|Home | About | Journals | Submit | Contact Us | Français|
Numerous health policy makers/researchers are concerned about the limitations of research being applied to support informed decision/policy making and the implementation of practical solutions. The aim of the Chaguo Letu project (which means our choice in Swahili) was to determine how local decision makers could apply a multimethod approach to make strategic decisions to effectively implement a Cervical Self-Sampling Program in Kenya.
A multimethod approach, involving participatory action research, scenario based planning, and phenomenology, was applied in conjunction with two tools to identify relevant factors (negative or positive) that could impact Cervical Self-Sampling Program implementation. A total of 107 stakeholders participated in interviews, focus groups, workshops, and informal interactions. Content analysis, an affinity exercise, and impact analysis were used to analyze data and develop robust strategic directions and supporting implementation strategies.
A total of 57 factors thought to impact the implementation of the Cervical Self-Sampling Program were identified and grouped into 13 thematic categories. These themes were instrumental in developing 10 strategic directions and 22 implementation strategies deemed necessary to implement a technically viable, politically supported, affordable, logistically feasible, socially acceptable, and transformative Program.
This study made three conclusions: 1) there is political will and a desire to improve cervical screening across Kenya, but in a period of dynamic change resources are constrained; 2) implementing the Program in urban/rural settings is logistically feasible, but the majority of Kenyan women could not afford screening without some form of a subsidy, and 3) self-sampling is perceived to be much more socially acceptable than the current Pap screening process. The Chaguo Letu study went beyond the traditional strategy development process of determining “what” needs to do done by describing in detail “how” the Program should be implemented to be relevant and accessible to all Kenyan women at risk of cervical cancer. This work could potentially facilitate communities of practice and knowledge sharing when addressing other types of health decisions in other low resource settings beyond Kenya.
Over three decades ago, Susman & Evered  spoke about a crisis in organizational science in which scholarly research did not relate to the real world of practicing managers. This situation has persisted over time with numerous health policy makers and researchers expressing their concerns about the limitations of research contributing to evidence based decision/policy making [2–5]. Some organizations have emphasized the importance of building the capacity of policy and decision makers to effectively access and use relevant research findings [6–8]. However, progress has been slow; many health organizations have little capacity to operationalize research to strengthen health systems [2, 9, 10].
Some scholars believe that health leaders would achieve greater benefit from science if the research agenda addressed contextual and implementation issues and if researchers and other users collaborated to define the research agenda, allocate resources, and implement the findings [11, 12]. The Participatory Action Research (PAR) methodology advocates for the integration of research into practice and broad-based participation in the implementation of a locally driven agenda [13, 14]. By including local experts (with the required tacit, explicit, and cultural knowledge) in the decision-making process, ownership of decisions could be strengthened, thereby increasing the probability that they will be acted upon [6, 15–17].
This study’s aim was to determine how local decision makers could apply a multimethod approach to make good strategic decisions on how to implement a Cervical Self-Sampling Program (CSSP) in Kenya. This is of particular importance given that 12.92 million women in Kenya aged 15 years and older are at risk of cervical cancer (CC). Since less than 3.2% of women aged 18–69 years are screened, it is not surprising that 4802 Kenyan women are diagnosed with CC annually and 2451 die from the disease [18, 19]. Minimizing the risk of making bad CSSP implementation decisions was mitigated by implementing the Scenario Based Planning (SBP) method  (to ensure a logistically feasible program) and the Existential Phenomenology (EP) method  (to promote a culturally sensitive program that would be adopted by Kenyan women).
Several studies in the United States [22–24]; Netherlands ; Canada ; India, Nicaragua and Uganda , and Kenya , have investigated specimen self-collection to determine if it can increase the effectiveness and coverage of CC screening. These studies generally agree that self-sampling could be a useful component of screening programs, especially where women have limited access to screening, or are reluctant to undergo screening due to cultural beliefs. Jones et al.  reported that 96% of low-income women were very/somewhat comfortable collecting their own sample. Despite organized outreach in the Netherlands, over 30% of women do not respond to screening invitations involving clinician specimen collection . In a rural community in Newfoundland Canada, cervical cancer screening increased by 15.2% where self-collection was available . Rositch et al.  showed that 82% of Kenyan women would feel comfortable using a self-sampling device. Although all of these studies suggest that self-sampling could be useful to increasing overall coverage of cervical cancer screening, there is limited evidence in the literature guiding decision makers on how to implement an appropriate CSSP in low-resource settings. This research study investigated how to fill this gap.
