Dismantling the signposts to public health? NHS data under the Health and Social Care Act 2012
BMJ 2012; 344 doi: https://doi.org/10.1136/bmj.e2364 (Published 26 April 2012) Cite this as: BMJ 2012;344:e2364
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Michael Soljak has failed to understand how the Health and Social Care Act 2012, in abolishing area-based structures and transferring most health service responsibilities to non-geographically based CCGs, as well as some to local authorities, will limit the availability of routine data to monitor the health service and the population, including inequalities in access to services and outcomes. His narrow focus on well documented problems with the decennial census suggests that he is unaware of the increasing extent to which Neighbourhood Statistics and the Index of Multiple Deprivation (IMD) draw on a range of more timely population-based administrative and survey data, to which we alluded. Even this fuller range of data and readily available look-up tables to convert postcodes to output areas and IMD scores cannot compensate for variation within output areas and the potential for practices to game the system by cherry picking the most healthy and wealthy patients from the more deprived output areas.
Some estimates of list inflation, to which he referred, compare general practice registrations with ONS’ population estimates and show not only that the average list inflation was around 5% but that it varied up to 30% in some PCT areas. And there are other comparators. In Manchester, only 78% of names and addresses held on GP registers could be matched with equivalent records held on the council tax system.
The developments in the use of data from general practice systems which Soljak describes are not new datasets and do not supersede data collections which have been lost or have deteriorated following cuts in ONS and NHS expenditure. It may be that general practices could provide data, some of which are more up to date than the census, but this ignores the fact that these data will be seriously incomplete in respect of the local residents living in an area. Moreover, data will only be recorded when people register and updated when they consult, with data being missing or out of date for those who do not consult. The ethnicity study in Scotland is completely irrelevant as South Asian ethnicity was identified retrospectively from surnames rather than being self-reported. Furthermore, the fact that GP and hospital systems all use classifications based on ICD10, is no guarantee that the ICD codes will be allocated or derived in the same way, as Hendrik Beerstecher points out.
It is essential to be able to continue to measure and monitor populations on a consistent basis, over time and in a way that is comparable for different population groups. The abolition of the NHS and the switch from contiguous administrative areas to a system where nobody has ultimate responsibility for monitoring and meeting the health care needs of all residents and where GP practices are incentivised to dump high risk patients, means that, as in low income countries, those who do not receive care will no longer be counted or measured.
Competing interests: No competing interests
Dear Editor,
Soljak may be placing too much trust on QOF data generated in primary care.
I responded to the article of Dixon et al with an example graph of the local diabetes prevalence (1). Even though our practice was an extreme outlier for diabetes prevalence in 2004 the local Primary Care Trust never questioned or checked our diagnostic criteria which determine our diabetic register.
If you add a non-diabetic to the register two things happen: Firstly your payments for the whole disease area increase as they are prevalence weighted and secondly your achievement of targets immediately improves, saving you from having to try to treat someone with actual diabetes to target.
Local diabetes prevalence was linked to deprivation but this link has nearly disappeared in the 6 year period covered by the QOF data.
The obesity indicator is payable in full the second a single patient is recorded with a BMI over 30. Not every registered patient attends every year and we only take height and weight if there is a clinical indication. The obesity 'register' of our practice is meaningless.
The introduction of Snomed-CT was promised within 2 years when we changed our clinical system supplier in 2003. However the only coding systems used for the QOF data (QOF business rules) are READ2 and CTV3. (2)The ICD 10 system that may be in use in hospitals is not used for data collection in primary care.
Pay formulae are wholly dependent on the database that was used to develop them. The Technical Steering Committee for the GMS pay formula review found huge shifts in funding depending on the database used to determine patient 'need' adding to existing problems with equitable distribution of resources. (3,4)
The concept of active patients as patients that are registered with a GP may not be valid for determining health need. We frequently register patients that have been in the area for years but never registered. They often present with immediate and neglected health needs. List inflation probably provides some amelioration of funding deficits in areas of high population turnover and simply removing this may not resolve inaccurate allocation of area funding.
(1) http://www.bmj.com/content/343/bmj.d6608?tab=responses
(2) http://www.pcc.nhs.uk/business-rules-v22.0
(3) http://www.nhsemployers.org/SiteCollectionDocuments/frg_report_final_cd_...
