Abstract
Scientific evidence showing the benefits of early initiation of antiretroviral therapy (ART) prompted World Health organization (WHO) to recommend that all persons diagnosed as HIV positive should commence ART irrespective of CD4 count and disease progression. Based on this recommendation, countries should adopt and implement the HIV “Treat All” policy to achieve the UNAIDS 90-90-90 targets and ultimately reach epidemic control. Attaining this goal along the HIV treatment cascade depends on the laboratory to monitor progress and measure impact. The laboratory plays an important role in HIV diagnosis to attain the first 90 and in viral load (VL) and HIV drug resistance testing to reinforce adherence, improve viral suppression, and measure the third 90. Countries in the Caribbean region have endorsed the WHO HIV “Treat all” recommendation; however, they are faced with diminishing financial resources to support laboratory testing, seen as a rate-limiting factor to achieving this goal. To improve laboratory coverage with fewer resources in the Caribbean there is the need to optimize laboratory operations to ensure the implementation of high quality, less expensive evidence-based approaches that will result in more efficient and effective service delivery. Suggested practical and innovative approaches to achieve this include: (1) targeted testing within HIV hotspots; (2) strengthening sample referral systems for VL; (3) better laboratory data collection systems; and (4) use of treatment cascade data for programmatic decision-making. Furthermore, strengthening quality improvement and procurement systems will minimize diagnostic errors and guarantee a continuum of uninterrupted testing which is critical for routine monitoring of patients to meet the stated goal.
Introduction
O
These targets required the use of HIV testing to diagnose persons and measure the first 90; CD4 cell counts to determine when to initiate ART and measure the second 90; and VL to determine if a person was virally suppressed and measure the third 90. 2 This was followed in October 2015 by WHO guidelines that recommended initiation of ART for all individuals living with HIV, irrespective of age and CD4 cell counts. 3 The WHO June 2016 guidelines further reaffirmed this, which led to the slogan “Treat All” 4 ; implying that all HIV-positive persons should be placed on ART.
The Caribbean region, like other regions and countries in the world, is committed to HIV Treat All. This approach was discussed and endorsed during three major regional meetings that brought together HIV Program Managers and various regional and international partners. The meetings include the: (1) First Latin American and Caribbean Forum on the Continuum of Care, Mexico City, May 26–28, 20145; (2) Second Latin American and Caribbean Forum on the Continuum of HIV Care, Rio de Janeiro, Brazil, August 18–20, 20156; and (3) Pan Caribbean Partnership Against HIV and AIDS (PANCAP) meeting on strategies to implement the new WHO 2015 consolidated guidelines, Port of Spain, Trinidad, March 15–17, 2016. 7
Collaborative efforts between governments and local and international partners in the Caribbean region resulted in tremendous improvements in laboratory services and systems. 8,9 However, significant gaps remain to be addressed to ensure effective testing and clinical monitoring coverage to meet the laboratory requirements of HIV Treat All. For example, 2015 data from Jamaica and Guyana show VL coverage of only 67% and 77%, respectively. 10 Although governments have committed to implementing The WHO Treat All recommendation, there is concern in the region on how this will be accomplished given limited resources, particularly regarding support for laboratory and other services. 7,11 –14 Laboratory gaps are rate-limiting factors that have been shown to affect efficiency in many settings. 12,14 In the midst of diminishing internal and external resources to support HIV programs, and the need for countries to do more with less, optimization of various laboratory operations, including the use of innovative and targeted models, will be key to ensuring improved efficiency to support the HIV response in the region. 15 To close these gaps, the use of strategies that measure and compare the output and costs of laboratory operations will be key.
In this review article, we examine the current state of preparedness of Caribbean laboratories to support implementation of HIV Treat All, identify gaps and propose strategies that countries could implement to improve efficiency, reduce cost of laboratory operations, and make more funds available to be used in achieving the stated goal.
