers cannot be tested, rapid diagnostic tests can be used reliably to ascertain HIV exposure among infants younger than four months of age. By contrast, when rapid diagnostic tests are used among infants and children four to 18 months of age, a negative result should not be considered as definitively excluding HIV exposure. If a child younger than 18 months is sick and the mother is not available for exposure to be assessed, a NAT should be performed regardless of the rapid diagnostic test result ().
HIV diagnosis among children older than 18 months
Five relevant studies showed that diagnostic accuracy among children older than 18 months using currently commercially available assays met existing WHO predefined standards for serology with an average sensitivity of 97.6% (95% CI 89.7–99.5%) and average specificity of 99.1% (95% CI 97.7–99.7%) (77). The risk of false-negative or false-positive results is likely to be limited and outweighed by the potential increase in uptake of testing, especially when following the national validated testing algorithms used for adults.
Implementation considerations for using rapid diagnostic tests among infants and children
Overall, using rapid diagnostic tests for infants and children will support making HIV testing available in rural and remote areas. Although the cost implications have not been formally assessed, rapid diagnostic tests are less expensive than serology laboratory-based assays (considering the total cost of testing rather than the cost of the tests alone) and likely to be cost-effective, as suggested by similar analyses conducted in the adult population (78) and on the use of rapid diagnostic tests to screen for syphilis and malaria (79–81).
Use of rapid diagnostic tests for HIV serology based on age, exposure status and breastfeeding practice.
Implementation considerations
Although the guidance for active case finding and facility-based testing and counselling among children has been in place since 2007, uptake of this recommendation has been poor. Issues around the legal age of consent and provider discomfort with disclosure have contributed to this lack of uptake, especially for adolescents and older children. A recent study in six primary clinics in Zimbabwe identified several other factors, including a perceived lack of importance attached to testing older children and a sense that testing was not warranted if children were asymptomatic (10). Lack of time and reagents and discomfort with approaching male caregivers were also noted as reasons for not testing. In addition, a WHO survey of health-care workers, policy-makers and programme managers from 17 countries found that almost half the respondents felt that testing children in immunization clinics would either be easy or very easy to do, suggesting that this policy is highly feasible to implement. Experience from countries that have been trying to roll out facility-based testing and counselling for children highlights the importance of thorough linkage to care and services for children who are exposed or infected. Linkage to care may be easier for children in inpatient settings than for those in busy outpatient clinics. The negative impact of HIV testing on the uptake of other essential childhood interventions, such as immunization, has been cited as an argument against integrating testing in immunization clinics (11). A study in the United Republic of Tanzania showed that, although integration of HIV testing resulted in an increase in immunization rates in urban centres, there was a decrease in rural facilities, possibly reflecting higher levels of stigma in rural communities (12).
2.8.6. Minimizing false-positive results by introducing an indeterminate range for infant diagnosis when using NAT
Recommendation (2018)
An indeterminate rangea of viral copy equivalents should be used to improve the accuracy of all nucleic acid–based early infant diagnosis assays
(strong recommendation, moderate-certainty evidence).
Source: Updated recommendations on first-line and second-line antiretroviral regimens and post-exposure prophylaxis and recommendations on early infant diagnosis of HIV: interim guidelines: supplement to the 2016 consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection (25)
- a
Indeterminate range: a range of viral copy equivalents that would be too low to be accurately diagnosed as HIV infected. The indeterminate range suggested is currently estimated to be approximately equivalent to a cycle threshold of 33 on the Roche COBAS® Ampliprep/COBAS® TaqMan® HIV-1 Qualitative Test v2.0 assay.
Indeterminate range cycle threshold equivalents of current nucleic acid infant diagnosis assays.
