Skip to main content

Archived Comments for: Persistent socioeconomic inequalities in cardiovascular risk factors in England over 1994-2008: A time-trend analysis of repeated cross-sectional data

Back to article

  1. Efforts to quantify the magnitude of inequalities must consider more carefully the implications of the patterns by which measures tend to be affected by the prevalence of an outcome.

    James Scanlan, James P. Scanlan, Attorney at Law

    10 March 2013

    As very few others have done, Scholes et al.[1] importantly recognize certain patterns by which relative and absolute differences between outcome rates tend to be systematically affected by the prevalence of an outcome. But they overlook that there are two relative differences (one in the favorable outcome and the other in the opposite, adverse outcome) and that two tend to change in opposite direction as the prevalence of an outcome changes. They also fail to recognize the implications of patterns by which measure are affected by the prevalence of an outcome with respect to efforts to determine whether the forces causing outcome rates to differ have increased or decreased over time.

    For reasons related to the shapes of distributions of factors associated with experiencing an outcome, the rarer the outcome the greater tends to be the relative difference in experiencing it and the smaller tends to be the relative difference in avoiding it. Thus, for example, as mortality declines, relative difference in mortality tend to increase while relative differences in survival tend to decrease; as rates of appropriate healthcare increase, relative differences in receipt of appropriate care tend to decrease, while relative differences in rates of failing to receive appropriate care tend to increase. As prevalence of a health or healthcare outcome changes, absolute differences tend to change in the same direction as the smaller relative difference. And, irrespective of the shapes of the underlying distributions, when an absolute difference and a relative difference have changed in opposite directions (as will frequently be noted), the unmentioned relative difference will have changed in the opposite direction of the mentioned relative difference and in the same direction as the absolute difference.[2-5]

    A useful illustration of these patterns is found in the study of the effects of folate supplementation on inequalities in folate level. The Scholes study¿s reference 13 discusses a finding that after such supplementation absolute differences between rates of low folate decreased while relative differences between rates of low folate increased. But were one to instead measure inequality in terms of relative differences in adequate folate, one would find the inequality to have decreased.[7] And it warrants note that in circumstances where observers relying on relative differences in favorable outcomes find dramatic decreases in inequalities, those relying on relative differences in adverse outcomes will generally find dramatic increases in inequalities (as reflected in Table 4 of reference 4).

    Observed patterns of changes in the two relative differences and the absolute difference do not invariably conform to the described, prevalence-related patterns. Other factors are at work, including changes in the strength of the forces causing outcome rate to difference. The purpose of health inequalities research is to determine whether the strength of those forces is increasing or decreasing. But only with a firm understanding of the prevalence-related patterns can such determinations be sound.

    A useful illustration that society¿s concern is not with differences between outcome rates per se, but with what those difference reveal about underlying forces, may be found in Table 3 of reference 5 (at 25). That table presents a situation where the issue of concern involves the degree of bias exhibited by four employers with different overall selection rate and where one would reach different conclusions about the comparative bias of the employers based on relative differences in selection or relative differences in rejection and still different conclusions based on absolute differences. But it makes no sense to maintain that one employer could be more biased as to selection and another as to rejection or that one employer is more biased in an absolute sense while the other is more biased in a relative sense. There can be only one correct answer as to the comparative ranking of the bias of the employers and one can derive that answer only with an understanding of the way each measure tends to be affected by the prevalence of an outcome (or by relying on a measure that is unaffected by the prevalence of an outcome, as discussed at 24 to 28 of reference 5).

    A useful step in putting health inequalities measurement on sound foot would be to eliminate the terms ¿relative inequalities¿ and ¿absolute inequalities¿ from the health inequalities research vocabulary. There are relative differences between favorable outcome rates, relative differences between adverse outcome rates, and absolute differences between rates. Those differences may provide useful clues to the strength of the forces causing rates to differ when such differences are interpreted with an understanding of the way they tend to be affected by the prevalence of an outcome. But the differences cannot themselves quantify the strength of those forces.

    References:

    1. Scholes S., Bajekal M, Hande L, et al. Persistent socioeconomic inequalities in cardiovascular risk factors in England over 1994-20000. A time-trend analysis of repeated cross-sectional data. BMC Public Health 2012, 12:129 doi:10.1186/1471-2458-12-129

    2. Scanlan JP. Can we actually measure health disparities? Chance 2006:19(2):47-51. http://www.jpscanlan.com/images/Can_We_Actually_Measure_Health_Disparities.pdf

    3. Scanlan JP. Race and mortality. Society 2000;37(2):19-35. http://www.jpscanlan.com/images/Race_and_Mortality.pdf

    4. Scanlan JP. The Mismeasure of Group Differences in the Law and the Social and Medical Sciences. Applied Statistics Workshop at the Institute for Quantitative Social Science at Harvard University, Oct. 17, 2012. http://jpscanlan.com/images/Harvard_Applied_Statistic_Workshop.ppt

    5. Scanlan JP. Harvard University Measurement Letter. Oct. 9, 2012. http://jpscanlan.com/images/Harvard_University_Measurement_Letter.pdf

    6. McLaren L, McIntyre L, Kirkpatrick S: Rose¿s population strategy of prevention need
    not increase social inequalities in health. Int J Epidemiol 2010, 39:372¿377.
    8. NHANES Illustrations sub-page of Scanlan¿s Rule page of jpscanlan.com: http://jpscanlan.com/scanlansrule/framinghamillustrations.html

    7. NHANES Illustrations sub-page of Scanlan¿s Rule page of jpscanlan.com: http://jpscanlan.com/scanlansrule/framinghamillustrations.html

    Competing interests

    None declared

Advertisement