The Caribbean Regional Context
The HIV prevalence among countries in the Caribbean is estimated to be 1.10% (0.9%–1.2%), with the highest prevalence of 3.2% found in the Bahamas. 16 The HIV epidemic in the region is concentrated mostly among key populations where one in three men who have sex with men (MSM) in Jamaica is HIV positive. 17 The Caribbean region is made up of an archipelago of island nations with cultural similarities, geographical proximity, and population mobility. This has promoted the concept of a shared relationship and establishment of regional bodies to address common issues. For example, the PANCAP is a regional body charged with the coordination of HIV and AIDS responses among countries in the region. 12 Laboratory services and systems in the region operate under the Caribbean Laboratory Referral Network with a Secretariat at the Caribbean Public Health Agency (CARPHA). Hence, this article addresses challenges and describes activities from a regional perspective that takes advantage of existing structures.
HIV Treat All and the Caribbean Region
Countries within the Caribbean region are well positioned to adopt HIV Treat All. A relatively low HIV burden per country provides a comparative advantage in terms of cost to enroll all people living with HIV/AIDS (PLHIV) into ART compared to countries with a much larger burden and fewer resources. The urgency for the region to move in this direction was recently emphasized by a regional ART outcome study that documented a high mortality rate associated with late ART initiation and recommended earlier ART initiation to avert this. 18 In fact, Brazil, which has a similar geopolitical structure to the Caribbean, became the first low- and middle-income country to adopt Treat All. 19 In addition, the regional response to HIV, coordinated by PANCAP, provides a platform for networking and leveraging resources available in other countries within the region. For example, published data from Barbados, one of the countries within the PANCAP network, show that 70.3% of patients on ART who had a VL result were virally suppressed in 2011. 9 Lessons learned from such a program could benefit other countries. In addition, Barbados has a highly functioning laboratory sample referral system that could benefit other countries in the region that have less laboratory capacity in expensive molecular testing. 8,9,20
In an effort to achieve sustainability, some Caribbean countries have increased domestic funding for several HIV prevention, care, and treatment services. 10 Despite their resolve to adopt Treat All, there is concern on how to achieve the stated goals with decreasing financial support from donors. For example, financial support from the U.S. Government for health initiatives in the region declined 31% from 2011 to 2014. 21 In light of this, the region needs to develop approaches that optimize HIV program service delivery to improve efficiency and reduce cost of operations.
Improve HIV Treatment Cascade Data to Identify Laboratory Gaps
The current HIV treatment cascades mapped to the UNAIDS 90-90-90 targets provide a framework to assess the effectiveness of a country's HIV prevention, care, and treatment program in a one stop shop. 22 These cascades can help determine laboratory gaps in a program. 22 Figure 1 is a hypothetical example of a country's HIV treatment cascade. It shows a gap between the presumed number of people who are HIV positive and those who know their HIV status to meet the first 90%. There could be many reasons for this gap, including laboratory gaps in access to quality HIV testing. 23 The cascade also shows a large drop in the number of patients on ART and those who are virally suppressed. This could be due to low VL testing coverage for ART patients (selective testing of patients failing therapy), leading to underestimation of viral suppression. Conversely, where VL testing coverage is high and a significant number of patients on ART have detectable VL results, other factors would explain the observed levels of viral suppression. In either case, there is a potential for misinterpretation given the way VL data are presented in the cascade. In addition to showing the number of patients on ART, the cascade should show the number of ART patients with VL results to avoid misinterpretation. Such data will also show laboratory gaps between the second and third 90s, providing an opportunity for targeted interventions to address these gaps. It has been difficult to determine what the gaps are in laboratory services in most countries in the Caribbean by analyzing treatment cascade data because some of the cascades are incomplete. 13,14 As Caribbean HIV programs consider options to improve efficiency, treatment cascade data will need to be updated and accurate. Closing gaps will require program improvement, innovations, and evidence-based interventions along the cascade. 24

A hypothetical example of a country's HIV treatment cascade showing gaps between each of the UNAIDS 90-90-90 targets. Ensuring completeness of data in this cascade is necessary as it can guide in identifying programmatic gaps to be addressed.
