If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Four disease modifying therapies (DMTs) use trajectories over 10 years were identified in a large sample of people with multiple sclerosis (PwMS) including long-term non-high-efficacy DMTs (39.2%), escalation to high-efficacy DMTs (31.2%), delayed start and escalation to high-efficacy DMTs (15.4%) and discontinued/ no DMT (14.2%).
•
Age, MS type, expanded disability status scale score, and the number of DMT switches were associated with cluster belonging.
•
PwMS belonging to the long-term non-high efficacy DMTs trajectory had fewer sickness absence and disability pension days towards the last year of the follow-up.
Abstract
Background
There is limited information on the trajectories of disease-modifying therapy (DMT) use and their association with sickness absence and/or disability pension (SADP) among people with multiple sclerosis (PwMS). The objective of the study was to identify trajectories of DMT use over 10 years among PwMS, identify sociodemographic and clinical factors associated with the trajectories, and to assess the association between identified trajectories and SADP days.
Methods
A longitudinal register-based study was conducted, on a prospective data set linked across six nationwide registers, assessing treatment courses of PwMS with DMTs for the 10 years following multiple sclerosis (MS) onset. The study included 1923 PwMS with MS onset in 2007–2010, when aged 19–56 years. In each 6-month-period, their treatment was categorized as before treatment, high-efficacy, non-high-efficacy, or no DMT. Sequence analysis was performed to identify sequences of the treatment categories and cluster them into different DMT trajectories. Cluster belonging, in relation to demographic and clinical characteristics, was assessed through log-multinomial regression analysis. The association of trajectories/cluster-belonging with SADP net days was assessed using generalized estimating equation (GEE) models.
Results
Cluster analyses identified 4 trajectories of DMT use: long-term non-high-efficacy DMTs (38.6%), escalation to high-efficacy DMTs (31.2%), delayed start and escalation to high-efficacy DMTs (15.4%), and discontinued/ no DMT (14.2%). Age, MS type, expanded disability status scale (EDSS) score and the number of DMT switches were associated with cluster belonging. The youngest age group (18–25) were more likely to be in the escalation to high-efficacy cluster. People with primary progressive MS were more likely to be in the delayed start or discontinued/ no DMT cluster. Higher EDSS scores were associated to being in the other three clusters than in the long-term non-high-efficacy DMTs cluster. Higher number of DMT switches were associated with being in the escalation to high-efficacy DMTs cluster but less likely to be in the delayed start or discontinued/ no DMT clusters. Descriptive analyses showed a trend of fewer mean SADP days among PwMS using non-high-efficacy DMT than the other clusters about 9 years after onset. PwMS in the escalation to high-efficacy and discontinued/no DMT clusters had more SADP days. PwMS in the delayed start and escalation to high-efficacy DMTs cluster, started with fewer SADP days which increased over time. SADP days adjusted through GEE models showed trends comparable with the descriptive analysis.
Conclusion
This study described the long-term real-world trajectories of DMT use among PwMS in Sweden using sequence analysis and showed the association of the trajectories with SADP days as well as sociodemographic and clinical characteristics.
). A global survey of more than 12,000 people with MS (PwMS) showed that 43% and 70% of the PwMS stopped working 3 and 10 years after MS diagnosis, respectively (
). Most have been used to treat relapsing-remitting multiple sclerosis (RRMS) with ocrelizumab becoming the first for primary progressive multiple sclerosis (PPMS) (
). Currently, treatment using DMTs involves alternative strategies of escalation therapy and induction therapy. In escalation therapy, historically a more traditional approach internationally, PwMS start treatment with non-high-efficacy DMTs, later changing to high-efficacy ones. In induction therapy PwMS receive more effective DMTs as early as possible to prevent accumulation of disability, with possible de-escalation if disease control is attained (
). Early treatment using high-efficacy DMTs is reported to have better effectiveness than non-high-efficacy DMTs in reducing the probability of relapse and the worsening of disability (
Treatment Escalation vs Immediate Initiation of highly effective treatment for patients with relapsing-remitting multiple sclerosis: data from 2 different national strategies.