(4) http://www.biomedcentral.com/1472-6963/10/156
Competing interests: No competing interests
Dismantling the signposts to public health? No, just updating them
Response to: Pollock A, Macfarlane A, Godden S. Dismantling the signposts to public health? NHS data under the Health and Social Care Act 2012. BMJ 2012; 344: e2364 doi: 10.1136/bmj.e2364 (Published 26 April 2012)
In their analysis article on the implications of the Health and Social Care Act 2012 for public health information Pollock et al raise some valid concerns, but also some which are spurious, and others which are inaccurate.1
They compare unfavourably population data derived from general practice registrations with annual population estimates derived from the decennial national Census, and state that practice registrations may be unreliable as denominators because list sizes are often inflated by inclusion of people who have deregistered, moved, or died. However list inflation is decreasing (the England average is 5%),2 and if general practice data is used for public health purposes, only that of active patients is useful anyway. Inter-Censal population estimates are highly dependent on the underlying methods used and are no gold standard. For example, the estimated 2010 resident populations of London Boroughs from two reputable sources, the Office for National Statistics and the Greater London Authority, vary by up to 18%.3
Pollock et al state that it will be more difficult to compile the data needed to monitor health impacts on NHS patients and local populations because denominators will be based on numbers of people registered with general practices rather than numbers of residents in geographical areas. However there is no requirement in the Health and Social Care Act 2012 that registered populations must be used for all epidemiologic analyses. The appropriate population denominator should be used for the type of intervention being studied. The effects of NHS-funded healthcare interventions should be looked for in registered populations, and the effects of resident-based interventions by local authorities or other Government Departments in resident populations. These analyses should include adjustment for population and healthcare confounders respectively.
Converting population data is very easily done using practice to residential area lookup tables. For example, we have converted registered population data on disease prevalence to resident-based populations to incorporate Census-based data,4 and vice versa to adjust for population factors in healthcare evaluations.5 6 As new healthcare and health determinant data sources become available, as is increasingly the case, such translation will be essential to use the data optimally. Estimates suggest that healthcare interventions have contributed around half of recent decreases in cardiovascular mortality.7 8 Ignoring the impact of healthcare on population health is a vestige of the McKeown era.
Pollock et al state that practices collect few socioeconomic data. Practices have never been expected to collect socioeconomic data such as National Statistics Socioeconomic Class, but they should be encouraged to do so to provide more up to date information than decennial Censuses. Ethnicity data is now widely collected by practices at registration and retrospectively (as well as on hospital admission), and is proving to be extremely useful in evaluating health inequalities.9 In some areas ethnicity is recorded in over 70% of the registered population, and ethnicity data collected from healthcare sources can now be used to supplement Census projections.10
Practices can also provide data on the prevalence of smoking, obesity and other risk factors in small populations.11 12 Some public health teams are already making extensive and innovative use of primary care data to shed new light on population health,13 and the new General Practice Extraction Service has the potential to make these data much more accessible.14
The authors’ statement that practice computer systems differ from each other and from hospital episode statistics (HES) in how they code clinical conditions is incorrect. Like HES, Read codes used in practice systems use the International Classification of Diseases version 10, and the transition to the “supercode” SNOMED-CT terminology, now a fundamental NHS Standard, will ensure that all coded primary and secondary healthcare data are fully compatible.
There is an information revolution in progress, but it appears that nostalgia or ignorance of new data sources will prevent Pollock et al from joining it.
References
1. Pollock AM, Macfarlane A, Godden S. Dismantling the signposts to public health? NHS data under the Health and Social Care Act 2012. BMJ 2012;344. Link: http://www.bmj.com/content/344/bmj.e2364 .
2. Primary Care Commissioning. Tackling list inflation, 2012. Link: www.pcc.nhs.uk/uploads/.../pcc_list_inflation_briefing_feb2012.pdf .
3. Greater London Authority. London Datastore. 2011. Link: http://data.london.gov.uk/datastore/package/office-national-statistics-o... .
4. Soljak M, Samarasundera E, Indulkar T, Walford H, Majeed A. Variations in cardiovascular disease under-diagnosis in England: national cross-sectional spatial analysis. BMC Cardiovasc Disorders 2011;11(1):12. Link: http://www.biomedcentral.com/1471-2261/11/12 .
