Insights from a Danish nationwide implementation of patient-reported outcomes in multiple sclerosis

Abstract

Background

Multiple sclerosis (MS) monitoring often relies on traditional methods like the expanded disability status scale (EDSS) and magnetic resonance imaging, which may not fully reflect patients’ health. Patient-reported outcomes (PROs) provide valuable insights into patients’ experiences of their disease.

Objectives

This study evaluated the Danish implementation of PROs for people with MS (PwMS) 3 years post-launch, comparing clinical and demographic characteristics between PRO users and non-users, and assessing PRO adherence.

Results

As of 1 January 2025, 2121 PwMS had signed up for PRO, compared to 14,125 who did not. PRO users were younger (mean age 50.6 vs. 53.2 years), more frequently female (73.6% vs. 68.0%), had a shorter disease duration (15.1 vs. 16.4 years), and were more likely to be on high-efficacy disease-modifying therapies (75.6% vs. 62.2%). Additionally, fewer PRO users had low EDSS scores (<3, 53.7% vs. 56.1%) and high scores (>6, 16.5% vs. 18.9%). Most enrolled patients lived in Region Hovedstaden (75.1%).

Conclusion

The study highlights differences in profile between PRO users and non-users, demonstrating the PRO platform's adaptability across different ages and disability levels in PwMS. However, these findings also highlight potential implementation challenges and emphasize the need for sustained commitment to support long-term use of PRO.

Introduction

Multiple sclerosis (MS) is a chronic immune-mediated disorder of the central nervous system, causing diverse neurological symptoms that require lifelong monitoring and management. However, the diverse MS presentations and heterogeneity in disease trajectories challenge the establishment of a quick yet clinically relevant and effective disease-monitoring routine. For many years, relapses and magnetic resonance imaging (MRI) activity have been the cornerstone of monitoring disease activity, while the expanded disability status scale (EDSS) has been the primary scale used to describe the accumulation of disability. While this information supports treatment decisions, it does not fully capture the patient's status or subtle disease development. Limitations of the EDSS have been extensively discussed in the literature,^1,2 and MRI, while valuable, is not without its limitations.^3–5 Consequently, additional modalities are needed in routine care to systematically monitor the full spectrum of MS-related symptoms.

Patient-reported outcomes (PROs) are defined as any report of a patient's health status that comes directly from the patient. ⁶ PROs provide a systematic method for collecting data on parameters that are otherwise difficult or time-consuming to cover. They offer critical insights into patients’ subjective experiences, enabling closer monitoring of disease progression, symptoms, and treatment efficacy.^1,7

Over recent decades, there has been a growing emphasis on personalized treatments and care, extending to the MS field.^8,9 The diversity among people with MS (PwMS) necessitates a patient-centered approach. This is supported by surveys and registry-based studies demonstrating the positive influence of communication between patients and neurologists on treatment satisfaction and adherence.^10,11 By combining objective health data with subjective outcomes, PROs can foster patient autonomy and empowerment, ¹² support cost-effective care, and enhance the overall quality of MS management.

In Denmark, we launched an online PRO platform for recording MS-related symptoms on 1 January 2022.

Experience with a national PRO solution in MS is limited, and no prior attempt has been made to implement such a system in Denmark. Thus, this study aimed to quantitatively evaluate the Danish implementation of PRO for PwMS 3 years after its launch by comparing clinical and demographic characteristics between patients who signed up for PRO and those who did not. We further describe user behavior by investigating PRO adherence, contributions from different questionnaires, and time of completion in relation to clinical visits.

Materials and methods

Study design and data source

We conducted a nationwide observational registry-based study in Denmark with a cross-sectional design, using data extracted on 1 February 2025. The study utilized data from the Danish Multiple Sclerosis Registry (DMSR). ¹³

Patient-reported outcome platform and measures

Since 1 January 2022, the Danish PRO platform has been integrated into the DMSR, enabling patients to complete electronic questionnaires prior to regular appointments. To utilize the platform, patients affiliated with one of the 13 Danish MS clinics needed to sign up online. This was done by filling out a short form with the patient's name, email address, and hospital affiliation.

Patients were informed about PRO and how to sign up by physicians during clinical visits, flyers and QR codes in waiting rooms, electronic reminders sent with invitations to scheduled clinical visits, online campaigns including interviews and an informative video distributed by the patient association, and student assistants available in selected clinics to support patients. There were no restrictions regarding age, sex, treatment status, phenotype, or geographic location for signing up for or using PRO.

The Danish PRO platform includes three internationally validated instruments. Two of these are specific to MS, the Multiple Sclerosis Impact Scale (MSIS-29) ¹⁴ and the 12-Item MS Walking Scale (MSWS-12). ¹⁵ The third is a generic measure, the Fatigue Severity Scale (FSS). ¹⁶ In addition to these validated PROs, a Danish MS consultation questionnaire and a smoking habits questionnaire were included. The selection of all five questionnaires was informed by focus group interviews with PwMS ¹⁷ and a workshop with MS-specialized clinicians to ensure alignment with patient and clinical priorities.