Early use of DMTs was also shown to be associated with better outcome in earnings and with longer time before increased use of sickness absence and/or disability pension (SADP) benefits (
). In Sweden PwMS using natalizumab, a high-efficacy DMT, showed a decrease in SA by a third and an increase in productivity after one year of treatment (
There is limited information on how treatment of PwMS with different types of DMTs looks over long follow-up periods and their possible association with SADP. The entry of newer DMTs to clinical use also warrants studying their pattern of use over time and identifying associated sociodemographic and clinical factors. Furthermore, sequence analysis (
), a relatively new method in health research for longitudinal analysis of categorical states, provides an illustrative approach to characterize different trajectories of DMTs use over time revealing treatment sequencing and changes in treatments. Sequence analysis can be used to illustrate individual trajectories in a manner suitable for quantitative analysis and to compare trajectories among different groups. One of the important features of sequence analysis is that it has a holistic perspective providing information on an entire trajectory rather than specific transitions. It also provides easily interpretable illustration of sequences and trajectories (
). These and it being easier for computations have been among its advantages in comparison to similar methods such as event history analysis and latent class analysis (
Longitudinal methods for life course research: a comparison of sequence analysis, latent class growth models, and multi-state event history models for studying partnership transitions.
). Using the sequence analysis method, the aim of the present study was to identify trajectories of DMT use over 10 years among PwMS, identify sociodemographic and clinical factors associated with the trajectories, and to assess the association between identified trajectories and SADP days.
2. Materials and methods
A prospective register-based cohort study was conducted on PwMS, sampled from the Swedish Multiple Sclerosis Registry (SMSreg) (
), with disease onset in 2007–2010 when aged 19–56. The follow-up period for each individual was divided into 6-month time windows starting from onset date. The follow-up covered a total of 10.5 years of the DMT use trajectories for each of the PwMS during the period from 2007 to the first half of 2021. The follow-up of SA/DP days covered 11.5 years for each of the PwMS within the period from 2006 to the first half of 2021.
2.1 Study population
Of a total of 2373 PwMS with onset in 2007–2010, exclusions were made for missing information (onset, diagnosis, and treatment start dates), PwMS with pediatric onset MS, and those not in working ages (<18 and >55 years at baseline (Y-1)) (to ensure PwMS remain in working-age throughout the follow-up). In addition, individuals not living in Sweden at baseline (Y-1), and those who died (n = 18) or emigrated during the follow-up were excluded. This resulted in a cohort of 1923 PwMS.
2.2 Data sources
Linked microdata on the included PwMS were obtained from 6 Swedish nationwide registers: SMSreg (
A modified Rx-Risk comorbidity index, which uses records of dispensed drugs to estimate comorbidity, was used. The validity of this method has been compared with other approaches of identifying comorbidity and has been used previously to assess comorbidity in PwMS (
Data regarding SA and DP were obtained from MiDAS. We used information on all SA spells >14 days and net SADP days (hereafter SADP days) were calculated by using gross days and the extent/percentage of SADP.
2.3 Sickness absence and disability pension in Sweden
In Sweden, all individuals 16 years or older with income from work, unemployment or parental leave benefits can apply for SA benefits from the social insurance agency in case of reduced work capacity due to disease or injury. SA and DP cover about 80% and 64% of lost income, respectively, up to a certain level. Both SA and DP can be provided on a part- or fulltime basis.
2.4 DMT states
DMT states were defined for each 6-month time period from MS onset date for a total of 10.5 years. DMT states were assigned to one of 4 categories: (1) a before treatment state, the time between MS onset and the first registered information/decision regarding treatment/no treatment; (2) a high-efficacy DMT state, a period with high-efficacy DMTs; (3) a non-high-efficacy DMT state, treatment with moderate to low-efficacy DMTs; and (4) a no DMT state indicates one of the following: a period with no DMTs since the time of registered treatment information (i.e., registered information of no treatment or treatment with non-DMTs), using no DMTs throughout the follow-up, or discontinuing a DMT. In cases of more than one state in the 6-month time period, the state with the longest duration was used.
The 16 DMTs in the present study were categorized as high-efficacy and non-high-efficacy DMTs based on classifications employed in recent literature reviews (
Comparative efficacy and acceptability of disease-modifying therapies in patients with relapsing–remitting multiple sclerosis: a systematic review and network meta-analysis.
