The 10-year operational experience of MSF in these four programs in Honduras, Guatemala, and Bolivia demonstrates that diagnosis and treatment of Chagas disease are feasible, relatively safe, and potentially effective in low-income, resource-constrained settings. Through the lessons learned from earlier studies 
and these MSF projects and their common, essential logistical components, we propose that this programmatic approach is feasible at the primary health care level and replicable in other Chagas-disease endemic countries and regions, even in periurban and remote rural areas. With proper coordination between different stakeholders focused on integrated health care services for Chagas disease, including national and regional programs, the diagnosis and treatment of the disease in early chronic phases (mainly indeterminate form) can be safely implemented and should be deemed necessary for affected populations 
Etiological treatment of Chagas disease can and should be integrated at the primary health care level because most patients are near primary health care services, and the majority of patients would be able to receive medical care at this level, taking into account the proportion of Chagas patients with the indeterminate form of the disease. In MSF's programs, this implementation was achieved in remote rural settings through the application of six central features and criteria: IEC, vector control, health staff training, logistics, screening/diagnosis, and treatment/compliance, with family-based support.
IEC was a chief component of program strategies and is vital to ensure treatment compliance and early detection of adverse events, especially when providing care for populations with differing cultures, practices, and modes of communication, among others. IEC was crucial for raising awareness in the general population about the disease (regarding transmission routes, clinical manifestations, and treatment and prevention possibilities) and inform patients and patient families that diagnosis and treatment services were available.
Vector control carried out by national programs was also an important program component and should be simultaneously implemented with patient access to diagnosis and treatment 
. MSF involvement varied as projects progressed, depending on the need and capacity of national authorities and other partners. After treatment, vector control was continued through the national programs, but regular spraying every 6 months was not always carried out. Community entomological surveillance occurred regularly, but spraying for vector control was irregular at times. Eliminating the vector from the environment and households of patients and those at risk is critical.
Health staff training and family IEC for family-based treatment monitoring were exceptional ways of both ensuring quality of diagnosis and treatment compliance, as well as engaging the family in the health care process. Diagnosis and treatment of Chagas disease in all the projects relied on well-trained health personnel to apply their medical skills to care for patients and to establish family commitment to treatment adherence and follow-up care. With minimum logistical capacity, especially support for outreach teams, our program experience may be replicable in other endemic areas.
The fundamental program component of screening and diagnosis used differing diagnostic protocols adapted to the contexts of each country/region. For diagnosis in our programs, the two tests selected were the two with minimum acceptable sensitivity and specificity (ideally 99–100%) and which could be feasibly implemented at the primary health care level 
. Filter paper blood samples were used in the earlier programs mainly for sensitivity and adapted ease of use (ie, no need for centrifugation, relatively easy to supply/refill, portability) in remote rural settings. We introduced the use of Chagas Stat-Pak RDT in the Bolivian programs and carried out a field evaluation using whole blood samples. Recent studies using this RDT have shown relatively low sensitivity (93–94%) compared with conventional tests 
, and this limited sensitivity must be considered in the use of this test. A whole-blood RDT with high sensitivity would be ideal for screening and diagnosis in resource-limited settings 
For treatment and compliance, a large number and proportion of patients started and finished treatment according to protocol in our programs, with over 90% of patients completing >55 days of treatment. Access to treatment, follow-up, and referral of complicated cases were successful elements of the protocol. Relatively low dropout before treatment and low default rates (mostly migrations of patients and adverse reactions) were observed. However, in one program, Sucre, about 9% of diagnosed patients did not start treatment. The main reasons for this were migration, reluctance to start treatment (after counseling and informed consent), pregnant or lactating mothers, and treatment being offered by MOH national programs. Still, overall we found that Chagas treatment and follow-up can be achieved with adequately trained, sensitized, and motivated health staff and family members in both rural remote and periurban settings. The family-based approach for daily drug administration and compliance was key for Chagas disease because of the length of therapy and occurrence of adverse events.
Drug treatment was safely administered in these four programs, with no deaths occurring due to adverse events. Despite this, nearly half of all patients had some type of adverse event, a few of which were severe, including 1 case of Lyell syndrome, and 1 case of Stevens Johnson syndrome. Although no previous studies of Chagas disease have reported either of these syndromes, these two cases must be viewed in the context of over 3,000 patients treated in the four programs, with no deaths in even the most severe cases. The majority of adverse event cases were treated with a reduced dosage of benznidazole (to the minimum dose of 5 mg/kg/day) or temporary suspension of treatment. The time of appearance, intensity, and clinical patterns of adverse events were not different than those observed in other experiences 
, except that we did not see any hematological reactions (ie, no clinical manifestations such as anemia, severe infection, or hemorrhage were observed to make us suspect detrimental effects on bone marrow). However, hematological reactions were only followed clinically, without routine laboratory testing, due to issues of practicality under field conditions. This therefore poses a limitation in that hematological adverse events cannot be completely excluded, especially since severe hematological reactions (such as bone marrow suppression) can be asymptomatic. Proximal neuromuscular adverse events presented later (after 35 days of treatment) compared with other adverse event types, demonstrating cumulative drug toxicity.
