Field Trials of Health Interventions

The book addresses the significant gap in readily available practical guidance for conducting field trials, particularly randomized controlled trials (RCTs), in low- and middle-income countries (LMICs). While many large-scale field trials have been successfully conducted in LMICs, detailed descriptions of the methods and procedures are often not published. This lack of accessible information forces new investigators to rely heavily on personal experience, leading to inefficient repetition of mistakes and increased costs. The authors aim to synthesize the experiences of seasoned researchers to create a practical and comprehensive guide to improve the design and conduct of such trials, emphasizing RCTs but acknowledging the relevance of the discussed issues to other field research in LMICs. The book seeks to help new investigators build upon existing knowledge and avoid common pitfalls, promoting a more efficient and less wasteful approach to research.

Conducting field trials in LMICs presents unique challenges. These trials are often complex and expensive, involving the long-term follow-up of hundreds or thousands of individuals. Successful completion requires meticulous planning. The lack of published details on essential procedures and methods further complicates the process, leaving researchers with limited practical guidelines. The authors emphasize the importance of learning from previous experiences to minimize costly errors. The book aims to bridge this knowledge gap by offering a practical and comprehensive guide that synthesizes the experiences of investigators who have conducted field trials in LMICs. The goal is to provide procedures and practices that have proven successful in these settings, thereby allowing new investigators to build on the collective knowledge rather than repeating past mistakes.

The book strongly emphasizes the importance of both scientific merit and ethical considerations in conducting field trials involving human participants. Scientific merit is underscored by the need for appropriate research methods aligned with the study's objectives and the incorporation of findings from prior research, including systematic reviews. The text advocates for thorough reviews of existing literature to avoid unnecessary duplication of research efforts. Ethical considerations are extensively discussed. These include the principles of voluntariness and informed consent, with a focus on the unique challenges posed by cultural contexts in LMICs and the need to address potential power imbalances, illiteracy, and undue inducement. Special attention is given to the ethical treatment of vulnerable populations and the careful consideration of potential long-term harms, even when short-term benefits exist. The authors highlight the need for strict ethical review processes and the importance of communication and feedback between the research team and participating communities.

This section delves into the practical aspects of implementing and monitoring field trials. It addresses the issue of participant compliance, noting that pragmatic trials may incorporate some degree of non-compliance for a more realistic measure of effectiveness, while explanatory trials should strive for high compliance. The text emphasizes the importance of continuously monitoring compliance, perhaps through methods like urine or blood tests to verify intervention intake. The authors discuss the significance of documenting compliance, especially in behavior-mediated interventions, such as breastfeeding promotion trials. The book highlights the role of communication and feedback between the research team and participating communities, which is crucial for both pragmatic and ethical reasons. It stresses the importance of maintaining open and transparent systems where possible, particularly when employing staff in areas with limited employment opportunities.

The text explains the crucial role of the trial sponsor – an individual or institution legally responsible for the trial's proper conduct, management, and financing. The responsibilities of the sponsor are detailed, including ensuring adherence to the protocol and managing the budget. The book discusses data analysis techniques, including the use of forest plots and considerations for comparable data between studies. It also highlights the importance of documenting changes in intermediate outcomes and creating impact models to link interventions to primary outcomes. The process of reporting findings is discussed, with particular emphasis on providing feedback to participants and communities, as well as collaborating with local health workers and authorities. The section also mentions the importance of considering the appropriate point at which to stop a trial based on sufficient evidence to support or reject the intervention's introduction into general health services.

This section highlights the significant role of food and nutrition in determining human health and disease, particularly in low-income and deprived populations within middle-income countries. Under-nutrition, ranging from life-threatening severe malnutrition (kwashiorkor and marasmus) to milder forms impacting disease susceptibility and outcome, is a major concern. Trials addressing these issues may involve interventions such as providing high-protein/calorie diets or specific micronutrient supplementation (iron, folate, zinc, iodine, vitamin A). These trials often require repeated visits over several years, with the frequency determined by the supplement type. Community-level interventions, such as food fortification and changes in agricultural practices or eating habits, are also considered as potential approaches to improve nutritional intake and reduce the burden of malnutrition-related diseases.