Treatment effectiveness of alemtuzumab compared with natalizumab, fingolimod, and interferon beta in relapsing-remitting multiple sclerosis: a cohort study.
) and expert opinions from neurologists. Accordingly, alemtuzumab, daclizumab, hematopoietic stem cell transplantation (HSCT), mitoxantrone, natalizumab, ocrelizumab, ofatumumab, and rituximab were categorized as high-efficacy. The non-high-efficacy DMTs were cladribine, dimethyl fumarate, fingolimod, glatiramer acetate, interferons (interferon beta-1a, interferon beta-1b, and peginterferon beta-1a), and teriflunomide.
2.5 Statistical analyses
Descriptive analyses of demographic and clinical characteristics were performed, through proportions, means, and chi-square tests to assess the distribution of PwMS by year of onset and by clusters resulting from the sequence analysis. In comparing demographic and clinical characteristics of PwMS by onset year, beside chi square tests, one-way analysis of variance (for continuous variable version of age) and Kruskal Wallis (as a non-parametric test to compare SADP days across the onset years) tests were also performed.
Sequence analysis was performed to determine trajectories of DMTs during the follow-up. This provided an approach to assess longitudinal sequences of categorical states, revealing different trajectories based on the dissimilarities among a set of sequences (
). Two to 10 clusters of trajectories were assessed for their quality to choose the best solution. The choice of the number of clusters was based on cluster quality assessment and on whether the clusters reflected meaningful real-world DMT use trajectories (Table S1). After selection, the clusters were named to depict their overall trend.
A log-multinomial regression analysis was performed to assess the association of different demographic (age and sex) and clinical (type of MS, comorbidity, first available expanded disability status scale (EDSS) score, frequency of DMT switch and MS onset year) variables with belonging to a cluster.
Mean SADP days in each cluster over the follow-up period were calculated and generalized estimating equations (GEE) with a negative binomial distribution was used to assess mean SADP days over time in the clusters. Unadjusted models and those adjusted for sex, age, type of MS, comorbidity, EDSS score, frequency of DMT switches and MS onset year were assessed. These demographic (age and sex) and clinical variables were selected for adjustment as they were considered crucial factors in multiple sclerosis in the progression and prognosis and were found to have association with SADP among PwMS. The analyses were performed using R version 4.1.1 (R foundation for statistical computing, Vienna, Austria), SAS software version 9.4 (SAS Institute Inc, Cary NC, USA) and Stata software version 17.0 (Stata Corp, College Station, Texas 77845, USA).
The sMethods section of the Supplement provides more details on the categorization of DMTs, on the steps followed in the sequence analysis including selection of the number of clusters as well as more information on the GEE models.
2.6 Ethics
The project was approved by the Regional Ethics Review Board in Stockholm, Sweden. In this type of study based on pseudonymized register data, patient consents are not applicable.
3. Results
Of the 1923 PwMS, 70.4% were female (Table 1). At baseline, 64.8% were aged 26–45 years, 49.1% were single with no children at home, and 86.9% had at least some high school education. PwMS with RRMS or secondary progressive multiple sclerosis (SPMS) constituted 91.1% while 6.6% had PPMS. The first available EDSS score was between 0 and 2.5 in 73.2% of the PwMS. All the demographic and clinical variables were distributed evenly among PwMS across MS onset years from 2007 to 2010, based on chi-square tests. Mean SADP days in the year before onset showed some variation by onset year.
Table 1Descriptive statistics and chi-square tests on distribution of sociodemographic and clinical characteristics of people with multiple sclerosis by onset year (n = 1923).
DMT: disease modifying therapy; EDSS: Expanded disability status scale; MS: multiple sclerosis; PwMS: people with multiple sclerosis; SD: standard deviation. Statistically significant results are shown in bold.
a one-way analysis of variance.
b relapsing-remitting and secondary progressive multiple sclerosis are grouped together.
