New Methodologies for Global Health Research: Improving the Knowledge,Attitude,and Practice Survey Model Through Participatory Research in Haiti

Introduction

Knowledge, attitude, and practice (KAP) surveys form the foundation for many health education initiatives and outcome improvement strategies around the globe. However, in both high- and low-income settings, a critical epistemological examination of whether the KAP survey model can generate in-depth and valuable knowledge that can be used for health program planning is so far lacking.

Currently, researchers and global health practitioners use KAP survey models to reveal insights necessary to health program design and implementation (Lawal & Taiwo, 2018). International development agencies such as the United States Agency for International Development (USAID) also employ KAP surveys for a variety of reasons such as acquiring a baseline assessment of the target population, planning community interventions, and ensuring future funding (Launiala, 2009; Norwegian Refugee Council, Save the Children, & World Vision, 2015; USAID, 2008). Widespread use of the KAP survey model has encouraged the publication and distribution of resource packs and guides by international organizations. The World Health Organization (WHO) resource pack, for example, outlines sample KAP questionnaires to help governments and response partners “rapidly obtain valuable and insightful information to tailor interventions to better address people’s needs at a community level, thereby contributing to the overall public health response” (WHO, 2016, p. 4). Researchers highlight that the KAP survey model allows them to better understand and document observed behavior (Ogoina et al., 2015), guide health education initiatives (Hejazi et al., 2010; Vanslyke et al., 2008), and inform health policy across a variety of social, cultural, and economic settings from low- (Hairi et al., 2003) to high-income countries (Smith & Umenai, 2000).

Following a KAP survey model may be an easy way to gather information quickly, but a critical evaluation of the survey’s methodological shortcomings has established this model’s limitations. As Launiala (2009) argues, even though KAP surveys are easy to conduct and are cost effective to perform, they rest on two assumptions: first, that the results can be generalized to a larger population, and second, that there is a direct relationship between knowledge and behavior. It is tenuous to assume that a sample is representative of the social and cultural attitudes of an entire population and especially that knowledge and attitudes are directly correlated to the decisions of a person or community to behave in a certain way. A KAP survey performed in Mongolia demonstrates that despite being powered to be “representative” of the country, participants were only offered choices such as “nothing at all,” “only know the term,” “a little,” and “very familiar,” when asked to self-report their knowledge of breast cancer (Yerramilli, Dugee, Enkhtuya, Knaul, & Demaio, 2015). Although the population may be representative, the data generated from questions such as these do not signify their understanding of the topic in any representative fashion. Furthermore, when KAP surveys are used as knowledge-generating tools by international development practitioners and global health researchers, the communities and populations that allegedly benefit from the research outcomes rarely assist with the definition of the research question, study design, and interpretation. This leads us to question the relevance of KAP survey findings for the community and its capacity to provide socially and culturally informed data as well as how much that knowledge affects real-world decisions.

Participatory research methodologies have for years sought to address criticisms leveled at public and global health research top-down approaches. Community-based participatory research (CBPR) projects, for example, stem from long-term collaborations between academic institutions and underserved communities to conduct heath research that is relevant to community needs and seeks to provider deeper and more critically informed knowledge of social and cultural behaviors (Minkler & Wallerstein, 2008). The methodology was developed to address power imbalances in knowledge production by realigning the researcher/researched power relations (Israel, 2005), and to ensure that the research objectives were determined in cooperation with the community to ensure it met their stated needs. According to public health researchers, although community partnerships at the center of CBPR can be long and complex to mature, these partnerships ensure that the studies are relevant to the community. CBPR methods also seek to build sustainable programs and partnerships, which are important to uphold when vulnerable communities are sacrificing time and scarce resources for the research partnerships (Kobetz et al., 2009; Parker & Kaufman, 2009).

CBPR models are also being employed increasingly in the Global South. In the Dominican Republic, a CBPR study was conducted by academic researchers from the United States in partnership with Dominican health professionals and Dominican community health workers to determine adequate indicators for improving maternity care (Foster et al., 2015). Mano a Mano in rural Bolivia practices a CBPR model to create new sustainable community health centers by working only with communities that request partnerships (Velasquez, Knatterud-Hubinger, Narr, Mendenhall, & Solheim, 2011, p. 309). They conclude that,

Failure to do this risks the sustainability of what is ultimately produced; history is replete with top-down “community projects” that have fizzled as soon as their principal investigator’s funding or grant ran out and/or a charismatic leader disengaged his or her associated effort(s).

