Article Text

Download PDFPDF
Original research
Increasing incidence and prevalence of multiple sclerosis in the Greater Hobart cohort of Tasmania, Australia
  1. Steve Simpson-Yap1,2,3,
  2. Roberts Atvars4,
  3. Leigh Blizzard3,
  4. Ingrid van der Mei3,
  5. Bruce V Taylor3
  1. 1Neuroepidemiology Unit, Melbourne School of Population & Global Health, The University of Melbourne, Melbourne, Victoria, Australia
  2. 2Clinical Outcomes Research Unit (CORe), Royal Melbourne Hospital, The University of Melbourne, Melbourne, Victoria, Australia
  3. 3MS Flagship, Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia
  4. 4Royal Hobart Hospital, Hobart, Tasmania, Australia
  1. Correspondence to Dr Steve Simpson-Yap, The University of Melbourne School of Population and Global Health, Melbourne, Victoria, Australia; steve.simpsonyap{at}


Background The Greater Hobart region (42.5°S) of Tasmania has consistently had the highest recorded prevalence and incidence rates of multiple sclerosis (MS) in Australia. We reassessed MS epidemiology in 2009–2019 and assessed longitudinal changes over 68 years.

Methods Cases recruited from clinic-based datasets and multiple other data sources. 2019 prevalence and 2009–2019 annual incidence and mortality rates estimated, and differences assessed using Poisson regression.

Results 436 MS cases resident on prevalence day were identified, and 130 had symptom onset within 2009–2019. Prevalence 197.1/100 000 (95% CI 179.4 to 216.5; 147.2/100 000 age standardised, 95% CI 126.5 to 171.3), a 36% increase since 2001 and 3.1-fold increase since 1961. 2009–2019 incidence rate=5.9/100 000 person-years, 95% CI 5.0 to 7.0 (6.1/1000 000 age standardised, 95% CI 4.7 to 7.9), a 2.8-fold increase since 1951–1961 and 65% since 2001–2009. 2009–2019 mortality rate=1.5/100 000 person-years, 95% CI 1.1 to 2.2 (0.9/100 000 age standardised, 95% CI 0.4 to 1.7), comparable to 2001–2009 (1.0/100 000) but reduced by 61% from 1951 to 1959 (2.1/100 000). 2001–2009 standardised mortality ratio=1.0 in 2009–2019, decreased from 2.0 in 1971–1979. Female:male prevalence sex ratio was 2.8, comparable to the 2009 value (2.6); incidence sex ratio (2.9) increased from 2001 to 9 (2.1). Comparisons with Newcastle, Australia (latitude=32.5°S) demonstrate a near complete abrogation of the latitudinal gradients for prevalence (ratio=1.0) and incidence (ratio=1.1), largely attributable to changing Hobart demography.

Conclusions Prevalence and incidence of MS continue to increase significantly in Hobart, alongside marked reductions in mortality and increased case longevity. The marked increase in incidence is of particular note and may reflect longstanding changes in MS risk behaviours including changing sun exposure, obesity rates, and smoking behaviours, particularly in females. Falling mortality contributes to increase longevity and prevalence, likely reflecting improved overall MS healthcare and implementation of disease-modifying therapy.


Data availability statement

Data are available on reasonable request.

Statistics from

Key messages

What is already known on this topic

  • The Greater Hobart region of southern Australia has one of the highest frequencies of multiple sclerosis (MS) and the highest in Australia, this consistently shown over repeated surveys since 1961.

What this study adds

  • We now demonstrate a continued increase in both prevalence and incidence, despite relatively static case ascertainment, while mortality fell and mean longevity increased.

  • The standardised mortality ratio has fallen to 1.0, possibly reflecting the impact of more highly effective disease-modifying therapies.

  • The ratio of prevalence and incidence ratios compared with mainland Australia have fallen to parity, possibly reflecting alterations in demography and changes in behaviour.

How this study might affect research, practice and/or policy

  • These results demonstrate continuing increases in the frequencies of MS in Hobart not solely attributable to increased case detection, providing valuable information in support of resource allocation and planning by federal, state, and private agencies.


