Evidence Acquisition and Evaluation for Evidence Summit on Population-Level Behavior Change to Enhance Child Survival and Development in Low- and Middle-Income Countries

Abstract

Recognizing the need for evidence to inform public health officials and health care workers in the U.S. government and low- and middle-income country governments on efficient, effective behavior change policies, strategies, and programs for child health and development, the U.S. government convened the Evidence Summit on Enhancing Child Survival and Development in Lower- and Middle-Income Countries by Achieving Population-Level Behavior Change. This article summarizes the background and methods for the acquisition and evaluation of the evidence used to the achieve the goals of the summit that is reviewed in other articles in this special issue of the Journal of Health Communication. The process began by identifying focal questions intended to inform the U.S. and low- and middle-income governments about behavior change interventions that accelerate reductions in under-5 mortality and optimize healthy and protective child development to 5 years of age. Experts were selected representing the research and program communities, academia, relevant nongovernmental organizations, and government agencies and assembled into evidence review teams. This was followed by the systematic gathering of relevant peer-reviewed literature that would inform the focal questions. Members of the evidence review teams were invited to add relevant articles not identified in the initial literature review to complete the bibliographies. Details of the search processes and methods used for screening and quality reviews are described. The evidence review teams were asked to comply with a specific evaluation framework for recommendations on practice and policy on the basis of both expert opinion and the quality of the data reviewed.

As a global community, we have made significant progress in reducing child mortality. Over the past 50 years, child mortality has been reduced by 70%. However, in rich and poor countries alike, the poorest and most disadvantaged children continue to miss out on life-saving, affordable interventions. In 2012, about 7 million children—most of them in Sub-Saharan Africa and South Asia—died before they reached their fifth birthday, the majority from preventable causes. In low- and middle-income countries, an estimated 200 million children younger than 5 years of age—more than 30% of the world's children—fail to achieve their developmental milestones, limiting their ability to contribute to their families and communities or obtain gainful employment (UNICEF, 2012).

On the basis of current rates of progress, many countries will not achieve United Nations Millennium Development Goals 4 and 5 to reduce child mortality and improve maternal health. The Child Survival Call to Action (http://5thbday.usaid.gov/pages/responsesub/event.aspx), is a forum that convened on June 14–15, 2012, in Washington, DC, that provides a roadmap for ending preventable child death. The achievement of this goal requires sustainable population-level behavior changes that impact maternal and child health. A range of behaviors need to be targeted (Fox & Obregón, 2014), such as healthy timing and spacing of pregnancy, the utilization of pre- and postnatal care services, better nutritional choices, utilization of bed nets, and accessing and practicing preventive health care such as immunization and prenatal care. Changing the behavior of health service providers also is a key element to a sustainable improvement in public health and the adoption of new and existing health care practices.

A renewed emphasis on the application of research and evaluation to inform strategic planning for low- and middle-income countries is integral to U.S. government efforts to improve health by promoting country-owned, effective, and sustainable interventions. To that end, the U.S. Agency for International Development (USAID) convened a series of evidence summits focused on important development challenges (http://www.usaid.gov/what-we-do/global-health/summits). The goal of the prior summits was to provide evidence-based expert recommendations on how to achieve some of the world's most intractable health and development goals, such as caring for children living outside families (Mahomes, Fluke, Rinehardt, & Huebner, 2012), reducing maternal mortality (Stanton, Higgs, & Koblinsky, 2014) and supporting community health workers (Naimoli, Frymus, Quain, & Roseman, 2012).

To address the problem of ending preventable child deaths, the U.S. government, in collaboration with UNICEF and other partners, convened the Evidence Summit on Enhancing Child Survival and Development in Low- and Middle-Income Countries by Achieving Population-Level Behavior Change forum on June 3–4, 2013, in Washington, DC. The Evidence Summit brought together leading researchers, development experts, and those implementing programs in the field to assess the evidence that is rigorous enough to inform policies, strategies, and programs relevant to behavior change for child survival and development in low- and middle-income countries. A linked goal is to identify evidence gaps that could shape the future research agenda. This Evidence Summit reflects USAID's commitment to evidence-based, innovative, efficacious, and sustainable U.S. government development efforts in partnership with other governments. The rapid deployment of new discoveries and innovations in public health for at-risk populations requires a continuum of learning from basic to operational research and a broad coalition of expertise and contributors across the U.S. government, academia, and policy and practice leadership from developing countries. This article describes the Evidence Summit process from inception to post-Summit activities. Other articles in this special issue of Journal of Health Communication present the findings and recommendations. The background and rationale for the selection of the topic for the Summit and the link between theory and practice is described in another article in this issue (Fox & Obregón, 2014).

