Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The trend in mental health-related mortality rates in Australia 1916-2004: implications for policy

The trend in mental health-related mortality rates in Australia 1916-2004: implications for policy Background: This study determines the trend in mental health-related mortality (defined here as the aggregation of suicide and deaths coded as “mental/behavioural disorders”), and its relative numerical importance, and to argue that this has importance to policy-makers. Its results will have policy relevance because policy-makers have been predominantly concerned with cost-containment, but a re-appraisal of this issue is occurring, and the trade-off between health expenditures and valuable gains in longevity is being emphasised now. This study examines longevity gains from mental health-related interventions, or their absence, at the population level. The study sums mortality data for suicide and mental/behavioural disorders across the relevant ICD codes through time in Australia for the period 1916-2004. There are two measures applied to the mortality rates: the conventional age-standardised headcount; and the age-standardised Potential Years of Life Lost (PYLL), a measure of premature mortality. Mortality rates formed from these data are analysed via comparisons with mortality rates for All Causes, and with circulatory diseases, cancer and motor vehicle accidents, measured by both methods. Results: This study finds the temporal trend in mental health-related mortality rates (which reflects the longevity of people with mental illness) has worsened through time. There are no gains. This trend contrasts with the (known) gains in longevity from All Causes, and the gains from decreases achieved in previously rising mortality rates from circulatory diseases and motor vehicle accidents. Also, PYLL calculation shows mental health-related mortality is a proportionately greater cause of death compared with applying headcount metrics. Conclusions: There are several factors that could reverse this trend. First, improved access to interventions or therapies for mental disorders could decrease the mortality analysed here. Second, it is important also that new efficacious therapies for various mental disorders be developed. Furthermore, it is also important that suicide prevention strategies be implemented, particularly for at-risk groups. To bring the mental health sector into parity with many other parts of the health system will require knowledge of the causative factors that underlie mental disorders, which can, in turn, lead to efficacious therapies. As in any case of a knowledge deficit, what is needed are resources to address that knowledge gap. Conceiving the problem in this way, ie as a knowledge gap, indicates the crucial role of research and development activity. This term implies a concern, not simply with basic research, but also with applied research. It is commonplace in other sectors of the economy to emphasise the trichotomy of invention, innovation and diffusion of new products and processes. This three-fold conception is also relevant to addressing the knowledge gap in the mental health sector. Background indicates that mental disorders are the seventh most Burden of disease studies indicate that the impact of expensive disease category in Australia [4,5]. Various mental disorders is considerable [1-3], while the latest dimensions of mortality associated with mental disor- Australian Institute of Health and Welfare (AIHW) ders are not trivial. There are some meta-analyses indi- report on relative health expenditures by disease groups cating the excess mortality associated with these disorders–both natural and unnatural causes increase the risk of premature death for mentally ill people [6,7]. * Correspondence: ruth.williams@vu.edu.au Also, a single international meta-analysis, focussing only † Contributed equally School of Economics and Finance, Victoria University, Sunbury Campus, on suicide, shows a heightened suicide risk is associated Australia © 2010 Doessel et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 2 of 10 http://www.anzhealthpolicy.com/content/7/1/3 with almost all mental disorders [8]. Another approach– There are several policy implications in applying the that of the psychological autopsy–has found that about PYLL metric. The most important is that government 90 per cent of people who die by suicide have at least policies usually are designed to affect a particular vari- one mental disorder at the time of death [9]. Several able or target. When forming policy, or evaluating exist- Australian studies that have examined various aspects of ing policy, such as evaluating the expenditure on mortality from mental disorders are now available, their Australia’s National Suicide Prevention Strategy [34,35], focus largely being on suicide [10-17]. it is important to use appropriate measures, as discussed The present study measures the time-trend in the elsewhere [24-26], of the variable being targeted. The mental health-related mortality rate. The term “mental PYLL metric provides relevant information for societal health-related mortality” is defined as the sum of deaths or policy issues, because it is a weighted measure (Table from mental and behavioural disorders and suicide. 1). This example shows clearly that the PYLL is a more Apart from the conventional headcount measure, some appropriate measure of premature mortality, from a studies have applied an alternative measure, the poten- societal perspective, than the (equal) headcount tial years of life lost (PYLL) to suicide [18,19]. The measure. PYLL metric originated in the 1940s for the evaluation Another policy implication relates to an argument of tuberculosis prevention programs, when it had from health economics. There has been considerable become apparent that headcount (only) measurement of concern about the rising absolute and relative expendi- mortality did not convey all the information relevant to tures of health services [36-38]. For example, the focus the prevention of tuberculosis mortality [20]. The PYLL from this perspective is that Australia’s expenditure on metric subsequently achieved prominence in the burden health in 2004-05 was 8 per cent of Gross Domestic of disease work of Murray and Lopez [21]. Currently, it Product (GDP), whereas in 1960-61 it was 4.1 per cent is routine practice for the Australian Bureau of Statistics of GDP [39]. The OECD average in 2003 was 8.8 per (ABS) [22] and the Australian Institute of Health and cent of GDP. Thus, comparatively, Australia’s position is Welfare(AIHW)[23]toreportbothheadcount and “in the middle”, between the United States (15.0 per PYLL measures of suicide. A small number of Australian cent) and the United Kingdom (7.7 per cent). Even Aus- analyses have applied both headcount and PYLL mea- tralia’s “middle” position is viewed with some concern, surement to suicide [24-26]. These studies show the as other countries in a lower position must use their added information gained by applying both headcount health resources differently from those above them [40]. and PYLL metrics. Various economists, governments and others (e.g. insur- The focus here is on providing historical and compara- ance carriers) have adopted a cost-containment view. tive (with respect to other diseases/conditions) analyses However, in the recent international literature on the through time, using headcount and PYLL measures. The economics of health services, the cost-containment measure of mental health-related mortality that we apply emphasis has been subject to re-appraisal. This re- here involves summing across the relevant ICD codes appraisal involves an examination of the contribution of through time both “suicide” and mortality from “mental the health sector in the totality of the economies of and behavioural disorders”.Wetakethese combined OECD-type countries. Scholars of the re-appraisal bring causes of death to approximate the (mortality) size of the a different emphasis–it is argued that due regard must problem of mental health-related mortality. be given to the gains to health that both public health The recent emphasis in burden of disease studies sug- programs and clinical medicine have wrought [41-45]. gests that measuring morbidity as well as mortality is In this context, accurate measurement of “the gains” is important. However, a limitation of the (Australian) bur- vital. The Discussion section below develops this point. den of disease work is its cross-sectional nature; that is, data have been constructed for only two years, 1996 and Methods 2003 [1,27]. This article examines 88 years of data. In order to extract data on mental health-related mor- The present study has a narrower focus, by measuring tality, annual data were summed across the relevant mortality in levels only; the focus does not extend to ICD codes for “suicide” and “mental/behavioural disor- studying its distribution. Measuring the (age) distribu- ders”, by five year age-groups, from the AIHW [23] and tion of mortality (due to all causes, or any cause) for the years prior to 1968, from ABS historical data through time is possible, by applying the analytical fra- [46-49]. The data set thus obtained provides a complete mework pioneered by Silber [28-30] and Le Grand enumeration, though underestimated [50], of mental [31-33]; there is one Australian study which has mea- health-related mortality in Australia; it is not a sample. sured the distribution of suicide [26]. Distribution topics Given the various changes in the structure of, and parti- in mental health-related mortality need further atten- cular codes in, the revisions of the ICD, we referred to tion, but here the focus is on levels only. Taylor’s detailed accounts to guide the data recording Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 3 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Table 1 Two approaches to mortality measurement: the measures increased due to this exercise, though the PYLL and the Headcount magnitude was very small both for headcounts and for Case The PYLL metric The count metric PYLLs, because there were so few cases in the “undeter- mined” category. Hence, no substantive conclusions (Given life expectancy US white women, drawn in that paper were found to be affected by that 1940s, of 69 years) under-reporting problem. Death of white 45 years of life 1 death However, other problems exist in the accuracy of sui- woman, aged 24 cide data. It is very important to realise that our PYLL Death of white 7 years of life 1 death calculations are based on the mortality coding underta- woman, aged 62 ken (and published) by the ABS. It has been known for Source: Adapted from Dempsey[20] some time that ABS data on suicide have been subject to misclassification [53], as well as other more general [51]. Australian data on suicide start from 1907, but errors indicated by De Leo [40], De Leo, Klieve and Mil- data on mortality from mental/behavioural disorders are ner [54] and Walker, Madden and Chen [55]. These available only from 1916. The data on All Causes, can- problems have become more important since 2000. cer and circulatory diseases are available from 1907. Harrison, Painter and Elnour [56] have undertaken an However, motor vehicle accident mortality data exist important study in which they re-worked the published only from 1924. ABS data for a single year, 2004. They found that the The above data are of a headcount kind, and they are