More Efficient HIV Testing Strategies to Reach the First 90
The HIV epidemic in the region is concentrated among key populations 25 that are hard to locate because of a high level of stigma and criminalization. 26 Under these circumstances, it is necessary to consider more innovative, efficient, and cost-effective approaches to reaching and testing them to attain the first 90. However, it is not uncommon for HIV programs in the region to test thousands of people without identifying a single new positive case (i.e., poor yield). This suggests that programs are not reaching undiagnosed positives or testing is not occurring within the right geographic locations or hotspots. This may contribute to late diagnosis and poor treatment outcomes. Evidence shows that about 40% of PLHIV in the region are diagnosed late with CD4 cell count below 200 cells/mm3. 13,14 The mass testing strategy described above is inefficient because it requires substantial time and human and financial resources. Hence, geographic prioritization, targeting specific populations, and innovative testing strategies such as partner notification and index testing 27 should be considered to improve testing yield. Attention should also be paid to the quality of testing to avoid misdiagnosis 28 –31 particularly in settings and populations with low HIV prevalence. Access to laboratory capacity for early infant diagnosis (EID) through the regional laboratory sample referral system 8 should be strengthened to ensure acceleration toward the first 90 among all age groups.
Scaling Up VL Testing to Reach the Third 90
Access to at least one VL result within the first 6 months of initiating ART is critical for clinical decision-making, and most importantly, to identify patients who are virally suppressed (VL <1,000 copies/ml). High rates of viral suppression are indicative of a successful ART treatment program. 1,32 Currently, most countries in the Caribbean have access to VL either nationally or through referral support from neighboring countries with that capacity. 8,20 However, a major challenge is VL access and coverage. Data show VL coverage rates for patients on ART of 67% in Jamaica, 77% in Guyana, and over 90% in Barbados. 10 Recent assessments show that some key challenges associated with low VL coverage in the region include prolonged turnaround time for results, poor procurement systems, machine downtime, weak VL sample referral network, and poor data management systems. 12 –14 This low coverage will result in low rates of viral suppression as few persons will have VL results. Viral suppression among all ART patients in select countries in the Caribbean ranges from 43% in Jamaica, 67% in Guyana, and 70.3% in Barbados. 8,10 This is substantially lower than the third 90 target. 2 There is the need to consider better strategies to improve VL sample referral networks and coverage to measure viral suppression. The WHO guidelines on implementing VL testing 30 and other recent approaches to improve testing efficiencies 33 are excellent tools to guide this process.
VL Counseling
VL counseling raises awareness on the importance of VL testing. This occurs at various levels, targeting different individuals. In particular, patients initiating ART for the first time should go through VL counseling to understand the importance of VL results for their treatment monitoring, when their samples should be collected, and how to interpret their VL result (e.g., what to do if their VL is not decreasing after a certain period on ART). 34 Patients engaged in VL counseling have been reported to do well and have higher rates of viral suppression than those who did not. 35 When patients are not virally suppressed within 6 months on ART, the recommendation is that they go through enhanced adherence counseling (EAC). 36,37 It is reported that only about 45% of patients who initiate ART achieve viral suppression after 3 years, as many are lost to follow-up. 38 In addition, a systematic review showed that 70.5% of patients not virally suppressed, re-suppressed following EAC. 37 Hence, patients within the Caribbean country programs with low viral suppression 9,10 are expected to benefit from EAC. A recent review article by Barrow and Barrow on HIV treatment as prevention in Jamaica and Barbados identified testing, ART coverage, and treatment adherence as key programmatic gaps. 39 Strategies to enhance VL counseling among ART patients in the region are important for ensuring improved viral suppression. This should include capacity building for clinicians, laboratory staff, and other healthcare staff so they understand the importance of VL testing, changes to CD4 testing guidelines, and how to interpret results.
CD4 and VL Dynamics
Over the years, CD4 has been used as the main clinical laboratory parameter for making decisions to initiate ART and monitoring treatment. However, a recent recommendation is to use VL only to monitor stable patients on ART, while CD4 should be used only for patients presenting late to care and where VL testing is not available. 40 Evidence shows that in situations where VL testing is available and patients are virally suppressed, CD4 testing should be scaled down or more targeted. 41 Scaling down CD4 in this situation can result in significant cost savings for laboratory operations. In fact, discontinuing routine CD4 testing for patients who were virally suppressed in the South Africa National HIV treatment program was expected to save about U.S. dollar 68 million. 42 In light of this, HIV treatment programs in the Caribbean could consider scaling down CD4 and scaling up VL testing to monitor stable patients. Cost savings from this strategy could be used to support expanded VL coverage.
