Using intervention mapping to design and implement quality improvement strategies towards elimination of lymphatic filariasis in Northern Ghana

Introduction The Global Strategy to Eliminate Lymphatic Filiariasis (GFELF) through Mass Drug Administration (MDA) has been implemented in Ghana since the year 2000 and transmission has been interrupted in 76 of 98 endemic districts. To improve the MDA in the remaining districts with microfilaria (MF) prevalence above the 1% threshold for the interruption of transmission, there is a need to identify and implement appropriate quality improvement (QI) strategies. This paper describes the use of intervention mapping to select QI strategies to improve an existing evidence-based MDA program in Northern Ghana. Methods Due to the complexities associated with implementing evidence-based programs (EBP) such as the lymphatic filariasis MDA and variability in the context, an initial assessment to identify implementation bottlenecks associated with the quality of implementation of lymphatic filariasis MDA in the Bole District of Ghana was conducted using a mixed methods approach. Based on the findings of the initial assessment, a context specific QI strategy was designed and operationalized using intervention mapping strategy in terms of seven domains: actor, the action, action targets, temporality, dose, implementation outcomes addressed, and theoretical justification. Results The initial needs assessment shows that the persistent transmission of lymphatic filariasis in the Bole District is characterized by high levels of refusal to ingest the drug, high levels of reported adverse drug reactions, low MDA coverage at community level, poor adherence to the MDA protocol and non-participants’ responsiveness. Conclusion This study has shown that it is feasible to develop a context specific QI strategy for an existing evidence-based intervention based on an initial needs assessment through stakeholder participation using the IM approach. However, working (towards) QI requires more time than is usually available in public health service. Sufficient theoretical knowledge of implementation research and experience with technical IM experts must be available.


Introduction
Over 120 million people have been infected with Lymphatic Filariasis (LF) worldwide [1,2]. This disease is caused by microscopic, thread-like worms (filarial) known as Wuchereria bancrofti, Brugia malayi or B. timori [1,2]. Lymphatic Filariasis is the second leading cause of permanent disability after leprosy [3] and is considered a Neglected Tropical Disease (NTD) [4].
In 1997, the World Health Assembly passed a resolution to eliminate LF by the year 2020. The global program to eliminate the disease was then launched by the World Health Organization in 2000 [5]. Towards achieving this target, an evidence-based intervention (EBI) that involved annual mass treatment with a single-dose of diethylcarbamazine (DEC) or Ivermectin (IVM) in combination with Albendazole (ALB) was employed by endemic countries. Multiple rounds of this treatment are necessary to reduce the microfilaria prevalence to zero [6].
In June 2000, Ghana became one of the first West African countries to implement the MDA strategy through the Ghana Filariasis Elimination Program (GFEP). The target of the program in Ghana was reducing the LF prevalence level to below 1% after 4-6 rounds of highcoverage (>80%) MDA [7].
After 15 years of implementing the MDA program in Ghana, a total of 76 districts out of the initial 98 endemic districts achieved the target [8]. The remaining 22 districts still have prevalence above the 1% threshold [8] and are now referred to as LF "hotspot" districts. Bole District in Northern Ghana is one of these districts. The 2016 microfilaraemia prevalence in Bole District was 1.9% with some communities having prevalence rate as high as 5.9%. A baseline study to assess the coverage and quality of implementation of an ongoing Mass Drug Administration (MDA) in the Bole District of Northern Ghana was conducted. The baseline results described in detail later in this paper indicated that, the persistent transmission of the lymphatic filariasis in Bole District is characterized by high levels of refusal to ingest the drug, high levels of reported adverse drug reactions, low MDA coverage at a community level, poor adherence to the MDA protocol and a lack of responsiveness by nonparticipants.
Based on the results of the baseline, a quality improvement (QI) intervention was proposed to improve the quality of implementation. Experts in the field of implementation research have emphasized the need for a rigorous, theory-based approach to select intervention strategies that link to the barriers. One such approach to selecting an evidence-based strategy is intervention mapping (IM).
Intervention mapping is a program planning framework for the development of theory and evidence-based health interventions [9]. Intervention mapping offers procedures and tools for establishing empirical and theoretical bases for selecting health intervention programs, for the translation of theory into actual intervention activities and materials, for the management of program adoption and implementation and for partnership between health system, researchers, priority groups and stakeholders [9]. It also describes the intervention development process. Intervention mapping enables the development of health interventions that may accomplish program objectives and match priority populations and intervention contexts. The IM strategy suggests that the priority population and program implementers should be involved in all stages of decision making to ensure that the intervention package is based on the needs of the community [9]. IM is also a planning approach that is grounded in evidencebased strategies, and as fundamentals for taking an ecological approach to assessing and intervening in health problems through promotion of stakeholder engagement at all levels [9]. IM has been effectively used to design, implement and evaluate interventions that showed substantial increase in application of disease prevention programs [10].
For Ghana to be able to meet the 2020 NTDs deadline, it is important to strengthen the lymphatic filariasis MDA delivery systems through context specific QI strategies to make the drug accessible to all eligible individuals in the endemic communities. In this paper, we describe how intervention mapping was used in selecting QI strategies to improve LF MDA outcomes in Bole District of Northern.

