A mixed methods approach to evaluating community drug distributor performance in the control of neglected tropical diseases

In January 2012 the momentum to combat the neglected tropical diseases (NTDs) was accelerated with the London Declaration on NTDs [1]. Donors and endemic country governments, private sector leaders, pharmaceutical companies and multilateral organizations made a unique commitment for a close partnership to control or, in certain cases, eliminate ten NTDs by 2020, in alignment with World Health Organization (WHO) targets and the Roadmap on NTDs [2]. The roadmap is a guide to the implementation of policies and strategies to combat each of the NTDs and the extraordinary commitments towards these by each stakeholder.

As the primary implementers of mass preventive chemotherapy (PC) campaigns for NTDs in sub-Saharan Africa, village-elected Community Drug Distributors (CDDs) are critical stakeholders in achieving the Roadmap [3–5]. These stakeholders volunteer their efforts to perform multiple fundamental tasks which can include diagnosis, creating census records, treatment and reporting. If these individuals, who are trusted and respected within their communities and whom essentially act as the front-line health facility, are not supported adequately then entire disease control programmes could be in jeopardy.

In Uganda, where all 112 districts are endemic for one or more NTDs [unpublished data, Uganda MoH Master Plan for the control NTDs (2011–15)], there is an overall health worker deficit of approximately 80 % [6] and only approximately 49 % of the population live within five kilometres from a health facility [7]. ‘Task shifting’, where tasks are delegated to less specialised health workers, including community volunteers, and Village Health Teams (VHT) have been introduced by the Ugandan Ministry of Health to bridge these gaps [7, 8]. The CDDs are part of the VHT, where the structure exists, and these individuals play a pivotal role in the implementation and success of the country’s integrated NTD Programme. In Uganda, stand-alone NTD control programmes existed which were then integrated, or co-implemented, from 2007 onwards to reduce duplication of resources and costs [9]. The stand-alone programmes also used CDDs for drug distribution and, in many cases, the same individuals participate under the current integrated NTD Programme.

There are several conflicting views in the literature as to how integration of several control interventions impacts community volunteers such as CDDs. One view is that co-implementation of interventions can effectively make it easier for community volunteers, because they only have one multi-layered intervention to be responsible for which results in increased community volunteer efficiency, reduced reporting requirements, increased intervention utilisation as well as reduced time and transportation costs [10–12]. In contrast, by requesting community volunteer to be responsible for an intervention with multiple and additional services, there is the view that they will become overburdened, resulting in increased time and a subsequent decline effect in their performance [13, 14].

Diaries have been used in many contexts as a research tool to monitor daily experiences and behaviours which are expected to change over time. Events that are recorded can be as wide-ranging as the ‘high-risk’ behaviours of commercial sex workers to that of the incidence of colds or household consumption and expenditure [15–17]. The benefit of using diaries for data collection is that the information is recorded in real time which greatly minimises recall bias. Furthermore, in low-income countries where many of the population are illiterate, pictorial diaries have been shown to be a particularly effective tool for measuring activities, such as, expenditure activities and daily frequency and consistency of stools [18–20]. Although there is substantial literature about the CDDs roles in community-directed treatment with ivermectin (CDTI) for onchocerciasis control and in trachoma and schistosomiasis control [3–5], there is currently no evidence of the CDDs role in integrated PC (IPC). Furthermore, although diaries and work logs are used extensively in research, there are no published data which refers to pictorial diaries recording information on daily activities of community volunteers in mass treatment programmes.

The aim of the current study was to determine the role and the work burden of the CDD in the Ugandan integrated NTD Programme which delivers IPC on an annual basis. This was achieved using a mixed methods approach including a Pictorial Diary (PD) which was designed to collect daily prospective data on the time spent by CDDs on both their routine activities and NTD activities. By recording time spent on NTD activities and assigning a monetary value the opportunity costs of CDD time were determined. CDD performance was estimated from a validation IPC coverage survey. Additionally semi-structured interviews, questionnaires and focus group discussions (FGD) captured qualitative and socio-demographic data. The results of this study can be used by those responsible for implementing health interventions using community volunteers. Also of importance, the evidence can inform those funding and donating drugs to large-scale NTD Programmes on how the partner countries are contributing to these community-based programmes.