Multiple sclerosis (MS) is a complex disease with both neurodegenerative and neuroinflammatory components.1 2 Environmental and lifestyle factors are associated with the onset of MS, and these are particularly important in those with an increased genetic risk for MS.3 4

The Greater Hobart region of Tasmania, Australia has been the site of three major MS epidemiological studies in 1951–19615, 1971–19796 and 2001–20097, showing consistently increasing trends in MS incidence and prevalence, particularly people aged over 60. The recently completed MS Atlas project found nearly 2.8 million people live with MS worldwide—a significant increase from the 2.3 million people diagnosed in 2013.8 9 Considering the increasing health economic burden of MS in Australia10 and the potential for overall societal MS costs to continue to rise, it is important to understand the changing epidemiology of MS in areas where there is a long history of well-conducted epidemiological studies.

Because MS diagnostic criteria have changed over time11 and access to sensitive imaging techniques (MRI) has increased, it has been anecdotally suggested that the increasing prevalence of MS is due to greater ascertainment of those with milder disease phenotypes, shorter times to diagnosis, and potentially increased misdiagnosis.12 To assess these questions and better inform our understanding of the epidemiology of MS in this area, we have undertaken a fourth comprehensive MS study within the Greater Hobart area of Tasmania, Australia, to determine the temporal trends in MS prevalence, incidence, and mortality over time, in a homogeneous and stable population of people at high risk of developing MS.


The Greater Hobart statistical region encompasses the city of Hobart and its surroundings, located at 42.5°S. Hobart is the capital city of the state of Tasmania, Australia and occupies an area of 1360 km2. Though this area has increased somewhat since 1961, the majority of the population resides in the area in Hobart city and the adjacent suburbs on the near and far sides of the Derwent River which has comprised Greater Hobart since 1961. The population of the area has increased from 113 952 in 1961 to 2 21 226 in 2019, primarily in Hobart city and adjacent suburbs in the centre of the area.

Case ascertainment

Since 1996, MS clinical services in Hobart have been centralised at the MS clinic at the Royal Hobart Hospital, which has provided MS care to the vast majority of people with MS living in Tasmania.

We previously conducted prevalence surveys in Hobart in 2001 and 2009.7 The 2009 database provided the baseline for this study. We first confirmed the status of all people recorded as having MS in Hobart in 2009 as being deceased, living with MS in Hobart, migrated elsewhere, or disease reclassified. We searched the Tasmanian Health Services digital medical record for all persons with an MS diagnosis living in southern Tasmania. We also searched the MS clinic appointment lists for 2016–2019 for any additional cases. All neurologists practising in southern Tasmania were surveyed to ascertain cases that were seen exclusively in the private system. Lists of people with MS who had consented to participate in MS research at the Menzies Institute for Medical Research were searched for additional cases. A consolidated list was then constructed, excluding duplicates and all diagnoses confirmed as per 201113 or 201711 McDonald criteria by investigator RA with input from BVT. These case ascertainment methods are broadly comparable with those undertaken in previous surveys at this location, including medical record review, correspondence with private neurologists and review of available ancillary research databases.5–7 Data on onset year, diagnosis year, MS type, sex, and place of residence were extracted. We also ascertained whether cases had migrated to the Greater Hobart region before or after the diagnosis of MS was made. Cases were considered incident if they had migrated to the region prior to the onset of MS, and as migrant cases otherwise. We collected mortality data on all MS cases who were resident in Hobart in 2009. We have no records of any person being diagnosed with MS in Hobart after the 2009 study who died before the prevalence day. Cases with MS in 2009 who migrated out of the study area were classified as emigrants.


Prevalence was calculated as the number of people with MS resident in the study area on prevalence day, 30 June 2019, divided by the estimated 2016 Census population. The annual incidence rate was calculated as the number of MS cases with onset while living in the study area during the period 2009–2019, divided by 10 times the estimated 2016 Census population. In addition, incidence rates were calculated as number of MS cases diagnosed while living in the study area during 2009–2019, divided by 10 times the 2016 Census population. The annual mortality rate was defined as the number of decedents with MS in the study area during the period 2009–2019, divided by 10 times the estimated 2016 Census population. Standardised mortality ratios (SMR) were estimated using national sex-specific mortality statistics for 2009–2019 to estimate age-specific and sex-specific mortality rates. These total mortality rates were multiplied by the number of prevalent MS cases resident in Greater Hobart in 2019 to estimate expected numbers of deaths. This was then evaluated in ratio against the observed deaths by the function (observed/expected).

