Rapid Responses to:

RESEARCH: Mark Ashworth, Jibby Medina, and Myfanwy Morgan Effect of social deprivation on blood pressure monitoring and control in England: a survey of data from the quality and outcomes framework BMJ 2008; 337: a2030 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

prevalence of prehypertension needs to be better characterised, and treatment goals better defined

oscar,m jolobe   (31 October 2008)

trends

Nigel Konzon   (21 November 2008)

Important inequalities still remain

Nick Payne   (25 November 2008)

Calculation of deprivation scores

Gary McLean, Catherine O'Donnell (University of Glasgow) on behalf of the GMS Impact Study team   (25 November 2008)

Is primary care equitable? A need for an accurate measure of socio-economic deprivation.

Mark Strong   (28 November 2008)

Interpreting patterns of changes in absolute differences between rates when common outcomes become even more common

James P. Scanlan   (7 December 2008)

prevalence of prehypertension needs to be better characterised, and treatment goals better defined 31 October 2008
oscar,m jolobe, retired geriatrician manchester medical society, c/o john rylands university libarary, oxford road, manchester M13 9PP Send response to journal: Re: prevalence of prehypertension needs to be better characterised, and treatment goals better defined	According to the authors, the most recent detailed national population survey for England conducted in 2003 revealed that the mean blood pressure(BP) in 8834 adults aged 16 or more was 131/75 mm Hg(1), a result which is well within the prehypertensive range of systolic BP 120- 139 and/or diastolic BP 80-90 mm Hg(2). Although the subsequent quality and outcomes framework specified a requirement to document BP only subjects aged 45 or more(1), it might still be worthwhile to put on record whether or not prehypertension was as prevalent in the English study age group 20-44 as it was in a comparable age group in the United States National Health and Nutrition Examination Survey for 1999-2000. In the latter survey prehypertension had a prevalence of 30.3% in the age group 20-29, and 34.1% in the age group 30-39(3). Prehypertension is worth documenting with a high degree of precision because it has an adverse long -term prognosis, being associated with an elevated risk of cardiovascular disease in subsequent years(4).What we also need to recognise is that the definition of adequate blood pressure control(BP equal to or less than 140/90 mm Hg) given by the authors(1)does not apply to diabetics, in whom tight control of BP to a target level below 130/80 mm Hg confers the benefit of relative risk reductions amounting to 32% and 44% for diabetwes related death and for stroke, respectively, in comparison with counterparts with diabetes who are on less tight BP control(5). These benefits were subsequently eroded when BP control became less tight following the completion of the United Kingdom Prospective Diabetes Study(5). However, given the fact that prehypertension has an adverse long -term prognosis irrespective of whether or not the patient has diabetes(4), the ultimated goal BP for diabetics as well as for non- diabetics ought to be optimum blood pressure, namely, BP < 120/80 mm Hg(2). That should be the future definition of adequate BP control. References (1) Ashworth M., Medina J., Morgan M Effect of social deprivation on blood pressure monitoring and control in England: a survey of data from the quality and outcomes framework British Medical Journal 2008:337:a2030 (2) Chobanian AV., Bakris GL., Black HR et al The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. The JNC 7 Report Journal of the American Medical Association 2003:289:2560-72 (3)Qureshi AI., Suri MF., Kirmani JF., Divani AA Prevalence and trends of prehypertension and hypertension in United States: National Health and Nutrition Examination Surveys 1976-2000 Medical Science Monitor 2005:11:C£403-409 (4) Kshirgar AV., Carpenter M., Bang H., Wyatt SB., Colindres RE Blood pressure usually considered normal is associated with elevated risk of cardiovascular disease American Journal of Medicine 2006:119:133-41 (5)Holman RR., Paul S., Bethel A., Neil HAE., Matthews DR Long-term follow-up after tight control of blood pressure in Type 2 Diabetes New England Journal of Medicine 2008:359:1565-76 Competing interests: None declared
trends 21 November 2008
Nigel Konzon, general medical practitioner sw97se Send response to journal: Re: trends	This excellent article reinforces our determination in inner city areas to provide high quality care, and makes a significant contribution to research on the differences between UK populations. However attribution of the changes recorded do not take account of trends occuring prior to the research. It would be useful to know when deprived communities started to narrow the gap in BP recording and control, was it really during the advent of QUOF or had this trend been taking place many years before QUOF was introduced? This study may reflect changes that were already happpening prior to QUOF, may show an acceleration of an existing trend, or may correlate directly with QUOF. I would value the authors reply, and the need for any further research. This article may well have an effect on discussions on the role of performance management in primary health care, and clarity of the real attribution of effect will be important. Competing interests: None declared
Important inequalities still remain 25 November 2008
Nick Payne, Senior Lecturer in Public Health School of Health and Related Research, Sheffield University , Sheffield S1 4DA Send response to journal: Re: Important inequalities still remain	Ashworth and colleagues record a useful increase in blood pressure monitoring and control in deprived populations. However, what is disappointing is that the recorded prevalence of the chronic conditions is so similar between the most and least deprived fifths of communities. Both results from large population surveys, and from mortality and hospital admission data, suggest that for the conditions examined, especially coronary heart disease (CHD), this is not the case. Indeed, prevalence of CHD may be as much as 2-3 fold higher in the more deprived population fifth. As this implies inequalities in detection or recording of important chronic conditions, it suggests that the reduction in health inequalities may not be as hopeful as might at first appear. Competing interests: None declared
