Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
addressing inequality
- benjamin dean (29 April 2005)
-
Widening health inequalities: sensible measures and sensible comparisons
- Anne M Low, Allan Low (5 May 2005)
-
Glasgow’s younger stroke patients: 20 years behind
- Fiona Wright, Peter Langhorne, David J. Stott, Graham Ellis. (16 May 2005)
-
Inequalities in life expectancy in England
- Madhavi Bajekal, Allan Baker (28 May 2005)
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
benjamin dean, dr oxford
Send response to journal:
|
I agree wholeheartedly with many of the excellent observations in the article (BMJ 2005;330:1016-1021). It is shocking that economic inequality is increasing under a labour government. Undoubtedly one must tackle economic inequality if one wishes to reduce health inequalities. At the end of the article it was mentioned that to tackle poverty more money must be raised from our tax system and redistributed. I agree with this but believe the key is how the money is used to benefit the poor. Inequality is increasingly and most disturbingly social mobility is remarkably low in Britain when compared to other developed countries (1). Improving education is a of key importance and redistributed money should be invested here. The current government health policies that place a responsibility on the individual are misguided and ineffective, demonstrated by the growing inequalities in Britain today. The latest government white paper (2) is full of policies that are doomed to failure because they do not address the causes of ill health. Is is a mistake to believe that changing the health system will address health inequality, it simply won't. Policies in many areas outside health need to be changed if inequalities are to be addressed. Wealth must be redistributed effectively in ways that tackle the root causes of inequalities that exist in our society (ie education). If this does not happen then inequalities will continue to increase, with problems such as crime and poor health becoming more rife. 1. Intergenerational Mobility in Europe and North America by Jo Blanden, Paul Gregg and Steve Machin at the Centre for Economic Performance, London School of Economics. It can be downloaded at: http://www.lse.ac.uk/collections/pressAndInformationOffice/newsAndEvents/archives/2005/LSE_SuttonTrust_report.htm 2. Public Health White Paper. Department of Health. viewable at http://www.dh.gov.uk/ Competing interests: None declared |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Anne M Low, Director of Public Health, Derwentside PCT Shotley Bridge Hospital, Consett, Co Durham, DH8 0NB, Allan Low
Send response to journal:
|
EDITOR - We agree with Shaw et al (1) that the method of measuring
inequality in health outcomes used in setting the government targets and
indicators is problematic. (2) We applaud the authors’ more "sensible"
approach to quantifying inequalities in life expectancy by using the slope
index of inequality (SII). However, in one respect the official approach
is more appropriate than that used by Shaw et al. Although not highly
publicised, the government targets are based on reducing relative gaps in
life expectancy and infant mortality. (3) The SII used by Shaw et al
measures absolute gaps.
The distinction between absolute and relative gaps becomes important when comparisons are made over time. Shaw et al state that the SII "has a further advantage that it is, by definition, unaffected by general increases or decreases in life expectancy over time (in this case the constant changes, but not the slope)". But increasing life expectancy over time, with no change in the slope (constant SII), actually implies a faster rate of progress for those at the bottom of the scale than for those at the top. This is illustrated in the Chart 1. The two lines show the SII (slope) at two periods of time. The upward shift of the line from period 1 to period 2, without any change in slope is achieved by a gain of 10 years of life expectancy for both groups. The rate of improvement for the poor group is 10/70=14%. The rate of improvement for the wealthy group is 12.5% (10/80). Chart 1: Constant slope (SII) means a greater rate of progress for the deprived group
Contrary to what Shaw et al say then, the interpretation of the SII is affected by increases or decreases in life expectancy over time. This why we have advocated the use of the relative index of inequality (RII) in order to make valid comparisons of magnitudes of inequality in health outcomes over time, as well as between different conditions and rates of service provision. (4) The RII is calculated as the absolute gap in health outcome across a population, from least to most wealthy as a proportion of the average rate of health outcome for that population. In the example above the RII for period 1 is 13.3% (10/75). For period 2 it is reduced to 11.8% (10/85). Because life expectancy has been increasing generally over time, small increases in the SII do not necessarily mean that progress has been faster for the affluent than the deprived groups. The RII does however faithfully reflect relative rates of progress. We have calculated the RII from the data provided by Shaw et al by simply dividing the reported SII by the average life expectancy across all poverty groups. Table 1 shows the difference in inequality changes, whether estimated as absolute or relative gaps. In this case the SII overestimates the magnitude of gap changes, because the increasing general trend is not accounted for.
