Abstract
Objective:
Determine the prevalence of multiple sclerosis (MS) in Australia in 2017 using MS-specific disease-modifying therapy (DMT) prescription data and estimate the change in prevalence from 2010.
Methods:
DMT prescriptions were extracted from Australia’s Pharmaceutical Benefits Scheme (PBS) data for January–December 2017. Percentages of people with MS using DMTs (DMT penetrance) were calculated using data from the Australian MS Longitudinal Study. Prevalence was estimated by dividing the total number of monthly prescriptions by 12 (except alemtuzumab), adjusted for DMT penetrance and Australian population estimates. Prevalences in Australian states/territories were age-standardised to the Australian population. Comparisons with 2010 prevalence data were performed using Poisson regression.
Results:
Overall DMT penetrance was 64%, and the number of people with MS in Australia in 2017 was 25,607 (95% confidence interval (CI): 24,874–26,478), a significant increase of 4324 people since 2010 (p < 0.001). The prevalence increased significantly from 95.6/100,000 (2010) to 103.7/100,000 (2017), with estimates highest in Tasmania in 2017 (138.7/100,000; 95% CI: 137.2–140.1) and lowest in Queensland (74.6/100,000; 95% CI: 73.5–75.6). From 2010 to 2017 using the median latitudes for each state/territory, the overall latitudinal variation in MS prevalence was an increase of 3.0% per degree-latitude.
Conclusion:
Consistent with global trends, Australia’s MS prevalence has increased; this probably reflecting decreased mortality, increased longevity and increased incidence.
Keywords
Introduction
Multiple sclerosis (MS) is a complex and multifaceted neurological condition with a range of environmental, behavioural and genetic factors implicated in the disease’s aetiology and clinical course. 1 The Atlas of Multiple Sclerosis estimate for the worldwide prevalence of MS has increased from 2.1 million people in 2008 to 2.3 million people in 2013. 2 In addition, the Global Burden of Disease Study for neurological conditions 2015 estimated that the worldwide prevalence of MS increased by 59% from 1990 to 2015, or to 2,012,000 people with MS (95% uncertainty intervals (UI) 1,866,000–2,167,000). 3 This increase can be attributed in part to an increased world population, increased survival (of both people with MS and the wider general population), and in some countries including Australia, to increasing MS incidence.2,4 It may also reflect improvements in the diagnosis and reporting of MS, and the establishment of clinical registries. 2
Several MS prevalence studies have been undertaken in Australia and these studies have generally used complex and resource-intensive methods to calculate prevalence. 5 However, in 2010, a resource-efficient estimate of the prevalence of MS in Australia was developed that used 2010 centrally recorded Australian Pharmaceutical Benefits Scheme (PBS) prescription data of MS-specific disease-modifying therapies (DMTs). 5 . The 2010 study reported the number of cases with MS in Australia was 21,283 people (prevalence 95.6/100,000) and that prevalence increased with increasing latitude. 5 Moreover, the results of this study were used in a prevalence-based cost of illness analysis of MS in Australia for 2010. 6
The aims of our current study were to (1) employ the prescription data methodology to estimate the prevalence of MS in Australia overall, and for each Australian state and territory at December 2017; and (2) compare our 2017 prevalence results with the 2010 results.
Materials and methods
The prescription methodology of this paper reflects the method employed in the 2010 prevalence study.5,6
Information sources
Prescription data and the usage of DMTs (DMT penetrance calculation)
The last decade has seen a revolution in DMTs to treat MS.7,8 DMTs act by modifying the activity of the immune system to reduce the frequency and severity of inflammatory/immune attacks targeting the central nervous system. DMTs prescribed for MS are exclusively prescribed to treat MS, in particular, people with relapsing-remitting MS. 7 The use of two DMTs concurrently is not recommended nor permitted under Australian universal healthcare funding prescription rules. 7
Australia’s PBS (and Repatriation PBS for Australia’s eligible veterans and their families) 9 is an important component of the country’s universal healthcare system that provides subsidised medications to all Australian citizens and permanent residents. 10 From March 2018, there were 12 approved DMTs for MS listed on Australia’s PBS. 11
DMTs prescribed for MS in Australia from 1 January to 31 December 2017 were alemtuzumab (Lemtrada), daclizumab (Zinbryta), dimethyl fumarate (Tecfidera), fingolimod (Gilenya), glatiramer acetate (Copaxone), interferon β-1A (Rebif 44 and Avonex), interferon β-1B (Betaferon), natalizumab (Tysabri), peginterferon (Plegridy) and teriflunomide (Aubagio). Unique identifiers for each DMT and the medication’s therapeutic mode of delivery were extracted from the Australian Government’s PBS website. 11
Using these identifier(s), the number of PBS prescriptions for January–December 2017 was then extracted from the Australian Government’s Medicare Australia website. 12 We assumed that if a person with MS filled the script in 2017, they were counted as a case for that year. In addition, it is also noted that a unique Australian Medicare identifier number is generated for each PBS prescription and each Medicare eligible individual. We did not include medications used ‘off-label’, or medications or treatments that may have been received or sourced overseas, and that therefore do not appear in the PBS data centralised repository.