Ethics statement
All study participants were adults. Permission was sought from all research participants through written consent. Ethical approval was given by the Human Research Ethics Committee (medical) of the University of the Witwatersrand, Johannesburg, South Africa (Clearance Certificate No. M170219) and the Ghana Health Service Ethics Review Committee (Clearance Certificate No. GHS-ERC: 04/02/2017).

Intervention mapping
Intervention mapping (IM) is a systematic process involving six key steps [11]. The six steps of IM process consist of 1) initial needs assessment, 2) developing program objectives, 3) selection of intervention methods and practical strategies, 4) producing program components and materials, 5) plan for adoption and implementation and, 6) designing an evaluation plan. The advantage of IM is its use of models, theories and engagement with key stakeholders to identify context specific intervention. IM is used as a planning process to advance theories and evidence-based health promoting intervention programs. In this study, an adapted context specific IM process was used. In order to develop a context specific QI intervention within an evidence-based program such as the lymphatic filariasis MDA, we started with an initial needs assessment using mixed methods approach. Based on the finding of needs assessment, we identified evidence-based and practical strategies backed by the concepts of the Global Program to Eliminate Lymphatic Filariasis and the Ghana Filariasis Elimination Program. The selection of the strategies was grounded on models, theories and evidence from other related studies. The QI strategies were developed into consolidated implementation package based on evidence and stakeholders' brainstorming sessions. Monitoring intervention and evaluation was the final step of the IM process. The planning of the monitoring and evaluation was guided by the QI strategies. The monitoring and evaluation were done to assess both the process and impact of the QI intervention on knowledge of LF, knowledge of the MDA and the MDA coverage. The subsequent sections stipulate five steps of IM process used to design a consolidated QI implementation package for LF MDA in Northern Ghana.
Before the commencement of the study, stakeholders' meetings were held with the managers of the NTDs program in Ghana at national, regional and district levels to gather support for the project. This preliminary work provided a foundation from which the researchers worked with existing NTD program and community leaders to develop the QI intervention.
Initial needs assessment. The initial assessment was conducted to identify implementation bottlenecks of the ongoing MDA. This strategy was employed to gain a deeper understanding of the critical issues affecting the implementation of the LF MDA from the perspective of the services providers (health workers) and the clients (community members).
The needs assessment involved: (a) 20 in-depth interviews with District Director of Health Service, Sub-district Heads of Health Service, Disease Control Officers and Community Drug Distributors, followed by 16 key informant interviews with Chiefs/Opinion leaders and noncompliant community members who were purposively selected, (b) surveys with 222 randomly selected community members and (c) analysis of secondary MDA coverage data of 150485 participants.
Key informant and in-depth interviews: Semi-structured interview guides were used for the data collection. They were translated from English to the local language before the field work began. To ensure that the versions in English were the same as those in the local languages, back-to-back translation strategy was used. The guides covered broad areas such as knowledge and understanding of LF, knowledge and understanding of the MDA process and implementation challenges of the LF MDA in the study district. All interviews were recorded using a digital voice recorder, translated and transcribed verbatim. Field notes were also taken which were turned into data documents for analysis. The transcribed data and field notes were inputted into a Microsoft Word processor and exported into NVivo Version 10 for coding. Framework method of qualitative data analysis was used [12,13]. This method involved familiarization, identification of thematic framework, indexing, charting, mapping and interpretation [12][13][14]. Explore and query functions in Nvivo were used to provide a descriptive perspective of the data as per framework method of qualitative data analysis [15]. The descriptive accounts were compared with the existing literature to provide an explanation and interpretation of the data. The results were summarized based on the themes that emerged from the data with key quotes from participants.
