5.5.7. Analyse 5 - Mise à l'échelle

L'analyse de mise à l'échelle permet de tester différents scénarios visant à étendre un réseau de fourniture de services existant ou de générer un réseau complet de fourniture de services s'il n'en existe pas déjà un.

Les scénarios pouvant être testés à l'aide de cet outil prennent en compte:

1. Zones d'exclusion dans lesquelles il est interdit de créer de nouvelles formations sanitaires (voir section 3.3.1.5)
2. Différents facteurs de pertinence répartis géographiquement qui sont combinés dans un indice de pertinence pour définir les sites les mieux adaptés à la localisation d'une nouvelle structure de santé. Ces facteurs peuvent être basés sur la distance par rapport à des éléments spécifiques, le temps de trajet entre ces éléments, la densité de la population ou la définition des zones dans lesquelles les structures de santé doivent être prioritaires par rapport aux autres (voir Section 3.3.1.10).

Une analyse de l'accessibilité physique est ensuite effectuée sur toutes les cellules raster présentant la valeur la plus élevée, en dehors des zones d'exclusion, dans la couche de distribution de pertinence.

Le site (cellule) couvrant la population la plus nombreuse est celui sélectionné pour l’emplacement d’une nouvelle structure de santé. Le type de structure placé sur ce site est ensuite décidé en fonction du contenu de la nouvelle table d’informations sur la structure de santé (voir section 3.3.2.3).

Ce qui précède est répété jusqu'à ce que:

1. Le nombre de nouvelles structures de santé définies par l'utilisateur ait été localisé; ou,
2. La couverture de population définie par l'utilisateur ait été atteinte.

La description ci-dessus n'est qu'un résumé du processus global appliqué lors de cette analyse. Veuillez vous référer à l’Annexe 7 pour plus de détails. 

L'ensemble des captures d'écran suivantes explique comment remplir les sections permettant de saisir les différentes couches et paramètres permettant d'effectuer l'analyse de mise à l'échelle:

Données en entrée:

 (1)  Sous “Select population layer (raster)”, sélectionnez la couche de format raster contenant la distribution spatiale de la population cible d'origine - "population". Cette couche sera utilisée comme dénominateur pour le calcul de la couverture de population obtenue grâce à l'analyse de mise à l'échelle.

(2)  Sous “Select residual population layer (raster)”, sélectionnez la couche de format raster contenant la population cible à couvrir via l'analyse de mise à l'échelle. Cette couche servira également de numérateur pour calculer la couverture de population atteinte à chaque itération de l'analyse de mise à l' échelle. Pour l'exercice, sélectionnez la couche de format raster contenant la distribution spatiale de la population résiduelle obtenue après analyse de la couverture géographique - "[geographic analysis 120m]”.

(3)  Sous “Select merged land cover layer (raster)”, sélectionnez la couche de format de raster contenant la couverture du sol fusionnée résultant de l'utilisation du premier outil (voir la section 0) - dénommé "landcover merged".

(4)  Sous “Select scenario table (table)”, sélectionnez la table de scénario de voyage que vous souhaitez utiliser. Dans le présent exemple, il s’agit de celui utilisé pour l’analyse précédente.

(5)  Sous “Select existing health facilities layer (vector)”, sélectionnez la couche vecteur contenant l'emplacement des structures de santé déjà existantes - ici, nommée "facility". Veuillez noter que ce champ, ainsi que les champs suivants de cet outil, n'apparaîtront pas dans la liste si vous avez coché l'option "Start with empty layer" sous "Options for the output layer [OUTPUT FACILITIES]" dans le menu Analysis settings" (voir après la figure suivante ci-dessous).

(6)  Dans “Select facility ID field (unique)”, sélectionnez le champ dans la table attributaire de la couche des structures de santé qui contient les identificateurs uniques des structures de santé.

(7)  Dans “Select facility name field (text)”, sélectionnez le champ dans la table attributaire des structures de santé qui contient le nom de chaque établissement de santé.

(8)  Dans “Select facilities capacity field (numeric)”, sélectionnez le champ de la table attributaire de la couche des structures de santé qui contient la couverture maximale, la capacité de chaque structure de santé.

