CHWs are supposed to be nominated by village health committee. As village health committees are often not established or functioning, candidates for CHWs are often suggested by the village leader and the organization that supports the CHW or an NGO makes the final selection. The selection is usually based on criteria including education level and position in the community. Candidates are often individuals that already provide informal health care in their community. CHWs can be of any gender and aged between 18 and 50 years. They should have sufficient education to read and write Myanmar (Burmese) language and preferably speak local dialects.
Malaria CHWs are trained to obtain simple history of malaria symptoms and use an RDT to screen for parasitaemia. Patients with a positive RDT are treated with the appropriate anti-malarial. Severe and complicated malaria patients are referred to the nearest hospital, and CHWs are instructed to report immediately to the health centre when there are unusual high occurrences of malaria or febrile patients in the community. In addition to malaria diagnosis and treatment they typically also provide health education and assist health staff during impregnation sessions of bed nets or the distribution of long-lasting insecticidal nets (LLINs) in the village. Some organizations have extended the role of their CHWs beyond malaria control to other diseases and interventions, implying additional costs and potentially higher health benefits. While such diversification can improve the efficiency of CHW programmes [
11], in this analysis the focus is restricted to their role in improving malaria diagnosis and treatment, and their cost is attributed fully to this aim.
Costing
An ingredients based micro-costing approach was used to develop a model of the annual implementation cost of a malaria CHW in Myanmar whereby all ‘ingredients’ necessary to implement a CHW programme are (1) identified, (2) measured and (3) valued. The cost model was constructed based on the activity cost centres for CHWs comprising training, patient services, monitoring and supervision, programme management, overheads and incentives. The model takes a provider perspective to help inform funders and programme managers on resource allocation and project planning. Costs to patients are excluded but considered to be very low since the CHWs are based in their community and services are usually free of charge.
A key aspect of this analysis was to consider how the cost of CHW programmes differs in more or less remote contexts. To this end, four remoteness strata were defined; easy, medium, difficult and very difficult to access. These strata are defined by incremental increases in travel costs and a change in the mode of monitoring and supervision; details below. The model is based on a set of key parameters, detailed in Table
1. Model parameters can be adjusted to observe the effect of programme variation or parameter uncertainty on programme costs.
Table 1
Community health worker costing model parameters
Economic |
Exchange rate | 810 | 699 | 1000 |
Discount rate | 3 % | 0 % | 8 % |
Training |
Number of trainers | 2 | 1 | 3 |
Number of facilitators | 1 | 1 | 3 |
Number of CHWs in initial training | 25 | 10 | 40 |
Number of days for initial training | 5 | 3 | 7 |
Number of CHWs active year | 3 | 1 | 5 |
Number of CHWs in refresher training | 55 | 30 | 90 |
Frequency of refresher training per year | 1 | 0.5 | 2 |
Patient services |
Number of tests performed by CHW per year | 135 | 20 | 250 |
Percentage of test positivity | 14 % | 1 % | 50 % |
Test price | 0.69 | 0.48 | 0.90 |
Village size | 500 | 100 | 671 |
Monitoring and supervision |
Number of field supervision per CHW* | 8 | 4 | 12 |
Number of meetings in health centre per year** | 12 | 8 | 12 |
Percentage of attendant in monthly meetings** | 80 % | 50 % | 100 % |
CHWs incentive |
Monthly incentive | 20 | 5 | 50 |
Incentive for negative test | 0.2 | 0.2 | 0.5 |
Incentive for positive test | 0.3 | 0.3 | 1 |
Overhead | 10 % | | |
Financial data on CHWs programmes were obtained from the 2013 financial reports of the Three Millennium Development Goal fund (3MDG) implementing partners that have been working on malaria in Myanmar. While the model uses data from the 3MDG reports, the analysis does not reflect any specific organisation or an average of the 3MDG programmes. Prices in local currency (Myanmar Kyat) were converted to US dollars using the mean interbank exchange rate for 2013 (1 US$ = 810 kyats) [
12].
CHWs receive an initial training, which is typically supplemented by annual refresher trainings, the number of which depends on the duration of the project. CHWs are initially trained in malaria prevention and behavioural change communication, treatment of uncomplicated malaria according to national treatment guidelines, and patient registration. Some programmes include malaria as part of a wider package of services, this analysis relates to malaria-specific CHWs.
All ingredients identified in trainings were incorporated into the model including: travel costs for trainers, facilitators and trainees; food and accommodation, rental fees for training venue; expenses for stationary and learning aids; and fees for the trainers and facilitators. Total training costs were estimated depending on number of days and number of volunteers participating in the training.