The Australian Multiple Sclerosis Longitudinal Study (AMSLS), established in 2002, is a large representative sample of over 3000 Australians with MS living in Australia.13,14 The percentages of people with MS using MS-specific DMTs (DMT penetrance) were calculated for Australia overall and for each Australian state and territory (where the data were available for all of the Australian States) from the DMT usage question on the AMSLS Medications and Disease Course Survey where participants were surveyed from November 2016 to March 2017. 15
Population and age standardisation
Population estimates for the overall Australian population and for each Australian state and territory were sourced from the Australian Bureau of Statistics (ABS) most recent estimates (September 2017). 16
To calculate the age-standardised prevalence estimates for each Australian state and territory and allow comparing the 2010 and 2017 results, the numbers of cases of MS by age categories were extracted from the Australian MS Society client databases for the states and territories for 2010 and 2017. The age structures of the general Australian population were sourced from the March 2010 (used for the 2010 calculations) and September 2017 ABS population estimates. 16
Remoteness classification
Sociodemographic data (including client postcodes for Australian Remoteness Areas) were extracted from the AMSLS for participants who responded regarding DMT usage. Australia’s classifications of Remoteness Areas were sourced from the ABS’ Australian Statistical Geography Standard (ASGS) remoteness structure. 17 Remoteness Areas divide Australia into five classes on the basis of a measure of relative access classified as Major Cities, Inner Regional Australia, Outer Regional Australia, Remote Australia and Very Remote Australia. 17
Statistical analyses
Continuous variables were expressed as a mean (standard deviation (SD)), and categorical variables were expressed as percentages. Stata 15 15.0/SE (StataCorp, College Park, USA) was employed for all statistical analyses.
Sociodemographic characteristics and DMT penetrance
The age and sex profiles, and DMT penetrance of AMSLS participants who responded regarding DMT usages were calculated for the overall sample and for each Australian state/territory. We also calculated the proportions of people with MS living in the Remoteness Areas of Australia using the respondent postcodes.
Calculation of crude prevalence using the PBS DMT prescription data
To estimate the crude prevalence for our study, the annual number of PBS and Repatriation Pharmaceutical Benefits Scheme (RPBS) scripts dispensed for the DMTs were divided by 12 as these were monthly prescriptions, except for alemtuzumab, which was prescribed and administered once annually. This estimate was then adjusted for DMT penetrance for Australia overall, and for each Australian state and territory. The ABS population denominators for each state and territory from September 2017 estimates were then used to calculate crude prevalence per 100,000 people.
Calculation of age-specific and age-standardised prevalence, and Poisson regression to compare 2010 and 2017 estimates
Age- and sex-specific case numbers were estimated as per the age/sex structure of the MS Society participants in each state and territory, these were applied to the total number of MS cases estimated in each state/territory. State and territory prevalence estimates in 2010 and 2017 were age-standardised to the 2017 Australian population using the direct method. 18 Poisson regression was used to compare the 2010 and 2017 prevalence calculations, estimating prevalence ratios.
Ethics approvals
Ethics approval for our study was provided by the University of Tasmania’s Health and Medical Human Research Ethics Committee. All data were de-identified for analyses.
Results
AMSLS participant characteristics and DMT penetrance
Table 1 presents the age and sex profiles, and the Australia-wide and state and territory DMT penetrance for the AMSLS participants who responded ‘yes’ or ‘no’ regarding the use of DMTs (n = 1699).