The Evidence Summit as a Process

The Evidence Summit on Population-Level Behavior Change for Child Survival and Development was not a single event but rather a multiyear process that culminated in the June 2013 Evidence Summit and continues with the implementation of recommendations and a research agenda developed during the Evidence Summit process (Figure 1). The procedure for developing the evidence for the summit was based on the methods used for earlier evidence summits in the series (Higgs, Stammer, Roth, & Balster, 2014; Higgs, Zlidar, & Balster, 2012). The initial planning for the Evidence Summit originated in the Bureau for Global Health at USAID and was tied to President Obama's Global Health Initiative, which called for interagency collaboration, innovation, and doing more of what works as guiding principles for U.S. government assistance in global health and as a follow-up to the Child Survival Call to Action: A Promise Renewed (http://apromiserenewed.org). After initial discussions, a core group of persons responsible for the overall organization and direction of the Evidence Summit was assembled, with members drawn from several USAID offices, UNICEF, the National Institutes of Health, and many nongovernmental organizations and professional societies. The number of active Core Group members changed over the course of this process but was typically 25–35 individuals. The formation of the Core Group was followed by a scoping exercise and consultation with external experts to inform the topic selection; the background and rationale for the topic selection were then summarized in an internal concept article. This led to the creation of focal questions to guide the evidence review process, the selection and appointment of external and internal experts to evidence review teams (ERTs), a systematic literature search of documents relevant to the focal questions, the screening and evaluation of these documents, and the identification of additional relevant material by members of the ERTs. A pre-Summit meeting was held to organize and motivate the ERTs, which subsequently prepared draft evidence synthesis papers and draft recommendations for presentation at the Evidence Summit. The final step of the Evidence Summit process—the revision of materials on the basis of feedback at the Evidence Summit and the implementation of an Evidence to Action Plan to act upon recommendations—is still under way. The following sections provide more details on these steps in the process, focusing primarily on the evidence gathering and review phases.

Figure 1 The Evidence Summit process from initial organization to activities (still ongoing) to implement the recommendations. ERTs = evidence review teams.

Getting Started

External Consultation and Development of Focal Questions

The background and rationale for selecting the topic of behavior change for child survival and development is discussed in other articles (Fox & Obregón, 2014; Pablos-Méndez & Fox, 2012). The Core Group met over a period of several months to draft a concept article describing the background for the Evidence Summit, the goals and anticipated outcomes, and a process for accomplishing the Evidence Summit objectives. Following this, the Core Group organized a scoping exercise to gather substantive input from key people in the field and to identify priorities that resonate with overall contemporary development thinking and the present state of play in the social and behavior change field. To collect data quickly, an online survey—conducted by the Communication Initiative Network—was completed by 97 people (out of 167 invitations) nominated by the Core Group and representing a broad range of public health areas, types of organizations, work roles, and 22 countries of residence. All had broad experience in global health.

On the basis of discussions and recommendations provided during the scoping exercise, the Core Group drafted six key focal questions to address during the summit process:

1.	What are the effective and sustainable interventions to promote and support behavior changes required for and by families, mothers, and other caregivers to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?
2.	What are the effective means to facilitate and empower communities to organize and advocate for interventions to achieve behavior and social changes that are needed to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?
3.	What types of sustainable health systems and policy supports are effective in producing behavior and social changes for and by primary caregivers, families, and communities that are needed to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?
4.	What are the effective and sustainable interventions that focus on gender dynamics as a means to promote and support behavior and social change that is needed to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?
5.	What are the effective and sustainable interventions that address stigma and discrimination as a means to promote and support behavior and social change that is needed to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?
6.	What are the effective and sustainable interventions that use advances in science and technology to promote and support behavior and social changes that are needed to accelerate reductions in under-5 mortality and optimize healthy and protective child development to age 5?

The first three focal questions address the question of behavior change from a socioecological perspective, investigating changes at the individual, community, and health systems levels. The final three focal questions address cross-cutting themes that the Core Group felt influence population-level behavior change.

