Key East of England    Public Health Indicators     Lead person:  Clare Humphreys

Background | Explaining the tartan rug | Updates in the current tartan rug | Future plans | Download tartan rug

Background

The East of England Regional Public Health Quarterly Surveillance is a report commissioned by the Director of Public Health in the East of England, Paul Cosford, to regularly monitor regional public health comparators at PCT level within the East of England and between the East of England and the rest of the country.

There are four key objectives:

• To give regular updates on regional public health comparators
• To provide an overview of Strategic Health Authority and Government Office on the health status of the region. In particular highlighting how PCTs are performing against their Changing Lives: Saving Lives Pledges
• To highlight progress on PCTs meeting their Local Delivery Plan (LDP) targets
• To provide a simple and clear presentation that can be understood by a wide audience.

The East of England Regional Public Health Quarterly Surveillance is displayed as a “tartan rug” (colour coded surveillance chart).

Explaining the tartan rug

The traffic light colour coding

The tartan rug is a popular way to display public health data. Its traffic light colour code visually illustrates whether the public health indicator in question is significantly worse than the average (red), not significantly different than the average (amber) and significantly better than the average (green). In this case the public health indicators are compared with the average in the East of England or in England, and the significance level in most cases is 99.8%.

The tartan rug is displayed in Excel format but can be printed off in three A4 sheets which cover the three sections of the tartan rug.

These are:

• Sheet 1: PCT public health indicators compared with the average for the East of England
• Sheet 2: PCT public health indicators compared with the average for England
• Sheet 3: The Meta-Datasheet – a description of how the public health indicators are calculated, the traffic light coding method, and limitations with the data.

There are 36 public health indicators displayed in the tartan rug. The majority are compared with both the East of England and England, and are based on 99.8% confidence intervals which are calculated using Poisson or Binomial Wilson’s Funnel plots. Two indicators, the inequality ratios for men and women, are not compared with the East of England or England, they compare whether the all cause mortality rate in the least deprived quintile are significantly different from the most deprived quintile. Where the data are not complete or the national average is not available, the tartan rug colour coding can not be completed as the confidence intervals cannot be calculated.

Some of the indicators are directly standardised rates. Most of the directly standardised rates have been age standardised with the European Standard Population as the standard population, with the exception of smoking attributable mortality rate which uses the East of England 2001 Census population as its standard. When comparing directly standardised rates you need to make sure they have been standardised with the same standard population. Most of erpho’s output, such as the Inequality Health Profiles, use the East of England 2001 Census population as the standard population, where as here the European Standard Population has been used (in order that they can be compared with the national average). If you compared the two directly standardised rates for the same disease you will find the latter to be a lot lower than the former because the European Standard Population is much more youthful that the East of England 2001 Census Population.

Local Delivery Plan – LDP targets

Nine of the indicators have LDP targets. If the PCT is not reaching its target for the same time frame it is highlighted by white font. This only indicates whether or not the PCT is reaching the LDP target set.  It is a very crude measure of achieving the target as it does not take into account the confidence intervals around the target.  For example, the indicator could be fractionally below the target and highlighted in white, but in reality it is not significantly different from the target.  If it is not reaching the target, it does not reveal how far away from the target it is.  It should be noted that some LDP targets do not appear to be very consistent.    This is highlighted in the January 2008 version only. 

Limitations with the Data

• Not having data at the right geographical level
  Some data are not available at the appropriate PCT level and therefore look ups have been used to convert the data to PCT level data. The Meta-Datasheet should be referred to, to check where this is the case.

Two indicators in particular should be viewed with particular caution. These are: MRSA Bacteriaemias per 10,000 bed days and number of C.difficile reports in people aged 65 per 10,000 bed days. They have been assigned to a PCT based on the area the acute trust is situated in using table 1, page 5, in the HPA report: Healthcare-Associated Infections for Reporting Period 1st January to 31st March 2007 and Comparisons with Previous Quarters East of England, Health Protection Agency Regional Epidemiology Unit.

• Incomplete data
  For some of the indicators in various PCTs the data are either missing or incomplete. This is the case for childhood obesity, percentage of terminations within ten weeks, breastfeeding initiation and smoking during pregnancy.
  
• Validity of the indicator
  • There are no data available for smoking prevalence at a regional level, therefore a synthetic estimate is used.
  • The obesity in adulthood prevalence indicator is not a true estimate of obesity in the population. This is because it is calculated from QOF (Quality Outcomes Framework) data which measure the percentage of patients aged 16 and over with a BMI greater than or equal to 30 in the previous 15 months out of the practice population aged 16 (which is derived from Prescription Pricing Division GP lists). A large proportion of people on the practice register have not had their height and weight measured and so even if they are obese they do not contribute to the numerator - only to the denominator, and hence the QOF obesity prevalence estimate is likely to grossly underestimate the true prevalence. Unfortunately the number of the practice population with the BMI recorded is not available and therefore it is not possible to come up with a more accurate estimate of obesity at present.
    
• Frequency of updates
  The majority of the indicators are only updated on an annual basis, and there usually is quite a lag time before the data become available. This means that although the surveillance is meant to be quarterly, only a handful of indicators will change quarterly and some of the data refer to the prevalence over two years ago.

Updates in the current tartan rug

The following indicators have been updated from the April 2008 version:

Future plans for the East of England Public Health Quarterly Surveillance

This is currently provisional until it is signed off by Paul Cosford. The data are the most up to date available on 1st July 2008. It will next be updated in October 2008. If you have any further questions or comments please contact Clare Humphreys (email: clare.humphreys@rdd-phru.cam.ac.uk phone: 01223 330363).