Research about attitudes toward cancer in Haiti is relatively new, as widespread treatment has only recently become available. To date, studies have often been written in English and performed without local grassroots participation (Gichane et al., 2017; McCarthy et al., 2017; Zahedi, Sizemore, Malcolm, Grossniklaus, & Nwosu, 2014). Furthermore, they draw on samples recruited from health clinics only, use closed-ended questions or are traditional KAP surveys. Other research has been performed in Haitian immigrant populations in Miami (Barbee et al., 2010; Ilangovan et al., 2016) and Boston (Allen et al., 2013), whose life experiences are significantly different from those still living in Haiti. Research has almost exclusively been conducted by nongovernmental organizations (NGOs), with a few done by international university-based researchers and none from Haitian-founded organizations.

In other low- and middle-income countries, research on women’s cancer is similarly being performed largely as KAP surveys (Gilani et al., 2010; Jahan, Al-Saigul, & Abdelgadir, 2006; Mena et al., 2014), initially written in languages (usually English) other than in the native language of the population being studied, or with limited involvement of local people in writing the questions (Bradley, Risi, & Denny, 2004; Yerramilli et al., 2015). We were only able to locate a handful of studies on women’s cancer in low- and middle-income countries that employed participatory methods, demonstrating a clear need for more research, and we found nearly double the number of articles that used KAP surveys as their primary method.

We contend that if global health and development organizations adopted a CBPR approach (or other participatory research approach) to researching and designing programs, the communities they serve would hugely benefit. The challenge for smaller community-based organizations (CBOs) and NGOs working in community health is to undertake this approach without multiyear funding and limited availability of time and human resources required to maintain long-term partnerships in a constantly evolving landscape of government and international donor priorities. However, the principles that guide CBPR and other participatory research methods such as community participation can and should still be adopted by global health organizations.

We aimed this article at global health professionals, and present a case study of how to incorporate participatory research methods into global health programming while keeping time and funding limitations in mind. To illustrate this methodology, we will discuss a research project conducted in Haiti in partnership with the Ministry of Health, civil society, and international organizations, and cancer patients.

The Need for Qualitative Research About Breast and Cervical Cancer in Haiti

In 2015, we spearheaded a research project to investigate the way that breast and cervical cancer are understood by people in Haiti through Innovating Health International (IHI), an NGO working to treat breast and cervical cancer care in the community since 2013. This research was prompted by the need to design an intervention to improve the cancer stage at patient presentation, thereby increasing survival rates. Despite the availability of cancer screening and treatment, nearly 80% of women in IHI’s program presented with advanced stage breast cancer, with a mortality rate of 19% in the first year after presentation. Furthermore, more than half of the patients in the program waited more than 12 months to seek medical care after first noticing the breast mass (DeGennaro et al., 2018). Breast and cervical cancer screening, epidemiology, and treatment availability in low-income countries, including the current situation in Haiti, is discussed in depth elsewhere (DeGennaro et al., 2018; Gomez et al., 2017; Santamaria, Tillyard, Taylor, Bernard, DeGennaro, 2016).

The causes of lack of engagement with health services by women with breast and cervical cancer in Haiti and the subsequent high numbers of late-stage presentations is relatively unstudied and unknown. Anecdotal observation, in-depth conversations with IHI cancer patients, and a KAP survey completed in 2013 (Majid & DeGennaro, 2013) revealed that patients may not seek care when they have symptoms of cancer because they fear rejection from the home and or their community, belief that treatment does not exist or is not effective, as well as economic concerns over cost. Similar concerns have been documented through studies in other countries.

Haitian partner organizations working in cancer care as well as the director of oncology in the Haitian Ministry of Health agreed that further research from a cultural and sociological perspective was needed to identify the social, gender, and cultural context of cancer in Haiti. Drawing the first principles of CBPR that require a research project to begin with a consultation of community needs and perspectives, IHI held an in-depth focus group with the support of the Haitian Cancer Support Society (GSCC) to explore the social and cultural aspects of cancer for women in Haiti, with cancer patients in treatment and recovery. This scoping study, generally defined as a project to map existing evidence to inform future research or practice (O’Brien et al., 2016; Pollock, Fernandes, Becker, Featherstone, & Hartling, 2016), helped set priorities and revealed the need for a larger research project that looked at and sought to address the structural causes of late-stage presentation of women with breast and cervical cancer in Haiti. One of the preliminary findings from the scoping study indicated that many women with breast cancer have a lot of fear about the disease and the impact that it will have on their life as part of a community. Family members, however, fear the loss of a loved one, their powerlessness in the face of the disease, and the huge financial burden to treat what many perceive as an inevitable death sentence. Therefore, we decided to launch a broader research project that would seek to investigate collective as well as individual narratives about cancer care with a view to using this information for program design and service delivery.

