Background
Asthma and Chronic Obstructive Pulmonary Disease (COPD) account for 80% of the total burden of chronic respiratory diseases (CRD) in Australia [
1] and represent a significant burden to the Australian health care system [
2,
3]. They are nevertheless under-diagnosed among the older population because of atypical presentation and co-morbidity [
2,
4,
5]. These diseases have similar manifestations in older patients - wheezing, chest tightness and shortness of breath - despite very different causes [
6,
7]. The estimated prevalence of asthma in Australian adults over 15 years of age ranges from 10 to 12% [
3], whereas approximately 9-12% of Australians over 45 years have symptomatic COPD [
8,
9].
The effects of patient demographic and clinical characteristics on mortality in the general population is well established [
10]. For example, being a male, of indigenous race, of low socio-economic status (SES) or living in a remote area has been shown to increase the risk of dying from any cause [
10]. However, relatively few studies have examined the risk factors for mortality in older patients (≥ 65) with asthma [
11,
12]. Also, the risk factors for mortality in older COPD patients have mainly been explored in patients who have been hospitalised for COPD [
13,
14]. This is perhaps surprising given relatively high prevalence of asthma and COPD, an ageing population demographic and substantial health care costs resulting from primary and secondary care of patients with these conditions [
8,
15,
16]. Further, the effects of patient characteristics on mortality and how these effects differ according to pharmacotherapy level remains uncertain.
Treatment with inhaled corticosteroids (ICS) is the cornerstone of CRD management [
17,
18] despite that large randomly controlled trials have had difficulty in establishing a beneficial effect of ICS on survival in COPD patients [
19]. Since distinguishing asthma and COPD in older patients with CRD is often challenging due to substantial overlap both in clinical manifestations and in the approach to disease management [
6,
7], studying the benefit of ICS on survival in older CRD patients overall remains important. We thus aimed to investigate (a) survival by pharmacotherapy level and (b) the effects of patient gender, race, SES, and residential remoteness on mortality due to any cause, overall as well as within groups of differing pharmacotherapy level, in older CRD patients in Western Australia (WA) during 1992-2006.
Discussion
We investigated the effect of patient characteristics on mortality in patients aged ≥ 65 years with CRD in WA, overall and by pharmacotherapy level. Our main findings included a relatively unfavourable survival of patients using only SABS (i.e. in the lowest pharmacotherapy level group), who experienced a slightly worse survival than the highest pharmacotherapy level group. We also found an increased likelihood of all-cause mortality in males particularly if they were in the highest pharmacotherapy level group.
A strength of this study was its whole-population design. We used data that were routinely collected by the WA and Commonwealth governments to ascertain patients with CRD. The accuracy of the WA administrative data and the record links produced by the WADLS has been found to be exceptionally high [
20]. However, as the data were collected for administrative purposes rather than research purposes, some consideration is warranted of the inherent limitations in patient ascertainment. First, our CRD study population included everyone at least 65 years old who had ever had a CRD medication dispensed during a 15 year period and thus the majority of our patient sample was selected on the basis of having had a CRD medication dispensed. This case ascertainment strategy appeared to overestimate the cases of CRD. However, based on data analysed by our group from 12 general practices in WA that included diagnostic information, out of 23,850 asthma/COPD prescriptions prescribed by general practitioners, 92% had asthma and/or COPD as reason for visit or prescription (unpublished data) This result indicates a strong correlation between CRD prescribing practice in WA and CRD diagnosis. Also, out of the 72% of patients selected in our study based on CRD medication use only, 30% were selected based on one single prescription. The majority of these 30%, or 56%, were within the SABs only pharmacotherapy level, which by definition includes patients more likely to require only occational treatment with medications. Thus the majority of our study population had established patterns of medication use indicative of a CRD problem. Consequently, by using this selection strategy we can be reasonably confident that with considerable specificity we have included all CRD cases in the older WA population who have ever seen a doctor for breathing problems. Furthermore, when we restricted our analyses to the patients who were selected based on medication use only and to patients selected based on records from all datasets we found the results led to essentially the same conclusions. It is therefore unlikely that the ascertainment strategy had any meaningful impact on our results.