Survey: The survey was conducted with 222 randomly selected community members. The sample size (n) was determined using the formula: n = (Z α/2 ) 2 � (p � q) / d 2 [16]. The sample size calculation was based on the assumption that 65% [17] of the population were aware of the MDA program and ingested the drugs, Z α/2 is the confidence level at 95% = 1.96, p is the coverage, q is (1-p) and d is the level of precision. The participants were randomly selected using simple random sampling strategy. A random number generator was used to select the desired number of participants using the LF community registers as sampling frame. Information gathered includes socio-demographic characteristics of participants, knowledge of MDA, understanding of the MDA process, stakeholder engagement before and during the MDA, reasons for refusal to ingest the drug and knowledge of LF. Other information collected includes knowledge of how LF is acquired, knowledge of signs associated with LF and knowledge of LF prevention and misconceptions about LF. These variables were measured by correct answers to signs associated with LF, LF prevention methods and how LF is acquired. Comprehensive knowledge of LF was estimated by the ability to identify correctly at least two major ways of preventing LF transmission, mode of LF transmission, and rejecting at least two misconceptions about how LF is acquired. A descriptive analysis was carried out in Stata version 14.2 and results were presented in tables.
Secondary MDA data: The secondary data include information on adverse drug reaction, MDA coverage, noncompliance and non-eligibility of 176 857 participants. From the available secondary data, this information on MDA coverage, noncompliance and non-eligible of 176 857 participants was extracted.
Selecting evidence-based and practical strategies. Results from initial needs assessment were used to identify the possible areas of intervention. The main aim of this intervention was to improve the quality and level of coverage of the LF MDA in the study area, hence we identified evidence-based, cost effective and context specific strategies which was acceptable for all stakeholders.
The QI strategies are based on the concept of the Global Program to Eliminate Lymphatic Filariasis (GPELF) [18] and the aim of the Ghana Filariasis Elimination Program [7]. The GPELF is an innovative, public-private partnership for health improvement. The selection of the QI strategies in this study was also grounded on evidence from other LF MDA studies such as improving service delivery strategies of the MDA, effective stakeholder engagement, and improving compliance at the level of the community and the individual level [7,18,19].
The success of LF MDA is shown to be backed by community sensitization and mobilization through evidence-based, multi-channel communication approaches with key messages on treatment and side effects combined with a high visibility if the MDA should be pioneered [17][18][19][20][21][22][23].
As evident in a study in Nepal by Adhikari et al, awareness of MDA campaigns had significantly increased the level of compliance among participants. Individuals who had knowledge of side effects during MDA campaigns had a lower prevalence of non-compliance [24].
Community leaders were involved, based on the World Health Organization's (WHO) recommendation that community involvement in health programs will not only help to remove the burden of dependency that characterizes health development work but also creates general awareness among local people about the need for their involvement in all forms of health development [25].
Evidence elsewhere shows that the involvement of local leaders, religious and political authorities in the MDA activities enhances the participation, coverage and compliance of MDAs [18,20,21,26].
A study conducted in 2017 in Southern Ghana shows that knowledge, attitude and practices relating to LF and the MDA improved after a health education intervention [7]. It has been suggested by Njomo et al [27] that for the LF elimination program to succeed, there is a need for aggressive community sensitization towards the disease and elaborate social mobilization [21]. Thus, bridging the knowledge gap with an appropriate educational intervention is cardinal to change perceptions and popularly held-beliefs which have no scientific basis.
The selection of the QI strategies also combined several complementary theories such as, the health belief model, social cognitive theory, the community empowerment model and the social networks and social support model. These theories were used as a guide in developing the interventions package. This is because the desired change is at both the provider and the individual levels. Hence, there is the need convincing providers and the endemic community to consider the LF drug as a viable intervention option to stop transmission of LF. These theories have been applied in similar public health interventions [11,[28][29][30]. A practical description of the evidence-based strategies for this QI intervention were based on a model adopted from Proctor et al 2013 [31].