The section discusses the complexities of vaccine trials, particularly in the context of LMICs. While initial vaccine trials for diseases affecting both high-income countries (HICs) and LMICs typically occur in HICs due to vaccine development and manufacturing locations, the applicability of these findings to LMICs may be limited due to variations in infection prevalence and nutritional status. The differing prevalence of other infections or nutritional deficiencies in LMICs can interfere with a vaccine’s mode of action, necessitating further trials in these settings, even after establishing efficacy in HICs. The increasing focus on vaccines against infectious agents prevalent primarily or predominantly in LMICs (malaria, visceral leishmaniasis, tuberculosis, HIV) means that the first major field trials to assess their efficacy are likely to be conducted in LMIC settings, underscoring the critical role of LMIC-based research in vaccine development and deployment.

This section focuses on interventions aimed at preventing disease through behavior change. While educational components are nearly always essential for effective deployment of health interventions, the level of effort required can range from providing basic information (e.g., immunization clinic schedules) to promoting deeper understanding (e.g., male circumcision and HIV prevention) and to inducing lifestyle changes (e.g., diet or sexual habits). Simply increasing knowledge and skills is rarely enough to achieve behavior change; capacity, willingness, and motivation are also crucial. The design and implementation of educational interventions, and more complex interventions in general, require thorough community-based research, utilizing methods detailed elsewhere in the book. The emphasis is on the need for careful investigation to understand the multifaceted nature of behavioral change and its role in disease prevention.

This section addresses communicable diseases transmitted through vectors (mosquitoes, tsetse flies, triatomine bugs, sandflies, ticks, snails) or intermediate hosts. A broad range of control measures are discussed, emphasizing the importance of understanding the vector's life cycle and environmental conditions. Interventions include insecticide/larvicide application, using biological agents, engineering techniques to reduce vector habitats, community involvement in eliminating breeding sites and deploying traps, housing and screening improvements, and strategies combining multiple methods to delay insecticide resistance. Assessing impact involves monitoring intermediate indicators (vector density) and the population's health status. Malaria vector control is used as a case study, highlighting diverse approaches such as breeding site control, larvicides, personal protection measures (netting, screens, repellents), and insecticide-treated bed nets (ITNs). The complexities of designing trials, particularly when communities rather than individuals are the unit of randomization, are emphasized, including the challenge of vector migration across community boundaries.

This section examines the use of drugs and interventions for disease prevention (prophylaxis) or disease management following infection. Examples include isoniazid prophylaxis for HIV-infected individuals to prevent tuberculosis (TB) and antiretroviral therapy for HIV to slow disease progression. Some drug interventions don't require individual diagnosis; community or group diagnosis might suffice (e.g., mass anti-helminthic treatment for schoolchildren). While most therapeutic or preventive agents are administered individually, community-level distribution is sometimes possible (e.g., water fluoridation, medicated salt). Mass treatment of school-age children in highly endemic areas, such as with anti-schistosomal drugs, can effectively control disease. The section underlines the varied approaches to drug-based interventions, from individual-level treatment to community-wide strategies.

This section focuses on chronic diseases, which may have infectious (HIV, TB) or environmental/other causes (cardiovascular disease, cancer). Many chronic diseases are not curable but manageable through education/behavior change and medication. Clinical care is more complex than for acute conditions. Interventions often include community screening, disease staging, laboratory tests, and long-term treatment plans. Trials for chronic disease interventions typically span several years or even decades to fully assess efficacy. The unique challenges associated with long-term monitoring, reliance on reliable laboratory results, and tracking individual clinical and laboratory findings over extended periods are highlighted.

Accurate diagnosis is crucial for effective disease treatment. While clinical symptoms and signs are often used, their imprecision is increasingly recognized, highlighting the urgent need for improved diagnostic tests that are simple, inexpensive, and accurate for many infectious and chronic diseases. Case-finding and treatment strategies heavily rely on suitable diagnostic tests. Studies measuring the sensitivity, specificity, and predictive values of diagnostic tests are essential to assess the effectiveness of case-finding and treatment interventions. The development and widespread adoption of rapid diagnostic tests for malaria, replacing microscopy or presumptive treatment, serve as an example of a significant innovation in disease control and the ongoing need for improved diagnostic methods and appropriate treatment for non-malarial fevers. The section emphasizes the critical role of diagnostics in improving healthcare outcomes.