Formation of ERTs

Central to the Evidence Summit process was the selection and organization of experts into six ERTs, which were organized around the six focal questions. ERT members with expertise in a specific topic were assigned to that ERT. Other members were divided among the six ERTs to achieve a balance of representation from the research and program communities and government agencies as well as expertise in child survival and child development.

Each ERT was asked to assess the evidence on behavior change interventions for child survival and development and make recommendations on policies, programs, and research. Experts were nominated by Core Group members. Many members of the Core Group also served on ERTs. The challenge in selecting ERT members was to achieve representation along several dimensions. These dimensions included discipline, work setting (research, practice, academia, relevant nongovernmental organizations, government agencies), and the Global North and South. Although the number of ERT members based in low- and middle-income countries was relatively small, many experts had extensive experience working and residing in those settings. The Core Group selected ERT members on the basis of their knowledge and expertise in child health, child development, behavior change, health systems, gender studies, marginalized groups, innovative technology, development, and other related topics. Several rounds of invitations were sent to prospective ERT members; many invited individuals agreed to serve in this capacity.

Peer-Reviewed Evidence Acquisition

Two strategies were used to acquire evidence for the Evidence Summit process: a systematic literature review conducted by public health professionals and a call for evidence issued to members of the ERTs. Although it was recognized that evidence existed in both the published and unpublished literature, the Core Group opted to limit the formal literature search to articles published in peer-reviewed, scholarly journals, anticipating that relevant documents from the gray literature would be submitted by experts in the field in the subsequent Call for Evidence. The Core Group also decided to exclude documents on the basis of research carried out in high-income countries because this research may have limited relevance to low- and middle-income countries. If relevant research from high-income countries was identified by experts in the ERTs, it could be submitted during the call for evidence process.

Literature Search

Figure 2 shows the results of the initial literature search. Knowledge Management Services conducted the initial literature search, compiled the database, and conducted the screening and initial review of the literature under contract arrangements with USAID in collaboration with the Core Group. The Core Group worked with Knowledge Management Services to select the search terms from keywords identified in relevant articles and consultation with experts in the fields of behavior change and child survival and development. The search strategy for the peer-reviewed literature combined terms for the different levels of behavior change, with terms for the problem health areas identified in the Child Survival Call to Action as having the greatest impact on child mortality and development, terms for gender, stigma and, discrimination, and innovation, and the names of countries and regions categorized as low- or middle-income by the World Bank. More detailed information on the search terms used for the various databases can be found on the Evidence Summit website (http://plbcevidencesummit.hsaccess.org/docs/public-documents/evidence-summit-process-15mar13.pdf?sfvrsn=2).

Figure 2 The initial literature search method and results of the search as well as initial screening steps used for the Evidence Summit. KMS = Knowledge Management Services.

The databases searched were PubMed, JSTOR, EbscoHost (Africa-wide, CINAHL, CAB, Business Elite, Global Health), SCOPUS, ScienceDirect, and the Cochrane Library (Figure 2). Where possible, the following exclusion criteria were applied to the database searches:

	Documents not published in English, Spanish, or Portuguese
	Documents published before January 1, 1990
	Documents not detailing an intervention that includes a measurement of at least one child or maternal health outcome, behavioral outcome, or child development outcome
	Documents focusing only on knowledge and attitudes without any identifiable changes in behavior, developmental, or health outcomes
	Research carried out in high-income countries
	Magazine or newspaper articles; letters to the editor; obituaries; commentaries and recommendations that are not based on thorough literature reviews; book reviews; job postings and historical accounts

An additional search with added terms on empowering communities was conducted for the ERT addressing focal question 2 (ERT2) after the initial search following the same protocols to supplement the number of available articles for this review.

Screening

After we removed duplicates, the initial search of the peer-reviewed literature yielded 2,697 documents. Knowledge Management Services then undertook a manual screening of the documents with two goals: to delete any remaining documents using the aforementioned basic exclusion criteria and to include those describing interventions directly related to the focal questions and that contained behavior, developmental, or health outcome data (Figure 2). The latter goal was achieved by including articles that described research on a social or behavior change intervention and included measurement of at least one of the following: (a) health outcome related to maternal health or child health and development; (b) a behavioral outcome that contributes to improved child survival, including behaviors relevant to maternal health, female or adolescent reproductive health, provider behavior (includes teachers or other individuals who have the potential to influence maternal/child health behaviors or outcomes such as spouses, mother-in-laws), and community-level behaviors; or (c) child development outcome for children under 5 years of age. The screening algorithm is provided in the supplemental online appendix. This initial screening process was completed through a review of abstracts and full-text articles where necessary and resulted in the retention of 891 articles. Full texts were obtained for all these documents.