 Scénario de voyage:

 (9) Comme pour l'analyse d'accessibilité (voir section 5.5.3), la section "Selected scenario table", à droite du panneau, permet d'importer le contenu de la table de scénario externe et / ou de modifier manuellement les informations indiquées dans les colonnes étiquette, vitesse et / ou mode.

 Le reste de la section "data input" ainsi que les sections "scaling up" du panneau peuvent ensuite être renseignées comme suit:

Données en entrée:

(1)  Sous “Select existing capacity table”, sélectionnez la nouvelle table d'informations sur les structures sanitaires (voir section 3.3.2.3) que vous avez importée dans le projet, celle qui est étiquetée "capacity scaling up". Veuillez noter que le contenu de cette table peut être créé manuellement, s'il n'est pas déjà importé dans le projet (voir le point (4) ci-dessous).

(2)  Le champ “Select existing suitability table” peut être utilisé pour sélectionner une table contenant les facteurs de pertinence utilisés dans une analyse de mise à l’échelle précédente. Ce champ indique [NO DATA] pour le moment, car nous n’avons effectué aucune analyse de mise à l’échelle pour le moment. 

(3)  Le champ “Select existing exclusion table” peut être utilisé pour sélectionner une table contenant la liste des zones d’exclusion utilisées dans une analyse de mise à l’échelle précédente. Ce champ est vide pour le moment, aucune analyse de mise à l'échelle n'a été effectuée pour le moment.  

(4)  Les tables de la section “Capacity table for new facilities creation” affichent le contenu de la nouvelle table d’informations sur les structures de santé sélectionnées au point (1) ci-dessus - «capacity scaling up». Vous pouvez créer ou modifier cette table en cliquant dans chaque cellule de la table. Des lignes peuvent également être ajoutées ou supprimées via les boutons "Ajouter une ligne" et "Supprimer une ligne".

Tables de mise à l'échelle:

(5) La section “Scaling up tables” est utilisée pour définir les facteurs de pertinence et les zones d'exclusion à utiliser dans l'analyse. L'utilisation de cette section est décrite plus en détail ci-dessous. 

(6) Les deux tables, dans le coin inférieur droit de la figure ci-dessus (vide maintenant), contiennent la liste des facteurs de pertinence et des zones d'exclusion définies par l'utilisateur. La méthode pour remplir ces tables est expliquée dans les deux sections suivantes.

Utilisation de la section "Tables de mise à l'échelle"

Cette section permet de définir les facteurs de pertinence et les zones d’exclusion dans lesquelles aucune nouvelle structure ne sera située au cours de l’analyse de mise à l'échelle (voir la section 3.3.1.5). La section suivante décrit en détail comment définir ces facteurs et ces zones.

 Indice de pertinence

AccessMod définit le site le plus approprié pour localiser un nouveau centre de santé en combinant les distributions spatiales de différents facteurs de pertinence dans une seule couche d’indice de pertinence.

Les facteurs de pertinence accessibles à l'utilisateur sont les suivants:

-     Sum of population within a radius (in kilometers): cette option créera une nouvelle couche dans laquelle la somme de la population cible (originale ou résiduelle) située dans un rayon donné de chaque cellule sera attribuée à la cellule en question. Ce facteur de pertinence permet, par exemple, de considérer les sites avec la plus forte densité de population voisine comme les plus appropriés pour localiser une nouvelle structure.

-     Euclidean distance from features (ex: routes, structures existantes): cette option créera une nouvelle couche contenant la distribution spatiale de la distance euclidienne à partir d'entités stockées dans une couche vectorielle (point, ligne ou polygone). Cela permet, par exemple, de considérer les sites proches des routes comme les plus appropriés pour localiser une nouvelle installation.

-     Travel time to/from features: cette option est similaire à la précédente, mais utilise le temps de trajet au lieu de la distance euclidienne. Cela permet, par exemple, de considérer les sites situés loin des centres de santé existants comme les plus appropriés pour localiser un nouveau centre.