The use of PDs in this study was a unique approach to determine the amount of time CDDs were allocating to their NTD Programme responsibilities in Uganda. The data show that to conduct NTD Programme activities a CDD spends, on average, two and a half working weeks per year. Subsequently the CDDs personal time, expenses and time available for work at home, in the shamba and other forms of livelihood are negatively affected. The CDDs that only have one package of drugs to deliver over the IPC campaign spent substantially less time (two weeks) on the NTD Programme than those who required three and a half weeks to deliver three rounds of treatment. Moreover, as CDDs took longer to deliver the treatment to targeted populations, their performance decreased. Mixed methods provided evidence to evaluate the ability and motivation of the CDDs to successfully and sustainably deliver an integrated programme for NTDs. Although the majority of CDDs would return to their role in the programme the following year, they were less motivated to participate in the NTD Programme than in other health interventions. Lack of incentives was a major contributor to feeling less motivated and undervalued.

There is a growing body of literature which highlights the importance of CDDs participation in the successful delivery and sustainability of PC for the NTDs and other health-related programmes such as home-based management of fever and Integrated Community Case Management [28–30]. Few, however, look at time spent by CDDs in distributing health packages [31]. Furthermore, the costs of CDD involvement in programmes is rarely valued nor included in programme economic evaluations [32, 33]. In this study, time allocated by the CDDs towards their NTD Programme duties were a significant factor affecting their performance. Unsurprisingly the number of households (on average 90 per CDD) and village population (on average 497 per CDD) served were associated with the amount of time spent of NTD Programme activities by the CDDs. Katabarwa and colleagues included the number of households served by a CDD in Uganda and Cameroon [31] and found that programme coverage improved when distributors served fewer households (< 20), worked within the kinship system and worked within a radius of 1 km of their activities. This study aligns with those findings and it is intuitive that if there are more households in a population then it will take longer to sensitise and treat all of them.

The number of drug delivery rounds over the IPC campaign for NTDs was also influential in how much time the CDDs spent on the programme. With fewer deliveries required by the CDDs, the time spent on NTD Programme duties was significantly lower. Information elucidated from the CDDs during the semi-structured interviews and FGD revealed that time spent on visits to collect drugs from health unit was increased with the number of drug delivery rounds expected of them i.e. one collection of drugs for each disease(s) targeted. In addition, for a fifth of respondents (21 %), inadequate drug forecasting resulted in insufficient quantities of drugs, at the health units, to reach target eligible population. CDDs thus returned frequently to the health units to check if more drugs had arrived from Kampala via the district headquarters. Consequently, the increased time spent on collecting drugs from health units, due to drug package or drug stock-outs, contributed to a poorer coverage performance by the CDDs. A further factor contributing to poorer performance could be the decreasing amount of time recorded on health education and mobilisation activities. These findings are in contrast to several studies where adding additional health interventions to the CDTI platform improved programme coverage [11, 34, 35] but, in agreement with the study in Cameroon and Uganda where the performance of the CDDs was compromised by added interventions [31].

Our study also estimated the opportunity costs of the CDDs involvement in the NTD Programme in Uganda. A limited number of studies have reported the opportunity costs of CDDs in health interventions [11, 24, 36]. The average two and half weeks volunteered by a CDD for NTD activities is equivalent to a salary of $35.71, based on local casual labour wages. For the same number of days a teacher would earn $49.06, a Health Assistant at a health centre would earn $75.19 and a District NTD Focal Point $121.11. When the average value of the CDDs opportunity cost is multiplied by the 61,000 CDDs that were trained under the NTD Programme that year (unpublished data, RTI International semi-annual report 2008/9), this amounts to US$2,178,310, which is a substantial contribution by the CDDs to the overall economic programme costs. In reality the CDDs received the equivalent of, on average, US$1.80 allowance for training and a t-shirt (US$1.86). If all were to receive a t-shirt, which was not the case in the study districts, the combined cost for allowance and t-shirts for the CDDs would be an estimated $223,260 which is a 10^th of the total value of the opportunity costs.