Patient services include cost of commodities provided to CHWs to deliver malaria intervention and control to the community. The ingredients for patient services include: CHW kits, RDT and treatments for uncomplicated malaria (ACT plus single dose primaquine for Plasmodium falciparum malaria and chloroquine plus primaquine once a week over 8 weeks for Plasmodium vivax malaria). Anti-malarial medication and RDT are purchased separately by 3MDG and provided to implementing partners; the wholesale procurement price is used for these items. Data on the number of malaria tests performed in fever cases in the community and the number of positive cases were obtained from programme reports. The expected cost of patient services is calculated as a function of testing and positivity rates.
CHWs are usually monitored on a regular basis. In easy and medium accessibility areas a focal health centre model is used for CHW monitoring and support meetings. Once per month CHWs in the catchment area of a health centre meet to receive supplies, feed back data and receive general supervision. The proportion of CHWs attending monthly meetings is assumed to be 80 %. Supervisors also visit the villages quarterly to evaluate the CHWs’ work and to solve any problems the CHWs encounter. The ingredients for this activity include: travel cost for volunteers and supervisors; food and accommodation allowance; rental fee for meeting venues; and per diems for both volunteers and supervisors. In difficult and very difficult to access areas, field supervision trips to CHWs are performed by a mobile health team every 6 weeks as it is too burdensome for CHWs to travel to the nearest health facility. The teams provide on-site training and home visits of patients to evaluate the quality of the services performed and monitoring of correct use of resources. There are three members in a mobile supervision team and at least three villages are supervised in one trip, requiring a total of 4 days per supervision trip. In addition the daily cost of transportation is raised incrementally with remoteness category. The ingredients for this field supervision include travel cost for round trips and food, accommodation and salaries for malaria field supervisors.
Annual management cost for each CHW was calculated based on a hypothetical mid-sized organisation supporting 55 CHWs in five townships. For this model, all programme management is assumed to be undertaken by Myanmar nationals (rather than more expensive international staff). A 10 % overhead cost was applied to all cost centres except incentives to reflect office and utility costs. Incentives were included either as monthly fixed costs or according to performance of CHWs as detailed in the scenario analysis.
Opportunity costs of CHW time were estimated to explore full economic cost of CHW programme. The time contributed by CHW for malaria program were estimated and quantified the monetary value by multiplying average monthly salary in Myanmar US$ 180 per month [
13]. These cost included time spent for training, patient service and time spent for being monitored by the supervisor. Time spent for each fever case by the CHW (24 min) was taken from a study in Ghana reporting the CHW working time in management of malaria in children [
14].
Sensitivity and scenario analysis
CHW programmes vary considerably between settings depending on a range of geographic, demographic, behavioural and programmatic factors. Scenario analyses were carried out to reflect the variation in field settings. Estimating annual cost of CHWs depends on the geographical location of the villages, epidemiology of malaria transmission, infrastructure such as accessibility of road condition, and transportation and other activities performed by the project.
Three dimensions of parameter variation are used in the scenario analysis (1) setting remoteness; (2) CHWs testing rate; and (3) financial incentives. These factors were identified as important to CHW programme costs during initial model development and through dialogue with programme managers and financial officers. The default number of RDT performed by each CHW is an average 135 per year, with an upper mean estimate of 250 tests and 20 tests as a lower estimate, reflecting a range of malaria testing in the CHWs reporting data (unpublished data, National Malaria Control Programme). The organizations implementing CHWs programmes in Myanmar use different incentive structures. The incentive schemes included here are (1) no incentive; (2) US$ 0.3 per RDT performed and US$ 5 per month; (3) US$ 0.5 per test and 0.5 per treated patient; (4) US$ 20 per month.
One-way sensitivity analysis was performed on all programme and economic parameters to assess their impact on the model. The upper and lower limits data are detailed in Table
1 and the results are presented in a tornado diagram.
The cost of financial remuneration (incentives) for CHW is a topic of interest to decision makers. The scenario analysis presents different incentive models and incentive amounts based on active CHW programmes in Myanmar. In addition univariate sensitivity analysis is presented on a range of values for monthly incentives and per test incentives. The changes in cost per CHW per year are explored as incentives varying from US$ 0 to 80 per month and US$ 0 to 4 per test. The incentive per test is more uncertain to the programme planner as it depends on the number of tests performed. This is included in the analysis, using testing rates sampled from a uniform distribution of the range in Table
1.