Ethical research requires scientific merit, ideally judged by an independent committee, ensuring the research methods align with the aims and consider prior research. Systematic reviews are increasingly mandatory to avoid needless duplication. Voluntariness is paramount, implying free participation, continuation, or withdrawal. Participants must have full knowledge of consequences and should not be coerced by investigators, officials, or others, nor should their socio-economic conditions influence their decision. Informed consent is a crucial part of this process, though the implementation needs to consider diverse cultural contexts in LMICs, addressing potential issues arising from gender inequality or illiteracy. Compensation must fairly cover costs but not unduly incentivize participation. The importance of thanking potential participants who do not meet inclusion criteria is stressed. The process of obtaining consent, even without a signature, needs to be carefully conducted to ensure understanding, opportunity for questions, and the freedom to refuse. Ethical review boards pay particular attention to vulnerable individuals (children, those with incurable diseases, the impoverished, etc.), ensuring no undue pressure to participate.

The provision of incentives for research participation requires careful consideration, balancing fair compensation for lost time and income with avoiding undue inducement. While providing snacks, travel reimbursement, or simple medications for minor ailments might be acceptable, monetary payments exceeding lost wages or expenses risk becoming coercive. Each case must be judged according to its merits within the local context. Local Ethical Review Committees (ERCs) carefully assess the level of compensation proposed, ensuring it's fair but does not pressure individuals into participation. Maintaining a balance between fairly compensating participants and minimizing potential intra-family or community conflicts is crucial to upholding ethical research practices. The discussion highlights the fine line between fair compensation and undue inducement, emphasizing the need for context-specific assessments.

The ethical considerations extend to the provision of healthcare during the trial. Some improvements in healthcare may be necessary for the scientific goals of the trial (e.g., improved diagnostics). However, it's crucial to consider if enhanced healthcare during the trial is sustainable beyond its duration; otherwise it may harm the community or create unrealistic expectations of local services. Improvements should ideally be designed to be maintainable with existing resources, possibly through staff training or enhancements to routine medical record systems. The authors acknowledge the difficulty of maintaining all enhancements post-trial but stress that any temporary enhancements shouldn't be disproportionate. The ethical responsibilities of trial organizers in maintaining close ties with the providers of routine healthcare are highlighted. The text delves into the complexities of ethical decision-making by contrasting the 'purist' view (that only using the global 'best' intervention is ethical) with the 'pragmatist' view (that excluding the best intervention might be acceptable in resource-constrained settings where that best intervention is unaffordable).

The authors address challenges in evaluating complex interventions. These interventions are defined as those involving multiple interacting components (educational, behavioral, etc.), posing evaluation difficulties due to standardization issues, context sensitivity, logistical difficulties, and complex causal chains between intervention and outcome. The section explores the use of group randomization and ways to improve its efficiency through stratification based on pre-intervention disease risks to ensure balance between trial arms. The discussion acknowledges that randomization is often not possible, leading to the use of non-randomized designs (before-after or pre-post), which have limitations due to the natural variability of many diseases and potential changes in data collection methods over time. Strategies for improving the plausibility of conclusions from non-randomized studies are explored, including collecting data at multiple time points, comparing trends with control populations, and accounting for confounding variables.

The section discusses the selection of appropriate endpoints for trials. It highlights the challenge when less severe outcomes appear sooner than primary endpoints, potentially raising ethical concerns about continuing the trial due to a lack of 'equipoise'. The importance of considering such scenarios during trial design is emphasized, suggesting that defining secondary outcomes carefully can prevent these ethical dilemmas. Determining when a trial should end is complex. The authors suggest that a trial should be continued until enough evidence is collected to support or refute the intervention’s wider adoption in healthcare, rather than solely focusing on the point where the difference between intervention and control groups becomes statistically significant. The need to establish both the degree and duration of protection is stressed, highlighting that trials may need to continue beyond the initial demonstration of efficacy to determine long-lasting protection.