The study design (Fig. 1) for this multimethod qualitative research study included three main components: the Participatory Action Research (PAR) methodology [13, 14], Scenario Based Planning (SBP) method , and Existential Phenomenology (EP) method [21, 29, 30].
Each of these three components made a unique contribution to enabling local decision makers to collectively make informed, practical, culturally sensitive strategic decisions on how best to implement a CSSP in Kenya.
The PAR methodology philosophically grounded the Project Team [the Principle Investigator (PI), Project Management Team (PMT), and Local Decision Influencing Participants (LDIP) members] as they progressed through four study phases: 1) teaming, 2) diagnosing and planning, 3) acting and reflecting, and 4) specifying learning [13, 14]. The Project Team refined three PAR principles to influence their behaviour and guide and direct them throughout the study: 1) collaborative participatory action to solve problems, 2) linking scientific understanding to social action, and 3) co-construction of knowledge through research to influence social improvement [13–15, 31–33]. The degree to which these principles were demonstrated was evaluated by the Project Team at the conclusion of each of four workshops with the use of a confidential evaluation tool. To facilitate compliance with the PAR principles, the PI embraced the concept of power sharing by providing the education and mentoring needed to enable study participants to act as co-researchers, build their decision-making capacity, and take responsibility for co-owning the research process.
The SBP method, with its perception analysis, scenario visualization and impact analysis tools and techniques, was applied to minimize the risk of making bad decisions regarding the design of the CSSP. With this objective in mind, Wulf, Meissner & Stubner’s  SBP method (Fig. 1) was adapted for this study. The method included six steps:
The EP method was applied to investigate the existential meaning of cervical self-sampling and determine the social acceptability of self-sampling by Kenyan women. Understanding what would influence adoption and rejection of the CSSP was crucial to its design. The EP approach [21, 29, 30] consisted of six steps:
Applying any one of these methods on its own would not have enabled the Project Team to design a relevant, sustainable CSSP. The integration of these three components contributed to the co-creation of the knowledge required to design the best CSSP solution. PAR promoted local collaborative decision making behaviour, SBP enabled the design of a logistically feasible CSSP taking all factors into consideration, and EP ensured that the CSSP would be socially acceptable.
To ensure efficacy, credibility and transferability of study results, a multimethod approach to data collection and analysis was applied [29, 40–42]. Investigating multiple constructs (i.e., cervical self-sampling, decision-making, and strategy development) requiring different analysis efforts supported the rationale for applying a multimethod approach . For these reasons, data collection and analysis methods, which are aligned with the PAR methodology [13, 14, 32], SBP method [27, 43], and EP method [21, 29, 30], were incorporated into this study.
Consistent with the qualitative research paradigm of constructivism  and PAR [14, 32], data collection and analysis was a collaborative, participative, reflective and iterative process. Repeated rounds of data collection and analysis took place before, during and after four Workshop events. Figure 2 shows which activities were carried out during each workshop by the PMT and LDIP members, plus the subsequent data collection and analysis PMT members conducted after each workshop. These outputs were then brought forward to the next workshop for everyone’s input and validation.
Data was collected through: formal interviews, focus groups, workshops, and informal interactions. Forms of data collected included: audio tape recordings of interviews, focus groups and workshops, transcriptions of the recordings, and textual descriptions of observations in a personal journal.