published ABS data for Australia underestimated suicide employed in the calculation of PYLLs. The ABS provides enumeration by 16 per cent. By far, the most important one account of the processes applied to headcount data source of error was the non-inclusion of coroners’ cases in order to generate PYLL data [52]. The following when they were incomplete at the time of ABS enu- equation conveys the method. It specifies the PYLL for meration. (See also Elnour and Harrison [57].) The ABS Suicide for year t, and an assumed life expectancy of 75 for some years has published a “Caution” relating to the years: quality of published “cause of death” data and has announced changes to its processes of coding and pub- AS lishing suicide data [58]. The Caution reads as follows: Suicide PYLL () 75 tD=− [ (75 A )] ss The level of recorded deaths attributed to suicide, and i =1 observed changes over time are likely to have been where Suicide PYLL(75) is the total PYLLs due to affected by delays in finalising a cause [59]. The ABS AS Suicide, age-standardised, at time period t; D is the has, from 2009, commenced a process of revising suicide data as more information, such as coroners’ decisions, number of Suicides per age group; A is the adjusted age becomes available. at death due to Suicide per age group; and i is the num- However, the impact of this inaccuracy in the data for ber of age groups for i = 1, ... n. In the present study, the 85+ age group is open in the this study, and thus for PYLL calculations in general, is raw data. We closed it at 100 years of age in our PYLL not known. This is because the age-distribution of that calculations. Also, we applied conventional age standar- under-reporting is not known. Until known, the implica- disation techniques, using the Australian Standard tion for PYLL measurement cannot be determined. Population 1991, both to the headcount and the PYLL data sets. Results Earlier in this Section the problem of under-estima- We report mortality rates (i.e. conventional rates based tion of suicide data, which is a well-documented pro- on headcount data) first. blem, was raised. We have, elsewhere, quantitatively Mortality rates: mental health-related and some investigated one implication of this problem for PYLL other causes measurement [25]. That very specific implication relates Figure 1 puts the mental health-related mortality trend to the eighth and subsequent revisions of the ICD, in perspective by presenting a line graph of the All which involved new codes for “deaths undetermined Causes mortality rate as well as comparative mortality whether accidentally or purposely inflicted” (ICD codes rates associated with some causes of death, viz.mental E980-E989 in ICD-8 and ICD-9, and Y10-Y34 in ICD- health-related (as defined here), cancer, circulatory dis- 10). We noted that non-inclusion of such deaths in that eases and motor vehicle accidents. Given the large dif- study [25], subsequent to 1968, could under-estimate ferences in age-standardised rates, the figure has two the contribution of suicide, and thus overall mental- parts, Part (a) showing a long-run decrease in the All related mortality. Hence, we re-analysed the post-1967 Causes mortality rate. This reflects the experience of data with these additional codes included. We found all many countries. Some argue that the nineteenth century Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 4 of 10 http://www.anzhealthpolicy.com/content/7/1/3 witnessed a transition phase, and then a period of vir- nearly 5 per cent of all causes, and has remained at that tually continuous decline in mortality, and that this approximate level in subsequent years. trend is an “epidemiological transition” or a “demo- While it is not our purpose to explain the trends graphic transition” [60-63]. Part (b) depicts the mortality depicted in Figures 1 and 2, it is useful to reflect on rates for four specific causes: circulatory diseases; can- some known factors. The rise of motor transport early cers; motor vehicle accidents; and mental health-related. last century resulted in increased accident mortality Circulatory disease is by far the largest single cause of through time, but governments eventually undertook death and cancer is the second most important; mental many interventions, such as seat-belt legislation, crash health-related mortality and motor vehicle accidents are helmets, safety-designed roads, drink-driving legislation “small” by comparison. Note that there are differing etc. Some of these interventions are efficacious [64]. temporal trends in these causes. The mortality rate for Similarly, for ischaemic heart disease and related condi- cancer has declined slightly over recent years. Mortality tions [65], medical treatments have improved and from circulatory diseases initially rose, but a large awareness has increased due to interventions undertaken decline from 1968 has occurred. Mortality arising from by governments, particularly in the form of health infor- motor vehicle accidents also was once rising, but it has mation. While the data on mortality depicted here do declined after 1978. The fall in mortality due to both not show the remarkable downturn in other causes of these causes has been substantial, but this is not the death, such as infectious diseases [66], the implementa- case with mental health-related mortality. The rising tion of public health measures was successful in these trend due to mental health-related mortality is note- areas [67]. In this context, note that the first year of the worthy. It was relatively high in the 1920s, but it fell allocation of Australian Government funds to suicide during the World War II period, and, since then, has prevention was 1995. risen, reaching a (local) maximum in 1996. The motor The message from both Figures 1 and 2 is clear: there vehicle accident trend exceeded that for mental health- is no evidence that the prevention of mental health- related mortality from the 1930s to the 1980s (and often related mortality is working (as measured by the sum- by a substantial margin), but since 1983 the mortality mation of deaths from mental/behavioural disorders and rate from motor vehicle accidents is less than that for suicide). That is, no evidence is found that the demo- mental health-related mortality. graphic transition is presently operative for mental- Percentage contributions relative to All Causes health related mortality. Figure 2 depict the proportionate “shares” (or the per- This conclusion is in line with those scholars who are centage contributions) for each of these causes of death concerned with analysing “avoidable mortality”.Suicide relative to All Causes through time. The calculation and deaths from mental/behavioural disorders have not involved the ratio of each of the above four causes of been included in any list of “avoidable death”.Wepost- death to All Causes, expressed as a percentage. pone further consideration of this point to the Discus- Figure 2 is also in two parts. With very large differ- sion section below. We now report the results from ences in the proportions for the particular causes pre- applying a PYLL metric. sented here, the two-part depiction is helpful. The scale The PYLL results of the Y-axis in Part (a) is in levels, and is logarithmic Figure 3 as with Figure 2 is concerned with comparison. in Part (b). The latter scale makes the detail about the However, we now focus on mental health-related mor- temporal trends in the “smaller” causes of death clearer. tality, and compare the two measures of mortality (the Part (a) of Figure 2 shows that, during the 1950s-80s, headcount measure and the PYLL measure) of this circulatory diseases accounted for at least half the causes aggregated cause of death. We have calculated the per- of death but that, since then, the relative fall in this centage shares of mental health-related mortality to All cause of death is very marked, and accounted for 35 per Causes mortality (measured by a headcount and by cent of deaths in 2004. The relative importance of the PYLLs). The percentage share for “The Headcount Mea- contribution of cancer to all mortality has increased. sure” in Figure 3 is the same as the “Mental Health- Cancer accounted for approximately 10 per cent of all Related” share in Figure 2, given that the measure of the deaths in the1940s,whereas thecomparablefigurewas rates of Figure 2 is the headcount. nearly 30 per cent of all deaths by 2004. Although the It is clear that both measures of mental health-related trends through time in the contributions of motor vehi- mortality exhibit the same general patterns through cle mortality rates and mental-health related mortality time. Both measures show that the relative (numerical) rates appear different from each other, the graphs in importance of mental health-related mortality has risen Part (a) are difficult to interpret. Part (b) clarifies the since the end of World War II. However, the PYLL picture. In 1996, the relative importance of mental/beha- measure (which reaches a local maximum of nearly 14 vioural disorders and suicide reached a maximum of per cent of All Causes mortality in 1998) indicates that Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 5 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 1 Age-standardised mortality rates for (a) total mortality (all causes combined), and (b) circulatory diseases, cancers, mental health-related mortality and motor vehicle accidents, Australia, persons, 1907-2004. * These rates have been standardised to the age distribution of the 1991 Australian population. ‡ Mental-Related Mortality includes Mental/Behavioural Disorders and Suicide. Sources: AIHW[23], CBCS [46-49], Taylor[52] Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 6 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 2 Shares of five causes of death to all deaths, measured by counts, Australia, persons, 1907-2004. The shaded vertical lines indicate the years of implementation of revisions of the ICD. Sources: AIHWAIHW[23], CBCS [46-49], Taylor[52] Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 7 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 3 Mental health-related mortality as a percentage of all causes measured by the count measure (no.) and the potential years of life lost (PYLL) measure, Australia, persons, 1916-2004. Sources: AIHWAIHW[23], CBCS [46-49], Taylor[52] the problems in the mental health sector are consider- indicates that the underestimation has been more severe ably larger than indicated by “The Count Measure” since 2000 [56]. Thedatainaccuracyissueisaddressed (nearly 5 per cent in 1998). in more detail above, and Figure 3 reflects this problem. It may be thought that the apparently large downturn The PYLL measure particularly is accentuating the data which can be observed in Figure 3 post-1998 in the inaccuracy of recent years. PYLLs trend is of some importance, to the extent that Thus, the conclusions that one draws about the the conclusions of the article are contradicted. In the numerical importance of mental health-related mortality following paragraphs we argue that any such position is depends on the mortality measure employed. The head- not the case. First, we have elsewhere reported the count measure is the typical measure applied to the results of estimating equations to time-series data, in an mental health sector: the results in Figure 3 clearly indi- article that focuses on the distribution of suicide [26]. In cate that this measure of mortality