Department of Health

Website of the Department of Health

Please note that this website has a UK government access keys system.

Synthetic Estimates of Healthy Lifestyle Behaviours

  • Last modified date:
    8 February 2007

Synthetic estimates with 95% Confidence Intervals (1) have  been prepared using 2000-2002 data from the Health Survey for England, the 2001 Census and other information, at the 2003 Census Area Statistics (CAS) ward and Primary Care Organisation (PCO) (2) geographic level for the following variables:

  • Prevalence of current smoking (at the time the data was collected);
  • Obesity of adults;
  • Binge drinking for adults;
  • Consumption of 5 or more portions of fruit and vegetables a day (adults);
  • Consumption of 3 or more portions of fruit and vegetables day (children).

Guidance on Use and Interpretation

These synthetic estimates were produced as part of a research project to test and produce area-level estimates of healthy lifestyle behaviours, which was carried out at the National Centre for Social Research. The estimates were produced in response to the twin requirements to develop small area estimates for Neighbourhood Statistics and to meet local public health information needs.

The synthetic estimates are not estimated counts of the number of people or prevalence of a  behaviour, e.g. smoking in a ward or PCO. They are estimates based on a model and represent the expected prevalence of a behaviour for any ward or PCO, given the demographic and social characteristics of that area. To interpret the estimates it is recommended that users adopt statements such as "given the characteristics of the local population we would expect a prevalence of approximately x% within PCO X".

The synthetic estimates should be considered with their associated Confidence Intervals when making comparisons of one area with another. Although the estimates have large Confidence Intervals (many of which overlap), it is possible to distinguish between those estimates that are significantly different from the national average and those that are not (at both ward and PCO level). The attached charts and User guide show those areas that have estimates significantly above, below or similar to the National estimate.

The estimates can be used to give an indication of the likely prevalence in an area and an indication of where to target resources. They can be used to examine prevalences in the most disadvantaged areas (e.g. the Spearhead PCOs) as compared with other groups.  The estimates should however, not be used for performance monitoring purposes.

Latest publications

The synthetic estimates and associated metadata were published on the Neighbourhood Statistics website , on 29th July 2005, as part of the Health and Care Domain. Further details on the methodology and a user guide are available below:

Background to Synthetic Estimates

The Department of Health/IC commissioned the National Centre for Social Research (Natcen) to produce estimates of healthy lifestyle behaviours, using information from the Health Survey for England. The aim of the project was to develop small area estimates for publication on the Neighbourhood Statistics website (this website is designed for users who need local area information to provide key public health information currently not available from any other source. It is managed by the Office for National Statistics (ONS)

A key requirement of the project was to provide robust estimates generated on a consistent basis for all areas of the country, which allow meaningful comparisons within and between local areas.

Statistical modelling techniques were used to produce the estimates (known as 'synthetic estimates') because the sample size of national surveys (such as the Health survey for England) is too small to provide reliable estimates.

Various stakeholders have been involved in the project to produce these estimates, including key policy and analytical stakeholders within DH and the Information Centre for health and social care, representatives from the Public Health Observatories providing both a user input and technical expertise, and methodology experts from Natcen, ONS and University academics. Further details of the project are available in the Summary and Technical Reports.

Contacts

If you would like further information on Neighbourhood Statistics, please contact the team on 020 7972 6570 or email:

For technical details of the modelling methodology employed to produce the estimates or any other information please contact Kevin Pickering:

Email:

Telephone: 020 7549 9565.

Footnotes

(1) Confidence Intervals have been prepared for each synthetic estimate. The interval reflects the range between which the true value is believed to lie, at a given level of confidence. The confidence intervals therefore represent the uncertainty in the modelling process. At the 95% confidence level, assuming that the model is a good representation of reality, the confidence interval is expected to contain the true value around 95 times out of 100. For example, if a ward estimate of current smoking is 49% and the 95% confidence interval is [32%,67%] we know that 95% of the time the true prevalence estimate for that ward (based on its local population characteristics) will fall within this range.

(2) Primary Care Organisation is the umbrella term used to describe the 300 Primary Care Trusts and 3 Care Trusts in England.

Access keys