Table 2 Summary of decisions required regarding important characteristics of geographical classifications

Be clear on specific objectives and purpose of the classification as this determines what is being measured

Is it remoteness, isolation, access, disadvantage, rurality or something else? If it is an access classification, then what aspect of access is being measured, and in relation to what service - (e.g. GPs as a measure of primary care)

The choice of algorithm or procedure for grouping similar clusters matters

Accessibility can be measured by distance to nearest service, service provider to population ratios, or increasingly sophisticated methods such as floating catchments and distance-decay

The criteria and cut-off points underpinning groups matters

How many groups do you want? At what point do you differentiate between groups? (e.g. Is the decision based on minimising within-group and maximising between- group variance, or is the number arbitrarily defined by convenience for the end-user?)

The choice of spatial units matters

RRMA is often criticised for its use of Statistical Local Areas (which can be large in rural areas), but the more extreme use of 1 km grids such as ASGC-RA is typically not an option for most data required.