A total of 916 unique sequences of treatments were identified among the 1923 PwMS. The 10 most frequent sequences were observed in 20.4% of the PwMS (Fig. S2). On average PwMS spent 4.5 years on treatment with non-high-efficacy DMTs, 3.2 years on no treatment (before treatment (1.8 years) and no DMT states (1.4 years)) and 2.8 years on high-efficacy DMTs. Cluster analysis of the sequences of DMT states showed that 4 clusters performed best (Table S1, Table S2, Fig. 1). These were long-term non-high-efficacy DMTs; escalation to high-efficacy DMTs; delayed start and escalation to high-efficacy DMTs and discontinued/no DMT clusters termed hereafter as non-high-efficacy, escalation to high-efficacy, delayed start and discontinuation clusters, respectively. The main features of each cluster are described briefly in Table 2.
Fig. 1Sequence index plots of the four clusters of DMT trajectories among people with multiple sclerosis [DMT: disease-modifying therapy] [The figure presents the DMT use trajectory of each of the PwMS in each cluster, sorted by the DMT category at the start of the follow-up.].
Table 2Main features of the four clusters of DMT use and respective sickness absence and disability pension days.
Feature
Long-term non-high-efficacy DMTs
Escalation to high-efficacy DMTs
Delayed start and escalation to high-efficacy DMTs
Discontinued/ no DMTs
PwMS per cluster,% (n)
39.2% (753)
31.2% (600)
15.4% (296)
14.2% (274)
Description
PwMS mainly taking non-high-efficacy DMTs throughout the over 10 years follow-up
Shorter time to initial treatment with non-HE DMTs then escalation to high-efficacy DMTs
Longer time to initial treatment, then using non-high-efficacy DMTs followed by escalation to high-efficacy DMTs
Discontinuation occurred after some time on non-high-efficacy DMTs or no DMTs were taken throughout
Differences by demographic and clinical variables
Age compared to overall sample
Comparable
Younger
Comparable
Older
Proportion of people with PPMS compared to overall sample
Lower
Lower
Higher
Higher
Proportion of PwMS with EDSS score of 6+
Lower
Comparable
Higher
Slightly higher
Proportion of PwMS across onset years compared to the overall sample
Comparable
Higher in 2010 (last/most recent cohort)
Comparable across 2007, 2009 and 2010 onset; higher in 2008 onset
Comparable
Mean sickness absence and disability pension days
The trend showed a steep increase during the 6 months since onset, followed by a relatively stable trend towards the end of the follow-up It showed a statistically significantly lower SA/DP days than the others towards the final two years of the follow-up
PwMS showed higher number of SA/DP days than in long-term non-high-efficacy DMTs cluster for most of the follow-up
PwMS showed fewer mean SA/DP days from around onset up to midway through follow-up. From there it showed a relatively larger increase towards the end of the follow-up
PwMS showed higher number of SA/DP days than in the long-term non-high-efficacy DMTs cluster for most of the follow-up
DMT: disease modifying therapy; EDSS: Expanded disability status scale; PPMS: primary progressive multiple sclerosis; PwMS: people with multiple sclerosis; SA/DP: sickness absence/ disability pension.
3.2 Association of demographic and clinical characteristics with cluster belonging
The log-multinomial regression analysis showed no statistically significant sex differences regarding cluster-belonging. However, among females those who had a pregnancy during the follow-up represented a higher proportion in the escalation cluster (nearly 10%) than the others (about 4% each). Age, type of MS, EDSS score, and number of DMT switches were associated with cluster-belonging. Specifically, the 18–25 age group was more likely to belong to the escalation and the delayed start clusters than to the non-high-efficacy cluster group (Table 3).
Table 3Distribution of PwMS by cluster and a mutually adjusted log-multinomial regression analysis on the association of demographic and clinical characteristics with belonging to the four clusters of DMT use trajectories.
