Matching Development of Point-of-Care Diagnostic Tests to the Local Context: A Case Study of Visceral Leishmaniasis in Kenya and Uganda

Step 1: Literature Review

A critical review of the existing literature on a disease and its relevant health care context is performed to gauge the disease endemicity, resource availability, and current diagnostic practices in the resource-constrained settings. Such critical analyses help in identifying the potential and the limitations of the existing diagnostic practices and in identifying stakeholders such as patients, health care staff, patient families, health care organizations, nongovernmental organizations, and government bodies, and thereby highlight the implementation needs for novel diagnostic solutions.

Step 2: Selection of an Endemic Resource-Limited Region for a Case Study

Beyond mere literature study, it is important to direct field research for a case study. Selection criteria for a fitting case should include (i) an endemic region that bears a high burden of a disease; (ii) an area that has an urgent need for POC diagnostic tests; and (iii) a politically stable conflict-free zone for safety and logistic capacity (roads and access to remote endemic areas) for field research.

Step 3: Field Research With Direct Observations and Interviews With Stakeholders

Despite an extensive literature survey, many aspects of diagnoses in remote endemic regions often remain unclear, such as the clinical algorithm for disease identification, the patient’s journey from becoming infected to getting treatment, the availability of laboratory equipment, and the level of trained personnel at the lowest most constrained level of the health care system. To obtain observations from the field for filling the knowledge gaps in the literature, a field trip to the selected endemic regions was necessary. The objective of such a field trip is to gather direct observations at various health care levels in that region and to carry out semistructured interviews in the field to obtain expert input from key stakeholders as identified from the literature survey and from advice of the locals.

Step 4: Create Gigamaps and Patient Journeys Based on Insights Gathered

To get a deeper understanding of the diagnostic practices for a disease, the assumptions made by the detailed literature survey (step 1) are analyzed in conjunction with the information that was obtained from semistructured interviews with the health care professionals (step 3). The observations are processed and organized to highlight the limitations of existing diagnostic practices. Since the health care context of a disease in resource-limited settings is complex with social and technical challenges and involves a wide range of stakeholders, understanding the corresponding diagnostic practices for a disease can be quite challenging. To deal with the complexity of such a multifaceted health care system, visual thinking tools are used to present the complex diagnostic problem at hand (define) and to identify the logical solution (ideate) in the form of use case scenarios. Visual Gigamaps are created to outline current diagnostic practices at various stages of the disease, and health care-seeking behaviors of patients from a health care provider’s point of view across different health care levels. Next, visual patient journeys are constructed within the selected disease endemic region. In our approach, these extended patient journeys represent a sequence of interactions between the patient, the health care system, and the stakeholders involved, and they consider both the technical, economic, and social factors.

Step 5: Create Use Case Scenarios

Next, we create various scenarios for the application of a diagnostic test within the selected region. Again, visual thinking is utilized to present every scenario to provide a clear understanding of the requisite diagnostic need and to suggest logical solutions.

Step 6: Validation of the Use Case Scenarios

Visualizations of the scenarios from step 5 are used to facilitate discussions with different stakeholders with an objective to critically select and thus validate the most valuable scenario through which a diagnostic test can meet the local needs of the end users and stakeholders. Thus, detailed discussions are conducted with the stakeholders on each scenario to obtain the most pressing diagnostic need and to define the priorities from the perspective of health care providers.

Step 7: Formulation of a CTPP

Finally, the stepwise approach collectively leads to the formulation of a CTPP for a diagnostic test that will be most effective for the selected disease-endemic region. The CTPP is formulated for the most urgent diagnostic need as viewed from the perspective of health care providers.