underestimates the that paper, where the reported equations are for rates, size of the problems associated with mental disorders in both on headcount data and on PYLL data, we find that Australia. We undertook PYLL calculations for the con- the trend is not cubic: there is no downturn in those tributions of the other three causes of death but we do data. In other words, the Ramsey RESET test indicated not report these results here due to space limitations. that the post-1998 decrease was not statistically signifi- These results are available from the authors on request. cant. Given that we subjected those equations to a full Attention should be paid to both headcount and PYLL raft of diagnostic tests of the residuals, and the stability measures because each sheds light on different aspects of the specification of the model, considerable confi- of the phenomenon. dence can be placed in those results. Figure 3 is not a depiction of rates, but the trend in Discussion the contribution of mental health-related mortality to As argued briefly in the Background section of this arti- All Causes mortality. However, in the Methods section cle, we suggest that it is misplaced for policy makers to above, we explain that, for various reasons, the number have a concern solely with health expenditure. of (published) suicides has been falling recently. Not Attention should be directed to both health expendi- even the ABS believes that “less suicide” is actually hap- ture and the value of the health outputs produced by pening. Rather, what is happening is inaccuracies in the the health sector. A statement by Nordhaus neatly cap- suicide data have been worsening recently. Recall that a tures this perspective, as follows: “The new view of re-working of the mortality data for 2004, the last year health economics should shape the way we think about of our analysis here, indicated that the underestimation health policy. In the early 1990s, the general hysteria of the published data for that year amounted to 16 per about rising health costs led many to believe that the cent. Furthermore, there is other evidence which health care system was wasteful and out of control and Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 8 of 10 http://www.anzhealthpolicy.com/content/7/1/3 should be reigned in” [p. 20] [68]. There is not the space resources to averting suicide and all mental-related mor- to describe or review the reappraisal studies, but note tality, or passivity towards this cause of death, more so that Davis et al. have argued cogently that this perspec- since it is classified as “unavoidable”.The argument in tive is important, andthatitdoesnot negate thecom- this paper refutes any such stance. mon criticisms levelled at the health sector, eg poor The second point is technical, and relates to having access, inappropriate treatment, issues arising with mar- relevant data available when studies to evaluate the effi- ketpower etc[69]. The point of thepresent articleis cacy of prevention strategies are sought. Given the sug- made in this context: information about the relative gestion that efficacious government policy can numerical importance of mental disorders is very contribute to reversing the mortality trend from mental important. disorders, could such an impact be detected statistically? These arguments suggest reflecting on the factors Bhattacharyya and Layton provide one example [64]. known already to contribute to the long-run decrease in The task involves detecting (post-intervention) whether mortality, which is characteristic of “the demographic or not there has been a reversal in the sign on the slope transition”. One key factor is knowledge of disease pro- variable of the equation for the mortality trend (i.e. cesses. For example, in nineteenth century England, the from positive to negative). The above results also indi- observational disposition of John Snow towards the cate the importance of observing the sign on the slope water supply (wells etc.) ultimately provided the relevant of all relevant trends: e.g. the age-standardised head- knowledge of water contamination for the prevention of count rate, etc. Thus, the appropriate technique exists cholera; and knowledge of efficacious therapies, such as for establishing empirically the impact of a government the Fleming-Florey “story” of the development of knowl- policy on prevention of mental health-related mortality. edge about antibiotics, is an example, of a different kind, Third, as mentioned above, mortality is but a partial, of how knowledge is applied. A more recent factor is and imperfect, measure of the health status of a com- technological change. Technological change involves munity. Thus, this point is a qualification. Since Zec- both life-saving technologies, such as organ transplanta- khauser and Shepard [76] outlined the Quality Adjusted tion, and “maintenance” types of technologies, such as Life Year (QALY) concept, it is regular practice to con- dialysis for end-stage renal disease. Some technological sider the quality of life associated with morbidity, along change enhances the productivity of curative and pre- with mortality. While the formation of time-series data ventive health services. This factor is far from trivial. sets of mental health-related morbidity is desirable, the Consider extra-corporeal shockwave lithotripsy, extra- “quality” of morbidity is a “gap” in health data-sets, and capsular cataract extraction and phaco-emulsification it is rarely discussed. The absence of such a time-series cataract therapy, which are three technologies of this data set on morbidity has induced us to have recourse kind. to mortality data: such data are available for a relatively Additionally, there are the technological changes that long period. The results presented in this paper clearly have occurred in diagnosis. For example, whereas once indicate the importance of determining the shape and there was just radiology, there is now also MRI, multi- direction in the long-term trend in mortality from this slice CT scanning, Dopler ultrasound, PET scanning, specific cause of death. Determining even this much gamma camera imaging etc. The point of this paragraph information is not trivial, even though the trend in the is to emphasise the implication of this paper which is morbidity from mental conditions cannot be determined that an appropriate focus in research about the mental over that period. health sector is to determine, and implement, the factors Our final point is a qualification. The argument in our that will contribute to the long-run decrease in mortality paper does not negate some other very important issues in the mental health sector. in the mental health sector. Such issues include unmet We now qualify the argument here with four points. need in the provision of mental health services [77-79], The first relates to the notion of “avoidable mortality” and the matter of people with mental disorders not or “amenable mortality”. Since Rutstein et al. [70], receiving efficacious, evidence-based treatments [80-82]. numerous scholars have formed lists of diseases/condi- This paper is an exercise in descriptive science and it is tions for which medical or societal interventions are effi- not our purpose to take any normative stances or to cacious [71-74]. Nolte and McKee provide a enter the debates about these issues. Rather, we seek comprehensive review [75]. It is noteworthy that suicide measurement approaches that will inform policy debates and mortality from mental/behavioural disorders are not better. included in any list of avoidable deaths: deaths from mental illnesses are classified as “unavoidable” in that Conclusions literature. This classification can be confused with an It is unbalanced to focus solely on rising health expendi- implicit attitude of resistance towards allocating tures, without valuing the improvements in health status Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 9 of 10 http://www.anzhealthpolicy.com/content/7/1/3 References broughtbythose expenditures:itisimportanttocon- 1. Mathers C, Vos T, Stevenson C: The Burden of Disease and Injury in Australia sider health gains alongside expenditures. A re-thinking Canberra, Australian Institute of Health and Welfare 1999. of the conventional wisdom about cost-containment 2. Mathers C, Vos ET, Stevenson CE, Begg SJ: The Australian burden of disease study: Measuring the loss of health from diseases, injuries and does not imply that various criticisms of the provision risk factors. Med J Aust 2000, 172:592-596. of health services no longer apply, but it does suggest 3. Vos T, Mathers CD: Burden of mental disorders: Australian and Global that determining the place of mental disorders in the Burden of Disease studies. Bull World Health Organ 2000, 78:427-438. 4. Australian Institute of Health and Welfare: Health System Expenditure on context of the reappraisal literature is relevant. The Disease and Injury in Australia 2000-01 Canberra, AIHW 2004. above results are unambiguous: there is no evidence 5. Australian Institute of Health and Welfare: Health System Expenditure on that long-run mental health-related mortality is falling Disease and Injury in Australia 2000-01 - Update Tables and Figures AIHW 2004http://www.aihw.gov.au/publications/index.cfm/title/10006, (accessed (as measured here by the summation of deaths from 9/10/2006). mental/behavioural disorders and suicide) and that sui- 6. Brown S: Excess mortality of schizophrenia. The British Journal of Psychiatry cide and mental/behavioural mortality has not contribu- 1997, 171:502-508. 7. Brown S, Barraclough B, Inskip H: Causes of the excess mortality of ted to the decline in the All Causes mortality rate in schizophrenia. The British Journal of Psychiatry 2000, 177:212-217. Australia. The epidemiological transition is yet to come 8. Saha S, Chant D, McGrath J: A systematic review of mortality in to the mental health sector. schizophrenia: is the differential mortality gap worsening over time?. Arch Gen Psychiatry 2007, 64:1123-1131. Adequate levels of (public and private) mental health 9. Harris EC, Barraclough B: Suicide as an outcome for mental disorders. A expenditure, when combined with efficacious mental meta-analysis. The British Journal of Psychiatry 1997, 170:205-228. health policy in Australia, will contribute to any lasting 10. Cantor C, Neulinger K: The epidemiology of suicide and attempted suicide among young Australians. Aust N Z J Psychiatry 2000, 34:370-387. decline in the mental health-related illness or mortality 11. Cantor CH, Neulinger K, de Leo D: Australian suicide trends 1964-youth trend. Finding services and treatments that “work” with and beyond?. Med J Aust 1997, 171:137-141. mental disordersisimportant, asisapplying them. 12. Caldwell TM, Jorm AF, Dear KB: Suicide and mental health in rural, remote and metropolitan areas in Australia. Med J Aust 2004, 181:S10-14. Thus, there is a need for both basic and applied 13. Wilkinson D, Gunnell D: Youth suicide trends in Australian metropolitan research on the causes of mental illness, and the devel- and non-metropolitan areas, 1988-1997. Aust N Z J Psychiatry 2000, opment of efficacious therapies. 34:822-828. 