Variable
Distribution of PwMS by DMT use cluster (n=1923)
Adjusted relative risk ratio (95% CI) of belonging to a DMT use cluster (n=1725)
Long-term non-high-efficacy DMTs (n=753)
Escalation to high-efficacy DMTs (n=600)
Delayed start and escalation to high-efficacy DMTs (n=296)
Discontinued/ no DMT (n=274)
Long-term non-high-efficacy DMTs (n=696)
Escalation to high-efficacy DMTs (n=582)
Delayed start and escalation to high-efficacy DMTs (n=255)
People with PPMS showed a higher risk of belonging to the delayed start and the discontinuation clusters than people with RRMS. PwMS with higher EDSS scores were at a higher risk of being in the 3 other clusters than the non-high-efficacy cluster. PwMS who had one or more DMT switches were at higher risk to be in the escalation than in the non-high-efficacy cluster. In contrast, these PwMS were less likely to be in the delayed start or the discontinuation clusters. PwMS with onset in 2009 and 2010 were more likely to be in the escalation cluster than the non-high-efficacy cluster compared to those with onset in 2007 (Table 3).
3.3 Trends of sickness absence and disability pension in the four clusters
Of the 1923 PwMS, 74.6%, 29.6%, and 79.3% had at least one occurrence of SA, DP, and SA or DP throughout the follow-up, respectively. Table 2 provides brief descriptions of the SADP trend in the clusters shown in Fig. 2. Similar trends were shown in the proportion of PwMS with at least 90 SADP days across the clusters (Fig. S3).
Fig. 2Sickness absence and disability pension days per 6-month period (observed mean and adjusted mean through generalized estimating equations (GEE)) across clusters over the 10-year follow-up [DMT: disease-modifying therapy; SA/DP: sickness absence and/or disability pension; triangular shapes: observed means; circular shapes: adjusted means; the GEE models are adjusted for sex, age, multiple sclerosis (MS) type, comorbidity, expanded disability status scale score, frequency of DMT switches and MS onset year.].
The final adjusted GEE model (Table 4) and the models from the unadjusted to those sequentially adjusted for sex, age, type of MS, comorbidity, EDSS score, frequency of DMT switches and MS onset year (Table S3), showed trends in SADP days across clusters which are comparable in their relative trajectories to the observed mean SADP days. Examples of the similarity are the relatively lower SADP days in the follow-up in the non-high-efficacy trajectory and the SADP days which start low and increase over time in the delayed start cluster which are generally comparable in the observed and the adjusted models (Fig. 2).
Table 4Generalized estimating equations outputs of sickness absence and disability pension by DMT clusters overtime adjusted for demographic and clinical variables.
Parameter
Adjusted (sex, age group, type of MS, comorbidity, EDSS score, number of DMT switch, MS onset year)
Estimate
Standard error
P-value
Intercept
2.05
0.244
<0.0001
Time
0.07
0.009
<0.0001
Cluster (Escalation to high-efficacy DMTs)
0.17
0.112
0.1331
Cluster (Delayed start with escalation to high-efficacy DMTs)
−0.70
0.220
0.0014
Cluster (Discontinued/no DMTs)
0.17
0.165
0.3067
Time*cluster (Escalation to high-efficacy DMTs)
0.04
0.013
0.0012
Time*cluster (Delayed start with escalation to high-efficacy DMTs)
Looking into the final adjusted model, PwMS in the delayed start cluster had lower SADP days than the others up to the second year from onset. Towards the end of the follow-up (from the 8th year onwards) PwMS in the non-high-efficacy trajectory had lower SADP days than those in the escalation and discontinuation trajectories.
3.4 Sensitivity analysis
In the sensitivity analysis performed by re-categorizing DMTs into 3 groups, adding a modest- efficacy group, 5 clusters were chosen with an added trajectory of escalation from low to modest-efficacy DMTs (368, 19.1%) coming mainly from non-high-efficacy and escalation clusters (Fig. S4). Cluster quality metrics showed relatively weaker cluster structures. Results of the log-multinomial regression analysis were comparable to the two-group DMT categorization (Table S4). Mean SADP days in the clusters showed similar trends to the main analysis, with an added trend among PwMS who switched to modest-efficacy DMTs which was close to those on long-term low-efficacy DMTs (Fig. S5).
An additional sensitivity analysis, excluding people with SPMS, showed similar results to the main analysis in the number of clusters identified, association of demographic and clinical characteristics to cluster-belonging and the trends of SA/DP days in the clusters.