Similar to other studies [11, 23, 37], CDDs in Uganda are fully aware that they were committing to unpaid work and testified that they assume the responsibilities of the NTD Programme for reasons of intrinsic and extrinsic motivation. Nevertheless, this study highlights the significant amount of time forgone by the CDDs on their regular income or subsistence duties, the out-of-pocket expenses incurred and due to feeling undervalued, a better incentive package would be justified. In a recent study Downs et al. describe a framework where if work complexity is increased, so is the demand for incentives (monetary and non-monetary), and efforts towards meeting these demands can lead to an increased perceived value of the CDD position which can motivate CDDs to attaining a higher level of performance [38]. The use of pay for performance, or results-based financing, is being increasingly tested and used in developing countries [39–41]. These systems involve rewarding individuals for reaching their targets, such as numbers treated or programme coverage attained with financial or materials to effect an improvement in performance. In the context of the NTD Programme, however, such systems might only be feasible for CDDs if there was a general health system shift to such a reward scheme. Additional areas of concern for such schemes are false reporting and the lowering and even removal of intrinsic motivation in participants [39]. In contrast to these reward schemes, and perhaps more realistic for the NTD Programme, would be to employ alternative distribution mechanisms to what is currently in place. The traditional kinship system where each CDD treats below 20 households, which Katabarwa and colleagues have shown to be successful in terms of CDD performance, workload reduction and involvement in other health activities, in comparison to CDTI [22] offers one alternative distribution mechanism. In addition the demand for monetary incentives was reduced under the kinship system. Alternatively, the NTD Programme could choose to support and strengthen the existing health system through the VHT [7]. In Uganda one VHT member should serve approximately 25 to 30 households [7]. To support the VHT the NTD Programme would need to work closely with other MoH department to establish the needs of the VHTs in the target districts, for example, the provision or loan of bicycles and joint treatment registers and ensure reporting practices were standardised and in-line with those of the VHT [7]. Either of these distribution mechanisms would mean CDDs would be targeting less than a third of households than they are currently required to. Nonetheless, both suggested mechanisms would cause an increased cost to the programme. In the case of the kinship system the numbers of CDDs requiring training would approximately double and the VHT system may also require more training and the periodic supply of support. However, the potential for increased continuity, sustained performance and a decrease in the opportunity costs and out-of-pocket expenses incurred by the CDDs, could be argued to far outweigh the expense.

The availability, capacity, acceptance, and indeed ownership of the programme by the CDDs may only be one facet of sustainability of an NTD Programme, but it is critical for both effectiveness and longevity of the control and elimination efforts. The role of the CDD in the integrated NTD Programme requires a substantial commitment of time to achieve treatment goals. Additional workload is created with increasing number of delivery rounds within an IPC campaign and hampered by untimely national drug logistics. CDDs, subsequently, fail to achieve their target treatment coverage and there is limited access to these safe, effective and donated drugs for vulnerable individuals and families. For their performance, whether good or poor there is no feedback or consequence for the CDDs from the NTD Programme. Ultimately if programme coverage of 75 % and over is not being achieved as a minimum then the NTD Programme is not being effective and treatment will need to continue indefinitely and morbidity control and elimination goals for 2020 will not be reached. In order to utilise resources effectively and efficiently improved drug supplies and logistics to the districts and alternative support systems to the CDDs are imperative. This would involve investment in financial and non-financial resources yet would pay dividends in programme performance and sustainability. Affordable programmes with high impact are extremely desirable in the African context and continued high coverage of treatment with community participation and ownership are essential elements in their success.

ALB, albendazole; CDD, community drug distributor; CDTI, community directed treatment with ivermectin; IPC, integrated preventive chemotherapy; IVM, ivermectin; FGD, focus group discussion; PC, preventive chemotherapy; PD, pictorial diary; PZQ, praziquantel; NTD, neglected tropical diseases; WHO, World Health Organisation; VHT, village health teams; ZIT, azithromycin