The next screening step (Figure 3) was used to sort documents into tiers of evidence, health area, and the six ERT focal questions. This manual screening was based on assessment of the full texts of the documents. The full text of this second screening protocol appears in the supplemental online appendix. The assessment began by reapplying the screening criteria from the first screen but this time with the added benefit of examining the full text of all the articles. This eliminated an additional 159 documents for a total of 732 that were included in the initial bibliography. The assessment by tier was designed to give the ERTs some guidance in prioritizing documents for review on the basis of the nature of the outcomes reported (Fox & Obregón, 2014). The top tier of documents (Tier 1) contained at least one maternal or child health or child development outcome, such as infant mortality. Tier 2 consisted of studies that did not report a health or development outcome but did report a change in behavior or community empowerment. Tier 3 documents reported changes in attitudes, knowledge, or beliefs only. Documents that did not report any of these types of outcomes were assigned to Tiers 4 or 5. ERTs were advised to prioritize their work by focusing initially on Tier 1 and 2 documents.

Figure 3 Results of the second screening process, the call for evidence, and an added search and screening conducted for the evidence review team that was considering community-level interventions (ERT 2).

During this second screening, documents were also assigned to one or more ERTs on the basis of an assessment of their relevance to the ERT focal question. Some documents were assigned to more than one ERT. Documents were coded by health area from the following list identified in the Child Survival Call to Action (see Fox & Obregón, 2014): (a) healthy timing and spacing of pregnancy, (b) nutrition/Vitamin A/breastfeeding, (c) neonatal health, (d) handwashing/prevention of diarrhea, pneumonia, and acute respiratory infections, (e) malaria, (f) prevention of mother to child transmission of HIV, (g) immunization, and (h) healthy early child development. Most of the ERTs organized their work by assigning members to review the documents that were relevant to a single health area, with the reviews by health areas combined in the final evidence reviews.

Call for Evidence

To supplement the articles identified through the literature search, members of the ERTs were invited to submit documents they felt would help address the focal questions. This took place through a formal call for evidence that began before the pre-Summit meeting. It used an online document submission process in which ERT members were asked a series of questions about the documents they were submitting. The Call for Evidence submission protocol is included in the supplemental online appendix. In selecting articles for submission, the ERT members were encouraged to review the exclusion criteria that had been applied to the initial literature search, but they were not required to adhere to those criteria. They were also advised that purely descriptive articles had not proved useful in previous USAID-supported evidence summits and that priority should be given to the following types of articles:

	Relevant documents that were not published in scientific journals (e.g., agency reports, other gray literature) and thus would not have been found in a search of journal publications
	Documents of relevance to low- and middle-income countries, even though they may describe work done in high-income countries
	Relevant documents published before January 1, 1990, that would have been excluded from the formal search

After removal of duplicates and documents that were already in the database, 332 unique documents were formally submitted through the call for evidence, resulting in 1,064 total documents that were initially made available to the ERTs. The added search, on empowering communities, yielded an additional 695 articles plus 8 others that were found in paper bibliographies. After screening the results of this added search, 18 articles were included in the final database of 1,082 articles (Figure 3). During the review process, ERT members were encouraged to cite other documents that they felt were relevant without submitting them through the formal call for evidence process. A workspace on the Internet was made available to the ERTs that contained full texts of all the documents and an index that helped ERT members identify documents that matched the evidence tier and health area they were working on. This web site also provided background documents on the Evidence Summit, a methodology for the literature search process, and a place where working documents could be shared.