Setting Up a Project Advisory Board

CBPR methodologies concentrate on research that is conducted with particular communities (Etowa, Matthews, Vukic, & Jesty, 2011). Setting up a community advisory board (CAB) is a staple methodological approach of CBPR in accordance with its defining principle that the methodology “requires a partnership comprised of committed community members, organizational representatives, and academic researchers” (Rhodes, Malow, & Jolly, 2010, p. 176). Community members, therefore, must have an active role in the production of knowledge. Most academic proponents of CBPR state unequivocally that the partnership that is set up between the community and the academic institution must constitute ongoing commitments of all members. Equally, CBPR researchers recognize communities as social entities that go beyond criteria such as geographical location. In fact, for many researchers, people who share common areas of interest or action also constitute a community.

For our purposes, we found that the level of commitment and resources required to set up a long-term CAB exceeded the capacities of a global health organization that receives inconsistent and project-specific funding. Furthermore, it is difficult to justify ongoing costs associated with research to donors who want to fund health programs, in spite of the fact that research is essential to successful implementation. To honor the participatory methods of CBPR ¹ but to account for the constraints of an operational NGO, the first action taken by IHI was to set up an advisory board that was active for the duration of the project that would help design a study in line with the priorities of cancer patients, women’s organizations, as well as medical organizations familiar with the challenges of cancer care in Haiti. The board was active for 12 months and was funded as part of a project implementation grant. All members of the advisory board participated in providing feedback on the process through monthly in-person meetings of the group and regular emails. Board members frequently brought additional people to participate in one or two meetings as well, further diversifying the group’s responses. The advisory board also oversaw the collection and analysis of data, as well as assisting in the write-up of a comprehensive set of recommendations on the basis of the findings.

The advisory board activities for IHI’s research project falls short of CBPR’s requirement, and the board members also did not strictly comply with the definition of community, but the organizational makeup of the project advisory board reflected the need for a wide range of expertise and standpoints that had become evident in the project scoping study. To design a set of research questions that could adequately explore the complexities of breast and cervical cancer in Haiti, women’s organizations with expertise in gender-based violence and economic development as well as organizations working in medical care made up the advisory board. After careful consultation and drafting of terms of reference, the board was made up of representatives of five different community organizations: IHI, a community-based NGO called J/P Haitian Relief Organization (J/P HRO), Femmes en Democratie—a national network of women’s groups, the Group de Support Contre le Cancer (Haitian Cancer Support Group), and the Haitian Ministry of Health’s Oncology Department. The board was mostly comprised of women and also had breast cancer survivors. The inclusion of women and survivors from different parts of the country ensured that the survey questions addressed their concerns and could be understood by them.

Redesigning the Scope of the KAP Survey Model

The advisory board’s first objective was to design a research project by adapting the KAP survey model through a participatory and consultative process. This meant first analyzing the shortcomings of the KAP survey model and subsequently addressing them through collective consultation.

As previously discussed, global health organizations produce a huge amount of knowledge about the health issues faced by the communities they serve. Global health organizations are often subject experts that have a permanent institutional foothold on the ground and share knowledge in the form of published academic research but mostly through other forms of content such as organization reports and briefings (Delisle, Roberts, Munro, Jones, & Gyorkos, 2005). These research findings, along with anecdotal observations and program participant feedback, inform the policy and design of health programs all over the world (Eisenberg, 2001). Health data that explore community needs are mostly generated through rapid needs assessment and tools such as the KAP surveys.