Table 1: The widening of gaps in life expectancy 1992-4 to 2001-3:
absolute versus relative measures
Both sexes Males Females
SII - absolute gap 4.3% 3.8% 4.0%
RII - relative gap 2.0% 0.9% 2.2%
If similar calculations were made for the other official headline target of infant mortality, the SIIs would underestimate the magnitude of gap changes, as infant mortality rates are on a general downward trend. It is good to see "sensible" methods for estimating magnitudes of inequalities in health being adopted. When comparing magnitudes of inequality it is important to be sensitive to the distinction between absolute and relative health gaps. Relative measures, such as the RII, are particularly powerful, as they can be used to compare the size of inequality gaps for indicators based on very different scales of measurement. Thus the magnitudes of inequality can be compared between one health condition and another or between a health condition and the rate of access to a related service. The latter comparison is particularly relevant in health equity audit studies. (5) References: 1 Shaw M, Davey Smith G, Dorling D. Health inequalities and New Labour: how the promises compare with real progress. British Medical Journal 2005;330:1016-1021. 2 Low A. Mind the gap: overcoming inconsistencies in health inequalities measurement. Paper presented at the UKPHA 13th Annual Forum, Sage, Gateshead. 2005. 3 Department of Health. The national health inequalities targets. Available at www.dh.gov.uk/assetRoot/04/07/82/24/04078224.pdf. Last accessed 15 April 2005. 4 Low A and Low A. Measuring the gap: quantifying and comparing local health inequalities. J Public Health 2004;26:388-395. 5 Low A. Health equity audit: a quantification approach. Paper presented at the UKPHA 13th Annual Forum, Sage, Gateshead. 2005. Competing interests: None declared |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Fiona Wright, Specialist Registrar Academic Section of Geriatric Medicine, Level 3, Centre Block, Royal Infirmary, Glasgow, G4 0SF, Peter Langhorne, David J. Stott, Graham Ellis.
Send response to journal:
|
We note your recent coverage of health inequalities (1). Rothwell et al (2) have demonstrated a drop in the age-specific incidence of stroke in Oxfordshire over the past 20 years. This trend may not be reflected in more deprived areas. Using the WHO definition of stroke (excluding subarachnoid haemorrhage) we identified individuals on admission to hospital (including early deaths and transfers) and on referral to a well-established stroke clinic or geriatric medical service (including day hospital) from August 2000 to July 2002. This included all potential stroke patients from 6 post code areas in East Glasgow which has some of the highest levels of deprivation and ill health in the United Kingdom (1,3 ). We reviewed 1,030 referrals; 753 were classified as definite stroke of which 478 had no previous symptoms of stroke or transient ischaemic attack. In comparison with the OXVASC study (Table) we observed similar total stroke incidence rates but higher rates among younger individuals and a lower mean age (68, SD 13 years). Although the overall incidence had fallen when compared to OCSP (1981-84) the rates of stroke in the under 65s were comparable. Although our data fall short of the gold standard of stroke incidence studies set by Rothwell and colleagues, these limitations do not explain our apparently higher rates of stroke at younger ages and the higher prevalence of smoking. It is striking then, that younger patients (<65) in Glasgow not only have a higher incidence of stroke than their peers in Oxfordshire but also have a comparable incidence of stroke to that observed 20 years ago. It seems probable that the trends in stroke incidence in more deprived populations reflect a legacy of socioeconomic deprivation that lags considerably behind those of Oxfordshire. Smoking appears to be a consistent part of that legacy.