Relevance Review

During the Call for Evidence process, ERT members were given the opportunity to conduct additional relevance reviews of the 732 documents obtained through the literature search. This was facilitated by an online survey that could be filled out by ERT members, with the results included in data packets made available to them on the workspace. The full text of this ERT relevance review is included in the supplemental online appendix. The following are the purposes of this review:

1.	Begin to familiarize ERT members with the evidence.
2.	Identify the types of interventions that are the focus of each document.
3.	Determine ERT member assessment of the relevance of the document to the focal question of their ERT.
4.	Determine whether the document could be helpful to the ERTs concerned with gender issues and stigma and discrimination.
5.	Determine ERT member assessment of the contribution made by the document to estimating the sustainability of the intervention for wide adoption in low- and middle-income countries.
6.	Make a preliminary identification of documents that ERT members believed would be critical to support recommendations coming from their ERT concerning evidence-based practice and policy.

The implementation of this relevance review differed from ERT to ERT. Some chose to use it for all of the documents assigned to them while others used the review more selectively. In some cases more than one ERT member was asked to review each document. In general, the relevance review process achieved the added purpose of informing ERT members about their task.

Quality Review Process

After the screening and sorting processes were completed, ERT members were provided an online instrument for assessing the quality of the studies reported in the articles derived from the literature search and screening process and obtained through the call for evidence. The quality review instruments were derived from previous evidence summits (Higgs et al., 2012; Higgs et al., 2014). The quality review was designed to accommodate review articles, reports of qualitative work, and reports of quantitative research, using survey logic to direct respondents to questions appropriate for the document type. Experience from previous summits suggested that a short and simple quality review tool would be more practical for the purposes of this process, and so the question set was revised to a set of 20 questions. Depending on the manuscript type, only a subset of these 20 questions were applicable (see the supplemental online appendix for the text of the quality review questionnaire). ERTs differed in the extent to which they used the document quality reviews, and one of the ERTs (ERT 1) used a different set of quality items than those that were provided. ERTs also differed in the approach they took to assigning members to undertake the quality reviews. ERTs were provided with scores for items ranking quality that were generated by turning the quality review responses into a numeric score of 3 (strongest), 2, 1, or 0 (weakest), then dividing these scores by the total number of questions answered. When the reviewer felt the item was not applicable to the document under review, these nonapplicable scores were censored, meaning that a nonapplicable response did not reduce the overall quality score. Articles falling in the bottom third of the quality scores were ranked as good, those in the middle third were ranked as better, and those in the top third were ranked as best. In summary, ERT members were provided two different sources of information related to the quality of the documents they were asked to review. The tier assignment reflected the type of outcome and the quality score results came from the ERT document reviews.

Africa and South Asia Consultations

In parallel with the work of the Evidence Summit process, two consultations were held to discuss regional evidence for population-level social and behavior change interventions, one for Africa and one for South Asia (Figure 1). Two-day meetings were held in conjunction with each consultation. For both consultations, regional evidence was gathered and case studies identified that were used to obtain regional feedback and to lay the groundwork for the final results of the Evidence Summit. These consultations provided background materials for ERT members as they engaged in the evidence review process. In addition, some of the persons who participated in the two consultations attended the Evidence Summit and provided feedback to the ERTs.

The Population-Level Behaviour Change for Child Survival in Africa meeting was held on January 16–17, 2013, in Addis Ababa, Ethiopia, and was sponsored by UNICEF and USAID. An evidence review for this meeting on social and behavior change research was performed by the Center for Communications Programs of The Johns Hopkins University Bloomberg School of Public Health. A preliminary presentation on this review was provided to ERT members who attended the pre-Summit meeting. The review findings were complemented by presentations on the basis of successful interventions that had been used in Africa. This consultation meeting was held in conjunction with a high-level meeting of African Leaders for Child Survival: A Promise Renewed (http://www.unicef.org/infobycountry/ethiopia_67656.html and http://www.unicef.org/ethiopia/ET_APR_MESSAGES.pdf) where there was an opportunity to present some findings on social and behavior change to this audience and pave the way for scaling up effective interventions that have been shown to work in Africa and that emerge from the Evidence Summit.

The Consultation on Social and Behaviour Change for Enhancing Child Survival in South Asia was sponsored by UNICEF, USAID, and the Population Council on January 29–30, 2013, in New Delhi, India. A presentation was made of an evidence review for research carried out in South Asia conducted by the Population Council. An expanded and revised document summarizing this evidence was provided to ERT members during their evidence review process that they could draw upon to supplement their own reviews.

