This study used mixed (quantitative and qualitative) methods to identify and explore community (demand-side) and health system (supply-side) challenges which could potentially affect MDA coverage.
Our results showed low levels of awareness about MDA and lower levels of awareness of NTDs in both counties. We found that 48% of people reported taking the MDA tablets during the last round, with the most common reasons for not adhering being because they were not around when it was being delivered or were not informed about the delivery. This emphasises the importance of ensuring MDA timing and organisation is cognisant of daily and seasonal activities in specific communities [19,20,21,22]. As part of the capacity- and awareness-building target of the NTD roadmap to 2030, awareness-generation activities to educate and inform the endemic communities are deemed essential , and positive community leader influence and community trust is vital for effective healthcare delivery . Indeed, we found that community leaders encouraged the community to listen to the message from town criers, which was where most respondents reported getting information from in the household survey. More could be done to enhance the role of community leaders in supporting MDA, which should, in turn, generate trust and community ownership of MDA, with likely benefits for coverage and adherence [24,25,26].
We found that the Liberian CDDs spent an average of 16.04 workdays in the MDA round, similar to the 13.31 workdays a year on NTD activities including MDA in Uganda . Our quantitative findings indicate that Liberian CDDs incurred direct and opportunity costs for taking part in MDA and that for some CDDs, these costs were not adequately compensated by allowances, resulting in a mean cost to CDDs of $11.90. Placing this into context, the minimum monthly wage for civil servants in Liberia is $37.52 (5600 Liberian dollars). This led to frustration among CDDs, which were reflected in our qualitative findings and have been identified in other settings [11, 26,27,28].
From our KIS survey, we discovered that communities were expected to compensate CDDs via providing gifts in kind such as a cup of rice or a small fee in exchange for medicines to CDDs. Some health staff described that community members were often unable or unwilling to do this due to their level of income and/or socio-economic status and so avoided the CDD. With the high incidence of extreme poverty in the communities receiving MDA and possible deterrence [29,30,31], this leads us to question the viability of such an approach on equity and efficiency grounds and a potential risk to the achievement of NTD targets on equitable access to healthcare 
Our observations of CDDs during their MDA activities identified that two tasks accounted for a large proportion of their time, namely measuring and drug administration, and record-keeping and registration. However, qualitative findings indicated that transportation for CDDs was also a challenge, and they found it difficult to move within communities to distribute drugs. Motorcycles were limited and insufficient to reach out and serve communities, making the workload greater for CDDs as they walked longer distances, taking up more time to distribute drugs and incurring higher direct and opportunity costs by the end of the MDA. Support to compensate CDDs for the costs incurred during their role in MDA delivery and for more efficient transport would likely bring benefits in terms of greater satisfaction and retention, supporting the NTD road map for 2021–2030 .
Creating awareness and educating households on MDA and its benefits and side effects of drugs was relatively time-consuming for CDDs at each household. Hence, low awareness increased the workload and consequently the opportunity cost incurred by the CDDs in our study and in others [21, 26]. CDDs could be better supported during MDA with access to materials to support conversations about the reasons for the MDA and drug safety information to reassure households [24, 32, 33].
Having an adequate supply of quality-assured medicines and an efficient supply chain at the community level is critical for the effective allocation and distribution of medicines. However, as in other studies, we found that drug shortages made it difficult for CDDs to effectively distribute drugs to the target population . Similarly, delayed distribution due to drug shortages confused both CDDs and the communities, hampering adherence when the MDA happened . Furthermore, even though most CDDs attended at least one training session, many had limited training on how to complete reporting forms, and consequently, they submitted incomplete records which in turn compromised programmatic-level medicine estimates . These findings highlight the critical importance of NTD monitoring and evaluation mechanisms and suggest that investment in improved CDD training may yield programmatic benefits .
Recall bias is inherent in any study relying on household surveys, and this study is no exception. The recruitment of female CDDs was challenging due to fewer females acting as CDDs in the study communities because of pre-existing gender norms, competing domestic priorities, and reduced literacy levels compared to their male counterparts [29, 30]. We also faced capacity challenges that affected the quality of some of the data collected. Analysis of the household survey results on NTD and MDA awareness revealed many missing or no responses. On investigation, it was found that this was due to confusion and lack of understanding of the diverse range of definitions and local terminologies used to refer to NTDs and MDA in the different communities. Unfortunately, due to internet connectivity problems and the community process of data collection, this was not picked up until it was too late to alter the survey tool. This may have led to an underestimate of NTD and MDA awareness in our study.
MDA activities were measured in this study by following the CDDs and using a stopwatch to record time spent on each activity. Unfortunately, time spent by CDDs in houses was only recorded in minutes (not minutes and seconds) which may have affected results. In Uganda, a similar study used pictures and drawings to describe the entire day of the CDD during NTD activities, bringing in more context to the opportunity costs borne by these CDDs . This approach might have worked well in the Liberian context; however, it would not have yielded a quantitative estimate of costs per house, highlighting the benefits of complementary quantitative and qualitative research in understanding health systems.
The data collection limitations of this study highlight the need for continued investment in health research capacity strengthening in low- and middle-income countries. Nevertheless, this work represents a first effort at conducting a mixed-methods health economic study of NTDs in Liberia and thus a major step forward in health systems research for the country.