Strengthening Regional Capacity for HIV Drug Resistance Testing
Drug-resistant HIV can be transmitted at the time of infection or be acquired due to poor adherence to therapy. Recent data from more than 50,000 patients on ART in 111 countries showed a median prevalence of transmitted HIV drug resistance of 2.8% in sub-Saharan Africa and levels that reach 7.6% in Latin America and the Caribbean. 43 Hence, primary HIV drug resistance may pose more serious treatment challenges in the Caribbean region than in other areas. In fact, genotypic data from one of the HIV treatment sites in Jamaica showed that 12.6% of treatment-naive patients had clinically significant mutations. 44 Although these data are limited to one site, it is possible that this may mirror the situation at other sites in Jamaica. The Centers for Disease Control and Prevention (CDC) in collaboration with the Jamaica MOH is currently conducting an HIV drug resistance surveillance study in Jamaica using the WHO approved guidelines. 45,46 Additional regional efforts are under way to support HIV drug resistance surveillance in other countries. 46 This includes validation of genotypic platforms to support routine testing for patient monitoring. 47 Sustainability of these services is critical for the efficient functioning of HIV drug resistance surveillance and treatment programs in the region.
The median 12-month ART retention rate is 75% in the Caribbean, with rates as low as 53% in some countries. 10 The region has also observed high rates of attrition or “lost to follow-up” where only 50% of patients are retained in care following entry into care. 10 ART patients who have stopped therapy are at risk of developing HIV drug resistance, which has implications for follow-up regimens. 43,45 As the region implements strategies to reengage lost ART patients, there is the need to ensure availability of HIV drug resistance services to test and make informed clinical decisions that will positively impact treatment outcomes.
Laboratory Data Management Systems
Accurate collection, analysis, and presentation of laboratory data are key to assessing the effectiveness of program implementation. Poor quality data contribute to poor decision-making. For example, an HIV-negative person may be placed on ART because of misdiagnosis. Two of the three UNAIDS 90-90-90 targets along the Continuum of HIV treatment cascade depend on laboratory-generated data. 24,45 The two targets include the first 90 (know status) and third 90 (viral suppression). Adequate generation and use of laboratory results is still a big problem because of weak data collection systems. Challenges with interoperability between laboratory and treatment sites hamper efficient data generation, reporting, and use. This gap results in outdated or incomplete data for the treatment cascades of many countries. This has resulted from programs not being able to link patient's laboratory results with their treatment record to quantify retention along the continuum. 13,14 Without appropriate data system in place, distinguishing number of tests performed from actual number of people tested, which could guide in monitoring number of samples rested, could be a challenge. There is a need to use affordable laboratory information systems, particularly ones that have an interface between the laboratories and treatment sites. The incorporation of a unique patient identification code across all systems would minimize challenges with samples and mismatched results, as well as improve data sharing at the programmatic and national level.
Intercountry and Intracountry Sample Referral Systems
Functional laboratory sample referral networks have been shown to improve program efficiency by decreasing turnaround time and cost of operation and increasing the number of samples collected and tested. 48 Past collaboration between partners and host countries in the region led to the establishment and operationalization of the Caribbean Regional Laboratory Referral Network that supports the collection and transportation of samples from countries with less capacity to those with higher capacity for testing. 8 Countries within the Organization of Eastern Caribbean States (OECS) without molecular testing capacity have used this system to collect and ship samples to Barbados for testing and reporting of results for VL and EID within the stipulated turnaround time. 8,20 This intercountry laboratory referral system has worked well and should be sustained. However, there is also the need to review and strengthen national or intracountry laboratory referral systems in countries with functional molecular testing capacity to support the collection and transportation of samples from treatment sites to the laboratories for testing. Such an approach should start with site assessments to determine needs and challenges. This would include mapping all HIV care and treatment sites, determining which laboratories have the capacity for VL testing and other key laboratory parameters. Sample referral routes should be selected based on proximity and logistics rather than for administrative or political reasons, as the most important consideration is to improve efficiency and reduce the costs of the referral system. For example, Jamaica has four Regional Health Authorities (RHA) with several HIV treatment sites; however, VL testing occurs only at the National Public Health Laboratory (NPHL) in Kingston. 13 A sample referral network within this country is needed to ensure the collection and transportation of VL samples to the NPHL for testing. Trinidad and Tobago has five RHA with five treatment sites and two laboratories that offer VL testing. 14 A VL sample referral system in this country should consider the most effective approach in mapping treatment sites to these laboratories. Once laboratories with capacity to support VL testing and other HIV care and treatment parameters, including CD4, are mapped out, other considerations to support efficient functioning should be strengthened. To improve VL and EID sample collection, packaging, and transportation, the region should consider the use of dried blood spots as an option. 49 Also since this involves movement of samples across borders, government policies that will enhance material transfer agreement across borders should be developed.