Developing the QI strategies. This stage of the IM process involved stakeholder engagement meetings. The purpose of the stakeholder discussion was for the dissemination of the initial needs assessment results and to map out tailored, discrete, need and evidence-based implementation strategies to address the specific contextual bottlenecks identified in the initial needs assessment.
The stakeholders included the District Director of Health Services, the District Health Promotion Officer, the District Disease Control Officer, the District Public Health Officer and Sub-district heads of health services from Bole District. Others involved were sub-district disease control officers, NTD focal persons, local government representatives and journalists from two radio stations.
The stakeholders were placed into seven groups to brainstorm the barriers outlined in Table 1. They were tasked with coming up with ideas that would translate into an implementation strategy to overcome the barriers identified during the needs assessment stage. The initial needs assessment shows that the persistent transmission of lymphatic filariasis in the Bole District is characterized by refusal to ingest the drug due to the fear of adverse drug reactions, poor knowledge and misconceptions of the disease, reported adverse drug reactions, poor community mobilization and sensitization, non-adherence to the direct observed treatment strategy, poor adherence to the MDA protocol and non-participants' responsiveness as detailed in S1-S6 Tables. The barriers and needs identified are summarized in Table 2 The ideas from the seven groups (shown in Table 2) were compiled into a consolidated QI implementation package as described in Table 4 using an implementation specifying model adopted from Proctor et at 2013. The consolidated QI implementation package is also informed by evidence and theories identified in selecting evidence-based and practical strategies. Staff of the Bole District Health Management, Sub-district Heads of Health Service and opinion and community leadership agreed that the consolidated QI package had the potential to improve the coverage and quality of implementation of the ongoing LF MDA.
Monitoring and intervention evaluation. Monitoring and evaluation of the QI intervention is the final Step of IM. This step of the IM allows for sympathetic formative evaluation to assess both the process and impact of the intervention and whether changes need to be made [32,33]. The monitoring and evaluation step was planned to assess the activities before, and during the implementation of the QI strategies based on the planned consolidated QI implementation package. As part of the monitoring and to ensure successful implementation of the QI strategies, a QI implementation team with representation from the district and each subdistrict was formed. The general responsibility of the QI team was to ensure the smooth execution and sustainability of the QI strategies. A WhatsApp group was created to link members of the QI team. This social media platform was used to facilitate the sharing of ideas and challenges encountered during the implementation of the QI strategies. The WhatsApp platform also allowed the principal investigator to review progress, troubleshoot problems and reinforce compliance to the QI strategies during the implementation process. The sub-districts heads of health services were also task to submit a weekly written implementation report. The evaluation includes collection of qualitative (key informant and in-depth interviews) data on MDA activities before and during the implementation of the intervention, knowledge of LF, and Knowledge on MDA. The Participants of the evaluation qualitative study include District Director of Health Service, Sub-district Heads of Health Service, Disease Control Officers and Community Drug Distributors, and Chiefs/Opinion leaders. Secondary quantitative data on MDA coverage, noncompliant, non-eligible and adverse drug reaction was extracted from the  • 3 weeks of preparation before the distribution of drugs are required. Onemonth period is required for re-visits before submitting reports to region.
• Duration of MDA should be extended in order to accommodate revisits and reduce absenteeism. MDA period should be changed from the rainy season to the dry season.
• A follow up visit should be carried out after the mass distribution.
• Appeal for extension of MDA period to meet the demands of the people. Appeal for the MDA activities to be carried out during the dry season. Group proposed early notification for proper preparation.
• Maintain a yearly calendar on MDA.
The training of drug distributors should be in the dry season.
• (Continued ) database of the Bole District Health Administration from 70591 and 79894 individuals for the pre and post intervention phases of study respectively.