Applying the principle of triangulation , a series of analytic processes were conducted in support of four Workshops (Fig. 2). Initially, data analysis focused on conducting content analysis  of outputs from SBP 3600 Stakeholder Feedback interview transcripts to identify factors (negative or positive) that could impact implementation of a CSSP. The similarity between factors was then assessed as part of an affinity exercise  to categorize factors into thematic groupings. Following the design of potential self-sampling scenarios, the phenomenology method of analysis [21, 29, 30] was applied to identify situational and emotional units of significance for adoption of cervical self-sampling phenomenon. As a last step, a final round of content analysis, aggregating all study findings, was conducted to identify CSSP strategic directions and supporting implementation strategies.
Brock University, Action Africa Help International (AAHI), and the University of Nairobi linked their resources and scientific backgrounds, with LDIP’s, to ensure the design of a culturally sensitive, practical approach for cervical self-sampling. With this goal in mind, AAHI applied a purposeful, criterion sampling approach to select the LDIP’s based on: a) existing relationships, b) recommendations from respected key informants, c) fit with desired characteristics, e.g., familiar with cervical cancer, possess technical and/or content expertise, and d) ability to influence solution adoption.
Nine PMT members along with 13 LDIP’s comprised the Project Team (PT), 18 of which were Kenyans (Table 1). Although this was a large team, it proved to be manageable and was necessary to provide confidence in the study results; it ensured that sufficient numbers of relevant stakeholders were engaged in CSSP design during workshops, and promoted future adoption of the CSSP. Other urban participants included eight subject matter experts. In addition, 97 rural participants from four Kenyan counties provided valuable input in the consultation process regarding the social acceptability of the CSSP (Table 1). Determining which counties to focus on and which interview and focus group participants to include were based on their proximity to Nairobi, plus AAHI and University of Nairobi’s relationships that could be leveraged.
The Project Team chose to name this study the Chaguo Letu Project. Chaguo Letu is a Swahili term that means “our choice”.
Applying content analysis  on the Round One 360° Stakeholder Feedback results identified 57 factors thought to have an impact (positive or negative) on CSSP implementation. Grouping these factors as part of the affinity exercise  produced 13 thematic categories (Fig. 3).
Figure 4 shows the results of plotting the 57 factors on the Impact and Uncertainty Grid . Factors with high impact ratings and high to moderate uncertainty ratings, were identified as the critical uncertainties serving as the basis for CSSP strategic directions. Factors that had high to moderate impact ratings and moderate to low uncertainty ratings were incorporated within the CSSP supporting implementation strategies. The remaining factors were identified as secondary elements of least importance.
A total of 107 participants (97 rural and 10 urban) responded to six binary questions asked during five focus groups and 10 interviews (Table 2). The purpose of these questions was to provide background information on the current level of CC knowledge and experience, plus social acceptability of self-sampling by urban and rural Kenyan women. The tabulation of responses for question three shows that all but 10 rural women found cervical self-sampling socially acceptable.
The phenomenology analysis  conducted by four PMT members during two working sessions reduced all interview/focus group statements and quotes into 14 units of significance, followed by clustering these into five themes. Following in-depth reflection on the units of significance by four PMT members, it became apparent that the clusters had an affinity to two main broad categories. Three thematic clusters represented units of significance dealing with situational variables, while two themes emphasized emotional variables (Fig. 5). Examples of quotes (by theme) include the following:
Figure 5 shows that both situational and emotional variables impact the decision to adopt or reject cervical self-sampling. Situational units of significance could be described as precursors to initiating the decision to adopt or reject cervical self-sampling. If these variables are not mitigated, a woman would automatically reject the option of self-sampling. A woman would need to know about cervical self-sampling, can afford the test, and have the physically means and knowledge as to how to collect her own cervical sample, to proceed to the step of deciding if this is something she wants to do. Providing these situational units of significance are addressed, a woman is then in the position to contemplate how she “feels” about cervical self-sampling. This is when emotional units of significance come into play. The level of anxiety and distress she is feeling, how embarrassed she would be to collect her sample, and her husband’s or partner’s perception about this test could influence her decision to either adopt or reject cervical self-sampling.