Background
In 2016, WHO recommended that HIV virological testing be used to diagnose HIV infection among infants and children younger than 18 months of age and that ART should be started without delay while a second specimen is collected to confirm the initial positive virological test result (3). Confirmatory testing is critical because of the risk of false-positive results, potential contamination with maternal blood (and virus), specimen mislabelling or mix-up, laboratory or cross-sample contamination and an observed trend of low detection of HIV among both mothers and infants because of increased exposure to maternal treatment and enhanced prophylaxis (82). The potential for false-positive results is of particular concern in settings in which the mother-to-child transmission rate is less than 5%, since the positive predictive value of highly sensitive nucleic acid–based technologies may decrease to nearly 70% (83). However, in some countries in sub-Saharan Africa, less than 10% of infants with an initial positive test result receive a confirmatory test, potentially resulting in a significant proportion (12.5%) of infants starting lifelong treatment unnecessarily.
Although a variety of causes may result in a false-positive test result, most infants with false-positive test results have low levels of viraemia; however, guidance is limited on how to interpret low levels of viraemia detected in early infant diagnostic assays. Test results that the nucleic acid–based technology reports as detectable are generally considered to be positive, relying on the thresholds of detection provided by the manufacturers. To ensure that infants do not start lifelong treatment unnecessarily, various approaches have been considered. Guidelines in the United States of America suggest that infants should not be considered HIV-positive unless they have the equivalent of at least 5000 viral copies/mL (84), and South Africa has introduced an indeterminate range that requires further testing before a definitive diagnosis is provided and treatment is initiated (85).
Previous WHO guidelines do not specifically address this growing concern about false-positive test results. Further, there is currently no specific recommendation on what level of viraemia should be considered a true positive result among infants.
Supporting evidence
Systematic review
A systematic review of 32 studies using any indeterminate range found 14 753 non-negative test results, of which 2436 (16.5%, 95% CI 15.9–17.1%) were classified as indeterminate (86). One study reporting the final diagnoses of indeterminate cases found that 76% of infants with an initial indeterminate test result were negative on retesting, suggesting that most infants were not HIV-infected despite the initial non-negative test result. These data indicate that, in countries not implementing an indeterminate range to manage early infant diagnosis test results, up to 12.5% of non-negative results could be falsely positive on initial testing, and those infants could potentially start lifelong treatment unnecessarily.
The optimal indeterminate range is considered to be the approximate equivalent of a cycle threshold of 33 on the Roche COBAS® Ampliprep/COBAS® TaqMan® HIV-1 Qualitative Test v2.0 assay. This represented a balance between the proportion of infants living with HIV that would be incorrectly identified as indeterminate (about 8–13%) and the proportion of HIV-uninfected infants that would potentially start treatment unnecessarily (about 2–7%). The cycle threshold values vary by the assay used and cannot be directly applied between technologies or assays. Because true-positive infants with low levels of viraemia generally have less rapid disease progression and need to be followed up until final diagnosis is ascertained, the Guideline Development Group determined that having a higher proportion of true-positives incorrectly classified as indeterminate than false-positives is more acceptable, since all indeterminate test results should be followed up and repeat tested as soon as possible before lifelong treatment is administered.
Implementing an indeterminate range will support more accurate nucleic acid–based early infant diagnosis. Fewer HIV-negative infants will probably start unnecessary lifelong treatment, since most false-positives will fall within the indeterminate range rather than being identified as positive. This will limit confusion and challenges in interpreting potential subsequent discordant test results if the infant was classified as positive and already initiated treatment. Finally, in addition to reducing unnecessary treatment by limiting false-positive results, an indeterminate range will promote increased attention to confirmatory testing of all non-negative test results and trigger corrective action to minimize potential contamination at the point of collection or in the laboratory. See .
Possible harm identified includes the potential requirement for additional specimens, which could result in delays to treatment initiation and the associated risk of loss to care for HIV-positive infants with indeterminate results on the first sample. However, infants with low levels of viraemia (the small proportion of HIV-positive infants who would fall within the indeterminate range) are expected to progress more slowly to morbidity and mortality (87). Further, implementing an indeterminate range may reduce the confidence of health-care workers in early infant diagnosis testing programmes if high rates of resampling and retesting are required.
