2            Methods

2.1    The English Hospital Episodes Statistics (HES) database

The HES database at Bristol University holds information on patients who attend hospitals in England, either as a day case or as an in-patient. From 1991 to 1999 there were approximately 97 million records on the database. Each record in the database relates to one 'Finished Consultant Episode' (FCE). This is the period of time an individual spends under the care of one NHS consultant. The information held includes the age and sex of the patient, area of usual residence, reason for admittance to hospital and procedure to be performed. Information is available for financial years from 1991 to 1999, meaning that the data for 1999 include April 1999 to March 2000; updates tend to be about one year out of date. Further information is available from the Department of Health at the website:  http://www.doh.gov.uk/hes/

Diagnoses were coded to ICD 9 (1991–4) and ICD 10 (1995 onwards) and procedures were coded to OPCS4. Perfect matching between codes is not always feasible, although for surgical procedures it appears to be good. The codes used in the following report are described in Appendix 1.

Private hospital procedures are excluded from HES as there is no requirement for these hospitals to provide routine data. It has been estimated that about 20% of CABGs and 10% of PTCAs are funded privately (South West Consortium of Health Authorities, 2001).

2.2  Ward level area deprivation

The variation in hospital utilisation by individuals in different socio-economic circumstances was explored using area-based data on material deprivation at the time of the 1991 census. Wards in the South West region were ranked on the basis of Townsend score (Townsend et al., 1988) and then grouped into quintiles. The Townsend score is a composite score based on the following four items:

•     Unemployment – unemployed residents over 16 as a percentage of all economically active residents aged over 16.

•     Overcrowding – households with 1 person per room and over as a percentage of all households.

•     Non car ownership – households with no car as a percentage of all households.

•     Non home ownership – households not owning their own home as a percentage of all households.

The index uses Z scores to standardise the component variables. The Z score is the 'observation' (percentage or proportion for the ward on a given measure) minus the mean observation divided by the standard deviation (for England and Wales). Scores used in these analyses were obtained from the Census Dissemination Unit, based at the University of Manchester (Census, 2001a). Figure 3 presents deprivation quintiles for wards in the South West region.

2.3    Urban/rural residence

For the basis of initial analyses the ONS ward classification was used to identify rural areas (Wallace and Denham, 1996). This classification was derived from 1991 census data and it groups areas into clusters by measuring similarities on a whole range of variables, synthesizing these dimensions into a single classificatory structure. Areas classified under this scheme as rural or rural fringe (approximately 32% of all wards in the South West region) were grouped together as 'rural', all other areas were classified as 'urban' (approximately 68%). Data used in these analyses were obtained from the Census Dissemination Unit (Census, 2001b). Figure 4 presents rurality in wards in the South West region.

2.4  Ward boundaries and populations

The majority of analyses presented involve rates, such as admission or operation rates. Calculating such rates requires estimates of populations over time, and a process to adjust for differences between the age and sex distribution of populations when making comparisons either between places, or over time. These estimates were made using extrapolations from the 1991 census for under-enumeration and by re-constituting regional boundaries so that consistent geographic areas were compared over the 1991–99 period. All rates were directly standardised for age using the 1998 population as the standard.

2.5 Assessing equity of provision of care

A major priority of the NHS is to provide health care equitably – avoiding what has been called a 'postcode lottery' whereby the chances of receiving care are related to where you happen to live. Describing equity requires some idea of need, as the NHS is expected to provide equal care on the basis of equal need. Ideally, need for revascularisation would be defined by counting the number of people with coronary heart disease who would benefit from revascularisation. However, no routine recording of morbidity CHD is conducted in the South West region so it is necessary to assess the need for revascularisation using proxy measures. In this report acute myocardial infarction admission rates and unstable angina admission rates have been used in comparisons with revascularisation rates to make an approximate adjustment for need. Acute myocardial infarction admission rates have been shown to correlate with prevalence of angina and with mortality (Payne and Saul, 1997). More recent data for 1999, presented in Figure 5, show the strong correlation of acute myocardial infarction rates and ischaemic heart disease standardised mortality ratios at district level. Acute myocardial infarction admission rates therefore provide a reasonable means of gauging the variation in need for health care between different groups. Unstable angina rates also show a significant correlation with standardised mortality ratios at the district level (figure not shown) and are used in assessing need by age-group and by deprivation where they appear to be more conservative indicators of need as rates decline, rather than increase, at older ages.

When relating need to service use, the ratio of acute myocardial infarction or unstable angina admission rates to revascularisation rates has been used. If this ratio shows a similar value across comparison groups (e.g. deprivation, age groups) it can be inferred that there is relative equity of provision to need.

Need might be assessed by indicators more closely related to demand for care such as waiting lists, but these are affected by local circumstances of supply and organisation and do not reflect the population need, as defined by burden of disease.

The Public Health Observatory is part of the South West Observatory, a wider Regional intelligence function, currently supported by the South West Regional Assembly, the Department of Health, the Department for Education and Skills. Government Office South West, the South West of England Regional Development Agency and the Environment Agency.