14. National Health and Medical Research Council, Commonwealth Department of Health and Aged Care: National Youth Suicide Prevention Strategy: Setting the Evidence-Based Research Agenda for Australia (A Literature Review). Acknowledgements Canberra 1999. The authors acknowledge the assistance of Roman Scheurer and Mario Gryl 15. de Leo D: Why are we not getting any closer to preventing suicide?. Br J in the preparation of this article. The authors also gratefully acknowledge Psychiatry 2002, 181:372-37. the helpful comments of two anonymous referees. Needless to say, any 16. de Leo D, Heller TS, eds: Suicide in Queensland 1991-2004, Mortality Rates remaining errors are the responsibility of the authors. and Related Data Brisbane, Australian Institute for Suicide Research and Prevention 2004. Author details 17. de Leo D, Heller TS: Who are the kids who self-harm? An Australian self Australian Institute for Suicide Research and Prevention, Griffith University, report shool survey. Med J Aust 2004, 181:140-144. Mt Gravatt, Australia. Queensland Centre for Mental Health Research, School 18. Gunnell D, Middleton N: National suicide rates as an indicator of the of Population Health, The University of Queensland, Australia. School of effect of suicide on premature mortality. Lancet 2003, 362:961-962. Economics and Finance, Victoria University, Sunbury Campus, Australia. 19. Yip PSF, Liu KY, Law CK, Law YW: Social and economic burden of suicides in Hong Kong SAR: a year of life lost perspective. Crisis 2005, 26:156-159. Authors’ contributions 20. Dempsey M: Decline in tuberculosis: the death rate fails to tell the entire DD and RW had the idea of conceptualising the issue considered here. This story. Am Rev Tuberc 1947, 56:157-164. framework was outlined to HW, an early writer on measurement and data 21. Murray C, Lopez AD: The Global Burden of Disease: A Comprehensive on mental health sector outcomes. All authors have made substantial Assessment of Mortality and Disability from Diseases, Injuries and Risk Factors contributions to conception and design, or to acquisition of data, or analysis in 1990 and Projected to 2020 Cambridge, Harvard School of Public Health and interpretation of data. They have all been involved in drafting the manuscript and revising it critically for important intellectual content. All 22. Australian Bureau of Statistics: Causes of Death: Australia 2004 Canberra, ABS three authors have given final approval of the version to be published. Each author has also participated sufficiently in this article to take public 23. Australian Institute of Health and Welfare: State & Territories GRIM (General responsibility for the content of this article. Thus, all authors contributed to Record of Incidence of Mortality) Books Canberra, AIHW 2006. this paper. 24. Doessel DP, Williams RFG, Whiteford H: A re-assessment of suicide None of our employers had any role in the conception and conduct of the measurement: some comparative PYLL-based trends in Queensland, study, the analysis of the data, or the first draft and subsequent revisions of Australia, 1920-2005. Crisis 2009, 30:6-12. the manuscript. All authors have reviewed and approved the final 25. Doessel DP, Williams RFG, Whiteford H: Policy-appropriate measurement manuscript, and take responsibility for the content of the paper. The work of suicide: headcount vs Potential Years of Life Lost, Australia, 1907- reported here is independent of all our employers. 2005. Archives of Suicide Research 2009, 13:87-99. 26. Doessel DP, Williams RFG, Robertson GJR, Whiteford H: Measuring an Competing interests Aspect of the Distribution of Mental Health: Suicide in Australia, 1907- The authors declare that they have no competing interests. 2003. Victoria University School of Applied Economics Working Paper Series, WP2006.9 2006. Received: 24 March 2009 27. Begg S, Vos T, Barker B, Stevenson C, Stanley L, Lopez A: The Burden of Accepted: 7 January 2010 Published: 7 January 2010 Disease and Injury in Australia, 2003 Canberra, AIHW 2007. 28. Silber J: Some applications of uncertainty theory. Soc Sci & Med 1982, 16:1663-1666. Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 10 of 10 http://www.anzhealthpolicy.com/content/7/1/3 29. Silber J: ELL (The equivalent length of life) or another attempt at 63. Easterlin RA: How beneficent is the market? A look at the modern history measuring development. World Dev 1983, 11:21-29. of mortality. The Reluctant Economist Cambridge: Cambridge University 30. Silber J: On inequality before death and life table: summary measures. PressEasterlin RA 2004, 101-138. Genus 1988, XLIV:25-39. 64. Bhattacharyya M, Layton A: Effectiveness of seat belt legislation on the 31. Le Grand J: Inequalities in health: some international comparisons. Eur Queensland road toll - An Australian case study in intervention analysis. Econ Rev 1987, 31:182-191. J Am Stat Assoc 1979, 74:596-603. 32. Illsley R, Le Grand J: The measurement of inequality in health. Economics 65. Young C: Mortality Pattern and Trends in Australia Canberra: AGPS 1976. and Health: Proceedings of a Conference London: MacmillanWilliams A 1987. 66. Metcalf AJ: The growth and development of public health services in 33. Le Grand J: An international comparison of distributions of age-at-death. Australia during 50 years. Med J Aust 1951, 1:45-51. Health Inequalities in European Countries Aldershot: GowerFox J 1989, 75-87. 67. Bodenheimer T: High and rising health care costs. Part 1: seeking an 34. Commonwealth of Australia: Living is for Everyone: A Framework for explanation. Ann Intern Med 2005, 142:847-854. Prevention of Suicide and Self-harm in Australia Canberra, Commonwealth of 68. Nordhaus WD: The health of nations: the contribution of improved Australia 2000. health to living standards. Measuring the Gains from Medical Research: An 35. Commonwealth of Australia: Living is for Everyone (LIFE) Framework Economic Approach Chicago: The University of Chicago PressMurphy KM, Canberra, Commonwealth of Australia 2008. Topel RH 2003, 9-40. 36. Newhouse JP: Medical care costs: how much welfare loss?. J Econ Persp 69. Davis K, Schoen C, Guterman S, Shih T, Schoenbaum SC, Weinbaum I: 1992, 6:3-21. Slowing the growth of U.S. health care expenditures: what are the 37. Glied S: Health care costs: on the rise again. J Econ Persp 2003, 17:125- options?. The Commonwealth Fund; mimeo 2007. 148. 70. Rutstein DD, Berenberg W, Chalmers TC, Child CG, Fishman AP, Perrin EB: 38. Pauly MV: Should we be worried about high real medical spending Measuring the quality of medical care. N Engl J Med 1976, 294:582-588. growth in the United States?. Health Aff 2003, Web Exclusive. 71. Holland WW: The “avoidable death” guide to Europe. Health Policy 1986, 39. Australian Institute of Health and Welfare: Health Expenditure Australia 2004- 6:115-117. 05 Canberra: AIHW 2006. 72. Holland WW, ed: European Community Atlas of “Avoidable Death”. 40. Organization for Economic Cooperation and Development: Health at a Commission of the European Communities Health Services Research Series No. Glance: OECD Indicators 2005 Paris: OECD 2005. 3 Oxford: Oxford University Press 1988. 41. Cutler D, Richardson E: Measuring the health of the U.S. population. 73. Nolte E, Scholz R, Shkolnikov V, McKee M: The contribution of medical Brookings Papers in Economic Activity: Microeconomics 1997, 217-271. care to changing life expectancy in Germany and Poland. Soc Sci & Med 42. Murphy KM, Topel RH: The economic value of medical research. 2002, 55:1905-1921. Measuring the Gains from Medical Research: An Economic Approach Chicago: 74. Korda RJ, Butler JR: Effect of healthcare on mortality: trends in avoidable The University of Chicago PressMurphy KM, Topel RH 2003, 41-73. mortality in Australia and comparisons with Western Europe. Public 43. Becker GS, Phillipson TJ, Soares RR: The quantity and quality of life and Health 2006, 120:95-105. the evolution of world inequality. Am Econ Rev 2005, 95:277-291. 75. Nolte E, McKee M: Does Health Care Save Lives? Avoidable Mortality Revisited 44. Cutler DM, Rosen AB, Vijan S: The value of medical spending in the London: The Nuffield Trust 2004. United States, 1960-2000. N Engl J Med 2006, 355:920-927. 76. Zeckhauser R, Shepard D: Where now for saving lives?. Law Contemp Probl 45. Hall RE, Jones CI: The value of the life and the rise in health spending. 1976, 40:5-45. Quart J Econ 2007, 122:39-72. 77. Andrews G: Meeting the unmet need with disease management. Unmet 46. Commonwealth Bureau of Census and Statistics: Population and Vital Need in Psychiatry: Problems, Resources, Responses Cambridge: Cambridge Statistics Melbourne: CBCS various 1916. University PressAndrews G, Henderson S 2000, 11-36. 47. Commonwealth Bureau of Census and Statistics: Australian Demography 78. Whiteford H: Unmet need: A challenge for governments. Unmet Need in Canberra: CBCS various 1922. Psychiatry: Problems, Resources, Responses Cambridge: Cambridge University 48. Commonwealth Bureau of Census and Statistics: Demography Canberra: PressAndrews G, Henderson S 2000, 8-10. CBCS various 1938. 79. Doessel DP, Williams RFG, Nolan P: A central dilemma in the mental 49. Commonwealth Bureau of Census and Statistics: Causes of Death Canberra: health sector: Structural imbalance. Cl Psych 2008, 12:57-66. CBCS various 1963. 80. Andrews G, the Tolkien II Team: Tolkien II – A Needs-based, Costed Stepped- 50. de Leo D: Suicide mortality data need revision [letter]. Med J Aust 2006, care Model for Mental Health Services Sydney: The University of New South 186:157-158. Wales at St Vincent’s Hospital and World Health Organization Collaborating 51. Taylor F: Guide to use of International Classifications of Diseases in Australia Centre for Classification in Mental Health 2006. Canberra: AIHW 1992. 81. Killackey E, Jorm A, Alvarez-Jimenez M, McCann TV, Hides L, Couineau A: 52. Australian Bureau of Statistics: Causes of Death: Australia 2004 Canberra: ABS Do we know what works, and if not why not?. ANZJP 2008, 42:439-444. 2006. 82. West JC, Wilk JE, Olfson M, Rae DS, Marcus S, Narrow WE, Pincus HA, 53. Cantor C, McTaggart P, De Leo D: Misclassification of suicide - the Regier DA: Patterns and quality of treatment for patients with contribution of opiates. Psychopath 2001, 34:140-146. schizophrenia in routine psychiatric practice. Psychiatr Serv 2005, 56:283- 54. De Leo D, Klieve H, Milner A: Suicide in Queensland 2002-04: mortality rates 291. and related data Brisbane: AISRAP 2006. doi:10.1186/1743-8462-7-3 55. Walker SM, Madden R, Chen L: Deaths due to suicide: the effects of Cite this article as: Doessel et al.: The trend in mental health-related certification and coding practices in Australia. ANZ J Publ Hlth 2008, mortality rates in Australia 1916-2004: implications for policy. Australia 32:126-130. and New Zealand Health Policy 2010 7:3. 56. Harrison JE, Pointer S, Elnour AA: A review of suicide statistics in Australia Canberra: Australian Institute of Health and Welfare 2009. 57. Elnour AA, Harrison J: Suicide decline in Australia: where did the cases go?. A NZ J Publ Hlth 2009, 33:67-69. 58. Australian Bureau of Statistics: Causes of death: Australia 2007 Canberra: ABS 59. Australian Bureau of Statistics: Suicides, Australia, 1994 to 2004 Canberra: ABS 60. Fries JF: Aging: Natural Death, and the Compression of Morbidity. N Engl JMed 1980, 303:130-105. 61. Lee RD: The demographic transition: Three centuries of fundamental change. J of Econ Persp 2003, 17:167-190. 62. Cutler D, Deaton A, Lleras-Muney A: The determinants of mortality. J Econ Lit 2006, 20:97-120. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australia and New Zealand Health Policy Springer Journals

The trend in mental health-related mortality rates in Australia 1916-2004: implications for policy

Loading next page...