Results
The study adopted intervention mapping strategy to come up with QI approaches for implementation in an existing EBP (LF-MDA program). An initial stakeholder engagement meeting involving officials of the NTDs program at National, Regional and District levels was organized during the formative stages of the study. The IM process comprised an initial needs assessment study, stakeholder engagement discussions in a systematic process grounded on models, theories and evidence from other related studies. As shown in Table 1, more than half (52.25%) of the survey participants were female. The mean age of the participants is 43.90 years (16.57 standard deviation). Majority of the participants (77.93%) were married. In terms of occupation, 71.17% of the respondents were farmers Towards elimination of lymphatic filariasis in Northern Ghana

Target (s) of the action
• Drug distributors in the endemic communities.
• People in the endemic communities.
• People in the endemic communities.

Temporality
• The drug distribution should start within one week after the training of drug distributors.
• Social mobilization and sensitization should start two weeks before and should continue during the MDA.
• Before, during and after MDA exercise.
• During the drug distribution

Dose
• The training of the CCDs should be detailed enough to equip them well for the MDA task and the training period should not be more than one day to enhance their active participation in the training.
• Each endemic community should have at least two social mobilization and sensitization exercises (community durbar, school education, information centre announcement, education at church and mosque, and or radio talk shows etc.) for the start of MDA and at least one during MDA.
• Every endemic community must have a community leader representing it.
• The distribution exercise should not be less than 1 week in the endemic district.
• Over 80% of the people in the endemic communities must be covered.

Implementation outcome (s) and effect
• Increase the level of adherence to LF MDA implementation procedures and participants' responsiveness.
• At least 15% increase in MDA coverage and reduction in the number of refusals (increased participant responsiveness).
• At least 15% increase in MDA coverage and reduction in the number of refusals (increased participant responsiveness).
• At least 15% increase in MDA coverage and reduction in the number of refusals (increased participant responsiveness).

Justification
• Researchers suggest that drug distributors are the interface between MDA programs and their targeted population hence their adequate training is crucial to the success of the MDA [18,20,34,35].
• It has been shown that evidencebased, context specific and multichannel social mobilization and sensitization is required for the LF elimination program to succeed [7,20,21].
• The MDA implementation process is very important for participants' responsiveness to the program [7,20]. The results of the brainstorming session (based on the results of the needs assessment) are presented in Table 3. The seven stakeholder groups suggested solutions ranging from the training of drug distributors (CDDs), community education, proper supervision, health education and community feedback, to the involvement of opinion/religious and community leaders. The selection of the final consolidated QI strategies was based on available evidence from literature as shown under the justification domain in Table 4.
The stakeholder engagement meeting for dissemination of baseline results and IM towards the intervention phase of the study was held in the district capital of Bole District. The purpose of the stakeholders' engagement was to disseminate the findings of the baseline study and to map out tailored, discrete, need-based implementation strategies to address the specific contextual bottlenecks identified in the preliminary study in order to improve the fidelity of the MDA and its coverage. Table 4 presents the consolidated QI strategies that was implemented. The researchers together with the stakeholders came up with the final cost effective and evidence-based QI strategies based on the model adopted from Proctor et al (2013) [31]. These strategies were informed by theories [11,[28][29][30], the concept of the GPELF [18] and evidence from other related studies [17][18][19][20][21][22][23]26]. The components of the implemented strategies include training of community-based drug distributors (CDDs), social mobilization and sensitization, involvement of community leaders and other stakeholders and the drug distribution process. Each of these strategies have seven (7) domains namely; actors, actions, target of action, temporality, dose, implementation outcome effects and justification.