Calculation of deprivation scores 25 November 2008
Gary McLean, Research Fellow University Of Glasgow, Catherine O'Donnell (University of Glasgow) on behalf of the GMS Impact Study team Send response to journal: Re: Calculation of deprivation scores	Editor, The paper by Ashworth et al is clearly of interest and adds to the growing literature on the impact of the QOF. It would indeed be welcome if the QOF is contributing to a narrowing of health inequalities. However, there is an issue with regard to the calculation of deprivation scores that needs to be raised before we can be sure that such reductions are a real effect. The authors note the ecological fallacy that the use of general practice postcodes as a proxy for the postcodes of registered patients at each practice may bring, but cite as supporting evidence work which stated that this method can provide a valid proxy for patient-level deprivation measures and may even underestimate the association between deprivation and all cause mortality.1 However, in that paper, the authors acknowledged that the lack of effect may be due to the location of the work in only one Primary Care Trust and that national level analyses may reveal greater levels of inequality with practice-level data. We have conducted such a national-level study, using QOF data from all general practices in England and Scotland for years 2005 – 2007. Analyses of the Scottish practices compared deprivation and ill health data assigned to either the practice postcode or the postcode of individual patients on the practice list and found that analyses based on practice postcode assigned data under-estimated the relationship between deprivation and ill health for both prevalence and quality care.2 For blood pressure this resulted in an overestimate of achievement levels for blood pressure recorded and controlled for CHD in the most deprived decile and underestimated achievement in the least deprived decile when using practice postcode to estimate deprivation compared to the postcodes of the practice population. This resulted in the practice postcode method showing no significant relationship between deprivation and achievement while the patient population postcode method showed a significant negative relationship between achievement and deprivation for both CHD blood pressure indicators. It is entirely possible that similar findings would occur in England and highlights the importance of using practice population-level wherever possible. The impact on ethnicity from using general practice postcodes as opposed to the practice population is not yet known and merits further investigation. Strong M, Maheswaran R, Radford J. Socioeconomic deprivation, coronary heart disease prevalence and quality of care: a practice-level analysis in Rotherham using data from the new UK general practitioner quality and outcomes framework. J Pub Health 2006; 28:39-42. McLean G, Guthrie B, Watt G, Gabbay M, O'Donnell CA. Practice postcode versus patient postcode: a comparison of data sources in England and Scotland. International Journal of Health Geographics 2008 Jul; 7: 37 Competing interests: None declared
Is primary care equitable? A need for an accurate measure of socio-economic deprivation. 28 November 2008
Mark Strong, MRC Fellow ScHARR, University of Sheffield, 30 Regent St, Sheffield, S1 4DA Send response to journal: Re: Is primary care equitable? A need for an accurate measure of socio-economic deprivation.	Dear Editor Quantifying the relationship between general practice level socio- economic deprivation, quality of care and health outcomes is key to determining whether primary care is being delivered equitably or not. Ashworth et al make the assumption that the level of socio-economic deprivation experienced by the population in the vicinity of the main building occupied by a general practice is a valid proxy for the level of deprivation of the practice population as a whole. They were forced into making this assumption because patients' postcodes are not openly available to researchers for reasons of data protection. The method of linking the location of a general practice building to a deprivation score is not ideal. As Ashworth et al point out in their discussion, we did find in our study in Rotherham that deprivation scores based on the location of the practice building were a valid proxy for the "true" practice population deprivation score, but only in the sense that there was a high correlation between them. We explain this in detail in a paper in International Journal of Health Geographics.[1] However, our study also reported that, by using the practice building location method, we underestimated the association between deprivation and mortality that was found when we used a score based on postcode data from the whole practice population. This underestimation problem has since been demonstrated by McLean et al in a similar but much larger study of Scottish practices, and I suspect that the same underestimation of association between deprivation and a health related measure may have occurred in the Ashworth study. There may in reality be a rather greater negative association between deprivation and blood pressure monitoring and control in English general practice than Ashworth et al were able to demonstrate. This, I believe, should lead us to be very cautious before we conclude that inequalities in blood pressure monitoring and control in English general practice have disappeared. [1] Strong M, Maheswaran R, Pearson T. A comparison of methods for calculating general practice level socioeconomic deprivation. International Journal of Health Geographics 2006;5:29 [2] McLean G, Guthrie B, Watt G, Gabbay M. O'Donnell CA. Practice postcode versus patient postcode: a comparison of data sources in England and Scotland. International Journal of Health Geographics 2008;7:37 Competing interests: None declared