Sociodemographic characteristics of people with MS on an Australian, and state and territory basis, and penetrance of disease-modifying therapies (DMTs).
AMSLS: Australian Multiple Sclerosis Longitudinal Study; CI: confidence interval; SD: standard deviation.
N = 1699 people with MS who responded to the Medications and Disease Course Survey of AMSLS regarding their current treatment with disease-modifying therapies (DMTs); n = 5 people who were not allocated to a state or territory but included in the overall calculation.
NT data unavailable; Australian Capital Territory and Northern Territory DMT penetrance calculated as an average of the states’ penetrance estimates.
Those people with MS currently being treated with DMTs were almost a decade younger than those not being treated with DMTs. The mean age at diagnosis was 41 (SD: 11) years, consistent across Australia’s states and territories. Females represented 78% of the total sample, again with similar proportions for the states and territories. Table 1 also shows that the mean (SD) number of years since diagnosis was 15 (SD: 9) years. In addition, people currently treated with DMTs were diagnosed 4 years earlier than those not using DMTs. Penetrance of DMTs was 64% (95% CI: 62–66) for Australia overall, ranging from 59% (95% CI: 48–69) in Tasmania to 69% (95% CI: 64–73) in Victoria.
PBS DMT prescription data
Table 2 (supported by Supplementary Table 1) provides the PBS and RPBS codes for the MS-specific DMTs, the types and number of prescriptions for MS-specific DMTs in Australia overall, and the Australian states and territories, from 1 January to 31 December 2017.
The number of Australian Pharmaceutical Benefits Scheme (PBS) and Repatriation PBS disease-modifying therapy (DMT) prescriptions exclusively prescribed for people with MS in Australian states and territories: January to December 2017 a .
Source: Medicare Australia Statistics (http://medicarestatistics.humanservices.gov.au/statistics/pbs_item.jsp).
Supported by Supplementary Table 1.
A total of 188,243 DMT prescriptions for 10 DMTs with 22 modes of therapeutic delivery were dispensed in Australia from 1 January to 31 December 2017.
Prevalence calculations based on the prescription data and latitudinal gradient
Table 3 presents the 2017 prevalence of MS and shows that based on a total Australian population of 24,702,900 people (September 2017 estimates) and a DMT penetrance of 64% (95% CI: 62–66), the overall crude prevalence of MS in Australia was estimated to be 25,607 people (95% CI: 24,874–26,478), or 103.7/100,000 (95% CI: 100.7–107.2) (Table 3).
Calculation of prevalence of people with multiple sclerosis (MS) based on Australian Pharmaceutical Benefit Scheme prescribed disease-modifying therapies (DMT), and age-standardised prevalence using the Australian MS Society age structures on a state and territory basis, and the Australian MS Longitudinal Study.
CI: confidence interval.
Australian Bureau of Statistics Population Estimates: September 2017.
ACT and NT penetrance and CI based on averages of all states.
Calculated from prescription bounds of Australia, not from summing the individual states and territories; significant increase from 2010 (21,283 people with MS) p < 0.001.
The presence of a latitudinal gradient was evident, with Tasmania, Australia’s geographically most southern state with a median latitude of 42.0 S, recording the highest prevalence for both unadjusted and age-standardised figures of 148.3/100,000 (95% CI: 138.2–159.1) and 138.7/100,000 (95% CI: 137.2–140.1), respectively. Queensland (median latitude 19.4 S) had the lowest crude and age-standardised prevalence of the states of 80.2/100,000 (unadjusted) (95% CI: 72.9–89.1) and 74.6/100,000 (age-standardised) (95% CI: 73.5–75.6). In addition, Victoria (median latitude 36.9 S) had a higher prevalence than New South Wales (the most populated Australian state with a median latitude of 32.2 S) (Table 3).
Figure 1 presents the line of best fit for the state and territory age-standardised prevalence estimates and the latitude of the state and territory capital cities demonstrating that the strong latitudinal gradient relationship between the states and territories persists in 2017.
Age-standardised 2017 prevalence estimates that reveal a persisting latitudinal gradient for Australia’s states and territories.