1. Literature Review on VL Endemicity and VL Health Care Context

Leishmaniases are caused by more than 20 different Leishmania species, which are parasites that can be transmitted to humans and other animals by the bite of infected female phlebotomine sand flies.^22,23 Worldwide, there are approximately 2 million new cases each year, and 556 million people are at risk of acquiring the infection.²⁴ VL (also known as Kala-azar) is the second-largest parasitic killer after malaria^25,26 and the most severe form of leishmaniasis because it affects the visceral organs, particularly the liver, spleen, and lymph nodes.²⁷ Although VL is curable, it remains a fatal disease because it is often left untreated due to its low index of suspicion by health care providers, late diagnosis, and inadequate case management, especially at an early stage in low-resource settings.²⁸ Furthermore, the initial symptoms of VL (e.g., persistent fever, weight loss, fatigue, and anemia) overlap with other febrile illnesses, such as malaria, and hence it is often misdiagnosed and treated incorrectly. If left untreated, VL is fatal within 2 years, due to severe anemia or secondary bacterial infections.²⁹ Furthermore, post-kala-azar dermal leishmaniasis (PKDL) is a skin condition that occurs after VL treatment due to persisting parasites in the skin.²⁹ Incomplete treatment is a major risk factor for PKDL.²⁹ Although PKDL lesions are typically self-healing, they pose as infectious reservoirs for sandflies²⁹ and cause aesthetic and psychological complications that affect the patient’s quality of life, especially young adults. VL is a complex disease that is further complicated by coinfections such as HIV.^28,30,31 Individuals with HIV who are immunosuppressed often present with more severe VL symptoms and require different treatment regimens.³¹ Other common coinfections in VL endemic regions include malaria³² and tuberculosis.³³ Thus VL needs to be diagnosed and managed on a case-by-case basis due to confounding conditions such as immunosuppression.¹³

VL is currently diagnosed by using either one or a combination of the following: (1) empirical clinical observations,³¹ (2) immunological rapid diagnostic tests (RDTs) and/or immunoassays, (3) molecular tests to detect the pathogen’s DNA in clinical samples, and (4) parasitological tests that require microscopic analyses of invasive splenic or bone marrow aspirations^34,31 (Figure 2). Except for clinical methods and RDTs, which can be less reliable, these diagnostic practices require expensive instruments, a stable source of electricity, a well-equipped laboratory, and an expert to operate; therefore, they are inadequate for use within resource-limited settings.^30,35 The most readily used test for VL diagnosis, especially in remote settings, is the rK39 RDT.²² However, since the rK39 RDT is an immunological test, it is less reliable because the sensitivity of the test differs between individuals,^29,36 and it cannot serve as a test-of-cure due to persisting antibodies after treatment.^30,37

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Current Methods Used for Diagnosing Visceral Leishmaniasis^a

Abbreviations: DAT, direct agglutination test; ELISA, enzyme-linked immunosorbent assay; IFAT, immunofluorescent-antibody test; KAtex, latex agglutination test; LAMP, loop-mediated isothermal amplification; PCR, polymerase chain reaction.^a After an empirical clinical observation is done, a screening test (usually, an immunological test) is administered, depending upon the availability.

2. Select a Region for a Case Study

Leishmaniases remain endemic in more than 98 countries with the majority of VL cases in South Asia (India, Nepal and Bangladesh), South America (Brazil), and the horn of Africa (Ethiopia, Somalia, South Sudan, Sudan, Kenya, and Uganda) (Figure 3).^31,38 Seven countries in particular (India, Brazil, Ethiopia, Kenya, Somalia, South Sudan, and Sudan), reported approximately 90% of the global cases of VL in 2015.²⁹ Although various VL control programs that focus on prevention and treatment are operational within these countries,²² the current rK39 RDT has been shown to have a poor performance in East Africa compared with India,^29,36 necessitating an improved diagnostic approach in East African countries.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Status of Endemicity of Visceral Leishmaniasis Worldwide³⁸

Considering the global spread of VL, our first challenge was to select a VL endemic region that was conducive for field research. Based on our selection criteria (see Methodology), we selected western Kenya and northeastern Uganda for our field research.