The domains
The actor. The actors used in this context are stakeholders who will deliver the QI strategies. The actors for this study include District and Sub-districts NTD focal persons, disease control officers, health promotion officers, local government representatives (assemblymen/ women) and the Principal Investigator (PI).
The action. The actions are the specific processes or steps and activities to be undertaken in a specified sequence. For this study, the actions for each of the four strategies are specified in Table 4. These actions include training activities, multi-channel communication strategies, involvement of community leaders and increase in duration of MDA activities for wider coverage of eligible population.
Targets of actions. An implementation strategy is a function of where it is directed to have impact. Hence, it is important to specify the target because it helps focus the strategy and suggests where and how an outcome of interest should be measured. The targets of actions in this study were community drug distributors, people in the endemic district and community and opinion leaders. A total of 162 community drug distributors where trained.
Temporality. The order or sequence in which a QI strategy is implemented is very important in the execution process of the strategy. This study specified when each strategy should be used as shown in Table 4.
Dose. These are details about the intensity of the implementation strategies required to have the needed effect. This ranges from number and duration of training sessions, number of social mobilization and sensitization activities per community for the duration of the drug distribution exercise. The training was for one day, but the number of social mobilization and sensitization activities varies from community to community. Implementation outcomes and effect. Implementation outcomes are a purposive and deliberate actions or activities to implement new strategies or practices. Implementation outcomes serve as indicators of the implementation success, act as proximal indicators of implementation processes and key intermediate outcomes [36]. Implementation outcomes therefore serve as necessary preconditions for achieving desired changes or results [37]. The implementation outcome in this study is increasing adherence to the LF MDA implementation process and involvement of the endemic community. Resulting in increasing the MDA coverage by 15% and reducing the number of refusals as shown in Table 4.
Justification. The selection of the implementation strategies in this study is based on a proactive needs assessment of barriers or factors affecting the implementation of the LF MDA in the study district. The strategies are also supported by current literature as shown in Table 4 and programmatic rationale through stakeholder engagement in the brainstorming exercise which is deemed appropriate to address the specific bottleneck posed by the implementation context.
The results of the secondary data as detailed in S1 Table shows that, the total MDA coverage for pre-intervention is 82.03% while that of the post intervention is 83.40%. The reported preintervention adverse drug reaction is 0.13% while that of post intervention is 0.02%. (See S1 Table for detail results). While the average community and sub-district MDA coverage in the pre-intervention is 82.67% and 81.32% respectively, that of post-intervention is 81.50% and 87.80% respectively as shown in Table 5.
There seems to be improvement in the knowledge about the disease and understanding of the MDA among the study participants and in the study, area as shown in the qualitative results shown in S7 Table.