4. Discussion
This 10-year prospective cohort study of working-aged PwMS, with onset in 2007–2010, on average spent 4.5 years using non-high-efficacy and 2.8 years using high-efficacy DMTs. Four clusters of trajectories of DMTs were identified. Age was associated with cluster belonging. People with PPMS were more likely to be in the delayed start and the discontinuation clusters. PwMS with higher EDSS score were also less likely to be in the non-high-efficacy cluster. More stable trends in SADP days, with a relatively lower increase over time, were found in the non-high-efficacy cluster, with fewer days than other clusters about 9 years after onset. The trends in the observed mean SA/DP days across the clusters were comparable to those of the findings from the GEE models which were adjusted for demographic and clinical variables.
We used sequence analysis to identify trajectories of treatment with DMTs. Sequence analysis has become an important method in social science studies and more recently employed in life-course studies, analysis of career pathways, and health trajectories (
) despite a slightly different design (treatment initiation was a starting point in the French study while onset date in our study). Three of the groups (first and second line DMTs, and no treatment) identified in the French study were roughly comparable with our non-high efficacy (39.2% of the PWMS), escalation (31.2% of the PwMS) and discontinuation (15.4% of the PwMS) clusters, respectively. They had no comparable findings for the delayed start (14.2% of the PwMS) cluster due to the different start points in the 2 studies.
The predominant use of non-high-efficacy DMTs (39.2%) among PwMS could partly be attributed to the onset time 2007–2010 and that entry of most high-efficacy DMTs into market and their use in Sweden increased in the latter half of the past decade (
). This may also explain why early use of high-efficacy DMT was not the most prevalent cluster. A similar increase in the use of mostly high-efficacy DMTs with time has been shown by a study in Australia (
). In relation to the substantial duration without treatment, assessment of the trajectories from MS onset rather than diagnosis time and possible treatment delay or not taking DMTs after diagnosis could have contributed. Considerable delay in MS diagnosis had been previously reported (
) and access to MRI. Another possible reason could be related to PwMS staying longer without or with only mild symptoms.
Older PwMS were less likely to be in the escalation and the delayed start clusters in comparison to non-high-efficacy cluster which was similar to the study from France (
). A meta-analysis of 38 randomized clinical trials showed the comparatively better efficacy of high-efficacy DMTs over non-high-efficacy ones decreased with age (
). In addition, the increased risk of adverse events among older PwMS in relation to high-efficacy DMTs, particularly immune cell depleting agents (alemtuzumab, cladribine, ocrelizumab) was also identified (
), immuno-modulating effects of DMTs are not expected.
People with PPMS, compared to RRMS/SPMS, were more likely to be in the discontinuation and delayed start clusters. These 2 clusters seem to reflect the features of PPMS where there is limited options for treatment (
) have very recently been used for SPMS with disease activity. The unavailability of drugs aimed at PPMS until recently could partly explain the trajectory of taking no DMTs for a long time or not taking them at all.
PwMS with higher first EDSS scores were more likely to belong to the escalation cluster, as would be expected by indication. Escalation could also be associated with relatively recent availability of high-efficacy DMTs in the market (
). The higher initial EDSS score in the delayed start cluster could be related to delay in diagnosis and treatment. The higher EDSS scores in PwMS who discontinued or did not take DMTs was congruent with the studies from France and Italy (
The relatively fewer mean SADP days among PwMS in the non-high-efficacy cluster towards end of follow-up could partly be related to treatment effectiveness and tolerability among PwMS in that cluster. On the contrary, the higher number of SADP days with steeper increase over time noted in escalation and discontinuation clusters could have followed worsening in disease severity leading to escalation/discontinuation. As discussed above, PwMS in these clusters were also more likely to have higher EDSS scores, which might explain more SA/DP days (
). Similarly, the SADP days trend in the delayed start cluster could indicate delayed diagnosis or worsening condition which may have necessitated treatment initiation and subsequent escalation. Also, the higher likelihood of PwMS in this cluster to have PPMS and higher EDSS scores could relate to more SADP days, as a previous study showed substantial SADP among persons with PPMS (
). In the GEE models adjusted stepwise for demographic and clinical characteristics, trends of adjusted mean SADP days were generally comparable to the trajectories shown in observed mean SADP days. PwMS in the non-high-efficacy cluster showed fewer mean SADP days than escalation cluster over the second half of the follow-up. Overall, the DMT use trajectories and the SADP days seem to reflect the underlying disease progression, considering treatment choices are determined accordingly and that SADP days follow different symptoms and stages of the disease progression.