-     Generic priority map: les couches de format raster contenant des valeurs entières peuvent être directement intégrées à l'index (voir la section 3.3.1.10). Cela donne un haut degré de flexibilité, la priorité pour la mise à l'échelle pouvant provenir de tout autre paramètre externe distribué dans l'espace, tel que les données sur la prévalence sous-nationale ou l'inégalité des revenus. Cela permet, par exemple, de considérer que les districts avec une prévalence particulière sont les plus appropriés pour localiser une nouvelle structure.

Pour chacun de ces facteurs, il est possible de définir si des valeurs plus élevées dans la couche de répartition des facteurs doivent être considérées comme plus ou moins pertinentes pour localiser une nouvelle installation. Cette option permet, par exemple, de s'assurer que les sites situés le plus près du réseau routier sont considérés comme les plus appropriés lors de l'utilisation du facteur de distance euclidien par rapport aux entités. Dans ce cas, ces sites présenteraient les valeurs les plus basses dans la couche de distribution des facteurs obtenue. 

AccessMod permet également l’attribution d’un poids spécifique à chacun des facteurs de pertinence à prendre en compte dans l’indice final. Cette possibilité offre un niveau de flexibilité supplémentaire à l'utilisateur qui souhaite combiner plusieurs facteurs de pertinence, tout en considérant que certains jouent un rôle plus important que d'autres lors de la définition des sites les plus appropriés pour la localisation d'une nouvelle structure de santé.

La table ci-dessous décrit les paramètres d’interface et d’entrée associés à chacun des facteurs de pertinence susmentionnés.

Sum of population within a radius	Euclidean distances from features	Travel time  from/to features	Generic priority map
(1) Under “Select a factor for the suitability index”, select the "Sum of the population within a radius" option from the pull-down list. (2) Set the radius (in kilometers) to be used. (3) A message indicating how many cells will be processed at each iteration will appear. (4) “Direction of prioritization” allows the analyst to define if higher values for this factor are more suitable or less suitable for locating a new health facility. (5) “Select factor weight” allows for the attribution of a relative weight to the factor. (6) “Select attribution layer” field selects the raster format population distribution that will be used to generate the factor distribution layer (see below). (7) The “Add” button allows you to add the factor to the list of suitability factors that will be used in the final index.	(1)  Under “Select a factor for the suitability index”, select the "Euclidean distances to features" option from the pull-down list. (2) “Direction of prioritization” allows the analyst to define if higher values for this factor are more suitable or less suitable for locating a new health facility. (3) “Select factor weight” allows for the attribution of a relative weight to the factor. (4) “Select available layer” selects the vector format layer that will be used to generate the factor distribution layer (see below). (5) The “Add” button allows you to add the factor to the list of suitability factors that will be used in the final index	(1) Under “Select a factor for the suitability index”, select the "travel time from/to feature" option from the pull-down list. (2) The “Type of analysis” field allows the analyst to specify if the travel time will be based on an isotropic or anisotropic approach. Selecting the anisotropic approach will result in a new field to open, in order to indicate the direction of travel (from, or to, the selected features). (3) The “Direction of prioritization” field allows the analyst to define if higher values for this factor are more suitable or less suitable for locating a new health facility. (4) “Select factor weight” allows for the attribution of a relative weight to the factor. (5) “Select available layer” allows selecting the vector format layer used to generate the factor distribution layer. (6) The “Add” button allows you to add the factor to the list of suitability factors that will be used in the final index.	(1) The “Select suitability analysis” factor takes into account a raster data set with priorities as values (2) The “Direction of Prioritization” field allows the analyst to define if higher values for this factor are moresuitable or less suitable for locating a new health facility. (3) “Select factor weight” field allows for the attribution a relative weight to the factor. (4) “Select available layer” selects the raster format layer containing the generic priority map (5) The “Add” button allows you to add the factor to the list of suitability factors that will be used in the final index.

