3.3.1.6. Health facilities location
AccessMod needs a vector format GIS layer containing the geographic locations of the health facilities to be considered in the physical accessibility, geographic coverage and referral analysis as well as scaling up analysis (when taken into account). These locations should be expressed in the same projection system as the DEM (see Section 3.3.1.1)
Depending on the analysis to be performed, the facilities in question can either all be placed in the same layer or in separated layers.
The attribute table of this/these GIS layer(s) should at least contain:
- A unique identifier to appear in the resulting tables generated by AccessMod (cat field in the example below).
- The maximum coverage capacity of the health facility when conducting the geographic coverage analysis (see Section 5.5.4). Annex 2 provides more information on this parameter and describes an approach to calculate it (capacity field in the example below).
In addition to the above, it is also recommended to keep the official name of the facility in the attribute table as per the example presented here below.
Example of attribute table for the health facility layer
It is important to remember that the accuracy of the health facility coordinates can have an impact on the result of the analysis (see Section 3.3.1.1). Priority should therefore be given to coordinates that have been collected using methods providing a high accuracy. A guidance document on primary data collection methods for facility geographic coordinates is provided by the Health GeoLab Collaborative (Ebener et al., 2016). Future guidance documents are planned to cover the extraction of coordinates from other sources, and these will be made available through the Health GeoLab Collaborative website.
Health facilities cannot fall on barriers to movement (i.e., “No Data” values in the merged land cover distribution grid) or on cells with speed of zero (considered as barriers). Such facilities will be tagged with, respectively, a “Yes” label in the “amOnBarrier” column and a “Yes” label in the “amOnZero” column of the "Selected facilities" table. They will either have to be unselected or their coordinates modified before being able to run the analysis.
The importance of the modification to be applied will depend on the resolution of the grid format datasets. The lower the resolution the higher the adjustment to be applied on the original coordinates. Even if this approach results in an error in the coordinates being used, the results obtained will themselves be consistent at the scale, and therefore resolution, of the analysis being conducted.
Limiting the fields in the attribute table of the health facility layer to those necessary to conduct the analysis can greatly reduce the size of the file and facilitate the analysis, especially when using a large number of facilities.