The strengths of this study include the long follow-up period, the use of sequence analysis method and the comprehensive data employed by linking several high-quality nationwide registers. As to the limitations, one concerns the consideration of individuals grouped in one cluster as the same, overlooking possible trajectory differences. The lack of baseline EDSS data for all PwMS necessitated using the first available score, which were taken after MS onset (median lag time=1.8 years). The focus on assessing long-term trajectories with earlier MS onset years prevented the possibility to observe recent cohorts more likely to initiate on high-efficacy DMTs. Another limitation could be related to not capturing time-varying covariates which may be associated to belonging to DMT use trajectory.
5. Conclusions
This Swedish 10-year prospective cohort study of identified 4 clusters of real world DMT use among PwMS, using sequence analysis, a relatively new method in health research. Age, type of MS, and EDSS score were important variables associated with the trajectories. The study also provides information on SA/DP trends in the different DMT trajectories showing fewer mean SA/DP days among PwMS on long-term non-high-efficacy DMTs than the others towards the end of the follow-up from onset to 10 years onwards. The study documented the trajectories of treatments among PwMS and how they relate to SA/DP. It also provides useful information for further studies of the individual trajectories, specific DMTs, and how they relate to work disability.
Funding/support
The project was supported by unrestricted research grants from Biogen. We utilized data from the REWHARD consortium, supported by the Swedish Research Council (VR Grant No.: 201700624). The design of the study, data collection, analyses, interpretations of data, and manuscript drafting were performed without involvement of the funding bodies. Biogen was given the opportunity to comment on the manuscript before submission. Open access funding provided by Karolinska Institutet.
FST: funded partly by unrestricted research grant from Biogen and Celgene/Bristol-Myers Squibb. AM: funded partly by unrestricted research grant from Biogen. CM: funded partly by unrestricted research grant from Biogen AH: declares no conflicting interests. KF: received honoraria for serving on advisory boards for Biogen, Merck, Roche and speaker's fees from Merck. HG: was employed by IQVIA; a contract research organization that performs commissioned pharmacoepidemiological studies, and therefore was collaborating with several pharmaceutical companies; previously funded partly by an unrestricted research grant from Biogen. AG: has received research support from Novartis. KA: had unrestricted research grants from Biogen. JH: received honoraria for serving on advisory boards for Biogen and Novartis and speaker's fees from Biogen, Merck-Serono, Bayer-Schering, Teva, and Sanofi-Aventis. He has served as PI for projects sponsored by, or received unrestricted research support from, Biogen, Merck-Serono, TEVA, Novartis, and Bayer-Schering. JH's MS research is also funded by the Swedish Research Council. EF: funded partly by unrestricted research grant from Biogen, and has received unrestricted research grants from Celgene/Bristol-Myers Squibb.
Acknowledgments
The authors acknowledge Emma Pettersson, a statistician at the Division of Insurance Medicine, Karolinska Institutet, for her support in the statistical analyses.
Treatment effectiveness of alemtuzumab compared with natalizumab, fingolimod, and interferon beta in relapsing-remitting multiple sclerosis: a cohort study.
Comparative efficacy and acceptability of disease-modifying therapies in patients with relapsing–remitting multiple sclerosis: a systematic review and network meta-analysis.
Longitudinal methods for life course research: a comparison of sequence analysis, latent class growth models, and multi-state event history models for studying partnership transitions.
Sequence analysis: where are we, where are we going?.
Sequence Analysis and Related Approaches: Innovative Methods and Applications. Springer International Publishing,
Cham2018: 1-11https://doi.org/10.1007/978-3-319-95420-2_1 ([online]Accessed atAccessed April 22, 2021)
Treatment Escalation vs Immediate Initiation of highly effective treatment for patients with relapsing-remitting multiple sclerosis: data from 2 different national strategies.
00960-9&id=gr1.jpg){kind=link}
00960-9&id=gr2.jpg){kind=link}