Each of the suitability factors defined by the user is stored as a separate record in the "suitability factors" table appearing in the Scaling up section of the module panel (middle part of the panel). This table looks like the following figure and contains the following fields:

• select: the checked entries in this column indicate the list of suitability factors being taken into account when creating the suitability index distribution layer. Any unchecked factor will not be taken into account.
• factor: suitability factor that has been selected, as follows:
  • popsum: Sum of population within a radius
  • dist: Euclidean distances from features
  • traveltime: Travel time from/to features
  • priority: Generic priority map
• layer: information identifying the layer used to spatially distribute the suitability factor. The information reported in this field can either be:
  • "rOutputPopulation", when selecting the target population that will be updated at each iteration (see the box below for more information)
  • "vOutputFacility", when selecting the health facility location layer that will be updated at each iteration (see the box below for more information)
  • xxx_yyy@zzz with xxx=data class (see Section 5.4.1); yyy= data tag and zzz= project name
• weight: weight attributed to this particular suitability factor. The higher the value the higher the influence of the factor compare to the other ones. For example, the weight of "2" applied to the sum of population within a radius (popsum) factor in the figure here above means that the influence of this factor is twice more important in the final index than the influence of the other factor considered here, the Euclidean distances from features (dist)
• options: list of the options set by the user, with:
  • "r": Value of the radius set for the sum of population within a radius suitability factor
  • "t": Type of analysis set for the travel time from/to features factor (iso= isotropic; aniso=anisotropic)
  • "d": Direction of travel set for the travel time from/to features factor (to= towards feature; from= from feature)
  • "p": direction of prioritization (hvms = higher values are more suitable; hvls= higher values are less suitable)

 Exclusion areas

 The exclusion areas are defined using the "Exclusion Areas" sub-section of the "Scaling up tables" section of the module panel, shown below:

The exclusions areas, used in the context of the scaling up analysis, can be defined in two ways, namely by using:

• The vector or raster format layers imported as exclusion areas in the project (see Section 3.3.1.5).
• The location of the health facilities, including the new ones being located at each iteration during the analysis.

 Please note that, in both cases, the user can define:

1. An optional buffer (expressed in kilometers) to be set around the exclusion areas
2. If new candidates should be kept inside or outside the final exclusion areas (i.e., the exclusion areas contained in the selected vector or raster layer, together with the defined buffer)

 The above can, for example, avoid locating new health facilities within a certain distance of already existing facilities - one of the constraints used in Burkina Faso.

 Each of the exclusion areas defined by the user is being stored as a separated record in the "Exclusion areas" table appearing in the Scaling up section of the module panel. This table looks like the following figure and contains the following fields:

• select: the checked entries in this column indicate the list of suitability factors being taken into account when creating exclusion area layer. Any unchecked factor will not be taken into account.
• layer: information identifying the layer used to define the exclusion area(s). The information reported in this field is the same as for the suitability factors table (see above).
• buffer: size of the buffer, expressed in kilometers, set around the selected exclusion area(s)
• method: indicate if the new facilities should be located outside (“keepOutside”) or inside (“keepInside”) the defined exclusion areas

For the purpose of the current exercise, we will consider two suitability factors and two exclusion layers as follows:

 Suitability factor 1:

For the first suitability factor, we would like for new facilities to be located in the most highly populated areas from the residual population distribution layer. For the sake of the exercise, we also consider that this factor is twice as important as the other suitability factor (see below).

 For this priority, we will be using the "Sum of population within a radius" suitability factor, and set the different parameters as follows in the "Suitability index" sub-section:

-       We will use a radius of 5 kilometers

-       For the "Direction of prioritization", choose "Higher values are more suitable"

-       Give a value of "2" for the "Select factor weight"

-       Under "Select available layer", choose [OUTPUT POPULATION] in order to use the residual population distribution layer updated after each iteration

 You can then press the "Add" button to add this combination of parameters as a suitability factor into the "Suitability factors" table.

 Suitability factor 2:

 For the second suitability factor, we also would like for new facilities to be located as close as possible to the road network. For this priority, we will be using the "Euclidean distances from features" suitability factor.

 To enter these settings through the "Suitability index" sub-section, please do the following:

-       For the "Direction of prioritization", choose "Higher values are less suitable"

-       Give a value of "1" for the "Select factor weight" weight

-       Under "Select available layer", choose the "roads" vector layer

 You can then press the "Add" button to add this combination of parameters as a suitability factor into the "Suitability factors" table.

 The "Suitability factors" table should then summarize the factors and parameters that you have chosen, as shown below:

We now turn to the "Exclusion areas" section to define the excluded zones in which no new health facilities should be located.