Interpreting patterns of changes in absolute differences between rates when common outcomes become even more common 7 December 2008
James P. Scanlan, Attorney Washington, DC 20007, USA Send response to journal: Re: Interpreting patterns of changes in absolute differences between rates when common outcomes become even more common	Ashworth et al.[1] find reason to be optimistic about the effects of pay for performance on healthcare inequalities in the fact that improvements in overall blood pressure monitoring and control rates since the reporting of performance indicators for primary care have coincided with the reduction or elimination of absolute difference between rates of socioeconomic groups. The authors overlook the ways that, solely for statistical reasons, standard measures of differences between rates tend to be systematically affected by the overall prevalence of an outcome. As favorable outcomes increase, relative differences in experiencing them tend to decrease while relative differences in failing to experience them tend to increase. Absolute differences tend to change in the same direction as the smaller of the two relative differences. Roughly, this means that, as rare outcomes become more common, absolute differences tend to increase; as already common outcomes become even more common, absolute differences tend to decline. Thus, during periods of overall improvements in health and healthcare, researchers who measure inequalities in terms of relative differences in favorable outcomes tend to find inequalities to be decreasing. Researchers who measure inequalities in terms of relative differences in the opposite (adverse) outcomes tend to find inequalities to be increasing. Researchers like Ashworth et al. who examine very common (increasing) outcomes in terms of absolute difference between rates tend to find inequalities to be declining. But researchers who examine relatively uncommon (increasing) outcomes in terms of absolute differences between rates tend to find disparities to be increasing. About a hundred references explaining the implications of these tendencies in particular settings, and why none of the standard measures of differences between rates can alone indicate whether an inequality is changing in a meaningful sense, may be found on the Measuring Health Disparities page (MHD) of jpscanlan.com. The nuances of the tendencies are discussed at length on the Scanlan’s Rule page of the same site. As it happens, in the U.S the principal study of the effects of performance measures on healthcare inequalities also relied on absolute differences between rates to measure inequality. But that study examined such differences with regard to very uncommon outcomes and found the inequalities to be increasing. As a result of the study, in the U.S. the perception is that pay-for-performance programs will tend to increase inequalities, the opposite of the finding of Ashworth et al. See discussion under the Pay for Performance tab of MHD. Though their analysis is concerned solely with absolute differences that are declining, Ashworth et al. regard their work as consistent with the “inverse equity hypothesis” formulated by Victora et al., which held that improvements in overall health would lead first to increasing inequalities but then later to declining inequalities. Victora et al., however, examined inequalities in terms or relative differences and focused on situations where improvements in health increased relative differences in adverse outcomes until the point where the advantaged group’s adverse outcome rate approached a “minimum achievable level” beyond which further improvements are difficult or impossible. The failure of Victora et al. to consider the way the observed patterns were driven by statistical tendencies raises questions as to the validity of their hypothesis.[2] But, in any case, the pattern described by Victora et al. whereby improvements in health may cease to cause relative differences in adverse outcomes to increase once the rate for the advantaged group approaches an irreducible minimum constitutes a substantially different phenomenon from that whereby improvements in health or healthcare may sometimes lead to increases, then later to decreases, in absolute differences. Approaches to measuring healthcare inequalities that are unaffected by overall prevalence are described under the Solutions tab of MHD. It warrants note that such approaches would likely lead to findings that the inequalities examined by Ashworth declined in a meaningful sense (that is, in the sense that there occurred changes not driven entirely be statistical tendencies, not in the sense that the changes were necessarily substantial). But the general issue of whether pay-for-performance programs will tend to affect healthcare inequalities in a meaningful sense in any setting is far more complicated than a simple question of whether the absolute difference between rates ¬–or any of the other standard measure of difference between rates – has increased or decreased. References: 1. Ashworth M, Medina J, Morgan M. Effect of social deprivation on blood pressure monitoring and control in England: a survey of data from the quality and outcomes framework. BMJ 2008;337:a2030. 2. Scanlan JP. “Inverse equity hypothesis” overlooks important statistical tendencies. Journal Review Dec. 2, 2008 (responding to Victora CG, Vaughan JP, Barros FC, et al. Explaining trends in inequities: evidence from Brazilian child health studies. Lancet 2000;356:1093-1098): http://journalreview.org/v2/articles/view/11009159.html Competing interests: None declared