Pre-Summit Meeting

The pre-Summit meeting was a technical working meeting held on February 14–15, 2013, in New York City. Members of the six ERTs came together to learn about the purpose of the Evidence Summit and its anticipated outcomes, discuss the initial review of the literature, and develop work plans for producing evidence synthesis articles and recommendations. The meeting consisted of the presentation of background material and a series of interactive roundtable discussions and breakout sessions to facilitate the attainment of the pre-Summit objectives. The Core Group outlined the expectations for the ERTs and for the products to be completed for the Evidence Summit in June. The products for the Summit were a draft evidence synthesis article and two presentations summarizing the evidence synthesis and providing recommendations for policy, practice, and research. ERTs were encouraged to continue adding relevant documents to the bibliography throughout the development of the evidence synthesis articles. They were also provided with a framework for evaluating the evidence and making recommendations on the basis of both evidence and expert opinion. This framework is subsequently described. The ERTs were also provided draft outlines for their review articles and began the process of modifying the proposed outline and assigning ERT members to work on sections of the review. In general, ERT members were asked to draft sections on the basis of their primary area of expertise (e.g., timing and spacing of pregnancy, nutrition, immunization).

Before the pre-Summit meeting, chairs, co-chairs, facilitators, and coordinators were appointed for each ERT. The chair and co-chair were selected to represent primarily the research and practice communities, respectively, although many had experience in both realms. The facilitator was a senior level person (most were employed by USAID) with content knowledge of the area who served as a resource to the ERTs, helped keep the process on course, and provided technical assistance on content questions. The coordinators were responsible for day-to-day management of the ERTs, arranging meetings, and providing support as needed. After the pre-Summit meeting, periodic ERT leadership meetings were held involving these four representatives from each of the ERTs and other Evidence Summit leadership.

The primary mechanism used by the ERTs to perform their review work was the online workspace which contained links to all of the review instruments, the call for evidence, and background materials on the summit process. The workspace also provided full texts of the review documents and an evidence packet for each ERT which contained identifying information on each of the documents, an abstract (when available), and the results of the relevance and quality reviews that were completed.

Evaluating Evidence and Making Recommendations

ERTs were provided an evidence framework to assist them in drawing conclusions and making recommendations. The framework was developed for the Evidence Summit on Protecting Children Outside of Family Care and is described in more detail in an article describing the methodology for that Evidence Summit (Higgs et al., 2012). Evidence standards have evolved from the medical field where physician decision making is based on randomized clinical trials which provides efficacy data from which health care providers extrapolate guidance for the individual patient. Global health decisions can often be the most complex because they are based not only efficacy at the individual level or within a specific context but also effectiveness at the community and population levels in differing locations and contextually varied environments. Furthermore, sustainability at the country level is critical for country ownership and long-term feasibility. For low- and middle-income country governments and donors, evidence on the feasibility of an intervention's implementation on a population level and its cost-effectiveness are critical to investment and resource allocation decisions that affect the sustainability of the intervention. These three streams of evidence typically result from different research approaches; thus, varying methodologies are needed to evaluate the evidence generated by each stream. Therefore, the evidence evaluation approach for this Evidence Summit allowed for mixed research methods to incorporate evidence targeted to efficacy, effectiveness, and sustainability.

Given the complexity of global health and development questions, using both the evidence base as well as the field experience of experts is important to developing evidence-informed recommendations for policies and programs that can maximize their impact. For some practices that have been widely and successfully implemented, there may not be rigorous controlled trials demonstrating their efficacy, much less their effectiveness and sustainability. It was also acknowledged that the quality of the research support for some interventions may not be very high, which is why ERT members were asked to rate the quality of the studies they were reviewing. In making recommendations for interventions to improve child survival and development, ERT members were advised to use both evidence and expert experience. They were also requested to consider the quality of the evidence and to make clear which recommendations relied more on expert opinion informed by field experience.

Evidence Summit

The Evidence Summit, held on June 3–4, 2013 in Washington, DC, was attended by 210 individuals. Participants included members of all of the ERTs and additional experts in both the science and practice of behavior change and child health and development.

At the Evidence Summit, each ERT provided an overview of their evidence synthesis and draft recommendations for policy, practice, and research. Participants were invited to identify other documents that the ERTs should consider in their reviews and to give written feedback on both the evidence syntheses and recommendations.