Data outputs from previous analyses provided the input for the impact analysis of each proposed scenario [49, 50]. This included identifying the risks and consequences of implementing each scenario, and the commonalities and differences between scenarios. This exercise led to the Project Team developing a generic CSSP scenario involving 10 steps (Fig. 6).
In addition, two unique applications were created—one to show how the CSSP could be implemented within a retail/clinical urban environment (Fig. 7), and one to demonstrate its application as a Community Health Volunteer scenario in rural settings (Fig. 8).
Content analysis of all aggregated data analysis outputs provided a mechanism to: translate data into meaningful information, gain insight through assessing scenario impact, construct new knowledge, and subsequently inform the Project Team to make appropriate CSSP implementation choices . As a result, the Project Team was well informed and able to finalize the CSSP strategic directions and supporting implementation strategies.
After extensive data collection and analysis of all factors identified, the Project Team determined that there were 10 recommended strategic directions, along with a total of 22 supporting implementation strategies that collectively provide sufficient clarity on how to successfully implement the CSSP.
Strategic Direction #1—Collector Validation: to acquire proof that the cervical self-sampling device has been successfully validated, has licencing and regulatory approval, and is safe and appropriate to procure and use within Kenya.
Strategic Direction #2—Endorsement: to gain the political will needed, and to ensure comprehensive endorsement of the CSSP (at all levels) by relevant stakeholders and key opinion leaders, from the outset of the Program.
Strategic Direction #3—Policies/Legal: to develop and approve all relevant policies and procedures for implementation of the CSSP, along with any related legal issues, and to ensure alignment of these CSSP policies with current government policies, standards, guidelines and regulations.
Strategic Direction #4—Cost/Funding: to ensure adequate funding to pilot and implement the CSSP, and to develop/implement funding strategies that will ensure that the cost of the CSSP is low and affordable by Kenyan women at all economic levels. This will enable all women to access the Program and realize the benefits of detecting cervical disease early.
Strategic Direction #5—Infrastructure/Human Capacity: to develop and make available the infrastructure that is required, along with sufficient numbers of appropriate, qualified resources to support the CSSP throughout both urban and rural/community environments in Kenya.
Strategic Direction #6—Scenario Logistics—Piloting and Implementation: to initially pilot and validate the logistical feasibility of the CSSP scenario design, followed by refinement of the design, and finally its implementation across Kenya.
Strategic Direction #7—Culture/Awareness: to design and implement culturally sensitive approaches to increase awareness about CC and the CSSP.
Strategic Direction #8—Education: to design and implement (in collaboration with government and relevant educational bodies and professional associations, such as the Kenya Medical Women’s Association) effective cervical self-sampling educational and instructional materials that will provide both trainers and women with the knowledge and skills they need to participate in the CSSP.
Strategic Direction #9—Innovative, Sustainable Cervical Cancer Screening/Value Proposition: to apply out-of-the-box thinking and investigate innovations, which have the potential to transform how women in Kenya get screened and tested for CC.
Strategic Direction #10—Impact Management: to develop and implement a CSSP impact management process that would incorporate data management components to facilitate Program monitoring, evaluation, and reporting, plus leveraging of CSSP data outputs for secondary use.
This study supported the findings of other studies showing that women are willing to collect their own cervical sample. All but 10 out of 107 Kenyan women participating in this study found cervical self-sampling socially acceptable. This is in line with studies conducted by Rositch et al.  in Kenya in which 82% of women were willing to self sample, and the works of Jones et al.  in which 96% found self-sampling comfortable and 79% would prefer to self sample the next time they were screened. Although this finding was important, the key contribution of this study was the determination of “how” to effectively implement a CSSP in Kenya.
The collaborative and participatory efforts  of the Chaguo Letu Project Team were instrumental in the development of 10 strategic directions and 22 supporting implementation strategies that need to be addressed to implement a cervical self-sampling program; one that has the greatest potential to be technically viable, politically supported, affordable, logistically feasible, socially acceptable, and transformative. This study supports the premise that all known variables need to be taken into consideration when planning for the implementation of a CSSP. Dealing with just some, rather than all, significantly jeopardizes success .