Cost and cost–effectiveness
Implementing an indeterminate range has been determined to save costs since minimum additional resources will be required to retest all non-negative specimens, especially those with an initial indeterminate test result, compared with the cost of unnecessary lifelong treatment.
A cost–effectiveness model compared the standard of care (no indeterminate range) to a variety of indeterminate range options and concluded that implementing an indeterminate range is far more effective than the standard of care across a variety of viral ranges (88). Since the prevalence, positivity and mother-to-child transmission rate at each testing time point decrease, the cost–effectiveness of an indeterminate range increases and saves more costs than no indeterminate range.
Equity and acceptability
Implementing an indeterminate range may improve equity by ensuring that HIV-negative infants do not start lifelong treatment unnecessarily. This may also enable access to available treatment for other infants correctly identified as HIV-positive.
A survey of values and preferences that included people living with HIV, health-care workers and programme managers found an indeterminate range to be highly acceptable. Most respondents preferred that the meaning of an indeterminate test result be clearly explained to mothers and other caregivers. The primary concern for all groups was the potential for confusion arising from inadequate explanations about the need to resample and retest the infant. However, as long as clear guidance on the meaning of an indeterminate test result is provided to mothers and other caregivers, there will be no uncertainty about the acceptability of implementation.
Feasibility
Implementing an indeterminate range is expected to be feasible, especially if indeterminate test results are repeat tested using the same specimen. In the survey of programme managers, slightly more than half indicated that their country already has a written standard operating procedure for requesting a second specimen for invalid test results. However, there were some concerns regarding the additional time required for repeat testing all indeterminate test results and the need for storing specimens at the laboratory.
To ease the acceptability, feasibility and implementation of an indeterminate range, a standard operating procedure has been developed to support and guide countries based on expert opinion and values and preferences surveys (24,89). This standard operating procedure suggests that all indeterminate tests be repeat tested on the same specimen, if and when available. Most indeterminate test results are expected to be resolved with a repeated test on the same specimen, which would alleviate the need for and delay incurred in requesting a new specimen from the infant (82). If the same specimen cannot be repeat tested, then a new specimen should be requested and tested as quickly as possible. Repeat testing of the same sample may not be possible with point-of-care or near point-of-care technologies when the sample is directly applied from the heel to the cartridge; however, in such instances a new sample should be taken and immediately tested to confirm a positive test result.
This recommendation to use an indeterminate range to support more accurate diagnosis of infants should be implemented for any nucleic acid–based test performed for infants younger than 18 months, including testing at birth and at six weeks of age.
Implementation considerations
Generally, early infant diagnostic assays measure the presence of virus using real-time nucleic acid–based technologies that often report cycle thresholds. The cycle threshold – the polymerase chain reaction (PCR) cycle when amplification is observed – is inversely correlated with the amount of virus in a sample.
Based on the meta-analysis and cost–effectiveness modelling, the Guideline Development Group assessed that detecting an approximate equivalent of a cycle threshold of 33 on the Roche COBAS® Ampliprep/COBAS® TaqMan® HIV-1 Qualitative Test v2.0 assay would be the most appropriate value to balance the risks and benefits of introducing an indeterminate range. Note that the cycle threshold values vary by assay used and cannot be directly applied between technologies or assays. Further, additional consideration may be necessary for countries using plasma as a sample type for infant testing rather than whole blood or dried blood spots, since the latter sample types typically capture and amplify intracellular nucleic acids that may increase detected viral levels.
Research gaps
Research priorities regarding using an indeterminate range include the need for more detailed evidence on the impact of implementing an indeterminate range in populations with increased drug exposure and enhanced infant prophylaxis, the time of testing (earlier testing near birth), various sample types, differences in prevalence and different virus subtypes. More research would be valuable on the best messaging for health-care workers and mothers and other caregivers and on the optimal standard operating procedure for indeterminate test results. Further, understanding the feasibility of implementing an indeterminate range with all available nucleic acid–based technologies for early infant diagnosis and in various programmatic settings will be critical.
Managing indeterminate test results: standard operating procedure. a Please refer to chapter 3 for further details on postnatal package of care. b Do not report as positive or initiate ART but maintain prophylaxis in accordance with current guidance.