Monitoring trial procedures is crucial for safeguarding participant welfare and ensuring adherence to the protocol. This responsibility often falls to a trial steering committee, with reports to a Data Safety Monitoring Board (DSMB). If no steering committee exists, the DSMB assumes a greater monitoring role, typically receiving detailed progress reports at each meeting and possibly conducting site visits. Day-to-day monitoring is usually handled by clinical trial monitors who report to the sponsor. The sponsor is responsible for reporting any concerns to the DSMB, via the steering committee if one exists. The process of monitoring serious adverse events (SAEs) is detailed, emphasizing the importance of timely reporting to the DSMB, classification of SAEs, and the use of blinded data analysis to avoid bias in assessing the distribution of SAEs between intervention and control arms. The DSMB's role in monitoring is described in detail, highlighting the complexities of this critical process and the need for clear expectations within a DSMB charter.

The section discusses data analysis techniques for interpreting trial results, including the use of forest plots to display effect measures and confidence intervals from multiple studies. Before combining data, comparability between studies is crucial, necessitating careful consideration of potential differences in effect based on subgroups. Subgroups should be defined a priori and incorporated into the review protocol, as seen in an example of a study on chewing substances in Asia, which stratified by geographical region to minimize confounding factors related to tobacco use. The text emphasizes the importance of considering potential differences in effects across subgroups and the need to define these subgroups in advance of the analysis. The importance of documenting changes in intermediate outcomes and using impact models to illustrate the causal pathway between the intervention and major outcomes is also discussed, highlighting the difference between simple models for biological interventions and more complex models for non-biological interventions. An example of an HIV prevention trial in Tanzania illustrates a complex intervention with multiple components and the challenges of interpreting the absence of an overall effect.

This section stresses the paramount importance of community engagement in the planning and execution of field trials. Clearly defining the 'community' is crucial, especially when considering diverse social and geographical factors like regular migration or work-based participation. The section emphasizes the need to understand and navigate the dynamics of multiple communities and potential conflicts between various community groups. It details strategies for identifying and engaging community leaders (village leaders, traditional healers, religious leaders, etc.), and suggests a multi-pronged approach for exploring community power structures and levels of trust. The critical role of front-line research staff in mediating between research institutions and communities is highlighted, recognizing their importance in 'doing ethics' in the field and fostering relationships. The text emphasizes the need for culturally appropriate compensation for participation, mechanisms for addressing harm, and strategies for minimizing intra-family and community conflicts. The importance of keeping local health workers informed of trial progress and final results is emphasized, underlining the ethical and practical benefits of such feedback and its role in fostering cooperation for future research.

Accurate budgeting is crucial for successful trial completion. Realistic estimates of resources (transport, staff salaries, equipment) and trial duration, including analysis and reporting time, are essential for calculating a feasible budget. Underestimating support can jeopardize objectives, while overestimation might hinder funding approval. The temptation to underestimate costs to improve funding chances is cautioned against, as experienced reviewers will likely identify inaccuracies. Trial duration is often underestimated, particularly when implementation depends on external partners (e.g., Ministry of Health). The text emphasizes the need for careful planning and accurate cost estimation to secure funding and ensure a high-quality trial.

This section offers guidance on preparing competitive grant applications. It advises applicants to research the work of funding committee members to anticipate potential questions and focus areas. Discussions within funding committees are confidential, but agencies typically provide feedback, particularly after rejections, to help improve future proposals. Direct contact between applicants and committee members is generally discouraged. Conflicts of interest are addressed, noting that committee members are typically recused if they or their institutions have a connection to the application. This section highlights practical strategies for preparing strong grant applications, increasing the likelihood of securing funding for research in LMIC settings. The importance of clear communication and professionalism in the grant writing and review process is implicitly underscored.

This section discusses practical aspects of mapping and data collection in field trials. It emphasizes the importance of defining the relevant communities and identifying their leaders and representatives. The complexities of defining communities are highlighted, considering factors like migration and work-based selection. The section advises using available data from national censuses to select appropriate study areas, acknowledging the potential for inaccuracies or outdated information (an example from Ghana is given). Census data can provide demographic information (age, sex, ethnicity, household size, density, mortality, fertility, migration rates), which is useful for planning and estimating potential loss to follow-up in longitudinal studies. The use of GPS devices and mapping techniques for spatial analysis, linking population data to risk factors (e.g., proximity to water sources for schistosomiasis), is discussed. The use of Geographical Information System (GIS) software and readily available online data sets (vegetation, rainfall, elevation) is also mentioned. The section advocates for the use of data from a variety of sources, utilizing technology to improve the accuracy and efficiency of data collection and analysis.