The Chaguo Letu study went beyond the traditional strategy development process of determining “what” needs to do done by describing in detail “how” the CSSP needs to be implemented, and by identifying exemplary practices and enablers of predictive and prescriptive decision-making [51, 53, 54]. The application of the 360° Stakeholder Feedback tools provided a structured, comprehensive approach for participants to predict which factors would have an impact, and the underlying uncertainties regarding CSSP implementation. Impact analysis of multiple implementation scenarios enabled decision makers to make informed decisions and prescribe the preferred option [53, 55], allowing participants to: 1) translate the data into meaningful information, 2) gain insight through assessing impact, 3) construct knowledge, and 4) subsequently inform project participants to make appropriate choices [51, 56].
Applying a multimethod approach  using the PAR methodology with the SBP method and analytic tools and techniques helped the Project Team make strategic, prescriptive decisions . Having a combination of one-on-one interviews, focus groups, and group participation at workshops, proved to be a successful approach to support collaborative decision-making. Incorporating the EP method delineated the phenomenon of cervical self-sampling and informed the SBP process with the cultural knowledge needed to design a socially acceptable self-sampling scenario .
Since review of the literature did not identify any similar studies to substantiate the application of this multimethod approach in other settings, the Project Team is recommending that further research be conducted replicating this approach in other low to middle income countries.
This study made three conclusions pertaining to the implementation of a CSSP in Kenya: 1) there is political will and a desire to improve many aspects of CC screening across Kenya, but in a period of dynamic change resources are constrained ; 2) implementing a CSSP in urban/rural settings across Kenya is logistically feasible, but the majority of Kenyan women could not afford CC screening without some form of a subsidy , and 3) the self-sampling approach is perceived to be much more socially acceptable than the current Pap screening process . Unfortunately, the consequence of implementing the CSSP would increase the demand for screening and therefore have a significant impact on the need for infrastructure, human resources, and treatment investments. The current Kenyan health system could not manage this additional burden independently at this time. This study shows a significant need for Public/Private Partnerships (PPP) to share the investment costs associated with implementing a CSSP, which is in line with the PPP Act .
Given the large investments required to implement the CSSP, the Project Team is recommending investigation of coupling self-sampling with biomarker-based point-of-care testing (currently under development) to radically transform the CC screening process from beginning to end . This approach could minimize the need for increased human resources and infrastructure strengthening .
We thank Action Africa Help International for collaborating with Brock University on leading the Chaguo Letu project in Kenya. Their provision of dedicated individuals to participate on the Project Management Team, and take on the responsibility for local project logistical activities, enabled the collection of the data needed for this research study. We also thank all of the local decision-making participants that were members of the Project Team; the CSSP strategic directions could not have been developed without their contributions.
This study was made possible through funding from the Canadian Institutes for Health Research (the Frederick Banting & Charles Best Graduate Scholarship) and Grand Challenges Canada Stars in Global Health grant, which was awarded to Action Africa Help International. The publication of the study results was not contingent on these organizations’ approval or censorship of the manuscript.
The datasets collected and/or analyzed during the current study are available from the corresponding author on reasonable request.
The full PhD thesis for this work will be uploaded onto the Brock University repository once the manuscript is published.
IP prepared the first draft of the manuscript. IP collected the data for the study and developed the study design with support from JC, CK and GO. IP, CK, GO, and JC all contributed to the data analysis and editing of the manuscript. All authors read and approved the final manuscript.
None of the authors have any competing interests (financial or non-financial) in this manuscript.
The Kenyatta National Hospital/University of Nairobi Ethics & Research Committee (KNH-ERC/A/69), and the Brock University Research Ethics Board (13-059) both granted formal ethics approval for this study.
A signed statement of consent was obtained from all study participants.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Irene Podolak, Email: ten.lleb@kalodopi.
Caroline Kisia, Email: gro.plehacirfanoitca@aisikc.
Gloria Omosa-Manyonyi, Email: ek.ca.ibnou@asomog.
Jarold Cosby, Email: ac.ukcorb@ybsocj.