Appendix A: Technical note on data analysis
Indicators
The table below provides a breakdown of indicators mapped in the supplementary health atlas and used to develop the local health index referred to in this report. These indicators have been routinely collated by the public health intelligence team to support local decision making and prioritisation.
Table 2: Compendium of health indicators available at ward level
| Health outcomes | Health determinants | Environmental exposures |
| Life expectancy at birth (male, female and statistical difference) | Index of Multiple Deprivation (including health domain) | Air quality: Nitrogen oxides (NOX) concentration |
| Years of life lost by cause | Rural or urban area | Air quality: PM10 particulate concentration |
| People reporting bad/very bad health in 2011 census | Density licensed premises | Air quality: PM2.5 particulate concentration |
| Emergency hospital admissions by cause | Density of fast food outlets | Indicative risk of radon |
| Rate of conceptions (all ages and under 18 years) | Households in fuel poverty | |
| Continuation of breastfeeding at 6-8 weeks | Children in poverty (IDACI) | |
| Proportion of children overweight or obese at reception and year 6 | Pensioners living alone | |
| Incidence of tuberculosis infections | Mothers smoking at time of delivery |
Analysis of health inequalities
The Public Health England inequalities calculation tool was used to generate measures of health inequality included in this report. Relative index of inequality (RII) was selected as a single measure as it gives both a relative measure of inequality within a population and can be used to track change in the level of inequalities over time. RII values included in this report were calculated using the comparison of highest and lowest score method.
Ward level measures of deprivation were derived from the IMD2015 LSOA data using the NHS general practice registered patient database which provides address point level population counts. The analysis was undertaken on request by a GIS specialist at South West and Central Commissioning Support Unit. Ward population estimates used in the calculation of the measures of health inequality were specific to the time period specific to the indicator measure. Other measures of health inequality are available on request, including absolute and relative gaps, slope index of inequality (SII), Gini coefficient, and concentration coefficient.
Derivation of a local health index
The local health index referred to in Section 3 of the report (specifically Figure 4) was calculated using the unweighted mean of the rank value of each health indicator listed in Table 2 above.
As this used indicators which were routinely available, rather than indicators selected on the basis of value of information, some information bias may be inherent to the index. Further development of a local health index is recommended based on an identified list of key indicators and with appropriate weighting for locally agreed values, principles and priorities.
Table 3: Ranking of electoral wards by summary of available health indicator data
| Ward Name | Q1 | Q2 | Q3 | Q4 | Q5 | Mode | Mean |
| Kings Chase | 15 | 1 | 0 | 2 | 2 | 1 | 1.75 |
| Staple Hill | 12 | 3 | 4 | 0 | 1 | 1 | 1.75 |
| Patchway | 11 | 3 | 1 | 2 | 3 | 1 | 2.15 |
| Dodington | 9 | 4 | 3 | 4 | 0 | 1 | 2.10 |
| Filton | 8 | 8 | 2 | 1 | 1 | 1 | 1.95 |
| Yate Central | 8 | 6 | 2 | 2 | 2 | 1 | 2.20 |
| Woodstock | 7 | 4 | 2 | 5 | 2 | 1 | 2.55 |
| Parkwall | 7 | 4 | 1 | 3 | 5 | 1 | 2.75 |
| Charfield | 6 | 2 | 4 | 4 | 4 | 1 | 2.90 |
| Almondsbury | 5 | 4 | 4 | 3 | 4 | 1 | 2.85 |
| Bitton | 4 | 7 | 3 | 4 | 2 | 2 | 2.65 |
| Siston | 4 | 5 | 2 | 4 | 5 | 2&5 | 3.05 |
| Winterbourne | 4 | 3 | 4 | 5 | 4 | 4 | 3.10 |
| Pilning and Severn Beach | 4 | 2 | 4 | 6 | 4 | 4 | 3.20 |
| Yate North | 3 | 8 | 5 | 3 | 1 | 2 | 2.55 |
| Chipping Sodbury | 3 | 3 | 7 | 4 | 3 | 3 | 3.05 |
| Westerleigh | 3 | 4 | 5 | 4 | 4 | 3 | 3.10 |
| Bradley Stoke North | 3 | 1 | 5 | 4 | 7 | 5 | 3.55 |
| Rodway | 2 | 12 | 4 | 1 | 1 | 2 | 2.35 |
| Oldland Common | 2 | 2 | 9 | 6 | 1 | 3 | 3.10 |
| Thornbury South and Alveston | 2 | 3 | 6 | 6 | 3 | 3 | 3.25 |
| Ladden Brook | 2 | 5 | 3 | 4 | 6 | 5 | 3.35 |
| Thornbury North | 2 | 5 | 2 | 6 | 5 | 4 | 3.35 |
| Boyd Valley | 2 | 2 | 6 | 6 | 4 | 3 | 3.40 |
| Cotswold Edge | 2 | 2 | 4 | 6 | 6 | 4 | 3.60 |
| Frenchay and Stoke Park | 2 | 3 | 4 | 2 | 9 | 5 | 3.65 |
| Hanham | 1 | 6 | 8 | 3 | 2 | 3 | 2.95 |
| Stoke Gifford | 1 | 6 | 6 | 7 | 0 | 4 | 2.95 |
| Frampton Cotterell | 1 | 4 | 7 | 7 | 1 | 4 | 3.15 |
| Bradley Stoke Central and Stoke Lodge | 1 | 4 | 5 | 4 | 6 | 5 | 3.50 |
| Downend | 1 | 3 | 4 | 7 | 5 | 4 | 3.60 |
| Bradley Stoke South | 1 | 4 | 1 | 8 | 6 | 4 | 3.70 |
| Emersons Green | 1 | 2 | 5 | 2 | 10 | 5 | 3.90 |
| Severn | 1 | 1 | 2 | 1 | 15 | 5 | 4.40 |
| Longwell Green | 0 | 4 | 6 | 4 | 6 | 3 | 3.60 |
Note that there is a risk of an inherent information bias in the evaluation of health needs between population groups (in this case electoral wards) where only routinely available data can be used. Higher levels of inequality may be present for health issues where indicator data is not routinely available at smaller geographic/population levels.