7.1 Levels of sophistication in facility location modelling

There are many application areas of GIS, where we need to decide how to best configure a network of facilities across a study area. Aside from health, retail applications of GIS are faced with much the same decision in siting new stores, warehouses and industrial premises. Emergency preparedness work similarly involves identifying the most suitable sites for emergency services. In planning health service provision, there are a variety of location-specific decisions that can be supported using GIS:

• The identification of the best site for a new health facility.
• The relocation of existing health facilities to new sites: For example, in many developed countries, older, city centre hospitals are being relocated to modern, out-of-town premises.
• Health facility closures and rationalisation of existing services: The closure of an existing health facility is as much a locational decision as the opening of a new facility.
• Changing levels of service provision at existing sites: A change in the mix of services provided at a particular site (e.g. the closure of accident and emergency facilities in one of two hospitals in a major city) is again a locational decision.
• Planning for population growth: It is also possible to use GIS to look at the likely changes in demand for services and caseload resulting from population growth (such as the provision of new affordable housing on ‘brown field ‘sites within a city).

 

All of these planning decisions can potentially be supported using GIS. There are a variety of ways in which GIS can be used to support health planning decisions such as these:

• Map creation: At its simplest, GIS can be used to develop maps to help support decisions about the provision of facility. Such maps may depict the market areas or catchment areas served by specific health facilities, or show measures of health service provision, such as physician: population ratios.
• Site availability and suitability analysis: There are clearly some sites that will be unsuitable for developing new health facilities, such as areas where any new building is heavily constrained by planning legislation or areas at risk of flooding. These unsuitable areas can be identified through map overlay within a GIS to narrow down the potential number of sites available for a new facility. This suite of techniques, involving the elimination of unsuitable / unavailable sites and the relative ranking of sites that potentially are suitable, is known as Multiple Criteria Evaluation (MCE) within GIS. MCE represents a more complex approach to health service location planning than simply providing the decision-maker with maps.
• Location-allocation modelling: Location-allocation modelling is a mathematical suite of techniques that can be used to optimise the spatial configuration of a health service network. In location-allocation modelling, a health service network is configured with a geographical pattern of facilities (hence ‘location’) and then demand (patients) are assigned to each of the facilities. This latter part is the ‘allocation’ step, hence the term location-allocation modelling. This technique represents the most sophisticated and complex approach to health service planning that is used with GIS.

 

In terms of spatial data, three types of information are frequently used in planning health services:

• Distances between facilities and population: Uptake of health services often displays ‘distance decay’ with consultations declining as distance to a given health facility increases. Distance often explains much of the observed geographical variation in consultation patterns, though its effect varies depending on the specific disease concerned. For this reason, distance is a key variable in most GIS-based studies of health service provision.
• Service provision: The location and nature of any existing service provision (e.g. in terms of capacity such as number of beds; staff levels; range of services offered, or quality of provision) is clearly also important.
• Demand characteristics: These describe how the population within an area is likely to make use of its health services. Demand characteristics may not only cover the geographical distribution of population, but other characteristics such as multiple deprivation, transport access, income, insurance coverage and ethnicity.

 

In addition to these core types of information, other criteria may also be important in locating health facilities, some of which may be difficult to represent in GIS.

---

---