A second consideration relates to the fact that the PBS data consisted only of records of dispensed drugs that were subsidised by the scheme. The vast majority of Australians over 65 years of age are eligible to receive a government concession card, entitling then to receive large PBS and MBS subsidies on medicines and other services. The coverage by these cards in 2004-5 was 90% at ages 65-75 and 95% at ages 75 + years [
25]. We expect that the coverage of older people with CRD has been even higher than this. We thus believe it is safe to assume that a very high proportion of CRD patients in this study were concessional beneficiaries. Hence, this would likely have had no more than limited impact on our study since all inhaled corticosteroids and long-acting beta agonist medications were subsidised under the PBS, and inhaled short-acting beta agonist medications and oral corticosteroids were consistently subsidised for concessional beneficiaries [
26]. Another issue concerning the use of PBS records in this study is that it is possible to buy inhaled SABs 'over the counter' from pharmacies in WA, which may make the findings of worse survival in the lowest pharmacotherapy level to appear to be due to less access to care. However, the vast majority of patients in the study would have been concessional beneficiaries. Thus seeing a doctor was free from out-of-pocket expense and there were strong financial incentives for the patients to buy SABs using a doctor's prescription instead of over the counter. In addition, the findings were adjusted for socio-demographic and locational disadvantage. We thus believe that access to care is an unlikely explanation for our findings.
The third possible limitation in this study concerns the inclusion of COPD patients with a mild disease in the highest pharmacotherapy level group. Exacerbations from COPD can occur at any severity and COPD patients may therefore be prescribed oral corticosteroids sporadically at any level of severity. We believe this would not have caused a problem in our study since we included in the highest pharmacotherapy level group those patients who were taking oral corticosteroids at a regular basis, using the criterion: > 1 prescription per year of oral steroids. Furthermore, despite exclusive oral steroid users being 51% of the patients in the highest pharmacotherapy level group; after excluding those users from this group, the results from our study remained unaltered.
The final limitation relates to the fact that we were unable to obtain data on smoking exposure in the patients. Smoking is a confounder in this study because it affects mortality in asthma/COPD as well being associated with the patient characteristics. The State and Commonwealth do not collect reliable data on smoking exposure as part of their routine administrative data sets. However, all our results were adjusted for factors such as socio-economic status, residential remoteness, gender, indigenous status, and co-morbidity, which most likely removed the majority of the confounding due to smoking.
We examined the survival outcomes of patients in our study according their level of pharmacotherapy and found that those using only SABs experienced the worst survival. This relatively unfavourable survival of patients not using ICS might reflect sub-optimal treatment in at least a sub-group of patients in this category [
4,
27‐
32]. That is, some patients in this group could benefit from an escalation of their therapeutic regime including the addition of ICS. This main finding of this manuscript is particularly interesting given the fact that large randomly controlled trials have had difficulty in establishing a beneficial effect of ICS on survival in COPD patients [
19].
We observed that males were almost twice as likely to die from any cause than females. These findings accord with previous studies of older asthma patients [
11,
12,
16] as well as reports on the general population [
10]. Further, our results indicate that the poorer survival of males was more marked in the highest pharmacotherapy level group. This may have reflected the tendency of men to seek medical care at a later stage in the disease process than women [
10,
33]. In fact, despite that women with asthma or COPD account for the majority of emergency department (ED) visits and hospitalisations for asthma [
34‐
39] as well as a large proportion of hospitalisations for COPD [
40], women who present to the ED have been found to receive more outpatient care and to have better pulmonary function than men [
34]. Hospitalised women with asthma have also been reported to experience less severe asthma than men [
35], evidenced by the higher incidence of ED intubation of male respiratory patients [
41].
We detected a relatively low proportion of indigenous patients in our study (0.6%), which was only slightly lower than the proportion of indigenous individuals among Western Australians aged 65+ years (0.8%) [
42]. Indigenous Australians experience significantly decreased life-expectancy and far higher mortality than other Australians [
10]. This was evident in our study of older patients with CRD, although the hazard ratio was not statistically significant, which may be due to small numbers. However, our results most likely reflect the persisting gap in access to adequate primary health care that indigenous people are faced with in Australia [
43]. For reasons such as cost, distance and lack of transport, indigenous people do not always seek medical assistance when needed [
43]. This can mean that any existing disease is more likely to become severe.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KE analysed and interpreted the data and wrote the article. JDE and CDJH designed the study. DBP, FMS, RR, JDE and CDJH contributed to the interpretation of data and revised the manuscript critically. All authors gave final approval for the manuscript to be published.