# Scaling up the treatment of severe acute malnutrition with community health workers: a geospatial coverage analysis in rural Mali | Human resources for health

### Study location and design

The intervention was conducted between October 2017 and October 2018 to compare the coverage and effectiveness achieved with CHWs treating children 6-59 months for simple SAM under different levels of supportive supervision in the three largest districts of the Kayes region. CHWs in Kita district received a high level of supervision (standard country supervision on primary health care plus nutrition-specific supervision provided monthly by Action Against Hunger staff), Kayes district received moderate level of supervision (country standard supervision provided monthly by Action Against Hunger Hunger staff without specific nutrition supervision) and Bafoulabé district considered control arm with standard supervision from health authorities without any additional support . Findings on the effectiveness achieved by CHWs under these different supervision models have been published elsewhere, showing that the level of supervision has a differential effect on the quality of treatment. [20]. The different models of supervision are not expected to have an effect on treatment coverage.

The main objective of this study was to comparatively assess SAM treatment coverage in these three districts of the Kayes region, before and after the inclusion of CHWs as new SAM treatment providers. supporting the SFs. In addition, other variables that may influence coverage were assessed in the three districts: (1) the geographical distribution of CS and CHW sites and the overlap between the areas covered by the two health care providers; (2) estimation of the number of children under five with geographic access to health care; (3) estimation of the number of children under five screened for SAM by community health volunteers (called Community Relays).

The study was approved by the Ethics Committee of the National Institute for Public Health Research (INRSP) in Bamako (Decision No. 13/2017CE-INRSP). The study protocol has been registered at the ISRCTN https://doi.org/10.1186/ISRCTN14990746.

### Assessment of treatment coverage

A baseline coverage assessment was conducted in August 2017 in all three districts before CHWs started treating SAM. The final evaluation was carried out a year later. SAM treatment coverage was defined as the proportion of children aged 6–59 months eligible for the treatment service out of the number of children aged 6–59 months who actually received the service. Coverage was assessed by the standardized Semi-Quantitative Access and Coverage Assessment (SQUEAC) methodology, which provides in-depth investigation and analysis of barriers and enablers to access [21]. The assessment was conducted in three stages: Stage 1, to identify areas of low and high coverage as well as reasons for coverage failure using routine program quantitative data and qualitative data collected before the investigation ; Step 2, to confirm the location of high and low coverage areas and the reasons for coverage failure identified in Step 1 by small surveys; Step 3, to provide an estimate of overall program coverage using Bayesian techniques.

A single coverage indicator was used to estimate SAM treatment coverage, which also includes recovering cases inside and outside the program using the following formula:

$${text{Single cover }} = , left( {{text{Cin}} + {text{Rin}}} right)/ , left( {{text{Cin}} + {text{Rin}} + {text{Cout}} + {text{Rout}}} right),$$

where Cin = current SAM cases in the program; Cost = current SAM cases, not in program; Rin = recovery of SAM cases in the program; Rout = SAM case recovery, not in program.

To analyze barriers to accessing SAM treatment, semi-structured interviews and focus groups were conducted with CHWs and nurses, community leaders and mothers/caregivers.

### Demographic data

The administrative structure of Mali has different levels ranging from the highest level to the lowest level. The highest level is the first-level administrative division, which refers to regions. The intermediate level is the second-level administrative division, called a district. And the lowest level is the third-level administrative division consisting of communes. A series of measures were taken to obtain the total population by district at the level of the third level administrative division. First, the 2018 United Nations Office for the Coordination of Humanitarian Affairs (OCHA) projections were assumed based on the 2009 census to obtain population data by village. [22]. Then we applied a population growth rate of 3% from the United Nations report “World Population Prospects”, to obtain the total population in 2019 [23]. Then, the population under five was estimated at 20% of the total, and the proportion of children with SAM at 2% (estimates used in the 2018 national SMART survey) [4]. Thus, the population of children under five obtained for the three districts was: 109,927 in Kita, 130,423 children in Kayes and 31,494 in Bafoulabé.

### Auxiliary data sources

According to the policy of the Ministry of Health in Mali, CHWs must be located more than 5 km from the referring CS, and they must cover a population of approximately 700 people within a radius of 3 km around the village site. [13].

Data on children screened for SAM come from district activity reports, including routine screening by community health volunteers and mass screening campaigns conducted during the study period. A case of SAM was defined as a child between 6 and 59 months with a mean arm circumference (MUAC)

### Mapping and geospatial analysis

ArcGIS Desktop 10.8 was used for mapping and geospatial analysis of factors influencing SAM treatment coverage achieved by adding CHWs as treatment providers [24]. The boundaries of the three health districts on the maps were established using information provided in the Humanitarian Data Exchange [25]. The 112 HFs and 152 CHWs in the area were georeferenced to analyze the spatial distribution of health service delivery. Of this number of health care providers, the district of Kita has 38 FS and 75 CHWs, Kayes has 52 FS and 43 ASCs and Bafoulabé, 22 FS and 34 ASCs. The total area in km2 covered by health services and the area not covered by either FS or CHWs were also calculated.

Influencing variables with a possible impact on treatment coverage were analyzed. First, the overlap between the areas covered by the healthcare providers was defined as a HF located less than 5 km from other FS, a CHW located within 5 km of the HF or a CHW located less than 3 km from another CHW. Second, the proportion of children under 5 without geographic access to a health provider was defined as children living more than 5 km from the nearest health center or CHW out of the total estimated number of children under five living in the health zone. Third, the proportion of children under five not covered by active screening, defined as children under five who have not been screened for SAM in their communities out of the estimated total number of children under five living in the health zone.

### statistical analyzes

Statistical analysis was performed with the online software Epitools (https://epitools.ausvet.com.au/). The proportions were compared by the chi-square test applying Yates’ correction when the expected cases were less than 5 in more than 20% of the cells. The Mantel-Haenszel chi-square test was applied to compare the final treatment coverage between the groups adjusted to the initial coverage providing the associated odds ratio (OR). A confidence level of 95% was applied to all analyses, taking into account the p-values ​​less than 0.05.