Len Smith

Applying Lee-Carter under conditions of variable mortality decline.

The Lee-Carter method of mortality forecasting assumes an invariant age component and most applications have adopted a linear time component. The use of the method with Australian data is compromised by significant departures from linearity in the time component and changes over time in the age component. We modify the method to adjust the time component to reproduce the age distribution of deaths, rather than total deaths, and to determine the optimal fitting period in order to address non-linearity in the time component. In the Australian case the modification has the added advantage that the assumption of invariance is better met. For Australian data, the modifications result in higher forecast life expectancy than the original Lee-Carter method and official projections, and a 50 per cent reduction in forecast error. The model is also expanded to take account of age-time interactions by incorporating additional terms, but these are not readily incorporated into forecasts.

On the measurement of relative and absolute income-related health inequality

In recent work on international comparisons of income-related inequalities in health, the concentration index has been used as a measure of health inequality. A drawback of this measure is that it is sensitive to whether it is estimated with respect to health or morbidity. An alternative would be to use the generalized concentration index that is based on absolute rather than relative health differences. In this methodological paper, we explore the importance of the choice of health inequality measure by comparing the income-related inequality in health status and morbidity between Sweden and Australia. This involves estimating a concentration index and a generalized concentration index for the eight-scale health profile of the Short Form 36 (SF-36) health survey. We then transform the scores for each scale into a measure of morbidity and show that whether the concentration index is estimated with respect to health or morbidity has an impact on the results. The ranking between the two countries is reversed for two of the eight dimensions of SF-36 and within both countries the ranking across the eight SF-36 scales is also affected. However, this change in ranking does not occur when the generalized concentration index is compared and we conclude with the implications of these results for reporting comparisons of income-related health inequality in different populations.

Improvements in Indigenous mortality in the Northern Territory over four decades.

Objective:To provide the first report of long‐term mortality trends over recent decades for an Indigenous Australian population. Very little information is available about improvements over time, or lack thereof, in the health status of Indigenous Australians.

More or less equal? Comparing Australian income-related inequality in self-reported health with other industrialised countries.

Objective: To measure the distribution of self–reported health by income in order to compare the level of health inequality in Australia with other industrialized countries.

It's time: record linkage — the vision and the reality

Objective : To examine reasons for the minimal use of Australian administrative health data for research and evaluation.

Comparing health inequalities among men aged 18-65 years in Australia and England using the SF-36.

Objective:To compare the distribution of the domains and summary scores of the SF‐36 health survey by occupation‐related social classes in England and Australia for employed males age 18–65 years.

Data for diagnosis, monitoring and treatment in Indigenous health: the case of cervical cancer

Deficiencies in the availability and quality of data on the health status of Indigenous Australians have long been recognised. For cervical cancer, data demonstrate a 2–5 fold greater incidence rate and an 8–10 fold greater mortality rate for Indigenous women compared to non‐Indigenous Australians. However, incidence and mortality data are only available for some states and there is little or no information available on the geographic or social distribution of risk, or the reasons for risk differentials. There are also little or no data on the utilisation of, or preferences for, screening services. Thus, while there is clearly a need for a cervical cancer control program specifically to target Indigenous women, current data are inadequate to inform planning and implementation, and current systems are inadequate to monitor effectiveness. This situation is the result of insufficient research and inadequate attention to recording of Indigenous status in routine data systems and applies to a greater or lesser extent across the spectrum of health of Indigenous Australians. Health workers across the spectrum in mainstream and Indigenous medical services have a shared responsibility for improving the availability and quality of data and ensuring the appropriate use of information necessary to achieve and monitor improvements in service delivery and health status of Indigenous people.

Infant mortality of Sami and settlers in Northern Sweden: the era of colonization 1750-1900

The study deals with infant mortality (IMR) that is one of the most important aspects of indigenous vulnerability. Background: The Sami are one of very few indigenous peoples with an experience of a positive mortality transition. Objective: Using unique mortality data from the period 1750–1900 Sami and the colonizers in northern Sweden are compared in order to reveal an eventual infant mortality transition. Findings: The results show ethnic differences with the Sami having higher IMR, although the differences decrease over time. There were also geographical and cultural differences within the Sami, with significantly lower IMR among the South Sami. Generally, parity has high explanatory value, where an increased risk is noted for children born as number five or higher among siblings. Conclusion: There is a striking trend of decreasing IMR among the Sami after 1860, which, however, was not the result of professional health care. Other indigenous peoples of the Arctic still have higher mortality rates, and IMR below 100 was achieved only after 1950 in most countries. The decrease in Sami infant mortality was certainly an important factor in their unique health transition, but the most significant change occurred after 1900.

Record linkage to advance Indigenous mortality statistics in Australia – sources of error and bias

This paper describes efforts made in Australia in the use of data linkage to enhance Indigenous mortality statistics. The extent of inadequacies of statistics sourced from death registration is discussed and the improvements made by data linkage are presented. Conceptual, methodological and data issues that may give rise to error and bias in such data linkage are discussed.

Geographic inequalities in mortality: comments on an article by Wilkinson et al.

The subject of geographic inequalities in mortality which Wilkinson et al. address in the June 2000 issue of the Journal1 is important and topical, and we (and presumably many other readers) spent much time and effort trying to interpret the surprising results they present. That task was not made easier by the failure of the authors to set the study in context. Obviously, a descriptive paper such as this need not address a specific hypothesis or test a particular theory, but it does need to review previous studies in the field and relate the findings to the public health context. Regrettably, your authors do neither, ignoring a definitive earlier study, and making no mention of the range of programs currently in place or in development aimed at addressing some known variations in mortality. These are serious omissions: if the authors had taken this background into account it might have alerted them to possible mistakes in the analysis. On page 226, the authors say, “we were unable to find recent published descriptions of the geographical distribution of all-cause and cause-specific mortality rates across Australia”. In fact, SK Jain2 published an extremely comprehensive analysis which examined variations in cause of death in States and Territories in 197192 and in Statistical Divisions and Subdivisions in 1991-92. How could both the authors and the referees have overlooked this significant publication on the very topic addressed by the article? The failure to provide a context extends to the technical level, resulting in mistakes not being recognised. Figures 2 and 3 present age-standardised rates but give us nothing to compare them with. It would have been far more informative to present standardised mortality ratios (SMRs) and their confidence limits, or at least to include the total Australian rate with which the State/Territory and Statistical Division rates could be compared. We tried to use the Australian rate to do this, but found that the total figure shown in Table 3 (7.35) differs from the total in Table 2 (6.98) (we assume the headings in Table 2 are wrong and should be the same as Table 1). In fact, all the figures for Australia in Table 3 appear to be wrong, since the rates barely fall within the range of State/ Territory values plotted in Figure 2. The fact that these figures are wrong casts doubt on the entire table. We do not have ready access to the original data, but to Letters to the Editor