But KAP surveys are often designed by international global health experts working with or for international organizations. In many cases, KAP surveys, such as the investigation into barriers to cancer care in Mongolia by Yerramilli et al. (2015), are cowritten by a panel of a mixed national and international medical experts. In most instances, people in underserved communities and/or those affected by the health issue being researched are not included in the panels or research design and implementation process. Still, the KAP is widely perceived to be an adequate method to gather satisfactory data to plan development and long-term health initiatives. USAID describes the KAP survey as,

A Knowledge, Attitude and Practices (KAP) survey is a quantitative method (predefined questions formatted in standardized questionnaires) that provides access to quantitative and qualitative information. KAP surveys reveal misconceptions or misunderstandings that may represent obstacles to the activities that we would like to implement and potential barriers to behaviour change. (USAID & Médecines du Monde, 2011, p. 1)

This statement reveals some of the common characteristics of KAP surveys: (a) it is often designed by international and local public health experts with their specific programmatic and informational needs in mind, (b) it is most often a survey comprised of a set of closed-ended questions, and finally (c) it aims to identify “misconceptions” about biomedical practices, so that they can be addressed and “corrected.”

These three cornerstones also present some of the most challenging conceptual underpinnings that the advisory board addressed by critically examining and redesigning the survey. First, KAP surveys often fail to look into the wider cultural context surrounding a health intervention (Hausmann-Muela, Ribera, & Nyamongo, 2003). As Launiala (2009) explains, specific concepts of KAP rely on an implied cultural understanding of the distinction between a form of biomedical or “scientific” knowledge, a culturally specific or traditional belief, and a way of behaving or cultural practice. “Knowledge” is presumed to be based on scientific facts and universal truths, which, therefore, implies “knowing” about biomedical information. In contrast, beliefs refer to traditional ideas, which are erroneous from the biomedical perspective, and which form obstacles to “appropriate” behavior. For example, in Médecines du Monde (2015), in a KAP training manual, this distinction is exemplified through the following survey questions; to understand a community’s level of knowledge about malaria, the example question reads, “Do you think mosquitoes might be responsible for malaria? Yes/No/Don’t know.” To discern a community attitude to tuberculosis, the question reads, “If you think you have been exposed to tuberculosis after contact with someone who was coughing, what would you do? Go see a doctor/Take traditional medicines/Go to a laboratory/Nothing.” To establish a community practice, the question asks, “Did you protect yourself by using a condom when you last had sex? Yes/No.”

This example KAP survey example highlights a number of methodological issues. First, the survey questions appear, on one hand, to be specific to the health intervention or problem and, therefore, “generalizable” to a population but the researchers fail to take other general questions into account. For example, the respondent is not able to answer that he or she would not go to a doctor because of the distance between his or her house and the nearest clinic or due to other potential barriers not listed in the multiple-choice answers on offer. For example, a traditional healer may be located much closer and, therefore, would be the first port of call, but not necessarily the primary preference. Equally, the respondent may not report to have had protected sex because of any number of circumstantial factors that are, in our view, difficult to generalize into a notion of “practice.” The margin of error in the results due to false reporting when respondents are asked a leading question or provide the answers in multiple-choice, closed-ended format is difficult to characterize as knowledge. The leading questions could also be considered judgmental or intended to educate the respondent instead of attempting to evaluate his or her biomedical knowledge. More important, any public health intervention designed on the basis of a result of such a KAP survey might fall short of addressing the most urgent needs. For example, an educational intervention based on these research findings could fail to address some of the largest barriers to women’s cancer care, and the organization would miss the opportunity to address other barriers by offering transportation fees or providing counseling, for example.

Furthermore, the narrow scope of the research question reveals the lack of importance placed on the wider sociocultural context that people, not simply diseases, inhabit. The preference for a purely quantitative research approach aggravates this issue because assumptions made by researchers about the cultural context cannot be corrected or amended through open-ended alternative questions. For example, if we look back at the earlier tuberculosis sample KAP survey question above, we can note how the closed-ended question not only limits the number of possible responses that the participant can choose but also that the options that are given are reflective of what the global health researchers presume to be the most important information for them to have “correct” attitudes and perceptions about tuberculosis. Barriers to accessing health care that may be influenced by complex issues linked to gender, race, or cultural dynamics are often downplayed or not investigated at all in KAP surveys designed by global health researchers, even though they may be of paramount importance to designing effective community–health interventions.