Study population

We identified all patients recorded in the DMSR with a diagnosis of MS at the data cut-off date (1 January 2025). Patients with an unspecified MS phenotype, who emigrated or died prior to the data cut-off date, or who were recorded as not followed in the clinic, were subsequently excluded. Patients in the resulting study population were stratified based on whether they had signed up for PRO between the PRO launch date (1 January 2022) and the data cut-off date (hereafter “PRO users” and “non-users”). PRO users who did not answer any questionnaire were still classified as PRO users. This decision was made because some patients registered after a scheduled clinical visit and chose to wait until their next visit to complete the questionnaires. Conversely, PRO users who only contributed PRO questionnaires in the context of a clinical trial were classified as non-users. Figure 1 illustrates the selection procedure.

Figure 1.

Patient disposition flowchart.

Patients’ administrative, demographic, and clinical characteristics

We retrieved from the DMSR the following patients’ administrative, demographic, and clinical characteristics, calculated at the data cut-off date: Patients’ MS clinic affiliation, region of the MS clinic, number and timing of clinical visits, age, sex, disease duration (time from symptom onset), time since MS diagnosis, phenotype [relapsing-remitting MS (RRMS), secondary progressive MS (SPMS), primary progressive MS (PPMS)], the most recent EDSS score within the preceding 2 years, and disease-modifying therapy (DMT) status. For patients on DMT, we categorized it as high-efficacy, moderate-efficacy, or others. A list of DMT efficacy categories is provided in Supplementary Table 1. Finally, we retrieved the number and type of completed questionnaires along with the date of PRO signup.

Statistical analysis

We reported characteristics of the study population using frequencies and proportions for categorical variables and means with standard deviations (SD) or medians with interquartile ranges (Q1-Q3) for continuous variables, as appropriate.

To investigate the association between key administrative, demographic, and clinical characteristics and the probability of being a PRO user, we fitted a generalized linear mixed model with a logit link function. Further details are provided in Supplementary File 1.

In addition, the same modeling approach was applied in stratified analyses by MS phenotype. From each model, adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were obtained. Adjusted predicted probabilities of being a PRO user for each age value were also obtained. Type III analysis of variance was performed to obtain global p-values for the main fixed effects.

We used frequencies and proportions to describe: (i) among PRO users, the percentage of patients answering ≥1 questionnaire, stratified by whether they had at least 1 visit; (ii) the relative proportion of type of questionnaire answered out of the total number of questionnaires submitted; (iii) among PRO users who replied ≥1 questionnaire and ≥1 visit, the percentage of patients who submitted 0, 1, 2, 3, 4, or all the questionnaire at each visit, and (iv) the type of questionnaire submitted at each visit. Finally, we calculated the median number of days between questionnaire submission and the next clinical visit, stratified by questionnaire type and visit.

All statistical analyses were performed using R Version 4.4.1. All p-values were two-sided, and p < 0.05 was considered statistically significant.

Results

Characteristics of patient-reported outcome users vs. non-users

A total of 16,246 PwMS fulfilled the inclusion criteria. By 1 January 2025, 2121 (13.1%) had signed up for PRO, whereas 14,125 did not (Figure 1). The latter group includes 532 (3.3%) PwMS who signed up for PRO but submitted questionnaires only in the context of clinical trials.

Table 1 outlines the characteristics of both groups. PRO users were slightly younger (mean age of 50.6 vs. 53.2 years), and the age distribution is illustrated in Figure 2. The female proportion was higher among PRO users (73.6% vs. 68.0%). Additionally, the median disease duration was shorter for PRO users (15.1 vs. 16.4 years), as shown in Figure 3. Among PRO users, we observed a higher proportion of patients with RRMS (79.2% vs. 69.3%), but both groups had the same median EDSS score of 2.5, with Q₁-Q₃ of 1.5–4.0 for PRO users and 1.5–4.5 for non-users. However, a slight difference was revealed when categorizing EDSS scores into three groups.

Table 1.

Outlines clinical and demographic characteristics for PRO users and non-users.