3. Field Research With Direct Observations and Interviews With Stakeholders

During our field trip to western Kenya and northeastern Uganda, we visited several Pokot tribal communities, health care facilities, and local organizations over 2 weeks in November 2018 (Figure 4). Our international team consisted of scientific re-searchers including a principal investigator, a postdoctoral researcher, and a PhD researcher, who are working together to develop innovative POC diagnostic tests for infectious diseases, and an industrial-design master student. Our local Kenyan team consisted of a public health officer, a research technician, and a research assistant. We engaged with county and subcounty officials, as well as with local health administrators and community health volunteers (CHVs). Our Ugandan team consisted of a medical doctor and a community health worker (CHW), and we engaged with key stakeholders such as the local chief in the Moroto district. A significant number of interviews were conducted as we followed the recommendations and advice of the locals about whom to speak to (Figure 4).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Number of Interviews Conducted with Stakeholders on Diagnosing VL in Western Kenya and Northeastern Uganda

Abbreviations: RDT, rapid diagnostic test; VL, visceral leishmaniasis.

In Kenya, VL testing and treatment can be accessed at the local health care facilities that are located in the Rift Valley region: the Kimalel health center and a newly constructed treatment facility at the Chemolingot subcounty hospital (both within the Baringo County), and the Kacheliba health center (within the West Pokot County). In Uganda, the Amudat hospital is the only VL treatment facility. VL testing can be accessed in Rupa subcounty and by a mobile CHW in the surrounding regions, and patients are referred to the Amudat hospital for further testing and treatment.

4. Create Gigamaps and Patient Journeys Based on Insights Gathered

A Gigamap of the health care system in western Kenya and northeastern Uganda was created based on the literature as well as on the insights obtained in the field. Seven phases were identified (Figure 5), from exposure to vectors at home (phase 1); being sick (passive) (phase 2); seeking care (active) (phase 3); getting diagnosed with VL (phase 4); getting to a treatment facility (phase 5); getting treatment (phase 6); and finally, to being treated and going home (phase 7). The Gigamap visualized the journey through the different levels of the health care system, starting from the rural setting (close to homesteads) and advancing towards more urban settings for treatment. The Gigamap also visualized the multiple stakeholders that are involved in the VL health care system.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Gigamap Diagram Showing the Journey of a Patient Who Has Visceral Leishmaniasis From Infection to Treatment in Western Kenya and Northeastern Uganda

Abbreviation: VL, visceral leishmaniasis.

Next, we created patient journeys (Figure 6). We related these journeys to the 7 phases defined in the Gigamap, to obtain a detailed overview of the challenges faced by a patient while seeking effective diagnoses and subsequent treatment. The 8 patient journeys (numbered as I to VIII) were a result of the information gathered in the field, and they represent a sequence of interactions between the patient, the health care system, and the stakeholders involved. A detailed description of patient story II can be found in the Supplement. Possible factors and barriers encountered during the patient journeys were identified. Several factors determined the progress of a patient through these different phases resulting in the least efficient (story II) and the most efficient (story VIII) journey from infection to the treatment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Phases of Patient Journeys That Represent the Number of Steps That Patients With Visceral Leishmaniasis Take From Infection to Treatment

An important observation during our study was the key role of the CHV/CHW as the closest link to the people in rural communities. In Kenya, CHVs and CHWs are trained on a variety of health issues including case definition, prevention, and control of common ailments, as well as nutrition and family planning. They facilitate access to health services through advocacy, outreach, referral, community education, informal mentoring, and social support.³⁹ Thus, their role is to identify patients at the homesteads (small clusters of homes) (Figure 6, phases 1 and 2) and refer them to the local health care facilities for further diagnoses (Figure 6, phase 4). In Uganda, we observed that CHWs were trained to perform VL diagnostics in the field using the rK39 RDT. Due to ongoing VL clinical trials, which raises funding availability, CHWs in Uganda also had access to a motorcycle to allow them to reach secluded homesteads.

In contrast, due to limited clinical trials and a subsequent lack of funding, we observed that CHVs in Kenya are involved in identification of patients and referring them to local health care facilities. CHVs are well-respected members of the communities, such as teachers and ministers, and are trusted by the locals. Furthermore, due to language barriers between the locals and researchers from abroad, CHWs and CHVs played a key role in most interactions with the locals, patients, and health professionals.

