Scottish Government

Fair Shares for All

ANNEX F: THE 'ARBUTHNOTT INDEX'

F.1 In the further analytical work that has been carried out following consultation on Fair Shares for All a large number of indicators was used to analyse the relationship between morbidity and life circumstances and the use made of different health services. Four indicators were found to be significantly more successful than other indicators in explaining the differences observed in the use of services between postcode areas. These indicators are:

• The standardised mortality rate among people aged 0-64 averaged over a 5 year period.

• The proportion of households having 2 or more deprivation indicators.

• The proportion of the population of working age claiming unemployment benefit.

• The proportion of the population aged 65 and over claiming income support.

The indicators based on mortality rates, unemployment and income support are annually updateable, while the indicator of deprived households is taken from the 1991 census.

F.2 These four indicators were combined into a single index (the 'Arbuthnott index') using the method of z scores.

• The values of each indicator are available for every postcode sector in Scotland.

• The mean and standard deviation of each indicator were also calculated.

• The values of the indicators were standardised by taking the difference between the actual value and the mean and dividing by the standard deviation.

• The standardised values for the four indicators are then added together to provide an overall z score.

• This z score has then been used as the 'Arbuthnott index' of the key aspects of morbidity and life circumstances that influence the relative need for healthcare.

F.3 The value of this index ranges from -6.45 (the area with the lowest levels of morbidity and deprivation) to +16.06 (the area with the highest levels of morbidity and deprivation). The distribution of postcode sectors using this index is shown in Figure 3.5 in Chapter 3 of this report. The distribution is skewed with a small proportion of sectors having levels of morbidity and deprivation that are well above the national average.

F.4 The method used to arrive at the overall index effectively gives each of the four indicators equal weight. The sensitivity of the index to alternative weighting systems was examined, and the results are summarised in Table F.1. The numbers attached to the term 'index' in this table indicate the relative weighting applied to the four indicators (in the order shown at paragraph 1 above). For example, the term 'index2224' indicates that the first three indicators were given an equal weight of 20% each, and the fourth indicator was given a weight of 40%.

Table F.1: Correlation Coefficients Between Indices Based on Alternative Weighting Patterns

The alternative indices based on different weighting patterns are very highly correlated with each other.

F.5 The 'Arbuthnott index' is also highly correlated with the Carstairs index though the four indicators used in Arbuthnott are quite different from those used in Carstairs. (The Carstairs index uses male unemployment, car ownership, overcrowded households, and social class.)

ANNEX G: GLOSSARY OF ACRONYMS

Abbreviations used for Health Boards