 
/lp/springer-journals/the-trend-in-mental-health-related-mortality-rates-in-australia-1916-CRnRm8UdJP

References (124)

Publisher
Springer Journals
Copyright
Copyright © 2010 by Doessel et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Public Health; Social Policy
eISSN
1743-8462
DOI
10.1186/1743-8462-7-3
pmid
20145728
Publisher site
See Article on Publisher Site

Abstract

Background: This study determines the trend in mental health-related mortality (defined here as the aggregation of suicide and deaths coded as “mental/behavioural disorders”), and its relative numerical importance, and to argue that this has importance to policy-makers. Its results will have policy relevance because policy-makers have been predominantly concerned with cost-containment, but a re-appraisal of this issue is occurring, and the trade-off between health expenditures and valuable gains in longevity is being emphasised now. This study examines longevity gains from mental health-related interventions, or their absence, at the population level. The study sums mortality data for suicide and mental/behavioural disorders across the relevant ICD codes through time in Australia for the period 1916-2004. There are two measures applied to the mortality rates: the conventional age-standardised headcount; and the age-standardised Potential Years of Life Lost (PYLL), a measure of premature mortality. Mortality rates formed from these data are analysed via comparisons with mortality rates for All Causes, and with circulatory diseases, cancer and motor vehicle accidents, measured by both methods. Results: This study finds the temporal trend in mental health-related mortality rates (which reflects the longevity of people with mental illness) has worsened through time. There are no gains. This trend contrasts with the (known) gains in longevity from All Causes, and the gains from decreases achieved in previously rising mortality rates from circulatory diseases and motor vehicle accidents. Also, PYLL calculation shows mental health-related mortality is a proportionately greater cause of death compared with applying headcount metrics. Conclusions: There are several factors that could reverse this trend. First, improved access to interventions or therapies for mental disorders could decrease the mortality analysed here. Second, it is important also that new efficacious therapies for various mental disorders be developed. Furthermore, it is also important that suicide prevention strategies be implemented, particularly for at-risk groups. To bring the mental health sector into parity with many other parts of the health system will require knowledge of the causative factors that underlie mental disorders, which can, in turn, lead to efficacious therapies. As in any case of a knowledge deficit, what is needed are resources to address that knowledge gap. Conceiving the problem in this way, ie as a knowledge gap, indicates the crucial role of research and development activity. This term implies a concern, not simply with basic research, but also with applied research. It is commonplace in other sectors of the economy to emphasise the trichotomy of invention, innovation and diffusion of new products and processes. This three-fold conception is also relevant to addressing the knowledge gap in the mental health sector. Background indicates that mental disorders are the seventh most Burden of disease studies indicate that the impact of expensive disease category in Australia [4,5]. Various mental disorders is considerable [1-3], while the latest dimensions of mortality associated with mental disor- Australian Institute of Health and Welfare (AIHW) ders are not trivial. There are some meta-analyses indi- report on relative health expenditures by disease groups cating the excess mortality associated with these disorders–both natural and unnatural causes increase the risk of premature death for mentally ill people [6,7]. * Correspondence: ruth.williams@vu.edu.au Also, a single international meta-analysis, focussing only † Contributed equally School of Economics and Finance, Victoria University, Sunbury Campus, on suicide, shows a heightened suicide risk is associated Australia © 2010 Doessel et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 2 of 10 http://www.anzhealthpolicy.com/content/7/1/3 with almost all mental disorders [8]. Another approach– There are several policy implications in applying the that of the psychological autopsy–has found that about PYLL metric. The most important is that government 90 per cent of people who die by suicide have at least policies usually are designed to affect a particular vari- one mental disorder at the time of death [9]. Several able or target. When forming policy, or evaluating exist- Australian studies that have examined various aspects of ing policy, such as evaluating the expenditure on mortality from mental disorders are now available, their Australia’s National Suicide Prevention Strategy [34,35], focus largely being on suicide [10-17]. it is important to use appropriate measures, as discussed The present study measures the time-trend in the elsewhere [24-26], of the variable being targeted. The mental health-related mortality rate. The term “mental PYLL metric provides relevant information for societal health-related mortality” is defined as the sum of deaths or policy issues, because it is a weighted measure (Table from mental and behavioural disorders and suicide. 1). This example shows clearly that the PYLL is a more Apart from the conventional headcount measure, some appropriate measure of premature mortality, from a studies have applied an alternative measure, the poten- societal perspective, than the (equal) headcount tial years of life lost (PYLL) to suicide [18,19]. The measure. PYLL metric originated in the 1940s for the evaluation Another policy implication relates to an argument of tuberculosis prevention programs, when it had from health economics. There has been considerable become apparent that headcount (only) measurement of concern about the rising absolute and relative expendi- mortality did not convey all the information relevant to tures of health services [36-38]. For example, the focus the prevention of tuberculosis mortality [20]. The PYLL from this perspective is that Australia’s expenditure on metric subsequently achieved prominence in the burden health in 2004-05 was 8 per cent of Gross Domestic of disease work of Murray and Lopez [21]. Currently, it Product (GDP), whereas in 1960-61 it was 4.1 per cent is routine practice for the Australian Bureau of Statistics of GDP [39]. The OECD average in 2003 was 8.8 per (ABS) [22] and the Australian Institute of Health and cent of GDP. Thus, comparatively, Australia’s position is Welfare(AIHW)[23]toreportbothheadcount and “in the middle”, between the United States (15.0 per PYLL measures of suicide. A small number of Australian cent) and the United Kingdom (7.7 per cent). Even Aus- analyses have applied both headcount and PYLL mea- tralia’s “middle” position is viewed with some concern, surement to suicide [24-26]. These studies show the as other countries in a lower position must use their added information gained by applying both headcount health resources differently from those above them [40]. and PYLL metrics. Various economists, governments and others (e.g. insur- The focus here is on providing historical and compara- ance carriers) have adopted a cost-containment view. tive (with respect to other diseases/conditions) analyses However, in the recent international literature on the through time, using headcount and PYLL measures. The economics of health services, the cost-containment measure of mental health-related mortality that we apply emphasis has been subject to re-appraisal. This re- here involves summing across the relevant ICD codes appraisal involves an examination of the contribution of through time both “suicide” and mortality from “mental the health sector in the totality of the economies of and behavioural disorders”.Wetakethese combined OECD-type countries. Scholars of the re-appraisal bring causes of death to approximate the (mortality) size of the a different emphasis–it is argued that due regard must problem of mental health-related mortality. be given to the gains to health that both public health The recent emphasis in burden of disease studies sug- programs and clinical medicine have wrought [41-45]. gests that measuring morbidity as well as mortality is In this context, accurate measurement of “the gains” is important. However, a limitation of the (Australian) bur- vital. The Discussion section below develops this point. den of disease work is its cross-sectional nature; that is, data have been constructed for only two years, 1996 and Methods 2003 [1,27]. This article examines 88 years of data. In order to extract data on mental health-related mor- The present study has a narrower focus, by measuring tality, annual data were summed across the relevant mortality in levels only; the focus does not extend to ICD codes for “suicide” and “mental/behavioural disor- studying its distribution. Measuring the (age) distribu- ders”, by five year age-groups, from the AIHW [23] and tion of mortality (due to all causes, or any cause) for the years prior to 1968, from ABS historical data through time is possible, by applying the analytical fra- [46-49]. The data set thus obtained provides a complete mework pioneered by Silber [28-30] and Le Grand enumeration, though underestimated [50], of mental [31-33]; there is one Australian study which has mea- health-related mortality in Australia; it is not a sample. sured the distribution of suicide [26]. Distribution topics Given the various changes in the structure of, and parti- in mental health-related mortality need further atten- cular codes in, the revisions of the ICD, we referred to tion, but here the focus is on levels only. Taylor’s detailed accounts to guide the data recording Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 3 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Table 1 Two approaches to mortality measurement: the measures increased due to this exercise, though the PYLL and the Headcount magnitude was very small both for headcounts and for Case The PYLL metric The count metric PYLLs, because there were so few cases in the “undeter- mined” category. Hence, no substantive conclusions (Given life expectancy US white women, drawn in that paper were found to be affected by that 1940s, of 69 years) under-reporting problem. Death of white 45 years of life 1 death However, other problems exist in the accuracy of sui- woman, aged 24 cide data. It is very important to realise that our PYLL Death of white 7 years of life 1 death calculations are based on the mortality coding underta- woman, aged 62 ken (and published) by the ABS. It has been known for Source: Adapted from Dempsey[20] some time that ABS data on suicide have been subject to misclassification [53], as well as other more general [51]. Australian data on suicide start from 1907, but errors indicated by De Leo [40], De Leo, Klieve and Mil- data on mortality from mental/behavioural disorders are ner [54] and Walker, Madden and Chen [55]. These available only from 1916. The data on All Causes, can- problems have become more important since 2000. cer and circulatory diseases are available from 1907. Harrison, Painter and Elnour [56] have undertaken an However, motor vehicle accident mortality data exist important study in which they re-worked the published only