5.2 Definitions of accessibility

The purpose of this object is to think about the concept of accessibility in the context of health care. Access to health services has become an issue of global importance because it affects the levels of care received by people in need which, in turn, has a fundamental effect on the health status of a population. A particular concern is that levels of access are very low in many parts of the world and this is contributing to severe and widespread public health problems, particularly in low-income countries. GIS has a role to play in measuring and modelling certain aspects of accessibility. An important first step is to try and define what we mean by accessibility in the context of health care, and explore some of the different factors that determine levels of access.

A dictionary definition of access is “the right or opportunity to use something” (Concise Oxford English Dictionary, 10th edition). In the context of health care, the terms access and accessibility are used in a variety of senses relating to the health services that are available to a given population and the way people use those services. The health services in question are often formal biomedical services such as doctors, hospitals, or clinics. However, the concept of access is important for other formal services that contribute to health and wellbeing such as social services, mental health care, and health education and to a wide range of informal sources of health care provided by family or community structures.

“The more accessible a system is, the more people should utilize it to improve their health.”

(World Health Organisation, 2000)

 

Health organisations and academics have sought to provide definitions of access and to identify different components of accessibility. Rogers et al. (1994) define optimal access as “ providing the right service at the right time in the right place.” Penchansky and Thomas (1981) consider access to be defined by five key elements: availability, affordability, accommodation, acceptability and accessibility. These are expanded on in the box below.

“In many low income countries, large segments of the poor still have no access to basic and effective care.”

(World Health Organisation, 2000)

 

 
The first four components discussed in the box above could be investigated from a spatial or non-spatial perspective. A spatial approach might be to map the variation in, for example, the price of a given service in different parts of a country to identify any interesting patterns. A non-spatial approach might be to compare these prices with possible explanatory factors such as levels of income. The last component, geographic accessibility, is explicitly spatial. To investigate, model, or measure geographical accessibility in a quantitative way requires the kind of spatial approach to which a GIS is ideally suited.

 

Two-Step Floating Catchment Models

One simple way of measuring access would be to take a set of polygons (e.g. health districts) and then overlay these on a point map of facility locations. If we had population head counts for each district and numbers of physicians for each facility, we could then calculate a physician: population ratio for each polygon as a measure of access. However, there are problems with doing this. Most obviously, the way in which the boundaries are constructed – their shape and size – could affect the physician: population ratio considerably (the so-called Modifiable Areal Unit Problem). A city centre hospital would have a very impressive physician: population ratio, for example.

An alternative might be to take detailed data on population and calculate a travel time or distance to the nearest facility. However, again there are problems with doing this. For example, a single facility could serve a very large population in a city centre. Even though people living next door to this facility would be very close, were they to use it, they would find staff hard-pressed to serve the large population living close by.
A more sophisticated approach has therefore been developed to modelling access which can readily be implemented in a GIS (Luo and Wang, 2003). This is known as the two-step floating catchment model (or 2SFCM). Let us assume that we have points representing facilities with counts of qualified health staff as attributes. We also have points representing demand, such as population-weighted centroids of census areas, with population counts attached in an attribute field. Given such data, the 2SFCM algorithm works as follows:

• Define a distance or travel time threshold for health facility attendance, e.g. 10km in a setting where people walk to facilities
• For each service / health facility location, undertake the first step of the algorithm:
• Identify the population-weighted centroids within the distance threshold of the facility
• Sum up the populations for these centroids
• Divide the number of physicians by the total population within the threshold distance to get a physician: population ratio
• For each demand location (the population-weighted centroids), undertake the second step of the algorithm:
• Identify the facilities within the distance threshold of the population-weighted centroid
• Sum up the physician: population ratio for each of these facilities (from the first step) and attach this as an attribute to the population weighted centroid.

 
The result is a measure of relative healthcare access that is more sophisticated than simply calculating the distance to the nearest facility. The original 2SFCM method has been extended to handle distances or travel times in a more sophisticated way (Luo and Qi, 2009). Rather than treating these in a binary way (i.e. 0 if beyond the threshold distance or 1 if within it), weights between 0 and 1 are used for distance in each of the two steps. The weights can be derived from a distance decay function, e.g. representing the probability of an individual attending a facility a given distance away.

---

---