Given these shortcomings, the advisory board made three key innovations to the KAP survey model. First, the breast and cervical cancer awareness survey was restructured in a way that the board felt was more intuitive to the average Haitian respondent. One of the key aspects of this restructuring was reorganizing the sections of the survey around subject areas instead of around different forms of perception and behavior. The survey was designed to have three main sections. In addition to standard demographic information, the first section addressed knowledge and community practices around breast and cervical cancer, the second section investigated knowledge and practices concerning health infrastructure generally as well as breast and cervical cancer services specifically. The final section was titled Kominikasyon (Communication) in Haitian Kreyol, and investigated the ways that communities had experiences of breast and cervical cancer as well as how they communicated about it. The women’s organizations on the board were the driver behind expanding the scope of the survey from individual perception to a wider query about the respondents’ experience of community behavior. Furthermore, they provided insight into the gendered aspects of cancer and, therefore, advocated that 25% of study respondents should be people who identify as men. The advisory board elected not to discuss sexual orientation of the respondents as part of this study.

Second, the advisory board designed a qualitative and quantitative mixed-methods survey that gave respondents the space to express their own understanding of breast and cervical cancer by answering open-ended questions. This approach is different to a standard KAP survey model that tends to prioritize quantitative data collection. As indicted in this statement in the training manual, “One advantage of a KAP survey is to allow, in a single survey, the collecting of a large amount of data that will be subject to statistical analysis (which qualitative methods of data collection do not allow)” (Médecines du Monde, 2015, p. 5). In contrast, the advisory board opted for a mixed-method approach that they felt held the considerable advantage of limiting the assumptions made about community attitudes by either experts or nonlocal people who might design the study. The board carefully balanced the surveys to include both closed- and open-ended questions. For example, instead of simply asking a respondent whether they knew the symptoms of breast cancer, the questionnaire also allowed space for the respondent to specify the symptoms in their own words. This allowed for a wide scope of qualitative responses, so that, instead of seeking to determine answers as simply “correct” or “incorrect,” the board could engage with the expressed perspectives as meaningful qualitative data. The board, comprised of mostly non–medically trained participants decided what constituted an “accurate” or “correct” answer, taking into account cultural interpretations of biomedical knowledge. This was done as an in-person discussion and finalized by group consensus. The approach also enabled the board to establish the margin of error for statistical significance and “correct” responses that would be used in a quantitative research study with only multiple-choice questions.

Third, the advisory board designed a set of follow-up questions for focus groups with participants grouped according to age and location. The purpose was to build a conversation around the qualitative responses in the questionnaire, so that focus group leaders could provide a more detailed summary of what was discussed and get a deeper look into wider sociocultural issues. Furthermore, advisory board members designed the follow-up session also as an ethical obligation to address any queries that the respondents may have had about cancer. The end of the focus group, then, served as a forum in which the focus group leaders could provide accurate information about breast and cervical cancer and available screening and treatment services to participants, ensuring that the community derived some immediate benefit from participation.

Data Collection

The project received ethics approval from the institutional review board of the University of Florida College of Medicine and the Haitian Medical Association’s Bioethics Committee. Participants were interviewed and focus groups were held in the Western, Northern, North Western, and Central Plateau geographic regions in accordance with the areas of work of the organizations on the advisory board. A total of 414 questionnaires and 35 focus groups were conducted over a 2-month period in four geographic regions of Haiti by employees or related volunteers from IHI, GSCC, Femmes en Democratie, and J/P HRO. Each group interviewed participants in the community in which they serve, increasing the speed with which the surveys could be completed and ensuring a broad sample of the population. Questionnaires and focus groups’ notes were completed by hand and then dispatched to Port-au-Prince where a data analyst inputted the responses into the University of Florida Research Electronic Data Capture system, a secure, web-based application designed for data capture and analysis (Harris et al., 2009). The open-ended survey responses were initially recorded as direct responses, then coded into main themes with the use of NVivo software. They were additionally transformed into nominal data for statistical analysis, as determined by the advisory board. In real time, the focus group leaders wrote down dominant themes, important quotes, and areas of disagreement for each focus group. The transcripts were then coded by grouping similar responses and questions into subject groups. From there, themes within the responses and questions were identified and transformed into nominal data by NVivo software.

Data Analysis

Drawing on the core principles of participatory research that any project must include a multidirectional exchange of information and learning and that research products must be shared among members of the board, the data were jointly analyzed and distributed among all board members. A subset of the board with statistical analysis skills aggregated the data, made it into pivot tables, graphs, and grouped results into visual graphics by thematic and word association. The collective responses to each question in the survey were disaggregated into pivot tables and reported according to gender, geographic location, education, and income, and in a written report and an oral group presentation.