Percentage of people with MS by remoteness category
Table 4 presents the number and percentage of people with MS residing in the various Remoteness Areas of Australia categories based on the postcode of participants of the AMSLS survey. Most people with MS reside in Major Cities of Australia (64%), followed by 25% residing in Inner Regional Australia, while Outer Regional Australia, Remote Australia and Very Remote Australia account for just over 10%. Table 4 also provides the state and territory breakdowns of the Remoteness Areas, revealing a similar pattern of most people with MS living in Major Cities of Australia.
The number and percentage of people with MS who participated in the AMSLS survey regarding DMT usage (‘yes’ and ‘no’) living in different Australian Remoteness Areas.
AMSLS: Australian Multiple Sclerosis Longitudinal Study; DMT: disease-modifying therapy.
Not applicable: Tasmania does not record a Major City.
Comparison of 2010 and 2017 results
Tables 5–7, and Figures 2 and 3 present results from Poisson regression analyses evaluating the change in prevalence estimates from 2010 to 2017.
Crude prevalence of the number of people with multiple sclerosis for 2010 and 2017, and changes for Australia overall and each state and territory.
CI: confidence interval.
Results in boldface denote statistical significance (p < 0.05).
Unadjusted prevalence as reflected in Table 3 for 2017, and 2010 results.
Population figures sourced from Australian Bureau of Statistics estimates for March 2010 and September 2017.
Rounded to two decimal places.
Australian age-specific MS prevalence change between 2010 and 2017 a .
CI: confidence interval.
Results in boldface denote statistical significance (p < 0.05).
Constrained to 18,217 and 25,022 Australian MS Society participants with age-specific information in 2010 and 2017, respectively.
Age-specific MS prevalence changes measured by prevalence ratios between 2010 and 2017, for Australia and each state and territory a .
NA: not available.
Results in boldface denote statistical significance (p < 0.05).
Estimates based on people with multiple sclerosis with age-specific data from MS Society data.
Comparison of the Australian prevalence for each age category in 2010 and 2017, where p < 0.05 for all age categories except 25–29 and 30–34 year age categories.
Comparison of the Australian prevalence for each age category for the Australian states and territories (Tasmania (TAS), Victoria (VIC), South Australia (SA), New South Wales and the Australian Capital Territory (NSW/ACT), Western Australian (WA) and Queensland (QLD)) in 2010 and 2017.
Overall, the prevalence of MS in Australia increased significantly from 2010 to 2017, where the absolute prevalence ratio increased by 9% over the interval (Table 5). The greatest changes were seen for the Australia Capital Territory (30% increase), South Australia (32% increase) and Western Australia (15% decrease), while the smallest changes were in Victoria and Tasmania (Table 5). Using the median latitudes for each state/territory, the overall latitudinal variation in MS prevalence was an increase of 3.0% (95% CI: 2.7–3.2) per degree-latitude.
The prevalence of MS increased across all age groups, with only those of 25–29 and 30–34 years old age groups not reaching significance, and the greatest increases observed among the highest three age categories (PR > 1.40 for age categories >70 years) (Table 6; Figure 2). Figure 2 shows a shift towards older age groups. Table 7 and Figure 3 show that similar results were also seen by state and territory, with age shifts most evident in New South Wales, South Australia and Western Australia, while Queensland showed the reverse, with the greatest increases among those under 54 years old. Tasmania and the Australian Capital Territory showed more erratic age-specific patterns of change, likely reflecting their smaller populations.
Figure 4 presents the case age proportional distributions of MS society data for the Australian states in 2017. This schematic representation of these age profiles suggests that Western Australia and Tasmania have an older age profile.
Age distributions of MS for the Australian states of New South Wales, Victoria, Queensland, Western Australian, South Australia and Tasmania, sourced from MS Society data.
Discussion
Using our prescription method, we found that the total number of people with MS in Australia significantly increased to 25,607 people in 2017 from 21,283 people in 2010. The national prevalence increased by 9%, and the latitudinal variation was consistent with previous findings, 2 the highest prevalence in Tasmania (latitude ~42 S) and lowest in Queensland (latitude ~19 S). We also found evidence to suggest that the age structure of people with MS changed since 2010.