The distance to a health care facility for diagnoses (Figure 6, phase 4) and treatment (Figure 6, phase 6) was found to be the most significant barrier that patients face when seeking health care. Diagnosis may take place at the homesteads by a CHW/CHV or at a health care facility, while treatment would take place at a higher-level facility such as a local hospital. Once diagnosed, either at the homesteads by a CHW/CHV or at a health care facility, patients in the East Pokot subcounty in Kenya that may have VL based on the rK39 RDT, currently travel approximately 80 km to the Kimalel hospital for another (confirmation) diagnostic test and for treatment, if needed. Due to the toxicity and costs of the treatment, patients are diagnosed multiple times to ensure that treatment is only prescribed when absolutely necessary (in contrast to antimalarial treatment that is prescribed more readily). The Chemolingot subcounty hospital treatment facility in Kenya is currently improving accessibility to treatment for patients in the East Pokot subcounty. However, in Uganda, patients from across the country need to travel to Amudat for VL treatment. Interestingly, many Turkanas in Kenya travel approximately 100 km to the Amudat hospital to seek treatment due to conflicts between their tribe and the Pokot in Kenya.

Overlapping symptoms with other febrile illnesses often lead to misdiagnoses based on empirical clinical observations (signs and symptoms) (Figure 6, phases 2–7). In VL endemic regions, acute fever is often associated with malaria and other prevalent tropical diseases. Malaria RDTs and treatment are generally readily available, which encourages their use. Furthermore, a lack of VL awareness, even at the health care centers, promotes such misdiagnoses. Health education influences the ability of patients and health care workers to recognize VL. Increased VL awareness will have a positive influence on the journey of the VL patient because it increases the likelihood of recognizing VL at an early stage. Furthermore, we learned that children play in the termite mounds that are preferred resting and breeding sites for sandflies (the VL vector), and males sleep outside at night, which further exposes them to sandfly bites. Important risk factors include area ecology (humidity, heat), vegetation (acacia trees), livelihoods (pastoralism and proximity to livestock), and general behavior (outdoor sleeping).

A significant lack of resources is apparent for the diagnosis and treatment of VL, particularly in terms of health care personnel, diagnostic testing, and financial resources. Despite traveling far distances, patients have no guarantee that diagnostic procedures will be practiced immediately once they arrive at a health care facility for diagnoses. We learned that technicians generally close their laboratories for a day or two when they need to be in the field or are receiving training. Due to a general shortage of staff, diagnostic procedures are often delayed. Furthermore, RDTs such as the rK39 for VL are not always available, which promotes misdiagnosis of VL. Midwives at the Rupa health center in the Moroto district in Uganda, who routinely perform RDTs for malaria and other illnesses, informed us that the supply of RDTs for VL is not consistent. Thus, symptomatic patients that do not respond to malaria medication are then referred to the Amudat hospital (120 km) for diagnoses and treatment. Finally, the cost of traveling to a health care facility is a significant barrier, particularly when patients are asymptomatic and/or are simply not convinced that they are sick due to a lack of awareness. In general, a lack of financial resources contributes to poor health-seeking behavior (Figure 6, phase 2).

We learned that many VL patients know that they are sick (Figure 6, phase 2), but do not actively seek health care (testing and treatment) due to the inaccessibility of health care facilities and the fear of encountering other tribes that may be hostile. We also learned that the Pokot and Turkana tribes in Kenya and Uganda are often in conflict, which makes it unsafe for members of either tribe to travel freely to seek health care. Such tribal conflicts inhibit patients from seeking health care and force them to travel to more distant health care facilities that are located away from a conflict zone. Such delays in seeking health care could further worsen the patient’s health, and they are often severely anemic and weak by the time they reach a health care facility. Furthermore, patients require screening for multiple infections, including VL, malaria, and tuberculosis, which is challenging when the patients are extremely weak. Stabilizing patients before starting them on VL treatment is also challenging because they may require blood transfusions and treatment for comorbidities, such as malaria, which require urgent attention before they can start VL treatment.