Laboratory Quality Improvement to Minimize Misdiagnosis and Attain Accreditation
Improving the quality of laboratory testing to reduce error and ensure efficient delivery of services is a critical but often neglected aspect of global laboratory system strengthening. 50,51 Literature points to several instances of HIV misdiagnosis. 52 Although data on HIV misdiagnosis are scarce in the Caribbean region, there was a published case in which a woman was erroneously diagnosed HIV positive and lived with this condition for over 2 years before she was correctly diagnosed as HIV negative. 53 The WHO in collaboration with other partners has stressed the medical, ethical, and social impact of misdiagnosis on delivery of health services and has made recommendations and produced guidelines to assist countries in improving their HIV diagnostic pathway. 27,29,30 Some examples include the use of innovative quality improvement tools such as the HIV Rapid Testing Quality Improvement Initiative (HIVRTQII), 27 following the WHO 2015 HTS guidelines 27 and retesting all presumed HIV positive persons before commencing ART. 30 Misdiagnosis and attendant consequences have the potential to seriously affect the efficiency of HIV treatment programs. As countries in the Caribbean consider moving toward Treat All, efforts should be made to adhere to these guidelines and recommendations.
Apart from assuring the quality of HIV diagnosis, quality testing is also important in other aspects of the laboratory. Hence, there is a need for overall laboratory quality improvement and accreditation of all sections of the laboratory, as this is the only benchmark that ensures that testing and quality are occurring according to international standards. 54 Reports have shown that laboratory quality improvement leads to reduction in the cost of testing by avoiding repeated testing and improving stock management. 51 In addition to improving the efficiency of laboratory testing, this cost saving can make additional resources available to improve program coverage and impact. In 2008, there was no accredited public laboratory in the Caribbean region despite the fact that about 90% of patients used public facilities. 8 However, collaboration with governments and global partners has led to the accreditation of several laboratories, as well as the establishment of the Caribbean Laboratory Quality Management Systems Stepwise Improvement Process (LQMS-SIP), a framework for implementing sustained quality systems and achieving accreditation in a stepwise manner. 8,55 These efforts should continue to ensure the availability of quality laboratory services to support the region's HIV response. In particular, ensuring the availability of quality assured and timely data will be key to constructing HIV treatment cascades that truly reflect the current situation, which is sine qua non to program improvement and impact.
Regional Procurement and Equipment Maintenance Challenges
Poorly managed procurement systems leading to waste and interrupted testing will affect the efficiency of HIV treatment service delivery, as well as increase the cost of laboratory operations. 55 Procurement issues are recurrent in almost all programs. However, careful identification and resolution of challenges could lead to tremendous improvement. A recent review showed continued procurement challenges within the Caribbean region. 11 This mirrors reports from the PAHO Treatment 2.0 visits to some countries in the region that linked weak procurement systems to poor HIV program implementation. 13,14 There is a need to address this through appropriate advocacy and policy change. Providing staff with training and tools on forecasting and stock management will be critical to minimize stock out and testing interruption. VL cost per test is still expensive in the region ranging from 40–50 U.S. dollars, which may limit program coverage. The region should explore other options to reduce the cost of reagents and consumables. This includes tapping into the UNAIDS/Roche Diagnostic Access Initiative 56 and Global Fund Framework Agreements that allow countries to access certain products at a discounted cost. Consideration should also be given to commodity standardization where there is planning and agreement to procure common items that address the regional context and the needs of small island nations. For example, it is easier to do a bulk procurement, share reagents and consumables between laboratories, and use the same engineers to service machines in different laboratories and countries. This will significantly improve laboratory performance, as well as reduce the cost of operations. 57
Point of Care Diagnostics
Several studies have demonstrated that decentralized community-based point of care (POC) diagnostics improve efficiency, reduce cost of operation and turnaround time, and promote retention in the HIV continuum of care. 28,58,59 Many HIV rapid testing and CD4 POC testing platforms are currently being used to support HIV programs. Improving access for key populations (MSM, female sex workers, and transgender) in the Caribbean region to improve testing yield and attain the first UNAIDS 90 will require extensive use of HIV POC testing. Because these individuals are stigmatized and criminalized, they are less likely to access centralized health services. 25,26 The use of POC testing to improve clinical monitoring and retention of patients on ART has been reported in other settings, including the Caribbean region. 59,60 These platforms should continue to be supported in the Caribbean region. WHO has approved the use of POC for VL and EID that are currently being validated in many countries. 61 Countries in the Caribbean should participate in these validations to ensure that the region is among those to use VL and EID POCs to support the Treat All initiative. Challenges have emerged in the use of POC platforms, particularly at the level of quality of testing and appropriate collection and use of data for decision-making. 62 The region should ensure that these issues are addressed as new POC platforms are rolled out.