Overall, the large number of children and adolescents treated and observed in the four programs (>3,100) provides valuable insight into drug safety for current Chagas disease drug treatment. Previous studies have reported experiences from lower numbers of patients 
. Of note, we observed sizeable variations in reported adverse events in the study locations, namely between the two programs in Central America (Honduras/Guatemala) and the two in South America (Bolivia). In recording adverse events and their severity in our four programs, observer bias no doubt played a role. The identification and classification of an adverse event is often dependent on the observing medical staff, and misclassifications were possible in the programs. We attempted to address this by defining mild, moderate, and severe adverse events based on whether treatment was temporarily interrupted or fully stopped. Other biases in adverse event profiles may exist, such as differences in early detection of side effects and more or less intensive medication and management for adverse events.
While a well-designed program should be able to minimize risks and ensure safe treatment, the lack of a non-toxic alternative drug remains a major obstacle to wider access to treatment for both adults and children. No pediatric formulation currently exists for benznidazole (nor nifurtimox, the only other drug used for treating Chagas disease), increasing risks of under- or overdosing in children. For the youngest patients, cutting tablets and mixing with water or other liquids for oral administration is difficult and has important pharmacological implications in terms of absorption and bioavailability.
The seroconversion rates detected in treated patients were relatively high in the Central American projects, Yoro, Honduras and Olopa, Guatemala, showing that therapy can clear T. cruzi
infection. However, seroconversion was far lower in the South American Bolivian projects in Entre Ríos and Sucre. The findings in Bolivia are similar to those reported from earlier studies in Argentina and Brazil 
. Also, seroconversion was detected earlier in the Central American programs compared with the Bolivian programs. Previous research has shown that in South America seroconversion is sometimes not detected until 5–7 years later 
. Thus, the higher and earlier seroconversion we detected in Central America supports previously reported findings 
and may have important public health implications.
The differences in seroconversion rates may be explained by a number of reasons. One primary explanation may be based on the presence of different parasite lineages in different geographic regions, with T. cruzi
type I predominating in Central America and T. cruzi
type II in South America, with varying degrees of overlap 
. Because of the potential differences in T. cruzi
subtypes present in Honduras and Guatemala compared with Bolivia, drug treatment effectiveness may have differed. Another factor to consider is the time between vector control activity and drug treatment, since cases (mostly asymptomatic) closer to the acute phase of the disease can possibly account for more rapid seroconversion. Also, statistical limitations of our data analysis may exist due to the varying age groups and varying times of post-treatment follow-up in the four projects, as has been examined in other studies 
. Finally, differences in immune response among populations may play a role.
Whatever the reason, the lack of a better marker for indicating parasitological cure is a major impediment to advances in treatment and development of more effective drugs 
. The observed differences between seroconversion rates in Central and South America highlight the need for further studies to confirm our findings and help improve etiological treatment protocols with dosages and duration adapted to the Chagas disease cycle in different geographic regions.
Since the start of our first Chagas disease program in 1999, which focused on young children, MSF has pushed to deliver diagnosis and treatment of this disease to wider and wider age groups. Over the past decade, treatment for Chagas disease has expanded from children <12 years old, to <15, then <18, and finally adults. This strategy has helped bring broader coverage of treatment delivery for Chagas disease.
Bolivia is the most highly endemic country in the world for T. cruzi
infection, with up to 1.8 million people believed to be infected 
. MSF currently has two active programs in Cochabamba, where Chagas disease treatment is integrated into primary health care and offered to adults as well as children and adolescents. Because of high prevalence in Bolivia, Chagas disease diagnosis and treatment remain an operational priority there for MSF.
MSF's programs, both past and present, highlight where and what the needs are for people affected by Chagas disease. In addition to increasing public awareness and patient access to existing diagnostics and drugs, the development of new, less toxic, more effective drugs; adapted pediatric formulations of treatments; and a reliable test of parasitological cure are all urgently required. Because Chagas disease and those afflicted with it are often neglected, medical care for this patient population should be implemented whenever and wherever possible, as MSF has demonstrated as feasible in these four programs, and research and development for the disease should be scaled up dramatically.