Key Data Sources for Public Health - Local Perspective - Irina Holland
0 likes602 views
National data has advantages like comparability between areas and standard indicators/targets, but local data provides more granular, timely, and useful information for public health. Local data sources in Somerset include population statistics, lifestyle surveys, screening and vaccination rates, deprivation indices, and disease registers from general practices. Examples shown how local cervical cancer and childhood vaccination data can identify poorly performing areas for intervention, and how smoking and deprivation data were used to target communities for smoking cessation programs.
Key Data Sources for Public Health - Local Perspective - Irina Holland
- 1. Key Data Sources for Public Health – the local perspective Public Health Information and Evidence Seminar Irina Holland NHS Somerset 4th November 2010
- 2. 2 Outline • Advantages and Disadvantages of national compared to local data • Local data sources • Examples using local data
- 3. 3 Advantages of national/regional data compared to local data • Comparability with other PCTs/LAs/… • Used for government indicators and targets • Can be used in conjunction with local data (eg adjusting populations to get estimates for small areas) • Other people do the work!
- 4. 4 Disadvantages of national/regional data compared to local data • Timeliness (often very old) • Granularity (often data not available at small geographies) • Geographies provided not always the ones required • Suppression • Availability • Not always as useful as it could be (eg not standardised, no confidence intervals) • No control of how the data is analysed (definitions, methods) • No local knowledge used in analysis and interpretation • Modelled data
- 6. 6 Public Health Data - 1 • Populations (including Births, and ethnicity) • Deprivation • Lifestyle • GP consultations • Prevalence • Screening • Cancer registrations • Immunisations • Hospital activity (inpatients, outpatients, A&E, referrals) • Ambulance statistics • Mortality
- 7. 7 Public Health Data - 2 • Teenage pregnancy • Crime and Community Safety (including Drugs and Alcohol) • Social Services • Benefits • Education • Emergency service activity • Environment • Housing • Income/Employment • Transport • …
- 8. 8 PCT Partners • Local Authorities (including Unitary Authorities) and all the associated sections (eg Children and Young People, Transport, Housing, Schools, Economic, Social Services) • Police • Sexual health services • Emergency services • Environment agency • Social Services • Voluntary Sector • Health and Wellbeing Partnership (HSNAG/JSNA) • Community Safety Partnership (inc ASPIRE) • Drugs and Alcohol Action Team • Somerset Intelligence Partnership • SINe •
- 9. 9 Examples using local data
- 10. 10 Cervical Screening • Routinely collected and reported on KC53 10 working days after quarter end at practice level, data only reported nationally at PCT level • Rates can be calculated at Federation (group of practices) level • Monitoring shows that – Some practices have higher inadequacy rates than others (interventions can be made at these practices) – the percentage of smears that are inadequate smears has fallen over time (partly as a result of improvement at poorly performing practices and partly with the introduction of the new technique)
- 12. 12 Vaccinations • Routinely collected and reported on COVER (Cover of Vaccination Evaluated Rapidly). Provisional data 10 working days after quarter end at treatment centre level, finalised data 2 months after quarter end. At least until recently data only reported nationally at PCT level • Rates can be calculated at Federation (group of practices) level • Monitoring shows that – Decline in MMR uptake following concerns over its safety. But if less than 95% of children are immunised against measles outbreaks will occur (concept of herd immunity) – Some areas are better at achieving high uptakes than others
- 13. 13 Vaccinations
- 14. 14 Deprivation • Indicators describe one or more aspects of deprivation (material, social, educational…) • IMD, Jarman, Carstairs, local indicator eg Somerset Health and Social Needs Analysis Group Child deprivation measure • IMD only published at Lower Super Output Area level • Can use local data and population weighting of LSOA scores to estimate deprivation for other areas and groups of people (eg schools, GP practices, wards) resulting in ability to identify inequalities and target resources.
- 15. 15 Lifestyle - Smoking • Collected through MIQUEST, a system that interrogates GP clinical systems, for National Service Framework monitoring • Related to deprivation • Smoking prevalence can be used to target communities with smoking cessation work. Either directly if area known to have a high prevalence or indirectly if area known to be deprived.
- 16. 16 Deprivation and Smoking prevalence
- 17. 17 Deprivation and Smoking prevalence
- 18. 18 Quality Outcomes Framework (QOF) • Is the annual reward and incentive programme detailing GP practice achievement results and contains Indicators relating to the care of patients • National data Includes prevalence (=number on various disease registers per list size) but no allowance made for the age/sex structure of the list • Not designed as a public health tool but can be adapted by standardising using local data • Can group practices into deprivation quintiles to look for inequalities
- 20. 20 Lifestyle • Some Somerset information available through national data based on national surveys – Most use modelling to provide estimates – Some use definitions that are not ideal for public health purposes (eg sport but not all physical activity) • Every 5 years or so Somerset public health surveys adults about lifestyle behaviours • Data can be compared over time appropriately by re- weighting data from previous surveys to reflect similar population distributions
- 21. 21 Lifestyle
- 22. 22 Equity of revascularisation • Coronary heart disease mortality by gender for those aged 35-64 • Revascularisaion procedures by gender for those aged 35-64 • The ratio gives an estimate of the number of procedures per death • Highlighted the gender inequality in provision of intervention – CHD had been regarded as a “male” disease and this might have led to under-referral by GPs
Editor's Notes
- #11: Timeliness, Availability, Granularity, Geographies, Local Knowledge
- #12: Timeliness, Availability, Granularity, Geographies, Local Knowledge Practice B is within Federation A, so although Federation is pretty much the same as the County average not all practices are achieving the same level – perhaps others could learn from best practice.
- #13: Timeliness, Availability, Granularity, Geographies, Local Knowledge Important to achieve high levels of immunisation to have herd immunity.
- #14: Timeliness, Availability, Granularity, Geographies, Local Knowledge Practice D is within Federation C
- #16: Example – also showing getting data from GP systems – mention MIQUEST
- #17: Using MIQUEST to get numbers with smoking status recorded and whether or not they are smokers for those aged 16+. Shown indicator is current smokers / number with smoking status known 3.5 fold difference between highest and lowest practice
- #18: Obvious gradient with more smoking in more deprived areas
- #19: Timeliness, Availability, Granularity, Geographies, Local Knowledge Also: For most of the conditions the “true” prevalence has been modelled based on scientific research and the population structures in practices, these are available on the NHS Comparators website. This is similar but isn’t quite the same as standardising within Somerset (which shows differences in levels that can’t be explained by differences in age and sex structure). The NHS Comparators work shows how complete diagnosis plus recording is (based on the assumptions in the model). Calculation of practice deprivation done locally using previous method, not shown here.
- #20: This is the result of standardising the QOF data for Dementia (for a selection of practices for clarity). The red line is the crude prevalence The blue columns are the standardised prevalence. The crude rate can be changed quite considerably, both toward the average (eg the first practice) and away from the average (eg the second practice). There is a 8 fold difference between the highest and lowest rates, which probably reflects diagnosis and coding issues as well as the true level of disease in the practice.
- #21: Availability, modelling, control, methods, geographies,… Can mention the report at …
- #22: Availability, modelling, control, methods, geographies,… Can mention the report at …
- #23: Availability, control, methods, geographies,…
- #24: Availability, control, methods, geographies,…