We are currently in the process of updating this chapter and we appreciate your patience whilst this is being completed.
Small-area analysis (SAA) permits the examination of data for groups, such as towns, which tend to be more homogenous in character compared with larger populations that are likely to be more diverse.
Uses of small area analysis
The methodology is often used in health services research to calculate the utilisation rate for a service in several geographic areas, such as counties, evaluate descriptive statistics, note differences in utilisation between counties, and attempt to seek explanations for these variations, such as service availability and doctors’ referral practice.1 Identification of areas of high utilisation help localise need and assist in the allocation of resources.
Small-area analysis provides the numerical data that allows researchers to consequently explore why, for example, utilisation rates of a service are much higher in one area than another. As well as greater need (for example higher levels of illness), other possible reasons for higher utilisation might include differences in illness behaviour, for example seeking medical help, diagnostic decisions of physicians or treatment decisions by physicians, as well as a greater availability of resources.
Other uses for small area data include:
- Increasing knowledge – for example, understanding socioeconomic variations across geographic areas
- Informing public policy
- Supporting decision-making – for example, the need for additional clinical services
Defining small areas
There are no hard-and-fast rules for defining small areas, as Smith2 points out, and they may range from less than an acre to thousands of square miles, and from no inhabitants to many millions. However, a small area would typically be a region for which data such as healthcare usage are readily available, for example a county, city or postcode area.2
The General Household Survey reported data at the level of Standard Regions — nine in England plus Wales and Scotland. However, the UK Office for National Statistics (ONS) often publishes population estimates by age and sex at local authority and ward level.
Super Output Areas
Small-area statistics are published by the ONS by Super Output Areas (SOAs). SOAs were developed following the 2001 UK Census to report Neighbourhood Statistics. SOAs are aggregates of adjacent output areas (OAs), which each have similar social characteristics. There are two categories of SOA, each of a different size. This permits a choice of scale for the publication of data, minimising the risk that local data that could be disclosive. For example, in a small area there may only be a handful of cases of a particular condition, which could subsequently be identifiable.
SOAs provide a basis for comparison across the country because the units are all a similar size. They are stable over time. The ONS has defined 34,378 ‘Lower Layer’ SOAs (LSOAs, typically containing 4 to 6 OAs, with a population of around 1500) in England and Wales and 7,193 larger ‘Middle Layer’ SOAs (MSOAs, built from groups of LSOAs, with an average population of 7200). There are no “Upper Layer” SOAs.
Sources of data
Carstairs provides a helpful introduction to SAA, and describes several studies that cover a wide range of health events to illustrate how the methodology can be used. These include comparisons of:3
- Perinatal mortality
- A&E attendances
- Cancer deaths
- Attempted suicides
- Vaccination uptake
SAA typically uses routine data sources (see Section 3, Health Information) and examples include:2
- Censuses and population estimations
- Administrative records such as birth and death notifications
- Hospital Episode Statistics and other healthcare utilisation data
SAA can be performed using three basic steps:4
- Identifying and defining the geographic boundaries of the areas of interest
- Estimating the amount of resources allocated to the population of each area, for example the number of hospital beds or GPs
- Calculating utilisation rates - these may be calculated for each area on a crude and age-adjusted basis, usually using the indirect method of standardisation. Rates represent events, not persons, as patients receiving the same service more than once are counted each time.
Two main techniques of analysis are typically used:
- Direct comparisons of rates between the areas of interest, to identify areas of high need
- Correlation analyses to establish general relationships between health indicators and social and economic characteristics
Most correlation studies tend to find that populations who are socially deprived also suffer greater health disadvantages.
Problems associated with small area analysis2,3
- Because the areas contain relatively small numbers of individuals, the observed rate may differ from the expected due to chance
- Difficulty allocating events to an area – the increasing use of patient postcode in administrative data means that this is becoming easier
- Absence of appropriate denominator data, for example, population size
- Populations may differ in their structure and size, making direct comparisons difficult. Standardisation techniques may facilitate this.
- Small populations naturally attract small numbers of events and this can pose problems in producing sufficient numbers for analysis, particularly where the phenomenon exhibits low incidence in the population
- The geographical boundaries of many small areas do not remain constant over time, which undermines the consistency of historical data series. Similarly the case definitions used in healthcare records may also change over time, or vary by locality.
- Routine sources of data may not be available to answer the question
- Errors are a problem in even the best censuses and administrative record systems. Their impact on data reliability is typically greater for small areas than large areas.
- Survey data are generally less reliable for small areas than large areas because sample sizes are smaller and survey responses more variable
- Particular events are more likely to disrupt orderly trends for small areas than large areas. For example, the construction of a large housing development, or the loss of a major employer, may sharply distort trends in a small area within a city, county, or state.
- In a small area the number of cases of a disease may be very small. This may create a risk that cases could be identified, potentially threatening patient confidentiality.
- Diehr P, Cain K, Kreuzer W, Rosenkranz S. Can small area analysis detect variation in surgery rates? Medical Care 1992; 30(6): 484-502
- Smith S. ‘Small-Area Analysis’ in Demeny P and McNicoll G (eds.). Encyclopedia of Population. Macmillan Reference, 2003. https://www.bebr.ufl.edu/sites/default/files/Research%20Reports/Pop_Encycl_Small_Areas_0.pdf - accessed 12/01/17)
- Carstairs V. Small area analysis and health service research. Community Medicine 1981; 3: 131-139
- Small-Area Analysis. http://mchp-appserv.cpe.umanitoba.ca/viewConcept.php?conceptID=1228 – accessed 12/01/17
© Helen Barratt 2009, Saran Shantikumar 2018