Exclusion area 1:

 First, we would like to exclude new health facilities from being located within 5 kilometers from an already-existing health facility.

 Set the parameters as follows, in the "Exclusion areas" sub-section of the Scaling up tables section, to include these exclusion areas in the scaling up analysis:

-       "Select exclusion areas": Choose [OUTPUT FACILITIES] in order to account for each new health facility being located after each iteration

-       "Set an optional buffer (km)": Enter "5"

-       "Choose exclusion method": Choose "Keep candidates outside the areas + buffer"

 You can then press the "Add" button to add this combination of parameters, and therefore include these exclusion areas, in the "Exclusion areas" table.

 It is important to emphasize the use of the [OUTPUT FACILITIES] option here, as the exclusion area attached to each new facility placed will have an influence on the location of the next newly-placed facility. 

Exclusion area 2:

 We then would like to exclude any new heath facility from being located within the polygon included in the vector format "exclusion" layer that we previously imported in the project (see Section 5.5.1).

 Set the parameters as follows, in the "Exclusion areas" sub-section of the "Scaling up tables section", to include this exclusion area in the scaling up analysis:

-       "Select exclusion areas": Choose the vector layer "exclusion"

-       "Set an optional buffer (km)": Enter "0"

-       "Choose exclusion method": Choose "Keep candidates outside the areas + buffer"

 You can then press the "Add" button to add this exclusion area in the "Exclusion areas" table.

 The "Exclusion areas" table should then summarize the factors and parameters that you have chosen, as shown below:

The following screenshots describe how the analysis settings and facilities selection sections are to be filled in order to run the scaling-up analysis:

Analysis settings

 (1) Select, under “Options for the output layer [OUTPUT FACILITIES]”, if you want to start the scaling up analysis using selected existing facilities or with an empty health facility layer. The former is checked by default when starting the scaling up tool. Selecting the later will result in the "Select existing health facilities layer (vector)" field and the other associated ones to disappear from the data inputs section.

Parameters for new facilities evaluation

(2) In “Type of analysis”, select whether you want an “anisotropic” or “isotropic” approach to be applied when identifying which type of facility should be placed in the most suitable site. We will choose “anisotropic” in the present exercise to take the impact of slope upon speed into account for the WALKING and BICYCLING modes of transportation.

(3) In “Direction of travel”, select the direction of travel of the patients. As in for the previous tools, this field will only appear if the "anisotropic" option has been checked in the previous point. As this is the case, for the exercise, choose "Towards facilities" to consider travel time in the direction of the health facilities.

(4) Under “Maximum travel time [minutes]”, specify the maximum travel time (in minutes) that AccessMod will use to draw the catchment area(s) for the most suitable site(s) identified at each iteration. The population located within the catchment area attached to the final suitable site will be used to define the type of health facility to be located on that site. We use 120 minutes for this exercise.

Computation limits

 The computation will stop as soon as one of the values set here are reached. Please note that a value needs to be indicated in each of these fields in order for AccessMod to run and that a value of zero indicated in any of these fields will be considered as 'no limit'.

 (5) Within “Percentage of population to cover (%)”, indicate the percentage of the population (expressed in %) you would like the new updated health facility network to cover through the scaling up analysis. Such percentages are calculated at each iteration over the study by dividing the new residual population distribution layer by the original population layer. A "0" value means "no limit". For the exercise, we wish to extend our facility network so that it covers at least 53% percent of the population, so you can specify "53" in this field. (Remember, our current facility network covers about 48% of the population within a 2-hour travel time. See the end of Section 5.5.4 for details.)

(6) Within “Number of new health facilities to locate [facility]”, indicate the number of new health facilities to locate during the analysis. This number corresponds to the maximum number of iterations that AccessMod will perform before stopping the analysis. A "0" value means "no limit", which is what we will be using for this exercise.

(7) Under “Maximum processing time (minutes)”, indicate the maximum processing time (in minutes) for the analysis. This parameter is being provided to allow for the user to fix a certain limit of time to the overall computation. A "0" value means "no limit", which is what we will be using for the exercise.

(8) Under “Add short tags”, give short tags to be attached to the different outputs of the analysis. We will use "scaleup53" for the present analysis.