5. Bang JY, Yadegarfar G, Soljak M, Majeed A. Primary care factors associated with cervical screening coverage in England. J Pub Health 2012. Link: http://jpubhealth.oxfordjournals.org/content/early/2012/03/16/pubmed.fds... .
6. Calderón-Larrañaga A, Carney L, Soljak M, Bottle A, Partridge M, Bell D, Abi-Aad G, Aylin P, Majeed A. Association of population and primary healthcare factors with hospital admission rates for chronic obstructive pulmonary disease in England: national cross-sectional study. Thorax 2011;66(3):191-96. Link: http://thorax.bmj.com/content/66/3/191.abstract .
7. Kottke TE, Faith DA, Jordan CO, Pronk NP, Thomas RJ, Capewell S. The Comparative Effectiveness of Heart Disease Prevention and Treatment Strategies. Am J Prev Med 2009;36(1):82-88.e5. Link: http://www.sciencedirect.com/science/article/pii/S0749379708007691 .
8. Ford ES, Ajani UA, Croft JB, Critchley JA, Labarthe DR, Kottke TE, Giles WH, Capewell S. Explaining the Decrease in U.S. Deaths from Coronary Disease, 1980-2000. N Engl J Med 2007;356(23):2388-98. Link: http://content.nejm.org/cgi/content/abstract/356/23/2388
9. Fischbacher CM, Bhopal R, Steiner M, Morris AD, Chalmers J. Is there equity of service delivery and intermediate outcomes in South Asians with type 2 diabetes? Analysis of DARTS database and summary of UK publications. J Publ Health 2009;31(2):239-49. Link: http://jpubhealth.oxfordjournals.org/content/31/2/239.abstract .
10. Hull SA, Rivas C, Bobby J, Boomla K, Robson J. Hospital data may be more accurate than census data in estimating the ethnic composition of general practice populations. Informatics Primary Care 2009;17:67-78. Link: http://www.ingentaconnect.com/content/rmp/ipc/2009/00000017/00000002/art... .
11. Szatkowski L, Lewis S, McNeill A, Huang Y, Coleman T. Can data from primary care medical records be used to monitor national smoking prevalence? J Epidemiol Comm Health 2011. Link: http://jech.bmj.com/content/early/2011/05/13/jech.2010.120154.abstract .
12. Szatkowski L, McNeill A, Lewis S, Coleman T. A comparison of patient recall of smoking cessation advice with advice recorded in electronic medical records. BMC Public Health 2011;11(1):291. Link: http://www.biomedcentral.com/1471-2458/11/291 .
13. Price S. Annual Public Health Report 2011: extending life in Islington. London: NHS North Central London, 2011. http://www.islington.nhs.uk/About-us/annual-public-health-report-2011.htm .
14. NHS Information Centre for Health & Social Care. General Practice Extraction Service (GPES). Leeds: NHS Information Centre for Health & Social Care, 2012. http://www.ic.nhs.uk/gpes .
Competing interests: No competing interests
Re: Dismantling the signposts to public health? NHS data under the Health and Social Care Act 2012
Michael Soljak has failed to understand how the Health and Social Care Act, in abolishing area based structures and transferring most health service responsibilities to non-geographically based CCGs, as well as some to local authorities, will limit the availability of routine data to monitor the health service and the population, including inequalities in access, services, and outcomes. His narrow focus on well documented problems with the decennial census suggests that he is unaware of the increasing extent to which Neighbourhood Statistics and the Index of Multiple Deprivation (IMD) draw on a range of more timely population-based administrative and survey data, to which we alluded.1 Even this fuller range of data and readily available look-up tables to convert postcodes to output areas and IMD scores cannot compensate for variation within output areas and the potential for practices to game the system by cherry picking the most healthy patients from the more deprived output areas.