Discussion
This paper has used the problem of persistently-high prevalence of LF in the Bole district of Ghana to demonstrate the step-by-step procedure for selecting and tailoring QI strategies to an evidence-based intervention program. The QI strategies take into account local contextual needs through needs assessment and stakeholders' engagement [38]. The IM method used in this study has galvanized stakeholders around common goals and generate consensus regarding improving an ongoing evidence-based intervention at a low-cost as shown in other studies [38,39]. The process emphasizes stakeholder contribution in coming up with a QI strategy [38,40].
As suggested in other studies [38,41], this article provides a clear description of the implementation strategies used as detailed in Table 4. In line with the article by Proctor et al (2013) [31], this article carefully names, conceptually defines, operationalizes strategies and specifies them according to features such as (1) the actor, (2) the action, (3) action target, (4) temporality, (5) dose, (6) implementation outcome [42] and (7) the empirical, theoretical, or pragmatic justification for use of the strategy [31]. In order to show a clear link between the strategies and the bottlenecks that they are intended to address, an IM process and Proctor et al (2013) implementation specifying model was used to clarify how and why carefully selected and tailored implementation strategy is proposed to work as suggested by other researchers [31,38].
Improving the quality of evidence-based interventions to achieve expected results has proven to be a challenging task for most health system programs [43][44][45][46][47][48][49][50][51]. Studies [49][50][51] suggest that barriers and facilitators must be taken into consideration and that different strategies may be required to overcome challenges associated with interventions, implementation contexts, and stages of implementation [38]. The intervention package of the current study is in line with this suggestion. The processes presented in this article have increase the rigor of selecting and tailoring an appropriate QI strategy as demonstrated by Powell et al (2015) [38]. This article also provides a step-by-step procedure for selecting and tailoring QI strategies and for engaging stakeholders who are very important in the success of the intervention. It further shows the links between strategies and barriers and other contextual factors.
Describing, implementing, and sustaining any evidence-based intervention is a complex undertaking [52,53]. This is because implementing quality improvement strategies (1) are usually multi-component and (2) must be adapted to local contexts. The local contexts in which these interventions are implemented are themselves complex because of multiple interacting levels with wide variation from one setting to another [52,54]. Hence, implementation research requires trans-disciplinary expertise which are not routinely part of the health system in Ghana. The emerging implementation research provides a systematic set of principles, methods and expertise to improve the quality and effectiveness of evidence-based interventions within the health system [52].
The move toward improving evidence-based health system programs with a more thoughtful and systematic application of implementation research will enhance the expansion and legitimacy of the field of implementation research. The scientific rigor of proposed QI studies, strengthening evidence-based interventions to achieve expected results and providing support for implementation of QI strategies will be enhanced.
The demonstration of this study was to provide a systematic method for selecting and tailoring QI strategies of evidence-based interventions to contextual needs in collaboration with stakeholders. This will provide an understanding of how, when, where, and why implementation strategies are effective in integrating evidence-based interventions and improving clinical outcomes in the health sector [55].

Limitations
Although the insight provided by this paper will be valuable to other low-resource settings seeking to use evidence-based QI strategies to improvement the quality of the lymphatic filariasis mass drug administration towards the elimination of the disease as a public health problem, it is important to note that, the QI strategies could not incorporate other issues such as incentives/motivation (monetary rewards) for CDDs, transportation related challenges and the timing of the MDA. These were part of the challenges identified during the initial needs assessment. These are decisions that could only be made at regional and national levels of neglected tropical diseases program due to the cost implications.

Conclusion
Although the overall MDA coverage in the study district increased from 82.03% in pre-intervention to 83.40% in the post-intervention, the social mobilization strategies such as (community durbar, gong-gong beating and announcements in churches and mosques) has not been effective in peri-urban and urban communities. In designing a QI intervention within a district, there is the need to consider sub-district and community level dynamics. The "one-sizefits-all" intervention does not work even within the same district.
The use of radio stations, information vans and posters for social mobilization and sensitization could have been more effective in peri-urban and urban communities as shown other studies [7,56,57]. Due to the financial implications, these methods of social mobilization sensitization (which the funding of this study could not support) were not included in the consolidated intervention package implemented.
The timing of the MDA which coincide with the farming season and Muslim fasting period leads to missing of eligible people and refusal to ingest the drug respectively.
For Ghana to meet the 2020 NTDs deadline, it is important to strengthen the lymphatic filariasis MDA delivery system through context specific QI strategies to make the drug accessible to all qualified individuals in the endemic communities. Hence, there is the need for concerted efforts to improve the quality of implementation and coverage of the ongoing lymphatic filariasis MDA campaigns in the Bole District of Northern Ghana.
This study has shown that it is feasible to develop a context specific QI strategy for an existing evidence-based intervention based on an initial needs assessment through stakeholder participation using IM approach. However, working according to QI requires more time than is usually available in public health services. Furthermore, sufficient theoretical knowledge of implementation research and experience with technical IM aspects has to be available.
Supporting information S1