Age was categorised in a fashion to make it comparable to previous studies, as presented in online supplemental tables.

Supplemental material

All population data were obtained from the relevant quinquennial Census or annual population estimates published by Australian Bureau of Statistics.

Longitudinal comparisons and age standardisation

Utilising previously published epidemiological statistics from the 19685 and 198814 studies, we extracted prevalence, incidence, mortality, and longevity statistics for 1961 and 1981, and for 1951–1960 and 1971–1979, as appropriate.

Age standardisation by the direct method,15 and separately by sex, was undertaken for current estimates, and to the extent possible using the data made available in the previous published reports. For comparison with 1961/1951–1960 prevalence and incidence measures, standardisation was to the 1961 Greater Hobart population. For 1981, standardisation was to the 1981 Greater Hobart population. For some comparisons, incidence and mortality rates were standardised to the 1954 Greater Hobart population, but the mortality rates for 1981 were not standardised because age-specific data were not available.

To allow comparison to more contemporary Australian data, we have also standardised to the 2016 Australian population.

Statistical methods

Poisson regression was used to assess differences over time and between groups in prevalence, incidence and mortality. Binary (0/1) covariates were included for each age group, sex, and study year with population entered as an offset. To assess trends over time, study year was included as a continuous covariate and the size and statistical significance of its coefficient are reported.

SMR confidence intervals based on the normal distribution were calculated using Byar’s approximations.15

All statistical analyses were carried out using STATA/SE V.16.0 (StataCorp).


On prevalence day, 30 June 2019, we identified 436 people with MS who met the McDonald criteria 2010 or 2017 and who were resident in Greater Hobart, comprising people who were also found in 2009 (n=199), who had symptom onset in 2009–2019 (n=125), who were diagnosed in 2009–2019 but had symptom onset before 2009 (n=48), who immigrated (n=60), and four who were missed in 2009. The age-standardised prevalence in 2019 was 147.2/100 000 (95% CI 126.5 to 171.3), a 48% increase from the 2009 prevalence of 99.6/100 000. The age-standardised incidence from MS onset also markedly increased from 3.7/100 000 person-years in 2001–2009 to 6.1/100 000 person-years (95% CI 4.7 to 7.9). Incidence from MS diagnosis was 7.7/100 000 (95% CI 6.7 to 9.0), persisting on standardisation to 7.6 (95% CI 6.1 to 9.6). Mortality on the other hand continued the decrease reported in previous studies, from 1.0/100 000 person-years in 2001–2009 to 0.9/100 000 here(95% CI 0.4 to 1.7). Full demographic details are shown in table 1 and a flow chart of the allocation of cases from 2009 to 2019 is shown in figure 1.

Table 1

Cohort characteristics of analysis sample and 2019 prevalence and 2009–2019 incidence and mortality rates

Figure 1

Flow chart of people with MS, 2009–2019. MS, multiple sclerosis.

Prevalence, 1961–2016

The 2019 crude prevalence was 197.1/100,000 (95% CI 179.4 to 216.5), reducing on age/sex standardisation to the 1961 Hobart population to 147.2/100 000 (95% CI 126.5 to 171.3) and to 190.0/100 000 (95% CI 188.3 to 191.8) on standardising to the 2016 Australian population (table 1). This standardised prevalence represents a 3.1-fold increase compared with 1961 and 36% increases since 2001(table 2). The standardised prevalence among females was 2.8 times that in the males, 216.9/100 000 (95% CI 181.7 to 258.9 vs 78.4/100 000 (95% CI 58.5 to 105.0), respectively. Other longitudinal prevalence statistics are shown in table 2.

Table 2

Change in crude and age-standardised prevalence over time, 1961–2016, overall and by sex

Age-specific and sex-specific prevalence in 2019 is reported in online supplemental table 1. The highest frequency age groups are those of persons aged 45–54 and 55–64 years. Figure 2 depicts changes in age-specific and sex-specific prevalence over time between 1961 and this study. There have been pronounced increases for females in all age groups. For males, the increases are more modest and mainly restricted to those aged 40 years or older, and with no further increase since 2009.

Figure 2

Age-specific prevalence estimates, 1961–2019. (A) All persons; (B) males; (C) females.