These results were shared and collectively analyzed by the advisory board and the respective research teams of each organization during the course of an intensive 1-day working group conference held in Port-au-Prince. In addition to the original board members; Haitian NGO, the Commission of Women Victims for Victims; and IHI’s cancer survivors’ support group were asked to join the data analysis and interpretation because the data revealed cross overs between local understandings of breast and cervical cancer and gender-based violence. The group discussed possible associations for most frequent qualitative responses, postulated hypotheses for trends, and attempted to resolve seeming discrepancies and overlaps between the qualitative and quantitative data.

Discussion

The complete findings of this study are discussed in detail in a separate results paper coauthored by the project leaders (Tillyard et al., 2018). We wish to highlight here some of the effects that our chosen methodology, a participatory approach to KAP survey design, data gathering, and analysis had on the results and findings of the study.

Most KAP survey projects tend to limit rather than open up the scope of participant responses through closed-ended questions, designed by public health professionals. Our approach followed the examples set by other participatory initiatives that created knowledge applicable to the specific context by drawing on the insight of nonmedical and ordinary people. For example, communities and researchers using CBPR in South Africa jointly created the concept of “cervical health,” which “acknowledged the impact on women’s bodies and lives of HIV-AIDS and STDs, sexual violence, poverty, and multiple social problems” (Mosavel, Simon, van Stade, & Buchbinder, 2005, p. 2577), instead of merely discussing cervical cancer.

First, the inclusion of women’s organizations as well as cancer patients and survivors on the advisory board was instrumental to shifting the approach to knowledge production. This method led to the research project uncovering social and cultural questions surrounding breast and cervical cancer in Haiti as well as barriers to care. This was achieved by including questions with an emphasis on collective, rather than individual, attitudes, and on social and structural barriers, rather than simply on health practices. Open-ended questions were key to revealing the wider social context of breast and cervical cancer in Haiti. For example, the study revealed that the most common cause attributed to breast cancer was being hit in the breast. The discussions by the focus groups helped to clarify that this statement was related to forms of gender-based violence in the home and the community. The findings similarly revealed a perceived link between causes of cervical cancer and promiscuity, with early sexual debut and frequency of sexual encounters but no mention of human papillomavirus (HPV) as the most cited causes. The space given to open-ended responses allowed for a set of findings that transcends conventional KAP survey preoccupations about correct and incorrect perceptions, and instead pointed to a set of issues that may be relevant to health programming. During the joint analysis of the research findings, the advisory board hypothesized that these findings might indicate that a link between the causes of cancer, fear of physical violence, and perceptions about promiscuity plays a role in key understandings of disease, and may also have serious implications for attitudes to seeking cancer care and the amount of support they may be offered by their partner or family.

Second, redesigning the KAP survey and the inclusion of a third section of the survey that was concerned with community, rather than individual, attitudes led to important data that further characterized the complicated cultural landscape of health care in Haiti. One of the crucial findings of the survey is that the wealthiest respondents demonstrated more knowledge across all questions with much greater frequency than the most educated. Although the more educated were not more likely to discuss illness with their family than others, the top 10% most wealthy did so at nearly double the rate of all others. The advisory board attributed these findings to the circulation of medical knowledge through transnational and diaspora networks, as family who send remittances back to Haiti not only increase their economic power but also foster more passage of biomedical health education and information from connections in the United States, Canada, and Europe. Furthermore, the study revealed that the poorest people, those who earn less than a dollar a day, did not report to be the least informed but in fact the second most informed in virtually every question. By contrast, the working poor reported the lowest knowledge levels overall. The advisory board’s analysis that the working class are working all the time and, therefore, have little to no contact with health services or international organizations’ services, lead the board to recommend for increased engagement with this demographic as well as workplace point of care medical programs. Again, the board’s diverse expertise enabled a much wider interpretation of data and richer overall understanding of data that might otherwise appear incongruous.