We observed that in general, VL patients also seek care from a traditional healer before considering visiting a health care facility because traditional healers are trusted, located closer to the homesteads, and often alleviate some of the initial symptoms. These deferrals worsen the patient’s health, thereby making the treatment more difficult. Unfortunately, critically ill patients often do not respond to the VL treatment, resulting in death. Such cases further strengthen the traditional beliefs and set a negative impression of health care facilities in the minds of health care seekers. Due to financial constraints and the cultural beliefs of the local population,⁴⁰ the health care facility usually needs to arrange the last rites because the family does not come to the hospital to take responsibility for the deceased.

We observed a traditional patriarchal society, in which males leave the homesteads for work, and females take care of the homes and the children. For males, the loss of income due to the time that is spent traveling to seek health care contributes to poor health-seeking behavior (Figure 6, phase 2). For females, being unable to leave children unattended at home contributes to poor health-seeking behavior. We learned that women from the Pokot tribes often need permission from their husbands to leave their homes before taking themselves or a child to a health care facility. Thus, unequal decision-making power in the household contributes further toward delayed diagnoses and VL treatment for women and children.

Given all these factors, the total number of steps a patient takes to seek treatment can vary significantly. For example, seeking care in phase 3 requires many steps in complicated patient journeys, such as for patient stories I and II (Figure 6). This may be due to several of the aforementioned factors. For example, a lack of resources in the unavailability of staff or RDTs may result in the patient returning home (Figure 6, phase 2) without receiving a diagnosis, and later traveling back to seek care (Figure 6, phase 3). Traveling between phases 1 and 2 may be further hindered by other factors such as distance between the homesteads and the health care facilities or a lack of financial resources. Conversely, a patient journey may be as simple as patient story III whereby a patient progresses easily from being sick (Figure 6, phase 2) to being treated and going home (Figure 6, phase 7).

5. Create Use Case Scenarios

Prompted by our extensive methodology, whereby we mapped the health care system in a specific endemic region and analyzed patient journeys, we sketched 6 use case scenarios based on the scope of application of a new POC diagnostic test (Figure 7). These different scenarios are based on considerations involving 3 elements: the characteristics of the new diagnostic technology, the contextual fit, and the local need. Each of the scenarios describes and visualizes a potential specific health care context in which a test based on a new diagnostic technology could be of added value to improve diagnostics in the health care system. Thus, each scenario represents a potential diagnostic setting.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Six Scenarios That Elaborate on the Scope of Application of a New Point-of-Care Diagnostic Test

Abbreviations: CHV, community health volunteer; CHW, community health worker; VL, visceral leishmaniasis.

6. Validation of the Use Case Scenarios

The aforementioned 6 scenarios were discussed with experts in the field, including a Kenyan public health officer and VL specialists at Médecins Sans Frontières, to identify the most urgent need based on the perspective of health care providers. Thereafter, we chose a researcher-centric approach to obtain a context-specific diagnostic need that would facilitate researchers to outline the technological requirements of a new POC test. Interestingly, 2 scenarios—screening and confirmation and a test-of-cure—were consistently identified as a priority for developing a CTPP, and we did not obtain any discrepancies in the opinions of the experts. The selection of these 2 scenarios was based on the following criteria: how well the scenario represents a diagnostic setting in the current health care context of VL; how well the scenario meets a local need in terms of VL case management; and how feasible it would be to implement a new POC test in the scenario.

The first scenario selected, “screening and confirmation,” is beneficial for a number of reasons. Given the poor performance of the serological rK39 RDT in east African countries, a more specific and more sensitive POC diagnostic test that can be implemented by end users with minimal training at the lowest level of the health care system is clearly required. Furthermore, it is crucial to consider the end user(s) of a POC diagnostic test because the level of training and availability of resources will influence the diagnostic setting. Thus, a simple, noninvasive, yet effective VL POC diagnostic test that can be used for initial screening and confirmation (scenario 1) would significantly improve VL case management because it would replace the rK39 RDT for initial screening and invasive splenic aspirations for confirmation. The added value of an effective screening and confirmation test is that patients could be screened more reliably at the lowest level of the health care system, for example, by a CHV/CHW who is closer to the homesteads, which would prevent patients from traveling unnecessarily to regional health care facilities.

