Ecological Studies: Use with Caution

editorials ecological studies: financial responsibility for patients. The first the need for reform address for correspondence use with caution example is in the UK where groups of GPs Primary care in many countries is martin roland (clinical commissioning groups) now have unrecognisable from 20 years ago. Institute of Public Health, Forvie Site, University of budgetary responsibility for the majority Countries are bound to continue to reform Cambridge School of Clinical Medicine, Box 113 of the healthcare budget to their patients their healthcare systems to deal with the Cambridge Biomedical Campus, Cambridge, including hospital and specialist care. The new challenges of ageing populations, and CB2 0SR, UK Background second example is in the US where the therefore changes to both the organisation e-mail: [email protected] Ongoing debates about the quality of NHS concept of ‘accountable care organisations’ and financing of primary care are inevitable. organisations have made ecological studies “Ecological studies in health services research are gives budgetary responsibility for defined GPs have proved to be both adaptable and references fashionable. One such study in the UK a powerful tool and with the wealth of organisational populations to providers of health care, entrepreneurial over many years. They will 1. Kringos D, Boerma W, Bourgueil Y,et al. considered the association between the The strength of primary care in Europe: an level data now available there are increasing numbers although in the US this is more likely to be need to show continued ability to adapt to a international comparative study. Br J Gen Pract average clinical quality of primary care a combination of generalists and specialists changing environment. 2013; DOI: 10.3399/bjgp13X674422. provided by primary care trusts and the of research questions where they are the study rather than primary care physicians having 2. Schoen C, Osborn R, Squires D,et al. A survey trusts’ rate of admission for coronary heart design of choice.” lead responsibility. martin roland, of primary care doctors in ten countries shows disease.1 An alternative to this ecological An anomaly of the UK’s current RAND Professor of Health Services Research, progress in use of health information technology, approach, which used data aggregated to Institute of Public Health, University of Cambridge less in other areas. Health Affairs 2012; 31(12): healthcare reform is that the CCGs are School of Clinical Medicine, Cambridge. 2805–2816. the level of a primary care trust, would responsible for purchasing specialist care have been to have used data for individual public3 the amount of data available is only epidemiological research also apply to 3. Kringos D, Boerma W, van der Zee J, but do not have responsibility for primary patients, and to have asked if there was an going to increase. ecological studies, for example in relation ellen nolte, Groenewegen P. Europe’s strong primary care care. It seems clear that they cannot systems are linked to better population health One strong advantage of using publicly to potential confounding (where two Director, Health and Healthcare, RAND Europe, association between the clinical quality of but also to higher health spending.Health Affairs manage a population budget without taking Cambridge. care an individual receives and their own available data in research is that there are domains of care appear associated, but an interest in both, and early indications 2013; 32(4): 686–694. chance of being hospitalised for coronary no problems with data confidentiality. When this is in fact simply because they are are that they are doing exactly that with, for 4. Starfield B, Shi L, Macinko J. Contribution of heart disease. By using aggregated data looking at associations between risk factors both associated with a third, confounding, provenance primary care to health systems and health. example, contracts that involve specialist, Commissioned; not externally peer reviewed. Milbank Q 2005; 83(3): 457–502. in ecological studies the relationship for and outcomes, linking information about variable). Where there is confounding by 14 primary, and community care. Currently 5. Friedberg MW, Hussey P, Schneider E. Primary individual patients is not directly explored, individual patients often requires extensive individual level variables (such as clinical neither hospital nor general practice competing interests care: a critical review of the evidence on quality although individual relationships may often ethical and governance approval. Linking diagnosis or disease severity case mix payment systems in the UK encourage The authors have declared no competing interests. and costs of health care.Health Affairs 2010; (correctly or incorrectly) be inferred from data at the organisational level, however, or sociodemographic variation), then if integrated care, and both need to change 29(5): 766–772. population-based analyses. Ecological does not, as usually the data have been individual level data are available for one to do so. DOI: 10.3399/bjgp14X676960 6. Vedsted P, Olesen F. Are the serious problems studies can either be descriptive, for published and are in the public domain of the measures of interest, accounting in cancer survival partly rooted in gatekeeper already. Ecological studies also allow us to for potential confounding at the ecological principles? An ecologic study. Br J Gen Pract example, exploring variation between 5 2011; DOI: 10.3399/bjgp11X588484. populations, or consider associations such look nationwide providing evidence that is level is possible. Information may also 7. Nolte E, McKee M. Variations in amenable as the example above. In health services potentially more generalisable than from be available at the organisational level mortality — trends in 16 high-income nations. research, where healthcare organisations studies considering individuals, but in only a about possible confounders, but individual Health Policy 2011; 103: 47–52. rather than individual patients are often small geographical area. A further strength level data are needed if individual level 8. Powell Davies PG, Williams AW, Larsen K,et al. the focus of inquiry, ecological studies are of ecological studies is that where data are confounders are the concern. Coordinating primary health care: an analysis of often an appealing tool. For example, a available the exploration of potential trends The importance of these last two the outcomes of a systematic review.Med J Aust 2008; 188(8 suppl): S65–S68. recent study looking at the features of over time can be considered with relative points (potential for ecological fallacy 9. Coleman K, Mattke S, Perrault P, Wagner E. general practices associated with lower ease. and unmeasured confounding), when Untangling practice redesign from disease coronary heart disease mortality was The strength of an ecological study for interpreting correlations observed at the management: how do we best care for the more concerned with the practice at an looking at associations at the institutional organisational level, cannot be overstated. chronically ill. Ann Rev Public Health 2009; 30: organisational level than with individuals.2 level (hospital, CCG or GP practice) is A simple, unadjusted, correlation of two 385–408. balanced by the fact that we cannot draw measures at the population level has the 10. Smith J, Holder H, Edwards N,et al. Securing strengths of this approach: the future of general practice: new models of conclusions about individual patients from potential for eye-catching headlines, such primary care. www.nuffieldtrust.org.uk/sites/ open data and organisation population data. We can tell whether GP as the study of the association between files/nuffield/130718_securing_the_future_of_ analyses practices where patients report a better chocolate consumption and winning a general_practice-_full_report_0.pdf (accessed 13 The availability of data describing NHS experience of care are those that achieve Nobel prize.6 However, the potential for Jan 2014). organisations has never been greater. A higher QOF targets,4 but this tells us ecological studies to lead into suboptimal 11. Casalino L, Wu F, Ryan A,et al. Independent practice associations and physician-hospital large volume of UK healthcare process and nothing about the association between policy-making is high; confounding and organisations can improve care management outcome data is becoming publicly available patient experience and the quality of clinical ecological fallacy mean that an unthinking for smaller practices. Health Affairs 2013; 32: from the Health and Social Care Information care at the patient level. Ecological studies analysis of associations at the organisational 1367–1382. Centre (http://www.hscic.gov.uk/) and via are at best hypothesis generating when rather than the individual level may have 12. Charlesworth A, Davies A, Dixon J.Reforming the government open data website (http:// considering individual level associations and far reaching consequences. Recently it payment for health care in Europe to achieve care is needed to avoid the risk of ecological better value. www.nuffieldtrust.org.uk/ www.data.gov.uk/) with GP practice, has been claimed that NHS hospitals that publications/reforming-payment-health-care hospital, and clinical commissioning group fallacy: assuming the associations that operate in a more competitive geographical (accessed 13 Jan 2014). (CCG) indicators available. Indicators cover exist at the population level persist at the environment have a

Load more