Incidence rate from MS onset, 1951–1959 to 2009–2019

The standardised annual incidence rate for 2009–2019 was 6.1/100 000 person-years (95% CI 4.7 to 7.9), a 2.8-fold increase from 1951 to 1959 and a 65% increase from 2001 to 9 (table 3). The sex ratio of incidence among females (9.0/100 000, 95% CI 6.7 to 12.1) and males (3.0/100 000, 95% CI 1.8 to 5.1) was 3.1 (95% CI 1.8 to 5.1), with each having increasing approximately twofold since 2001–2009. Other incidence characteristics are shown in table 3.

Table 3

Change in age-standardised incidence rate over time, 1956–2016, overall and by sex

Age-specific and sex-specific incidence rates are shown in online supplemental table 2 and figure 3. The age group with the highest incidence rate was 25–44 years (online supplemental table 2). Figure 3 depicts the pronounced increase in age-specific rates for adult females aged 20–49 years that occurred in the most recent period (2009–2019).

Figure 3

Age-specific incidence rates over 1971–1981 to 2009–2019: (A) all persons; (B) males; (C) females.

Incidence rate from diagnosis was 7.7/100 000 person-years (95% CI 6.6 to 9.0) (7.6/100 000 age standardised), this higher than the incidence rate from MS onset. The sex ratio of the incidence among females (10.8/100 000 PY, 95% CI 9.1 to 12.9) vs males (4.5/100 000 PY, 95% CI 3.4 to 5.9) was 2.6, similar to that from MS onset. Age-specific incidence rates are show in online supplemental table 3.

Mortality rates and longevity, 1951–1959 to 2009–2019

The average annual age-SMR for 2009–2019 was 0.9/100 000 person-years. This constitutes a 60% reduction from 1951 to 1959, and a 18% reduction from 2001 to 2009 (online supplemental table 4).

Continuing the trend noted in previous studies, the mean age of probands increased from 41.0 years in 1961 to 50.9 years in 2009 and 54.4 (95% CI 53.1 to 55.6) in this study.

The deaths observed were only 1% higher than the number expected based on age-specific mortality rates in the general population. The SMR was SMR=1.01, higher among males than females. The total SMR was slightly reduced from 2001 to 9 but was 50% lower than that in 1971–1981 (table 4).

Table 4

Standardised mortality ratio, Greater Hobart, 1971–1981 to 2009–2019


We have described the longitudinal epidemiology of MS in the Greater Hobart region (latitude 42.5°S) of Tasmania for the period 2009–2019, and how these characteristics changed since 1951–1959 and 2001–2009. In 2019, 436 persons with MS were found residing in the region, a prevalence of 197.1/100 000. Over 2009–2019, 130 cases had symptom onset while residing within the region, an average annual incidence rate of 5.9/100 000 person-years. Mortality rates continued to decline, reaching 1.5/100 000 person-years in this study with 34 recorded deaths in the cohort, while the mean age increased to 54.4 years. This high incidence alongside a decreasing mortality has led to the prevalence in the region continuing to be the highest reported in the Southern Hemisphere and among the highest worldwide.16 17

Hobart MS epidemiology over time

The standardised 2019 prevalence was 147.2/100 000, 36% higher than 2009 and 3.1-fold higher than 1961. Our results are consistent with Poskanzer’s precept that changes in prevalence reflects differences in incidence, migration, mortality, disease duration and case-ascertainment.18 We demonstrate a notable increase in incidence, with the age-standardised incidence rate during 2009–2019 of 6.1/100 000 person-years, 65% higher than 2001–2009 and 2.8 times greater than in 1951–1959. We show a continued decrease in mortality rates with the average annual rate for 2009–2019 only one-third the rate for 1951–1959, and a continued increase in case longevity. Changes in case ascertainment likely made little contribution for comparisons with 2009–2019, as there were not material changes in access to care over the period or in mean age of diagnosis over this interval. Similarly, but to a lesser extent, healthcare access and patient characteristics, as well as the area defining Greater Hobart, are all sufficiently comparable even in comparison to 1951–1961, so while there would be some increase in patient ascertainment over the total interval since 1951–1961, we believe the changes reflect genuine increases rather than better case detection. The less sensitive diagnostic criteria used in the two historical studies likely resulted in some underestimation of the prevalence and incidence at that time, and so the overall increases seen since 1951–1961 to 2009–2019 are accordingly exaggerated. Nonetheless, these results show appreciable and long-lasting increases in the frequency of MS in the Greater Hobart region over the 68-year period under consideration.