from 1924. ABS data for a single year, 2004. They found that the The above data are of a headcount kind, and they are published ABS data for Australia underestimated suicide employed in the calculation of PYLLs. The ABS provides enumeration by 16 per cent. By far, the most important one account of the processes applied to headcount data source of error was the non-inclusion of coroners’ cases in order to generate PYLL data [52]. The following when they were incomplete at the time of ABS enu- equation conveys the method. It specifies the PYLL for meration. (See also Elnour and Harrison [57].) The ABS Suicide for year t, and an assumed life expectancy of 75 for some years has published a “Caution” relating to the years: quality of published “cause of death” data and has announced changes to its processes of coding and pub- AS lishing suicide data [58]. The Caution reads as follows: Suicide PYLL () 75 tD=− [ (75 A )] ss The level of recorded deaths attributed to suicide, and i =1 observed changes over time are likely to have been where Suicide PYLL(75) is the total PYLLs due to affected by delays in finalising a cause [59]. The ABS AS Suicide, age-standardised, at time period t; D is the has, from 2009, commenced a process of revising suicide data as more information, such as coroners’ decisions, number of Suicides per age group; A is the adjusted age becomes available. at death due to Suicide per age group; and i is the num- However, the impact of this inaccuracy in the data for ber of age groups for i = 1, ... n. In the present study, the 85+ age group is open in the this study, and thus for PYLL calculations in general, is raw data. We closed it at 100 years of age in our PYLL not known. This is because the age-distribution of that calculations. Also, we applied conventional age standar- under-reporting is not known. Until known, the implica- disation techniques, using the Australian Standard tion for PYLL measurement cannot be determined. Population 1991, both to the headcount and the PYLL data sets. Results Earlier in this Section the problem of under-estima- We report mortality rates (i.e. conventional rates based tion of suicide data, which is a well-documented pro- on headcount data) first. blem, was raised. We have, elsewhere, quantitatively Mortality rates: mental health-related and some investigated one implication of this problem for PYLL other causes measurement [25]. That very specific implication relates Figure 1 puts the mental health-related mortality trend to the eighth and subsequent revisions of the ICD, in perspective by presenting a line graph of the All which involved new codes for “deaths undetermined Causes mortality rate as well as comparative mortality whether accidentally or purposely inflicted” (ICD codes rates associated with some causes of death, viz.mental E980-E989 in ICD-8 and ICD-9, and Y10-Y34 in ICD- health-related (as defined here), cancer, circulatory dis- 10). We noted that non-inclusion of such deaths in that eases and motor vehicle accidents. Given the large dif- study [25], subsequent to 1968, could under-estimate ferences in age-standardised rates, the figure has two the contribution of suicide, and thus overall mental- parts, Part (a) showing a long-run decrease in the All related mortality. Hence, we re-analysed the post-1967 Causes mortality rate. This reflects the experience of data with these additional codes included. We found all many countries. Some argue that the nineteenth century Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 4 of 10 http://www.anzhealthpolicy.com/content/7/1/3 witnessed a transition phase, and then a period of vir- nearly 5 per cent of all causes, and has remained at that tually continuous decline in mortality, and that this approximate level in subsequent years. trend is an “epidemiological transition” or a “demo- While it is not our purpose to explain the trends graphic transition” [60-63]. Part (b) depicts the mortality depicted in Figures 1 and 2, it is useful to reflect on rates for four specific causes: circulatory diseases; can- some known factors. The rise of motor transport early cers; motor vehicle accidents; and mental health-related. last century resulted in increased accident mortality Circulatory disease is by far the largest single cause of through time, but governments eventually undertook death and cancer is the second most important; mental many interventions, such as seat-belt legislation, crash health-related mortality and motor vehicle accidents are helmets, safety-designed roads, drink-driving legislation “small” by comparison. Note that there are differing etc. Some of these interventions are efficacious [64]. temporal trends in these causes. The mortality rate for Similarly, for ischaemic heart disease and related condi- cancer has declined slightly over recent years. Mortality tions [65], medical treatments have improved and from circulatory diseases initially rose, but a large awareness has increased due to interventions undertaken decline from 1968 has occurred. Mortality arising from by governments, particularly in the form of health infor- motor vehicle accidents also was once rising, but it has mation. While the data on mortality depicted here do declined after 1978. The fall in mortality due to both not show the remarkable downturn in other causes of these causes has been substantial, but this is not the death, such as infectious diseases [66], the implementa- case with mental health-related mortality. The rising tion of public health measures was successful in these trend due to mental health-related mortality is note- areas [67]. In this context, note that the first year of the worthy. It was relatively high in the 1920s, but it fell allocation of Australian Government funds to suicide during the World War II period, and, since then, has prevention was 1995. risen, reaching a (local) maximum in 1996. The motor The message from both Figures 1 and 2 is clear: there vehicle accident trend exceeded that for mental health- is no evidence that the prevention of mental health- related mortality from the 1930s to the 1980s (and often related mortality is working (as measured by the sum- by a substantial margin), but since 1983 the mortality mation of deaths from mental/behavioural disorders and rate from motor vehicle accidents is less than that for suicide). That is, no evidence is found that the demo- mental health-related mortality. graphic transition is presently operative for mental- Percentage contributions relative to All Causes health related mortality. Figure 2 depict the proportionate “shares” (or the per- This conclusion is in line with those scholars who are centage contributions) for each of these causes of death concerned with analysing “avoidable mortality”.Suicide relative to All Causes through time. The calculation and deaths from mental/behavioural disorders have not involved the ratio of each of the above four causes of been included in any list of “avoidable death”.Wepost- death to All Causes, expressed as a percentage. pone further consideration of this point to the Discus- Figure 2 is also in two parts. With very large differ- sion section below. We now report the results from ences in the proportions for the particular causes pre- applying a PYLL metric. sented here, the two-part depiction is helpful. The scale The PYLL results of the Y-axis in Part (a) is in levels, and is logarithmic Figure 3 as with Figure 2 is concerned with comparison. in Part (b). The latter scale makes the detail about the However, we now focus on mental health-related mor- temporal trends in the “smaller” causes of death clearer. tality, and compare the two measures of mortality (the Part (a) of Figure 2 shows that, during the 1950s-80s, headcount measure and the PYLL measure) of this circulatory diseases accounted for at least half the causes aggregated cause of death. We have calculated the per- of death but that, since then, the relative fall in this centage shares of mental health-related mortality to All cause of death is very marked, and accounted for 35 per Causes mortality (measured by a headcount and by cent of deaths in 2004. The relative importance of the PYLLs). The percentage share for “The Headcount Mea- contribution of cancer to all mortality has increased. sure” in Figure 3 is the same as the “Mental Health- Cancer accounted for approximately 10 per cent of all Related” share in Figure 2, given that the measure of the deaths in the1940s,whereas thecomparablefigurewas rates of Figure 2 is the headcount. nearly 30 per cent of all deaths by 2004. Although the It is clear that both measures of mental health-related trends through time in the contributions of motor vehi- mortality exhibit the same general patterns through cle mortality rates and mental-health related mortality time. Both measures show that the relative (numerical) rates appear different from each other, the graphs in importance of mental health-related mortality has risen Part (a) are difficult to interpret. Part (b) clarifies the since the end of World War II. However, the PYLL picture. In 1996, the relative importance of mental/beha- measure (which reaches a local maximum of nearly 14 vioural disorders and suicide reached a maximum of per cent of All Causes mortality in 1998) indicates that Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 5 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 1 Age-standardised mortality rates for (a) total mortality (all causes combined), and (b) circulatory diseases, cancers, mental health-related mortality and motor vehicle accidents, Australia, persons, 1907-2004. * These rates have been standardised to the age distribution of the 1991 Australian population. ‡ Mental-Related Mortality includes Mental/Behavioural Disorders and Suicide. Sources: AIHW[23], CBCS [46-49], Taylor[52] Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 6 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 2 Shares of five causes of death to all deaths, measured by counts, Australia, persons, 1907-2004. The shaded vertical lines indicate the years of implementation of revisions of the ICD. Sources: AIHWAIHW[23], CBCS [46-49], Taylor[52] Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 7 of 10 http://www.anzhealthpolicy.com/content/7/1/3 Figure 3 Mental health-related mortality as a percentage of all causes measured by the count measure (no.) and the potential years of life lost (PYLL) measure, Australia, persons, 1916-2004. Sources: AIHWAIHW[23], CBCS [46-49], Taylor[52] the problems in the mental health sector are consider- indicates that the underestimation has been more severe ably larger than indicated by “The Count Measure” since 2000 [56]. Thedatainaccuracyissueisaddressed (nearly 5 per cent in 1998). in more detail above, and Figure 3 reflects this problem. It may be thought that the apparently large downturn The PYLL measure particularly is accentuating the data which can be observed in Figure 3 post-1998 in the inaccuracy of recent years. PYLLs trend is of some importance, to the extent that Thus, the conclusions that one draws about the the conclusions of the article are contradicted. In the numerical importance of mental health-related mortality following paragraphs we argue that any such position is depends on the mortality measure employed. The head- not the case. First, we have elsewhere reported the count measure is the typical measure applied to the results of estimating equations to time-series data, in an mental health sector: the results in Figure 3 