2.8.7. Managing discordant results and treatment interruption
Since 2010, WHO has recommended initiating infants on ART after an initial positive NAT, while simultaneously collecting a confirmatory sample. The 2016 WHO consolidated HIV guidelines (3) suggest that if the second (confirmatory) NAT is negative, a third NAT, either qualitative or quantitative (viral load), should be performed before considering interrupting ART. The introduction of an indeterminate range should potentially reduce the number and proportion of infants with discordant test results (different NAT results on separate samples); however, guidance on how to conduct treatment interruptions is needed.
Several factors should be considered when assessing people for ART interruption after discordant test results (positive then a negative result) are followed by a third test with a negative result:
the infant ought to have no clinical signs or symptoms suggesting
HIV infection (
49);
a follow-up plan should be agreed on with the family, caregiver(s) and health-care staff; and
tracking information (phone, address, etc.) of the family and caregiver(s) should be collected and confirmed.
The following factors should be considered when following up any infant undergoing treatment interruption.
Active follow-up is needed to ensure that a potentially infected infant is retained in care and reinitiates treatment if viral rebound occurs.
Viral rebound among infants living with
HIV starting treatment early is expected to happen within eight months of interruption in >99% of cases (
90).
Infants who develop signs and symptoms indicating
HIV infection should undergo immediate testing.
Breastfeeding and continued risk of transmission require follow-up and appropriate testing throughout the period of risk until final diagnosis.
There is value in minimizing follow-up testing by leveraging existing opportunities for infant testing (based on the national infant testing schedule and immunization or well-child appointment schedules) until final diagnosis is ascertained.
Few countries have existing policies on how to interrupt treatment among infants with discordant test results. South Africa, for one, has implemented policies with intensive laboratory and clinical follow-up of these infants for 18 months (85). Both early infant diagnosis (qualitative) and viral load (quantitative) tests are performed at four weeks and three months and every three months after treatment interruption. However, since the likelihood of these infants being HIV-infected is low, a less aggressive eight-month approach is also reasonable to simplify the follow-up procedure: this is supported by emerging evidence on the timing of viral rebound among infants living with HIV treated early (55). In this case, both early infant diagnosis (qualitative) and viral load (quantitative) tests could be performed at four weeks, four months and eight months after treatment interruption (). Infants who test positive on any follow-up test in either protocol should reinitiate treatment in accordance with current guidelines (3), and a confirmatory sample should be taken.
Any standard operating procedures for interruption should be implemented considering the continuous risk of transmission resulting from breastfeeding and, once the intensive follow-up is completed (eight months after treatment interruption), the national infant testing schedule for HIV-exposed infants should be applied to ensure an appropriate final diagnosis. If breastfeeding has stopped before the end of the intensive follow-up, final HIV status can be defined with NAT performed at least six weeks after breastfeeding ends, as indicated in , scenario b.
Managing discordant results and treatment interruption. Early infant diagnosis and viral load at 4 weeks, 4 months, and 8 months after interruption
Research gaps
Several critical research gaps need to be addressed to fully inform the implementation of infant testing strategies. The impact of maternal treatment and infant prophylaxis may need to be assessed as drug exposure increases and vertical transmission decreases. Further, determining and evaluating the most effective approaches to retaining infants throughout the exposure period until final diagnosis will be critical. Tracking tools can ensure effective tracking of infants throughout the exposure period and including those negative at previous testing time points.
More experience and data are needed to assess the impact of adding virological testing at birth on the successful initiation of newborn ART, infant outcomes and uptake of virological testing at six weeks. This would also include the added value of integrating birth testing with BCG vaccination. The feasibility and acceptability of virological testing at birth also need to be further explored in the context of national programmes at different prevalence settings and in different epidemic contexts as well as for high-risk infants only.
The frequency of testing during breastfeeding and weaning should be explored to enhance early diagnosis in this period. This should be complemented to include the optimal timing and frequency of viral load testing of pregnant and breastfeeding women.