While the increase in prevalence follows logically from increased incidence and case longevity/decreased mortality, mechanisms underlying change in incidence are less clear. It is unlikely to be a function of increased case ascertainment, either as a function of improved diagnostic criteria or access to care, or greater awareness of MS by medical practitioners. While these changes did occur since 1951–1961, these characteristics have been effectively static since our previous study in 2001–2009. Instead, they may reflect factors from an earlier period only now becoming evident in their effects. Alongside genetic determinants,3 4 multiple environmental factors have been associated with MS risk, including sun exposure/vitamin D, body mass index/obesity, and smoking,3 4 ,19 the distributions of which have changed markedly over time. For example, childhood/adolescent obesity rates have increased dramatically in Australia over the last 40 years as they have done globally.20 Similarl+

Mortality rates and longevity continue the trends toward increased patient lifespan. This is particularly evident for SMR over the 1971–1981 to 2009–2019 intervals, showing a halving of expected mortality. Hirst’s 1985 study in South Wales estimated an SMR of 2.75.21 Koch-Henriksen evaluated national mortality in Denmark, showing a marked reduction in SMR from 4.09 in 1950–1959 to 1.79 in 1990–1999.22 These estimates are much higher than those seen here (1.01–1.10), but both precede our study by 1–3 decades. Importantly, our period of observation overlaps with the advent of the first and second-generation disease-modifying therapy (DMTs), while the Koch-Henriksen study concludes just as these DMTs began to become widely available. While further advancements in diagnostic methods for earlier MS detection also occurred in this interval, healthcare for people with MS only materially changed in terms of the introduction of these medications. Our findings suggest a beneficial impact of these DMTs on reduced mortality in people with MS, bringing their mortality on par with the general population, and a benefit of DMTs and overall improved healthcare for people with MS beyond their immediate impacts on clinical progression.

Of interest are the differences seen over time by sex, with increases in prevalence and incidence consistently higher in females. That prevalence increased more among females than males is partly attributable to the difference dynamics seen in incidence, which increased from rough parity in 1951–1961 and 1971–1981 to being more than double by 2001–2009 and 2009–2019. This increasing sex ratio has been described previously, and of interest this increase is more prominent at higher-latitude locations like that of Hobart.23–26 There is no reason to expect this increased incidence in females is attributable to bias of increased and earlier detection by sex, and may more likely reflect some of the factors which underlie the female preponderance for MS, such as hormonal or genetic effects.27–29 It is possible that these factors, including environmental and lifestyle, may interact with sex-specific characteristics to realise the changes in sex ratios observed here.30 Further investigation of these interactions should be undertaken, both to better our understanding of the pathophysiology of MS as well as to identify possible points of intervention.

MS epidemiology by place and time

Review of studies conducted over the period 1961–2019 reveals that prevalence and incidence continue to be higher in Hobart (43.5°S) than in Newcastle (32.9°S), but this gap may be narrowing (table 5). There is a remarkable abrogation of the latitudinal variation in prevalence, from twofold higher before 1996/2001 to confluency in 2009/2011. A similar pattern can be seen in the ratio of age-standardised incidence rates, reducing from 1.8 in 1981 to 1.2 in 1996/2001 and 1.1 in 2009/2011.

Table 5

Comparison of prevalence and incidence rate between Newcastle, New South Wales and Hobart, Tasmania, using summary data from epidemiological studies conducted over the period 1961–2019

The changes in prevalence and incidence between Hobart and Newcastle over this timeframe may partly reflect changed demography. Hobart has seen a marked ageing over time: 10–29 years comprised roughly one-third of the population in 1981 but this reduced to 20% by 2009, while in Newcastle this ageing has not been as appreciable. Thus, the age group at which MS most commonly presents is still well represented in Newcastle but is a relative minority in Hobart. It is also possible that adoption of sun-protective behaviours have been more pronounced in Newcastle than Hobart, resulting in an attenuation of the latitudinal differences in individual UV-load. There have also been appreciable increases in vitamin D supplement use, this a consequence of increasing societal awareness of vitamin D and vitamin D deficiency. Among people recruited soon after a first demyelinating event and following them over time in the Ausimmune Longitudinal Study, we showed a 24-fold increase in the proportions taking vitamin D supplements, from 1.4% at baseline to 34.8% 5 years later.31 This increase was not uniform across the study sites, however: though comparable at baseline, by 5-year review the proportion using these supplements was two times higher in Hobart than Newcastle (32.6% vs 16.7%).