* ≤0.01 References: 1) Mary Shaw, George Davey Smith, Danny Dorling. Health inequalities and New Labour: how the promises compare with real progress. BMJ 2005;330:1016-1021 2) Rothwell PW, Coull AJ, Giles MF, Howard SC, Silver LE, Bull LM, Gutnikov SA, Edwards P, Mant D, Sackley CM, Farmer A, Sandercock PAG, Dennis MS, Warlow CP, Bamford JM, Anlsow P for the Oxford Vascular Study. Change in stroke incidence, mortality, case-fatality, severity, and risk factors in Oxfordshire, UK from 1981 to 2004 (Oxford Vascular Study). Lancet 2004;363:1925-33. 3) Davey Smith G, Shaw M, Mitchell R, Dorling D, Gordon D. Inequalities in health continue to grow despite government’s pledges. BMJ 2000;320:582 Competing interests: None declared |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Madhavi Bajekal, Head, Morbidity & Healthcare Team Office for National Statistics, SW1V 2QQ, Allan Baker
Send response to journal:
|
In presenting results showing a widening in the gap in life expectancy (LE) between the most and least deprived areas in Britain over the 1990s, Shaw and colleagues argued that any study that does not use the recently released final revised population estimates will produce erroneous results, and cite as example a previous paper by one of us which showed a narrowing of the LE gap in England. Two issues arise from this. First, what impact population estimate revisions have on the conclusions. Second, whether this alone can account for the apparently different conclusions in the two papers. On the first issue, although the paper by Bajekal used the unrevised Census 1991 rolled forward population estimates, this mainly produces a difference of scale rather than in the direction of trends to the use of revised figures. We have checked whether the trend shown by Shaw et al. would have been different had they used the unrevised populations. We have found that their conclusions would still have differed from those in the Bajekal paper. The different conclusions therefore require another explanation. Both studies examined differential trends in LE between tenths of populations grouped by deprivation and covering almost the same period (1994 –1999 and 1992/4 -2000/3, respectively). We suggest that the reason for the discrepancy in the two findings is due to differences in geographical coverage (England vs Great Britain) and the spatial level of analysis (ward vs local authority (LA)). Using the same data as Shaw et al., but for England alone, we see no change in the LE gap between the LAs with the highest and lowest LE for males, and a relatively small increase in the gap for females (0.6 years) over the 1990s (Table1). Given that the width of the 95% confidence interval at LA level is about one year, these results suggest the LE gap in England remained stable over the period. Finally, even within England, results from a ward level analysis will differ from that at the LA level. There are considerable differences in both LE and social deprivation between wards within most LAs, and the socio-economic composition of wards also changes over time. This will tend to make differences in life expectancy between LAs both narrower and more stable over time than those for wards. Table 1: Local authority areas with the highest and lowest life expectancies at birth in England (with 95% confidence intervals). Males Females 1992-4 2001-3 1992-4 2001-3 Highest life expectancy (years) 78.0 (77.3 - 78.7) East Dorset 80.1 (79.3 - 80.8) East Dorset 83.1 (82.5 – 83.8) East Dorset 84.8 (84.3 – 85.4) Kensington and Chelsea Lowest life expectancy (years) 69.7 (69.4 – 70.1) Manchester 71.8 (71.4 – 72.1) Manchester 76.5 (76.1 – 76.8) Manchester 77.6 (77.0 – 78.3) Blackburn with Darwen Difference 8.3 8.3 6.6 7.2 Shaw M, Davey Smith G, Dorling D. Health inequalities and New Labour: how the promises compare with real progress, BMJ 2005, 330:1016-21. Bajekal M. Healthy life expectancy by area deprivation: magnitude and trends in England, 1994-1999. Health Statistics Quarterly 2005, 25:18-27. Competing interests: None declared |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