clearly indi- article that focuses on the distribution of suicide [26]. In cate that this measure of mortality underestimates the that paper, where the reported equations are for rates, size of the problems associated with mental disorders in both on headcount data and on PYLL data, we find that Australia. We undertook PYLL calculations for the con- the trend is not cubic: there is no downturn in those tributions of the other three causes of death but we do data. In other words, the Ramsey RESET test indicated not report these results here due to space limitations. that the post-1998 decrease was not statistically signifi- These results are available from the authors on request. cant. Given that we subjected those equations to a full Attention should be paid to both headcount and PYLL raft of diagnostic tests of the residuals, and the stability measures because each sheds light on different aspects of the specification of the model, considerable confi- of the phenomenon. dence can be placed in those results. Figure 3 is not a depiction of rates, but the trend in Discussion the contribution of mental health-related mortality to As argued briefly in the Background section of this arti- All Causes mortality. However, in the Methods section cle, we suggest that it is misplaced for policy makers to above, we explain that, for various reasons, the number have a concern solely with health expenditure. of (published) suicides has been falling recently. Not Attention should be directed to both health expendi- even the ABS believes that “less suicide” is actually hap- ture and the value of the health outputs produced by pening. Rather, what is happening is inaccuracies in the the health sector. A statement by Nordhaus neatly cap- suicide data have been worsening recently. Recall that a tures this perspective, as follows: “The new view of re-working of the mortality data for 2004, the last year health economics should shape the way we think about of our analysis here, indicated that the underestimation health policy. In the early 1990s, the general hysteria of the published data for that year amounted to 16 per about rising health costs led many to believe that the cent. Furthermore, there is other evidence which health care system was wasteful and out of control and Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 8 of 10 http://www.anzhealthpolicy.com/content/7/1/3 should be reigned in” [p. 20] [68]. There is not the space resources to averting suicide and all mental-related mor- to describe or review the reappraisal studies, but note tality, or passivity towards this cause of death, more so that Davis et al. have argued cogently that this perspec- since it is classified as “unavoidable”.The argument in tive is important, andthatitdoesnot negate thecom- this paper refutes any such stance. mon criticisms levelled at the health sector, eg poor The second point is technical, and relates to having access, inappropriate treatment, issues arising with mar- relevant data available when studies to evaluate the effi- ketpower etc[69]. The point of thepresent articleis cacy of prevention strategies are sought. Given the sug- made in this context: information about the relative gestion that efficacious government policy can numerical importance of mental disorders is very contribute to reversing the mortality trend from mental important. disorders, could such an impact be detected statistically? These arguments suggest reflecting on the factors Bhattacharyya and Layton provide one example [64]. known already to contribute to the long-run decrease in The task involves detecting (post-intervention) whether mortality, which is characteristic of “the demographic or not there has been a reversal in the sign on the slope transition”. One key factor is knowledge of disease pro- variable of the equation for the mortality trend (i.e. cesses. For example, in nineteenth century England, the from positive to negative). The above results also indi- observational disposition of John Snow towards the cate the importance of observing the sign on the slope water supply (wells etc.) ultimately provided the relevant of all relevant trends: e.g. the age-standardised head- knowledge of water contamination for the prevention of count rate, etc. Thus, the appropriate technique exists cholera; and knowledge of efficacious therapies, such as for establishing empirically the impact of a government the Fleming-Florey “story” of the development of knowl- policy on prevention of mental health-related mortality. edge about antibiotics, is an example, of a different kind, Third, as mentioned above, mortality is but a partial, of how knowledge is applied. A more recent factor is and imperfect, measure of the health status of a com- technological change. Technological change involves munity. Thus, this point is a qualification. Since Zec- both life-saving technologies, such as organ transplanta- khauser and Shepard [76] outlined the Quality Adjusted tion, and “maintenance” types of technologies, such as Life Year (QALY) concept, it is regular practice to con- dialysis for end-stage renal disease. Some technological sider the quality of life associated with morbidity, along change enhances the productivity of curative and pre- with mortality. While the formation of time-series data ventive health services. This factor is far from trivial. sets of mental health-related morbidity is desirable, the Consider extra-corporeal shockwave lithotripsy, extra- “quality” of morbidity is a “gap” in health data-sets, and capsular cataract extraction and phaco-emulsification it is rarely discussed. The absence of such a time-series cataract therapy, which are three technologies of this data set on morbidity has induced us to have recourse kind. to mortality data: such data are available for a relatively Additionally, there are the technological changes that long period. The results presented in this paper clearly have occurred in diagnosis. For example, whereas once indicate the importance of determining the shape and there was just radiology, there is now also MRI, multi- direction in the long-term trend in mortality from this slice CT scanning, Dopler ultrasound, PET scanning, specific cause of death. Determining even this much gamma camera imaging etc. The point of this paragraph information is not trivial, even though the trend in the is to emphasise the implication of this paper which is morbidity from mental conditions cannot be determined that an appropriate focus in research about the mental over that period. health sector is to determine, and implement, the factors Our final point is a qualification. The argument in our that will contribute to the long-run decrease in mortality paper does not negate some other very important issues in the mental health sector. in the mental health sector. Such issues include unmet We now qualify the argument here with four points. need in the provision of mental health services [77-79], The first relates to the notion of “avoidable mortality” and the matter of people with mental disorders not or “amenable mortality”. Since Rutstein et al. [70], receiving efficacious, evidence-based treatments [80-82]. numerous scholars have formed lists of diseases/condi- This paper is an exercise in descriptive science and it is tions for which medical or societal interventions are effi- not our purpose to take any normative stances or to cacious [71-74]. Nolte and McKee provide a enter the debates about these issues. Rather, we seek comprehensive review [75]. It is noteworthy that suicide measurement approaches that will inform policy debates and mortality from mental/behavioural disorders are not better. included in any list of avoidable deaths: deaths from mental illnesses are classified as “unavoidable” in that Conclusions literature. This classification can be confused with an It is unbalanced to focus solely on rising health expendi- implicit attitude of resistance towards allocating tures, without valuing the improvements in health status Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 9 of 10 http://www.anzhealthpolicy.com/content/7/1/3 References broughtbythose expenditures:itisimportanttocon- 1. Mathers C, Vos T, Stevenson C: The Burden of Disease and Injury in Australia sider health gains alongside expenditures. A re-thinking Canberra, Australian Institute of Health and Welfare 1999. of the conventional wisdom about cost-containment 2. Mathers C, Vos ET, Stevenson CE, Begg SJ: The Australian burden of disease study: Measuring the loss of health from diseases, injuries and does not imply that various criticisms of the provision risk factors. Med J Aust 2000, 172:592-596. of health services no longer apply, but it does suggest 3. Vos T, Mathers CD: Burden of mental disorders: Australian and Global that determining the place of mental disorders in the Burden of Disease studies. Bull World Health Organ 2000, 78:427-438. 4. Australian Institute of Health and Welfare: Health System Expenditure on context of the reappraisal literature is relevant. The Disease and Injury in Australia 2000-01 Canberra, AIHW 2004. above results are unambiguous: there is no evidence 5. Australian Institute of Health and Welfare: Health System Expenditure on that long-run mental health-related mortality is falling Disease and Injury in Australia 2000-01 - Update Tables and Figures AIHW 2004http://www.aihw.gov.au/publications/index.cfm/title/10006, (accessed (as measured here by the summation of deaths from 9/10/2006). mental/behavioural disorders and suicide) and that sui- 6. Brown S: Excess mortality of schizophrenia. The British Journal of Psychiatry cide and mental/behavioural mortality has not contribu- 1997, 171:502-508. 7. Brown S, Barraclough B, Inskip H: Causes of the excess mortality of ted to the decline in the All Causes mortality rate in schizophrenia. The British Journal of Psychiatry 2000, 177:212-217. Australia. The epidemiological transition is yet to come 8. Saha S, Chant D, McGrath J: A systematic review of mortality in to the mental health sector. schizophrenia: is the differential mortality gap worsening over time?. Arch Gen Psychiatry 2007, 64:1123-1131. Adequate levels of (public and private) mental health 9. Harris EC, Barraclough B: Suicide as an outcome for mental disorders. A expenditure, when combined with efficacious mental meta-analysis. The British Journal of Psychiatry 1997, 170:205-228. health policy in Australia, will contribute to any lasting 10. Cantor C, Neulinger K: The epidemiology of suicide and attempted suicide among young Australians. Aust N Z J Psychiatry 2000, 34:370-387. decline in the mental health-related illness or mortality 11. Cantor CH, Neulinger K, de Leo D: Australian suicide trends 1964-youth trend. Finding services and treatments that “work” with and beyond?. Med J Aust 1997, 171:137-141. mental disordersisimportant, asisapplying them. 12. Caldwell TM, Jorm AF, Dear KB: Suicide and mental health in rural, remote and metropolitan areas in Australia. Med J Aust 2004, 181:S10-14. Thus, there is a need for both basic and applied 13. Wilkinson D, Gunnell D: Youth suicide trends in Australian metropolitan research on the causes of mental illness, and the devel- and non-metropolitan areas, 1988-1997. Aust N Z J Psychiatry 2000, opment of efficacious therapies. 34:822-828. 