One key reason to support our findings is the increased survival of people with MS, as noted in the International Federation of MS Atlas of Multiple Sclerosis, and the Global Burden of Disease study regarding neurological disorders 2015.2,3 Indeed, our data support this notion as we identified that the increases were disproportionally higher in older participants. However, it is also possible that this could reflect changes in the recruitment of participants by the MS Societies, including the retention of existing participants and a somewhat slowed recruitment of people who are newly diagnosed with MS (and therefore relatively younger). The increase in prevalence may also reflect improvements in the diagnosis and reporting of MS, with modern diagnostic criteria enabling diagnosis on the basis of a single magnetic resonance imaging (MRI),19,20 rather than earlier clinical course-based criteria.
A further likely reason for the increase in prevalence is a posssible increase in incidence of MS as reflected in a recent study from the city of Newcastle in the state of New South Wales that found a dramatic rise in incidence for this region compared to that seen in prior surveys in this region (6.7 per 100,000 persons for the years 2001–2011). 4 The Newcastle study suggested that there have been several factors linked to the increase in MS incidence observed in different countries including Vitamin D and sunlight exposure, cigarette smoking and dietary changes and obesity; however, the study noted that an investigation of these changes in behaviour for people with MS in the Newcastle region was beyond the scope of the study. 4 The Newcastle study also suggested that the impact of dietary modifications and its effect on MS onset is a newer concept as a potential factor influencing MS incidence. 4 Similarly, the Atlas of Multiple Sclerosis 2013 suggested that the increase in global prevalence of MS is partially explained by an increased incidence, and that there is a call for some improvements in healthcare and support services, while it is also acknowledged that the global increase in MS groups and organisations is encouraging. Similarly, the Global Burden of Disease Study for neurological disorders 2015 3 suggested that the large and increasing numbers of patients with neurological disorders necessitate careful planning by governments and other healthcare providers to ensure adequate funding and staff for their treatment and rehabilitation services. 3 Overall, our prescription-based prevalence study found a significant increase in the number of people with MS in Australia from 2010 to 2017 and we support ongoing dialogue and action regarding adequate funding and staff for people with MS both in Australia and worldwide.
Of the environmental risk factors linked to MS, the latitudinal gradient of prevalence and incidence has been among the most consistent and striking findings in MS epidemiology. 21 Indeed, our recent meta-analysis revealed a positive association between latitude and MS prevalence with a 1.04 change in prevalence/100,000 per degree-latitude (p < 0.001). 22 Here, we substantiate that the latitudinal gradient exists in Australia, finding the age-standardised prevalence estimate was highest in Tasmania (138.7 per 100,000 (95% CI: 137.2, 140.1)), almost double that of Queensland (74.6 per 100,000 (95% CI: 73.5, 75.6) and Western Australia (87.7 per 100,000 (95% CI: 86.6, 88.9), and an average 3.0% increase in prevalence per degree-latitude, in line with regional studies indicating this gradient in Australia previously.23,24 Our key finding of a persistent latitudinal gradient is supported by a recent (March 2019) study that investigated over 125 million people from administrative healthcare records in the United States, including the use of prescription records for DMTs. 25 The study covered nearly 45% of the United States population, including 100% of those with publicly funded insurance. The study developed and tested an optimal algorithm that included prescription-release encounters for a DMT (approved by the US Food and Drug Administration) within a 1-year period. The paper’s approach accounted for the demographics of the national population, the sporadic follow-up for a chronic disease with young-adult onset, the different insurance providers within the healthcare system and the uninsured. 25 Importantly, this recent study found that the prevalence in the northern Census regions of the United States (Northeast and Midwest) was statistically significantly higher than in the southern Census region. The study also found that there has been a steady rise in the prevalence of MS over the past five decades, and that the prevalence of MS remains higher for women than men. 25 We identified that 89% of people with MS live in Major Cities or Inner Regional Australia, and that females represent three-quarters of the sample – these findings are similar to our 2010 findings.