After the Summit, the ERTs were asked to consider the comments and recommendations they received during the Summit and incorporate them into revised synthesis articles. In considering how to disseminate these findings, the leaders of the ERTs decided to produce six evidence articles, each exploring one of the focal questions. These six articles form the core of the reports on the Evidence Summit and are published in this special issue of Journal of Health Communication. In addition, a review of the evidence on the use of mass media for behavior change was commissioned during the Evidence Summit process, which used methodologies similar to those suggested for use by the ERTs. That evidence synthesis article is also included in this issue of the journal (Naugle & Hornik, 2014).

Conclusions

There is growing appreciation for the value of identifying evidence-based practice, policies, and strategies to inform global health decisions that are made by both the donor community and by low- and middle-income country governments and service providers. Further work to make this a reality in low- and middle-income countries is needed, but the demand for effective, efficient global health services requires action now. Evidence-informed policy and practice in global health necessitates quality evidence and standardized expert evaluation of that evidence using a transparent process (e.g., Oxman, Lavis, Lewin, & Fretheim, 2009). These challenges, such as population-level behavior change for child survival and development, are complex, multisectoral, transministerial problems. Research to address these problems would focus on the critical questions needed to update low- and middle-income country government policies and civil society organizations on sustainable, effective practices and strategies. The expertise used in designing and conducting studies would parallel the multidisciplinary nature of the problem. The evidence would address relevant contextual and subpopulation issues so countries could know what works in their settings with similar populations. Implementation science studies would follow standard guidelines for reporting that have yet to be developed.

In the process of implementing this evidence summit, several problems arose from the current state of development of research in this area to the development of evidence-based practice that meets these goals. The following are some of these problems:

	No broad consensus exists on a multidisciplinary conceptual model to link the individual-, community-, and systems-level domains to theories of change on how social and behavior change interventions lead to improvements in child survival and development.
	Research seldom addresses the continuum from changes in knowledge and attitudes to changes in healthy behaviors to meaningful changes in public health.
	Research seldom measures the effectiveness of interventions targeting communities, health systems, and policies. Methodological advances and scientific consensus on efficacy and effectiveness research at these levels are needed.
	Research methodologies for assessing interventions that address gender and stigma/discrimination issues are largely absent.
	Although several well-established guidelines for assessing the quality of controlled trials and other empirical intervention research exist (e.g., the CONSORT checklist and the GRADE approach), consensus is lacking on guidelines within implementation science for assessing the quality of program evaluations, qualitative research, and other sources of information on effectiveness of behavior change interventions.
	Consensus guidelines are missing for reporting effect sizes in behavioral intervention research.
	Consensus is lacking on a framework for deriving recommendations on evidence-based practices in the area of behavior change that are more applicable to the challenges faced in low- and middle-income countries.
	Furthermore, consensus is needed on both behavior change and health outcome measures that should be routinely included in effectiveness research and program evaluation. These common measurement tools could facilitate meta-analyses and comparative effectiveness research.

The lack of ideal evidence, or an ideal framework in which to evaluate evidence, should not deter the global health development and research communities from reviewing evidence regarding critical global health challenges. Doing so creates a transparent understanding of what is known, enabling better decision-making processes and more targeted and relevant research agendas while leveraging the tremendous resident expertise in the development, academic, and multilateral communities.

Supplemental Material

Supplemental data for this article (Screening, Call for Evidence, Relevance Review, and Quality Assessment Instruments) can be accessed on the publisher's website at http://dx.doi.org/10.1080/10810730.2014.918215.

Acknowledgments

The authors acknowledge the important role of Elizabeth Higgs, National Institute for Allergy and Infectious Diseases, whose vision for USAID evidence summits was the inspiration for the methodology for this summit and whose earlier publications on summit methodology served as the model for this article (Higgs, Stammer, Roth, & Balster, 2013; Higgs, Zlidar, & Balster, 2012). The authors also acknowledge the overall leadership of Elizabeth Fox at USAID, whose vision and support through the conceptualization and implementation of the evidence summit was essential. The following people also played important roles in developing and implementing the methodology for this evidence summit: Mario Bravo, Mary Sanitato, and Patrick Sheehan at USAID; Vera Zlidar and Neda Dowlatshahi at Knowledge Management Services; and Rafael Obregón at UNICEF. Willo Pequegnat from the National Institutes of Health and Alison Breland from Virginia Commonwealth University provided helpful comments on earlier versions of this article.

Notes

This article not subject to U.S. copyright law.