Some terms used in public health data.

You might also want: Director of Public Health Reports  Health profiles .  Needs assessments


Prevalence in these profiles refers to the proportion of a condition in a population. For instance 50% of GP patients being smokers, is the ‘prevalence’ of smoking in the GP practice population.

What is a directly standardised rate, or ‘DSR’?

Directly standardised rates compensate for the effects of age as a cause for variations in the data. For example, Coronary Heart Disease (CHD) is more likely to occur in an older population, and in an area with many older people you are bound to encounter a higher number of CHD diagnoses. Conversely a part of town with many students and not many elderly people would of course have less CHD.

But what if there was another reason for one area having higher CHD than another? Perhaps there is little access to open space or reasonably priced healthy food? Perhaps smoking rates are very high? In order to ask what else is contributing to variations in CHD rates between areas you need to compensate for age, if there are still significant differences then something else is causing the issue.

So comparing areas for CHD is only a reasonable thing to do if you correct for age differences in the two areas. This is what DSR does, it produces a rate as a result of weighting the age structure (and the CHD data) until it conforms to the European Standard Population profile.

That means a Directly Standardised Rate from Leeds can be compared to another in Bristol for instance, and only the people calculating the DSR need to know details of age structures in their city. DSRs can also be compared for areas of differing sizes, so a DSR for even a very small part of Leeds can be meaningfully compared against the overall Leeds DSR.  DSRs are expressed as a rate per 100,000.

What does ‘confidence interval’ (or ci) mean?

Ci or ‘confidence intervals’ describe the uncertainty around the exact “answer”, unless you have measured everyone, how can you be sure your data accurately reflects real life? There is an upper confidence interval (a high figure), and a lower confidence interval (a smaller number), and somewhere in-between is what would normally be thought of as “the answer”.

The calculation is complex but it can suffice to say that a confidence interval range is the range of numbers within which we know correct answer is most likely to be contained.

Why have confidence intervals?

There rarely is a single exact “answer”. If the range of possible values contained by upper and lower confidence intervals is wide, and you are comparing this to another value which is also surrounded by confidence intervals, how do you know that the two values are even different? They may be the same or very close.

A common approach is to say if the two confidence interval ranges (shown as black lines here) do not overlap then the two areas being compared can be said to be significantly different:

If they overlap though, it is usually said that the two values are not significantly different:

Index of Multiple Deprivation (IMD)

The IMD measures relative levels of deprivation in small areas of England called Lower Super Output Areas (LSOAs). The English Indices of Deprivation are a continuous measure of relative deprivation, therefore there is no definitive point on the scale below which areas are considered to be deprived and above which they are not. IMD scores and ranks have been produced for all LSOAs in England in 2004, 2007, 2010 and 2015.

Deprivation in Leeds is visualised using the IMD scores given to our LSOAs, we have worked with these and population locations to group Leeds into 5 areas which are made of MSOAs:

  • The most deprived fifth
  • The second most deprived
  • The middle fifth
  • The second least deprived
  • The least deprived

The most deprived fifth is a focus area for most public health work and is alternatively referred to as the “Deprived quintile”, remember that this definition is purely the most deprived fifth of Leeds, there is no comparison to England.

LSOA – Lower Super Output Area

These are geographic areas designed nationally to improve the reporting of small area statistics in England. LSOAs when originally generated had between 1,000 and 3,000 people living in them, with an average population of 1,500 people.

MSOA – Middle Super Output Area

These are geographic areas designed nationally to improve the reporting of small area statistics in England and Wales. MSOAs are built from groups of Lower Super Output Areas (LSOAs). The minimum population of an MSOA is 5,000 and the mean is 7,200 (when originally generated). There are 107 MSOAs in Leeds.