Human Capacity Building and Retention of Trained Laboratory Personnel
A trained and competent laboratory work force is critical to ensure short and long-term sustainability of programs in the region. This could be achieved through strengthening public health and laboratory institutions in the region that will eventually provide the needed human capacity building training. Governments in the region should continue to invest in institutions such as the University of the West Indies and Medical Laboratory training schools located in the different islands. There is also the need to engage international partners, as well as create Public Private Partners 50 with manufacturers, to ensure that resources are available to support these initiatives.
Monitoring and Evaluating to Determine Efficiency
Working smarter, and not harder, will be best realized if the efficiency of every program is measured. Efficiency, the effectiveness of program operations, can be measured by comparison of outcomes with costs. If the cost outweighs the outcome in relation to what is obtained from other programs, then the program is not efficient, and other cost-effective measures should be explored (Fig. 2). The period of measurement is also critical to avoid continued investment in unproductive or inefficient activities. This could be achieved by considering sequential process, outcome, and impact evaluations to determine achievement of goals and objectives during the entire laboratory operation. In particular, a goal of all laboratories should be to eliminate inefficiencies. This should be accomplished through periodic evaluation of laboratory services and systems and implementation of corrective actions as needed. Such an approach will reduce the cost of operations, as well as make additional funds available to increase program coverage. The laboratory should be included in health economics analysis that compares different service delivery models. This can further provide information on how efficiently the laboratory is functioning.

Proposed cost-efficiency model to support improved VL coverage. Assume the budget (A) is fixed for VL testing then increasing coverage (G) will be achieved by maximizing efficiency. A strategy with a high testing output (D) and few individuals tested (E) would not be considered efficient. An efficient testing service will be achieved when the mean number of tests per person (D/E) aligns with the expected ratio of tests per person. For example, if the national policy requires semi-annual VL testing of all ART patients, then a clinic with 500 stable patients would be expected to have an output of 1,000 tests (D) and 500 unique individuals tested (E)—this would translate to a ratio of 2.0. An inefficient service will have an output (D) of 1,000 or higher and a unique number of individuals tested (E) lower than 500. Using this example, if 350 unique individuals were tested, then the testing coverage (G) is 70% and the testing ratio 2.8 exceeds the expected ratio of 2.0. Improving coverage will require understanding and addressing the inefficiencies observed or cost drivers. This model can be applied to other tests. ART, antiretroviral therapy; VL, viral load.
Conclusion
To attain the UNAIDS 90-90-90 HIV treatment targets by 2020, and end the AIDS epidemic as a major public health threat by 2030, WHO has recommended that all adults, adolescents, and children diagnosed with HIV should commence ART regardless of CD4 count and disease progression (Treat All). To achieve this in the midst of diminishing health sector financing in the Caribbean region, there is a need to optimize laboratory services to improve efficiency, reduce costs of operation, and ensure the availability of more funds to support HIV testing and clinical VL monitoring coverage.
Footnotes
Acknowledgment
This research has been supported by the President's Emergency Plan for AIDS Relief (PEPFAR) through the Centers for Disease Control and Prevention (CDC).
Disclaimer
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention/Agency for Toxic Substances and Disease Registry or the U.S. Department of State.
Author Disclosure Statement
No competing financial interests exist.