There was a substantial increase of over 40% of people with MS using DMTs from 2010 compared to 2017. This possibly reflects an increased uptake of oral DMTs and the advent of more highly effective biological therapies in this time period. This suggestion is supported by another AMSLS study that investigated the DMT treatment patterns for DMT usage between 2010 and 2015 that found the usage of category 2 and 3 DMTs (e.g. more highly effective biological therapies) increased over the interval. 26
Strengths and limitations
Our current study employed the validated methodology developed in 2010 to calculate the prevalence of MS in Australia for December 2017 using PBS prescription data from 1 January to 31 December 2017. Importantly, using the same methodological approach to the 2010 study supported a direct comparison of the prevalence estimates. We have found limited studies employing this efficient method, including our 2010 study, a study of MS in the Lazio region of Italy, 27 and a national study in Austria in 2016. 28
Therefore, there are some strengths to our prescription method prevalence study and the key strength is the PBS prescription-based method for extracting medications used solely by people with MS, including the ascertainment of all prescription data Australia-wide from the one centralised and continually updated Medicare Australia database where a unique Medicare identifier number exists for each individual prescription and each individual. Another strength was the use of the same validated prescription method employed in 2010 that enabled a direct comparison of the results. Another strength is the robustness of the DMT penetrance calculations from the large representative AMSLS that informed the prescription-based prevalence calculations.
Nevertheless, there are also limitations to our study that also reflect the limitations of the same methodology outlined in our 2010 study; for example, we assumed 100% compliance with the DMTs when we divided the number of prescriptions by 12 to calculate the number of people using DMTs. Similar to our 2010 study, it is likely that a very small proportion of prescriptions were filled but not used, or that the person with MS switched or stopped the medication; nevertheless, most cases are stable on their DMT therapy and only a small number change each year. A recent study (2013) investigated the persistence of DMT usage for the Australian MSBase Clinical Cohort and found that patients with MS persisted on their first DMT for a median of 2.5 years and on subsequent DMTs for a median of 2.3 years. 29 Importantly, these results suggest that for the purposes of our prescription-based prevalence study that the persistence and adherence for DMTs over a 12-month timeframe is relatively stable. In addition, the 2013 Australian MSBase study also found that annualised switch or cessation rates were relatively low ranging from 4.4% (natalizumab) which is the most prevalent medication in our study representing over 32,000 prescriptions or 27% of our prevalence study’s sample scripts, to 9.5%–12.5% for individualised interferon-beta products that represented almost 8000 scripts or 7% of the prescription sample for our study. Accordingly, and coupled with the persistence rates reported above, these results further show the relative stability for our prescription method.
Another limitation is that there are geographical factors that could influence the calculated prevalence such as the influence of migration during 2017 both in and out of Australia as well as between different regions. However, the population in Australia is largely stable with migration and net interstate migration making up only a small amount of change between regions, and therefore, it is unlikely to affect the overall picture of people with MS particularly over a 1-year period. 30 Another limitation of our study is we assumed that if a person with MS filled a DMT prescription for 2017, they were alive and counted as a case, and therefore, mortality is not accounted for in our prescription-based methodology.
Another potential weakness to our study is the use of age structures for the age-adjustments from the Australian MS Society data, since our data on MS-specific DMTs from the Medicare database are not age specific. Our method thus presumes the age structure of the AMSLS cohort is comparable to the overall MS population in Australia. The similarity of the demographic and clinical characteristics of the AMSLS cohort to other MS cohorts in Australia and New Zealand has been demonstrated previously. 31 Moreover, our age-standardisation for each state and territory made use of those regions’ MS Society age structures. The MS Societies are a large population-based service for people with MS throughout each state and territory with good coverage and uptake; however, there may be some underestimation of some patient groups, including younger patients.
Conclusion
Consistent with global trends, the prevalence of MS in Australia has increased significantly from 21,283 people with MS in 2010 to 25,607 people with MS in 2017. In addition, the age structure of people with MS has shifted.
The use of prescription data to estimate prevalence is logistically economical, allowing timely estimations of prevalence that can be used in cost of illness studies and other health economics analyses. We suggest that this methodology should be explored for other countries where central repositories of prescription data and penetrance rates can be easily accessed in a timely manner to enable the rapid and accurate calculation of the prevalence of MS.
Supplemental Material
MSJ861270_supplemental_table – Supplemental material for Change in multiple sclerosis prevalence over time in Australia 2010–2017 utilising disease-modifying therapy prescription data
Supplemental material, MSJ861270_supplemental_table for Change in multiple sclerosis prevalence over time in Australia 2010–2017 utilising disease-modifying therapy prescription data by Julie A Campbell, Steve Simpson, Hasnat Ahmad, Bruce V Taylor, Ingrid van der Mei and Andrew J Palmer in Multiple Sclerosis Journal
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by funding from MS Research Australia.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