14. National Health and Medical Research Council, Commonwealth Department of Health and Aged Care: National Youth Suicide Prevention Strategy: Setting the Evidence-Based Research Agenda for Australia (A Literature Review). Acknowledgements Canberra 1999. The authors acknowledge the assistance of Roman Scheurer and Mario Gryl 15. de Leo D: Why are we not getting any closer to preventing suicide?. Br J in the preparation of this article. The authors also gratefully acknowledge Psychiatry 2002, 181:372-37. the helpful comments of two anonymous referees. Needless to say, any 16. de Leo D, Heller TS, eds: Suicide in Queensland 1991-2004, Mortality Rates remaining errors are the responsibility of the authors. and Related Data Brisbane, Australian Institute for Suicide Research and Prevention 2004. Author details 17. de Leo D, Heller TS: Who are the kids who self-harm? An Australian self Australian Institute for Suicide Research and Prevention, Griffith University, report shool survey. Med J Aust 2004, 181:140-144. Mt Gravatt, Australia. Queensland Centre for Mental Health Research, School 18. Gunnell D, Middleton N: National suicide rates as an indicator of the of Population Health, The University of Queensland, Australia. School of effect of suicide on premature mortality. Lancet 2003, 362:961-962. Economics and Finance, Victoria University, Sunbury Campus, Australia. 19. Yip PSF, Liu KY, Law CK, Law YW: Social and economic burden of suicides in Hong Kong SAR: a year of life lost perspective. Crisis 2005, 26:156-159. Authors’ contributions 20. Dempsey M: Decline in tuberculosis: the death rate fails to tell the entire DD and RW had the idea of conceptualising the issue considered here. This story. Am Rev Tuberc 1947, 56:157-164. framework was outlined to HW, an early writer on measurement and data 21. Murray C, Lopez AD: The Global Burden of Disease: A Comprehensive on mental health sector outcomes. All authors have made substantial Assessment of Mortality and Disability from Diseases, Injuries and Risk Factors contributions to conception and design, or to acquisition of data, or analysis in 1990 and Projected to 2020 Cambridge, Harvard School of Public Health and interpretation of data. They have all been involved in drafting the manuscript and revising it critically for important intellectual content. All 22. Australian Bureau of Statistics: Causes of Death: Australia 2004 Canberra, ABS three authors have given final approval of the version to be published. Each author has also participated sufficiently in this article to take public 23. Australian Institute of Health and Welfare: State & Territories GRIM (General responsibility for the content of this article. Thus, all authors contributed to Record of Incidence of Mortality) Books Canberra, AIHW 2006. this paper. 24. Doessel DP, Williams RFG, Whiteford H: A re-assessment of suicide None of our employers had any role in the conception and conduct of the measurement: some comparative PYLL-based trends in Queensland, study, the analysis of the data, or the first draft and subsequent revisions of Australia, 1920-2005. Crisis 2009, 30:6-12. the manuscript. All authors have reviewed and approved the final 25. Doessel DP, Williams RFG, Whiteford H: Policy-appropriate measurement manuscript, and take responsibility for the content of the paper. The work of suicide: headcount vs Potential Years of Life Lost, Australia, 1907- reported here is independent of all our employers. 2005. Archives of Suicide Research 2009, 13:87-99. 26. Doessel DP, Williams RFG, Robertson GJR, Whiteford H: Measuring an Competing interests Aspect of the Distribution of Mental Health: Suicide in Australia, 1907- The authors declare that they have no competing interests. 2003. Victoria University School of Applied Economics Working Paper Series, WP2006.9 2006. Received: 24 March 2009 27. Begg S, Vos T, Barker B, Stevenson C, Stanley L, Lopez A: The Burden of Accepted: 7 January 2010 Published: 7 January 2010 Disease and Injury in Australia, 2003 Canberra, AIHW 2007. 28. Silber J: Some applications of uncertainty theory. Soc Sci & Med 1982, 16:1663-1666. Doessel et al. Australia and New Zealand Health Policy 2010, 7:3 Page 10 of 10 http://www.anzhealthpolicy.com/content/7/1/3 29. Silber J: ELL (The equivalent length of life) or another attempt at 63. Easterlin RA: How beneficent is the market? A look at the modern history measuring development. World Dev 1983, 11:21-29. of mortality. The Reluctant Economist Cambridge: Cambridge University 30. Silber J: On inequality before death and life table: summary measures. PressEasterlin RA 2004, 101-138. Genus 1988, XLIV:25-39. 64. Bhattacharyya M, Layton A: Effectiveness of seat belt legislation on the 31. Le Grand J: Inequalities in health: some international comparisons. Eur Queensland road toll - An Australian case study in intervention analysis. Econ Rev 1987, 31:182-191. J Am Stat Assoc 1979, 74:596-603. 32. Illsley R, Le Grand J: The measurement of inequality in health. Economics 65. Young C: Mortality Pattern and Trends in Australia Canberra: AGPS 1976. and Health: Proceedings of a Conference London: MacmillanWilliams A 1987. 66. Metcalf AJ: The growth and development of public health services in 33. Le Grand J: An international comparison of distributions of age-at-death. Australia during 50 years. Med J Aust 1951, 1:45-51. Health Inequalities in European Countries Aldershot: GowerFox J 1989, 75-87. 67. Bodenheimer T: High and rising health care costs. Part 1: seeking an 34. Commonwealth of Australia: Living is for Everyone: A Framework for explanation. Ann Intern Med 2005, 142:847-854. Prevention of Suicide and Self-harm in Australia Canberra, Commonwealth of 68. Nordhaus WD: The health of nations: the contribution of improved Australia 2000. health to living standards. Measuring the Gains from Medical Research: An 35. Commonwealth of Australia: Living is for Everyone (LIFE) Framework Economic Approach Chicago: The University of Chicago PressMurphy KM, Canberra, Commonwealth of Australia 2008. Topel RH 2003, 9-40. 36. Newhouse JP: Medical care costs: how much welfare loss?. J Econ Persp 69. Davis K, Schoen C, Guterman S, Shih T, Schoenbaum SC, Weinbaum I: 1992, 6:3-21. Slowing the growth of U.S. health care expenditures: what are the 37. Glied S: Health care costs: on the rise again. J Econ Persp 2003, 17:125- options?. The Commonwealth Fund; mimeo 2007. 148. 70. Rutstein DD, Berenberg W, Chalmers TC, Child CG, Fishman AP, Perrin EB: 38. Pauly MV: Should we be worried about high real medical spending Measuring the quality of medical care. N Engl J Med 1976, 294:582-588. growth in the United States?. Health Aff 2003, Web Exclusive. 71. Holland WW: The “avoidable death” guide to Europe. Health Policy 1986, 39. Australian Institute of Health and Welfare: Health Expenditure Australia 2004- 6:115-117. 05 Canberra: AIHW 2006. 72. Holland WW, ed: European Community Atlas of “Avoidable Death”. 40. Organization for Economic Cooperation and Development: Health at a Commission of the European Communities Health Services Research Series No. Glance: OECD Indicators 2005 Paris: OECD 2005. 3 Oxford: Oxford University Press 1988. 41. Cutler D, Richardson E: Measuring the health of the U.S. population. 73. Nolte E, Scholz R, Shkolnikov V, McKee M: The contribution of medical Brookings Papers in Economic Activity: Microeconomics 1997, 217-271. care to changing life expectancy in Germany and Poland. Soc Sci & Med 42. Murphy KM, Topel RH: The economic value of medical research. 2002, 55:1905-1921. Measuring the Gains from Medical Research: An Economic Approach Chicago: 74. Korda RJ, Butler JR: Effect of healthcare on mortality: trends in avoidable The University of Chicago PressMurphy KM, Topel RH 2003, 41-73. mortality in Australia and comparisons with Western Europe. Public 43. Becker GS, Phillipson TJ, Soares RR: The quantity and quality of life and Health 2006, 120:95-105. the evolution of world inequality. Am Econ Rev 2005, 95:277-291. 75. Nolte E, McKee M: Does Health Care Save Lives? Avoidable Mortality Revisited 44. Cutler DM, Rosen AB, Vijan S: The value of medical spending in the London: The Nuffield Trust 2004. United States, 1960-2000. N Engl J Med 2006, 355:920-927. 76. Zeckhauser R, Shepard D: Where now for saving lives?. Law Contemp Probl 45. Hall RE, Jones CI: The value of the life and the rise in health spending. 1976, 40:5-45. Quart J Econ 2007, 122:39-72. 77. Andrews G: Meeting the unmet need with disease management. Unmet 46. Commonwealth Bureau of Census and Statistics: Population and Vital Need in Psychiatry: Problems, Resources, Responses Cambridge: Cambridge Statistics Melbourne: CBCS various 1916. University PressAndrews G, Henderson S 2000, 11-36. 47. Commonwealth Bureau of Census and Statistics: Australian Demography 78. Whiteford H: Unmet need: A challenge for governments. Unmet Need in Canberra: CBCS various 1922. Psychiatry: Problems, Resources, Responses Cambridge: Cambridge University 48. Commonwealth Bureau of Census and Statistics: Demography Canberra: PressAndrews G, Henderson S 2000, 8-10. CBCS various 1938. 79. Doessel DP, Williams RFG, Nolan P: A central dilemma in the mental 49. Commonwealth Bureau of Census and Statistics: Causes of Death Canberra: health sector: Structural imbalance. Cl Psych 2008, 12:57-66. CBCS various 1963. 80. Andrews G, the Tolkien II Team: Tolkien II – A Needs-based, Costed Stepped- 50. de Leo D: Suicide mortality data need revision [letter]. Med J Aust 2006, care Model for Mental Health Services Sydney: The University of New South 186:157-158. Wales at St Vincent’s Hospital and World Health Organization Collaborating 51. Taylor F: Guide to use of International Classifications of Diseases in Australia Centre for Classification in Mental Health 2006. Canberra: AIHW 1992. 81. Killackey E, Jorm A, Alvarez-Jimenez M, McCann TV, Hides L, Couineau A: 52. Australian Bureau of Statistics: Causes of Death: Australia 2004 Canberra: ABS Do we know what works, and if not why not?. ANZJP 2008, 42:439-444. 2006. 82. West JC, Wilk JE, Olfson M, Rae DS, Marcus S, Narrow WE, Pincus HA, 53. Cantor C, McTaggart P, De Leo D: Misclassification of suicide - the Regier DA: Patterns and quality of treatment for patients with contribution of opiates. Psychopath 2001, 34:140-146. schizophrenia in routine psychiatric practice. Psychiatr Serv 2005, 56:283- 54. De Leo D, Klieve H, Milner A: Suicide in Queensland 2002-04: mortality rates 291. and related data Brisbane: AISRAP 2006. doi:10.1186/1743-8462-7-3 55. Walker SM, Madden R, Chen L: Deaths due to suicide: the effects of Cite this article as: Doessel et al.: The trend in mental health-related certification and coding practices in Australia. ANZ J Publ Hlth 2008, mortality rates in Australia 1916-2004: implications for policy. Australia 32:126-130. and New Zealand Health Policy 2010 7:3. 56. Harrison JE, Pointer S, Elnour AA: A review of suicide statistics in Australia Canberra: Australian Institute of Health and Welfare 2009. 57. Elnour AA, Harrison J: Suicide decline in Australia: where did the cases go?. A NZ J Publ Hlth 2009, 33:67-69. 58. Australian Bureau of Statistics: Causes of death: Australia 2007 Canberra: ABS 59. Australian Bureau of Statistics: Suicides, Australia, 1994 to 2004 Canberra: ABS 60. Fries JF: Aging: Natural Death, and the Compression of Morbidity. N Engl JMed 1980, 303:130-105. 61. Lee RD: The demographic transition: Three centuries of fundamental change. J of Econ Persp 2003, 17:167-190. 62. Cutler D, Deaton A, Lleras-Muney A: The determinants of mortality. J Econ Lit 2006, 20:97-120.

Journal

Australia and New Zealand Health PolicySpringer Journals

Published: Jan 7, 2010

There are no references for this article.