Facilities selection

 (9) As in the previous analysis, you can select in this section the facilities on which the analysis will be performed. Please note that:

• such selections will only be used if the "Start using selected existing facilities" option has been checked earlier (Step 1)
• the table will still appear in the panel, even if the "Start with empty layer" option was checked

 Keep all the facilities selected for the purpose of the present exercise.

Validation

(10) The validation section (indicated to the right) should indicate that all fields have been correctly filled in (with a green “OK”). If this is the case you can hit the "Compute" button to launch the analysis. If this is not the case, the "Compute" button will still be red and you will have to go through the warning and error message to find out what needs to be adjusted.

A transparent window with some text and a progress bar will appear in front of the panel while the analysis is being conducted. Please wait until this window disappears to continue using AccessMod. This analysis can take a large amount of time to be performed because of the iterative way of processing each health facility, and depends on the number of suitability factors and exclusion areas that have been selected, as well as the size of the study area.

 Once the window disappears and the analysis is done, go back to the Data module to check that the eight output datasets that have been generated are as follows:

1. scenario processed class: Table containing the travel scenario that has been processed.
2. population residual time class: Raster format layer containing the residual population. This is the population still not covered by the scaled-up network obtained in this analysis.
3. facility scaling up class: Vector format layer containing the new health facilities that have been computed, and located, plus the original ones in case the "start using selected existing facilities" option has been checked in the analysis settings.
4. capacity processed class: Table containing the values of the processed capacity table.
5. result scaling up analysis class: Table containing the results of the scaling up analysis.
6. catchment scaling up class: Vector format layer containing the extent of the catchment area for each newly located facility.
7. exclusion processed class: Table containing the parameters and parameter values included in the exclusion table.
8. suitability processed class: Table containing the parameters and parameter values included in the suitability table.

 As you did for the previous analysis, archive all these results, export and unzip them in order to open and visualize the results.

 We will concentrate here on the main results of this scale up analysis, namely the location and types of the new located health facilities, and the statistics associated with the scaling-up analysis.

First, open in your GIS software:

• The vector format layer containing the locations of the new health facilities located during the analysis plus the original ones in case the "start using selected existing facilities" option has been checked in the analysis settings. This layer is found in the output folder named "vector_facility_scaling_up_scaleup53".
• The vector format layer containing the catchment areas for the new health facilities located through the analysis. It is found in the folder named "shape_catchment_scaling_up_scaleup53".

In Figure 10 below, we show a layer composition (made in QGIS) showing the resulting health facility layer containing the 10 original health facilities (in red) and the new 7 health facilities (in blue), together with their respective catchment areas (light blue). The road network and the exclusion area used during the analysis have also been added here to facilitate the visualization of the results. We can see in this Figure that seven new health facilities have been located in the south-eastern part of the area, and one health facility has been located in the northern part. For the last facility, we can see that it is has been located outside of the exclusion area, yet its catchment extends over this exclusion area. 
Error rendering macro 'captioneditem' : Plugin license error: EXPIRED

If you open the attribute table of the new health facility vector file in your GIS software, you will get the following table (here displayed in QGIS):

In this table, you can see that the first 10 lines contain the information for the existing health facilities, while the next 7 lines contain the information for the new health facilities that have been located through the analysis. Compared to the original attribute table of the existing health facility layer, there are 4 new columns that have been added, which are the 4 right-most columns in the screenshot above:

• amName: Temporary name attributed to each new health facilities by AccessMod.
• amid: Unique identifier attributed to each new health facility by AccessMod. Please note that this identifier does also correspond to the order in which the new health facilities have been located during the analysis.
• amCapacity: Maximum coverage capacity attached to the new health facility through the analysis, based on the field selected from the new health facility information table (see Section 3.3.2.3).
• amLabel: The health facility type attached to the new health facility through the analysis, based on the field selected from the new health facility information table (see Section 3.3.2.3).

 The statistical results from the scaling up analysis are found in the Excel file named "table_result_scaling_up_analysis_scaleup53.xlsx". This table contains the following columns. (Note that the table is split in two parts in the following screenshot.)