The second scenario, “test-of-cure,” is useful because it would replace cumbersome procedures (i.e., microscopic analysis of invasively obtained splenic aspirations or molecular tests such as polymerase chain reaction) that are currently used for test-of-cure. Serological tests cannot serve as test-of-cure owing to persisting antibodies after VL treatment. Thus, a POC diagnostic test that can serve as a test-of-cure (scenario 2) would significantly improve VL case management by replacing invasively obtained splenic aspirations. The added value of a test-of-cure is that relapse of the disease, which occurs in approximately 10% of VL patients, could be detected at an early stage after treatment. Therefore, the test-of-cure scenario fills a critical gap in VL case management.

Scenarios 3 to 6 were excluded from further development of a CTPP. A screening day, as depicted in scenario 3, would be challenging to implement because one cannot ethically test everyone in a community, especially not asymptomatic patients, while symptomatic patients are covered in scenario 1. Additionally, it is difficult to get people to travel to a central location for a screening day because financial constraints or daily routines of herding cattle or taking care of the homesteads often restrict travel. Community testing, as depicted in scenario 4, is dependent on a multiplexed POC test that tests for multiple diseases simultaneously. Multiplexed tests are often more expensive and very few, if any, have been routinely used in the field.

Similarly, integration with the malaria journey, as depicted in scenario 5, is also dependent on a multiplexed POC test for VL and malaria. A well-known multiplex test from DIAMED for VL and malaria is relatively expensive and not routinely used in Kenya and Uganda. By contrast, malaria RDTs are affordable and are routinely used in the field. A follow-up test, as depicted in scenario 6, would be challenging to implement due to the nomadic nature of the inhabitants that we encountered.

A unidirectional problem-solving approach, instead of an interactive design-thinking approach that requires multiple iterations, would have yielded possibly 1 or 2 predictable scenarios instead of the detailed thorough 8 scenarios that were obtained in this study. Thus, after multiple iterations with a Kenyan public health officer and Médecins Sans Frontières, we concluded that scenarios 1 and 2, a test for screening and confirmation and a test-of-cure, are pivotal for VL management as they meet the local need and are feasible to implement. Furthermore, the CTPP developed for the 2 selected scenarios could be broadly applicable to the other 4 scenarios.

7. Define the CTPP

Within each of these 2 selected scenarios, we identified variables that clarify how the diagnostic setting influences the features of the diagnostic test.

A CTPP such as formulated in Figure 8 presents the key features of a product that would fit a particular local health care context. The key features that a CTPP takes into account include variables such as why a test is performed (scope); where the test will take place (geographical location); which level of the health care system (such as at a home or a hospital); who the end user is (target user of the test); when the test will take place in the patients’ health-seeking pathway (diagnostic moment); and how the test will be conducted (operational characteristics of the test). In this study, the product is a new POC diagnostic test for VL in rural Kenya and Uganda. A CTPP is an efficient mapping tool that can guide the R&D of a new POC diagnostic test by determining the local needs, the end user(s), the intended use, and the adequate features of the test. We conclude that the approach presented in this study led to the development of a CTPP wherein the essential features of a POC diagnostic test for VL in Kenya and Uganda were identified more quickly compared with a TPP. The added value of a CTPP is that it does not impose overly stringent guidelines on the researcher during the early stages of R&D. Instead, a CTPP aims to guide the research and not limit the potential of the new POC diagnostic test that is under development.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

(A) CTPP for a VL Point-of-Care Diagnostic Screening-and-Confirmation Test, (B) CTPP for a VL Point-of-Care Diagnostic Test of Cure

Abbreviations: CHV, community health volunteer; CHW, community health worker; CTPP, concept target product profile; VL, visceral leishmaniasis.