Some estimates of list inflation, to which he referred, compare general practice registrations with ONS’ population estimates and show not only that the average list inflation was around 5% but it varied up to 30% in some PCT areas2 but the picture is no better when other population-based comparators are used. In Manchester, the PCT has been working with the City Council to match addresses held on GP registers with those in the Local Land and Property Gazetteer which is essentially a composite of the Council Tax and Electoral Registers. Only 78% of names and addresses held on GP registers could be matched with equivalent records held on the Council Tax system.3 Based on this level of accuracy, they estimated that there were about 117,000 patient records where the accuracy of the data was questionable. This has significant financial implications, given that both local authority and NHS funding formulae are based to some degree on population estimates or counts.3
In contrast, the substantial differences of up to 18% between the two sets of population estimates used in London, produced by ONS and the GLA, are well recognised. This situation, which is far from ideal, arises because they use different methods and assumptions.4, 5 For example, ONS uses the International Passenger Survey to factor in the impact of migration, whilst the GLA uses local authority housing capacity plans in its method of forecasting. When the first results of the ONS 2011 census are published in July this year, it will be possible to assess which sets of estimates are closest to the actual new count. However estimates based upon practice registered populations where the incentives are in place for risk selection and cherry picking will see a radical departure from a comprehensive population focus.
Soljak’s account is not of ‘new data sets’. Instead it is a description of developments in the use of current data from general practice systems, mainly for research purposes. These data do not supersede data collections which have been lost or deteriorated following cuts in ONS and NHS expenditure and which Soljak appears to be unaware of. It may be that general practices could provide additional data, some of which are more up to date than the census but this ignores the fact that these data will be seriously incomplete in respect of the local residents living in an area. Moreover, they will only be recorded when people register and updated when they consult, with data being missing or out of date for those who do not consult. The Scottish study he cited is completely irrelevant as it was a register-based study in which South Asian ethnicity was identified retrospectively from surnames rather than being self-reported.6 Furthermore, the fact that GP and hospital systems all use classifications based on ICD10, is no guarantee that the systems will allocate or derive the ICD codes in the same way, as Hendrik Beerstecher points out.7 His picture of the realities of local primary computing shows that there is some way to go before facilities in routine practice catch up with research.
It is essential to be able to continue to measure and monitor populations on a consistent basis, over time and in a way that is comparable for different population groups. The abolition of the NHS and the switch from contiguous administrative areas to a system where nobody has ultimate responsibility for monitoring and meeting the health care needs of all residents and where GP practices are incentivised to dump high risk patients has consequences for data collection. It means that, as in low income countries, those who do not receive care will no longer be counted or measured.8, 9
References
1. Department of Communities and Local Government. The English Indices of Multiple Deprivation 2010. London: Department of Communities and Local Government, 2011 http://www.communities.gov.uk/documents/statistics/pdf/1871208.pdf. Accessed 13 June 2012.
2. Primary Care Commissioning. Tackling list inflation, 2012. http://www.pcc.nhs.uk/tackling-list-inflation.
3. Getstats. Health data under threat, May 2012. Response from Neil Bendel. http://www.getstats.org.uk/2012/05/02/will-nhs-stats-be-less-reliable-in....
4. Greater London Authority. London Datastore. 2011. Link: http://data.london.gov.uk/taxonomy/categories/demographics.
5. Office for National Statistics. Annual mid year population estimates 2010. Newport: ONS, 2011. http://www.ons.gov.uk/ons/rel/pop-estimate/population-estimates-for-uk--.... Accessed June 13 2012.
6. Fischbacher CM, Bhopal R, Steiner M, Morris AD, Chalmers J. Is there equity of service delivery and intermediate outcomes in South Asians with type 2 diabetes? Analysis of DARTS database and summary of UK publications. J Publ Health 2009;31(2):239-49. Link: http://jpubhealth.oxfordjournals.org/content/31/2/239.abstract .
7. Beerstecher H. All that glitters. Rapid Response to Pollock AM, Macfarlane AM, Godden S. Dismantling the signposts to public health? NHS data under the Health and Social Care Act 2012 BMJ 2012;344:e2364.
8. Setel PW, Macfarlane SB, Szreter S et al. A scandal of invisibility: making everyone count by counting everyone. Lancet 370 (9598) 1569-1577, 2007.
9. Guardian. 31st May 2012. GP practice ‘offloaded vulnerable patients to save money’. http://www.guardian.co.uk/society/2012/may/31/gp-health-reform-cost-cut
Alison Macfarlane, City University London A.J.Macfarlane@city.ac.uk
Allyson Pollck, Sylvia Godden, Queen Mary University of London
Competing interests: No competing interests