• amid: Unique identifier attributed to each new health facility by AccessMod. Please note that this identifier corresponds to the one reported in the "amid" column of the resulting health facility layer
• amRankComputed: Order in which the new health facilities have been located during the scaling-up analysis
• amName: Temporary name of the new health facility attributed by AccessMod and corresponding to the name reported in the "amName" column of the resulting health facilities layer
• amTravelTimeMax: Maximum travel time, in minutes, set for the analysis
• amPopTravelTimeMax: Total population located within the maximum travel time set for the analysis (amTravelTimeMax)
• amCorrPopTime: Pearson correlation coefficient between the set of travelling times (t, from 0 to the maximum travel time) and the corresponding covered population within this time step (i.e., the sum of population in all cells that are located between t and t+1 of travelling time). The correlation can give a rough estimate of how to population is distributed through space as we move outward from the health facility. As an example, a large positive value (e.g. 0.421 in the first line of the above table) means that the population is relatively uniformly distributed as you expand outwards from the facility. A strong negative value would mean that there is more population closer to the health facility than there is far away from it. A correlation close to zero means that there is no tendency in how the population is spread within the catchment area. Therefore, this correlation should really just be used as a relative indicator, as we do not provide the significance of the correlation in AccessMod (You should use proper GIS analysis outside AccessMod if you need statistical insights about how the population is spatially spread within the catchments).
• amLabel: Health facility type attached to the new health facility through the analysis, based on the field selected from the new health facility information table. This field contains the same information as the "amLabel" column of the resulting health facilities layer.
• amCapacity: Maximum coverage capacity attached to the new health facility through the analysis, based on the field selected from the new health facility information table. This field contains the same information as the "amCapacity" column of the resulting health facilities layer.
• amTravelTimeCatchment: Travel time to reach the maximum extent of the catchment area attached to the health facility. This value is either:
  • Equal to the maximum travel time set for the analysis (amTravelTimeMax) when the maximum coverage capacity of the facility has not been reached within that given time (120 min in this case); or,
  • Smaller than the maximum travel time set for the analysis (amTravelTimeMax) when the maximum coverage capacity of the facility is reached before reaching amTravelTimeMax.
• amPopCatchmentTotal: Population located within the catchment area for the travel time reported in the amTravelTimeCatchment field
• amCapacityRealised: Part of the maximum coverage capacity of the health facility that is being realized (used) based on the total population located in the catchment area for the travel time. This value is:
  • Equal to the maximum coverage capacity of the health facility (capacity) when this value is reached before the maximum travel time (amTravelTimeMax).
  • Smaller than the maximum coverage capacity of the health facility (capacity) when such capacity is not reached within the maximum travel time (amTravelTimeMax).
• amCapacityResidual: Part of the maximum coverage capacity of the health facility that is not being used. This value is calculated by finding the difference between the maximum coverage capacity (capacity) and the realized capacity (amCapacityRealised). This value is therefore equal to 0 when the maximum coverage capacity is being reached within the given travel time.
• amPopCatchmentDiff: Part of the total population located in the final catchment area (amPopCatchmentTotal), which is not covered by the facility due to lack of coverage capacity. This value is calculated by finding the difference between the total population in the catchment area (amPopCatchmentTotal) and the realized capacity (amCapacityRealised). This value is therefore:
  • Equal to 0 when the maximum coverage capacity is not reached within the given travel time; or
  • Greater than 0 when the maximum coverage capacity is reached within the given travel time
• amPopCoveredPercent: Cumulative geographic coverage expressed in percent. This value corresponds to the percentage of the initial population distribution grid that is being covered after each iteration (i.e., after each new health facility is located)

Of particular interest after the scaling-up analysis are the values of the cumulative geographic coverage reported in the "amPopCoveredPercent" column. The content of this column allows for the identification of the percentage of the population covered after each iteration, i.e., after having located an additional health facility in the study area, taking the suitability factors and exclusion areas into account.

 As we specified that we wanted to reach 53% percent of geographic coverage of the total population in this analysis, the scaling-up analysis had to locate 7 new health facilities before reaching 53%. With these 7 new health facilities, the updated network of health facilities now covers 53.19% of the population (circled in red in the previous figure).