Third, it could be argued that our approach to reformulating the KAP survey is cumbersome as it seemingly makes a “lightweight” tool that is designed to provide uncomplicated, high-level, quantitative data about a particular health problem into a more laborious and complicated research process. However, our study shows that the existing KAP survey approach is flawed because it seeks to discern “correct” and “incorrect” perceptions through self-reporting without accounting for cultural perceptions or additional barriers researchers failed to offer as possible options. Previous research using KAP surveys in Haiti and other countries is, therefore, not able to define answers with specificity beyond Western biomedical research conceptions (McCarthy et al., 2017; Zahedi et al., 2014). For example, a KAP study conducted in Pakistan on 1,000 randomly selected women demonstrated high rates of knowledge about breast lump as symptom of cancer, but low rates of action in response. In addition, they found that 82% of women in Pakistan can recognize breast feeding as protective against breast cancer, which has questionable value when asked as a yes/no question, and all but one of the risk factor questions have “yes” as the “correct” answer. The only “no” question was “Is breast cancer caused by magic or evil spirits?” (Gilani et al., 2010, p. 207). Our mixed-methods study enabled us to assess the extent to which KAP surveys fulfill their objective of providing accurate quantitative data by asking participants who declared that they knew a risk factor, symptom, or complication of breast or cervical cancer to then provide at least one correct answer. Our study shows that participants correctly self-reported knowledge of breast cancer risk factors (14%), complications (86%), and symptoms (40%), and cervical cancer risk factors (60%), complications (82%), and symptoms (39%) of the time. More women knew where to get cervical cancer screening than what the test was, and most could not identify where in their body the cervix is. If the KAP survey model provides statistically inaccurate data, and given, as we have seen, that this is the study’s primary function, we make the case that a mixed-method or qualitative approach is not only useful but also essential to provide, supporting data to understand the health challenges experienced by the population. Similarly, without including fundamentals of participatory research that include local viewpoints, the conclusions drawn and, therefore, the interventions designed may miss the primary cause of certain health behaviors.

Finally, drawing on the CBPR commitment that research must have the potential to increase individual and community capacity (Rhodes et al., 2010), the findings of this research project were used to spearhead a number of follow-up initiatives. The qualitative and quantitative data were used to design a breast and cervical cancer education and communication tool kit that contained resources for doctors, midwives, community health workers, and cancer patients. The tools were codesigned with the GSCC, cancer patients, and their family members through a human-centered design approach. The tool kit is available for open-access download online at KanseAyiti.com. Based on our results, IHI applied for and received a grant to implement HPV-based cervical cancer screening for working women in the garment industry factories in Port-au-Prince from the Canadian government. Other programmatic initiatives in development include strengthening links between local organizations working in gender-based violence and cancer treatment facilities.

We have presented an actionable and replicable method that can be adopted by global health organizations. This methodological innovation does not, however, go as far as to challenge or disrupt dominant power relations or address colonial legacies and the role these play in forming and perpetuating health inequalities, although this is undoubtedly the next step (Darroch & Giles, 2014). Further reflection on the relationship between health research, structural health inequalities, and disregard for local and nonexpert knowledge is necessary to answer the broader and most important question, what does public health research say about the historical and cultural legacies of colonialism?

Conclusion

We present the steps taken by a global health organization to build on and improve one of the most commonly used research tools in global health: the KAP survey model. We hypothesize that the widespread use of this method can be attributed to the limited funding and resources available for community organizations, NGOs, and universities to conduct research in contrast, for example, to high-resource academic institutions, for-profit companies, or governments researching public health issues. The shift in the academic and governmental spheres toward CBPR models is at once a huge improvement but is somewhat exigent for small CBOs or NGOs to replicate. Consequently, we present a research methodology that draws on principles of CBPR as a first step to addressing asymmetrical power relations in the knowledge production and program planning of international organizations. Although we stop short of adopting all the principles of CBPR, we instead use it as the inspiration for methodological innovation.

CBPR may not be possible for all groups seeking to evaluate behavior and behavioral changes due to financial or time constraints. However, to rely solely on quick-fire surveys such as KAP has serious shortcomings as we have elaborated in this article. We have shown the clear benefits of a mixed-method data enquiry that accounts for time and budget constraints of operation organizations.

Forming an advisory board that included nonmedical experts, women’s organizations, and cancer survivors was key to expanding the focus of our enquiry and making it relevant to community and patient needs. Second, designing a mixed-methods survey enabled a much closer look into the wider cultural context of disease. Last, the role of the advisory board in interpreting and setting programmatic priorities using the data ensured that the exercise retained value and meaning for the community.

We conclude by highlighting how valuing local knowledge and processes of mutual learning reveals the imperative connections between social and medical issues. It also goes beyond simply documenting barriers to treating breast and cervical cancer in Haiti; it is about the care and ethics that inform knowledge production and improve methodologies for creating global health data. We maintain, that more work must be done to craft suitable methods for organizations based in the field that embrace the principles of CBPR methodologies and related methods such as participatory action research and community-based research. In particular, we stress that community-led research design, collection, and analysis is fundamental to successful health program implementation.