This study was conduced in a stable, geographically confined and well-educated population with centralised healthcare provision facilitating thorough case ascertainment. Building on three previous studies has enabled longitudinal evaluation over nearly 70 years of the epidemiology of MS in this region. While some patients with a progressive MS phenotype or very mild disease may not engage with neurology care, the use of a comprehensive digital medical record database that enables rapid search of cases would have facilitated their inclusion in this study despite lack of recent attendance at a MS clinic. In respect of misdiagnosis of disease, the six MS cases from the 2009 study that had to be reclassified subsequently demonstrate this is possible. However, with increasing use of MRI and the improved MS diagnostic criteria, such misclassification should be uncommon.

While the area which comprises Greater Hobart has been broadly consistent throughout the period of study described here, increasing in area but not so much included population since 1961, we do acknowledge some potential for patient migration within the state to reside closer to the tertiary medical centres in Hobart city. This is supported by the fact that while only five incident cases emigrated during 2009–2019, 60 cases immigrated, which may include intrastate migration to be closer to the city, which would contribute to prevalence. That said, there is no reason to expect an especial inward migration in the 2009–2019 period and the similarly increased incidence would suggest a genuine increase, rather than just local migratory effects.

Another consideration is the impact of changing diagnostic criteria. While the 2001–2009 and 2009–2019 both used the McDonald diagnostic criteria,13 32 33 the 1951–1961 McCall study5 used the Alison/Millar criteria34 and the 1971–1981 Hammond study6, the Poser criteria.35 The more recent surveys’ utilisation of the McDonald criteria which incorporate paraclinical evidence, particularly MRI, in a primary role enabled better capacity to detect cases with greater sensitivity and accuracy than the earlier studies whose diagnostic criteria relied solely or primarily on clinical symptom-based evidence. Particularly, these improvements in diagnostic sensitivity enable diagnosis at an earlier age as it is not solely necessary to rely on two or more clinical episodes for diagnosis, this shown here where the average age at diagnosis decreased from approximately 46 years in the 1951–1961 survey5 to 39.6 years here. Another element is the effect of improving treatment and particularly the advent of more highly effective disease-modifying therapies which would motivate neurologists to diagnose patients sooner to enable their access to these treatments. The combination of diagnostic criteria able to diagnose after a single episode plus paraclinical evidence alongside neurologist motivation to diagnose sooner to enable patient medication access may have contributed to some of the increases in prevalence seen since 1951–1961 and even since 2001–2009.

Another potential issue may lie in how we assessed cases in this study. Rather than conducting a new review of medical records to identify cases, we started with those cases on file from our previous 2001–2009 survey, validating their diagnoses and residence statuses, and then adding to this newly identified cases. While this has the benefit of both increasing our included sample size as well as potentially catching some cases who were previously resident in the study area but who since emigrated or were deceased, this would have led to greater case ascertainment than were possible in earlier surveys and thus, realised a higher prevalence than might have been the case using equivalent methods.


This study demonstrates a significant and important increase in MS prevalence and incidence in a well-studied MS population in Hobart, Tasmania, Australia. We noted also a significant continuing fall in mortality and an overall ageing of the MS population. This has important implications for healthcare providers and funders because MS is a costly illness to treat and has a high personal and societal cost. If these changes are reflected in other well-studied populations, public health measures to decrease the drivers of MS risk should be implemented in high-risk populations such as that of Hobart.

Data availability statement

Data are available on reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by Tasmanian Health and Medical Research Ethics Committee (ID=HOO17858).


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors SS-Y: data management, statistical analysis, manuscript drafting; RA: collected data, data management, critical review; LB: statistical guidance, critical review; IvdM: critical review; BVT: project conception, critical review; All authors have reviewed the article and approve it for submission. BVT accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.

  • Funding The Multiple Sclerosis Research Flagship is supported by a grant from the Australian Medical Research Future Fund: Emerging Priorities and Consumer Driven Research.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.