	Total (N =  16,246)	PRO users (N = 2121)	Non-users (N =  14,125)
Age, years
Mean (SD)	52.8 (13.8)	50.6 (12.5)	53.2 (13.9)
Sex, n (%)
Female	11,170 (68.8%)	1561 (73.6%)	9608 (68.0%)
Male	5075 (31.2%)	559 (26.4%)	4516 (32.0%)
Disease duration, years
Median [Q1, Q3]	16.0 [8.72, 25.0]	15.1 [8.25, 23.5]	16.4 [8.81, 25.5]
Time since MS diagnosis, years
Median [Q1, Q3]	12.2 [5.96, 20.5]	11.5 [5.32, 19.4]	12.4 [6.05, 20.5]
MS course, n (%)
RRMS	11,473 (70.6%)	1679 (79.2%)	9793 (69.3%)
SPMS	3139 (19.3%)	308 (14.5%)	2831 (20.0%)
PPMS	1633 (10.1%)	133 (6.3%)	1500 (10.6%)
EDSS a
Median [Q1, Q3]	2.50 [1.50, 4.50]	2.50 [1.50, 4.00]	2.50 [1.50, 4.50]
EDSS, n (%) a
0–2.5	7603 (55.8%)	1108 (53.7%)	6494 (56.1%)
3–5.5	3501 (25.7%)	615 (29.8%)	2886 (24.9%)
≥ 6	2530 (18.6%)	340 (16.5%)	2190 (18.9%)
On DMT, n (%)
Yes	10,466 (64.4%)	1720 (81.1%)	8746 (61.9%)
No	5779 (35.6%)	400 (18.9%)	5378 (38.1%)
Type of treatment, n (%) b
High	6738 (64.4%)	1300 (75.6%)	5438 (62.2%)
Moderate	3603 (34.4%)	383 (22.3%)	3220 (36.8%)
Other	125 (1.2%)	37 (2.2%)	88 (1.0%)
Region of the clinic, n (%)
Region Hovedstaden	5217 (32.1%)	1593 (75.1%)	3624 (25.7%)
Region Syddanmark	3783 (23.3%)	176 (8.3%)	3607 (25.5%)
Region Midtjylland	3711 (22.8%)	173 (8.2%)	3537 (25.0%)
Region Nordjylland	1495 (9.2%)	96 (4.5%)	1399 (9.9%)
Region Sjælland	2039 (12.6%)	82 (3.9%)	1957 (13.9%)

MS: multiple sclerosis; EDSS: expanded disability status scale; PRO: patient-reported outcome; RRMS: relapsing-remitting MS; SPMS: secondary progressive MS; PPMS: primary progressive MS; DMT: modifying therapy.

^aCalculated on 13,634 patients with available EDSS; ^bCalculated on 10,466 patients on treatment.

Figure 2.

Age distribution at data cut-off date, stratified by PRO status. Dashed vertical lines indicates the median location. A density plot is overlaid on the bar plot, where each bin represents 1 year. PRO: patient-reported outcome.

Figure 3.

Disease duration distribution at data cut-off date, stratified by PRO status. Dashed vertical lines indicate the median location. A density plot is overlaid on the bar plot, where each bin represents 1 year. PRO: patient-reported outcome.

PRO users were more likely to be on treatment (81.1% vs. 61.9%) and to use high-efficacy DMT (75.6% vs. 62.2%). Finally, there were geographical variations: Among all PRO users, 75.1% were followed in an MS clinic in Region Hovedstaden (the Capital Region), while only 25.7% of non-users received MS care in Region Hovedstaden.

Probability of being a patient-reported outcome user

After excluding PwMS without an EDSS evaluation in the last 2 years, as a proxy for a clinical visit, a total of 13,634 PwMS (2,064 PRO users) were included in the main analysis. Male sex [OR (95% CI): 0.718 (0.636–0.810), p < 0.001] and longer time since MS diagnosis [0.989 (0.982–0.996), p = 0.003] were negatively associated with the probability of being a PRO user. DMT status was the stronger predictor of being a PRO user (p < 0.001), with PwMS on moderate DMT [0.611 (0.528–0.707)] or not on treatment [0.482 (0.408–0.569)] having significantly lower odds than PwMS on high-efficacy DMT. There was no statistically independent association between region of the clinic (p = 0.712), MS phenotype (p = 0.229), and EDSS (p = 0.321) [Figure 4(a)]. There was a significant non-linear effect of age (p < 0.001) [Figure 4(b)], with the probability of being a PRO user increasing to reach a peak around 60 years of age.

Figure 4.

Forest plot showing adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for the association between administrative, demographic, and clinical characteristics and PRO enrollment (A), estimated from a multivariable generalized linear mixed model with clinic included as a random effect. Estimates are presented on a logarithmic scale. Global p-values are derived from type III analysis of variance. The lower panel (B) illustrates the adjusted predicted probability of PRO enrollment across age, modeled using a natural spline (3 degrees of freedom). Red labeled points and estimates indicate significant associations. PRO: patient-reported outcome.

It is worth noting that there was substantial between-clinic variability in the probability of being a PRO user, as reflected by a relatively large random intercept variance (SD > 1), corresponding to a high intraclass correlation coefficient (ICC).

Stratified analysis by MS phenotype (sample size: RRMS = 10,306 [2,064 PRO users]; PPMS = 1,040 [122 PRO users]; SPMS = 2,288 [293 PRO users]) confirmed the results obtained in the overall analysis in the RRMS group [Supplementary Figure 1(a) and (b)]. In PPMS [Supplementary Figure 2(a) and (b)] and SPMS [Supplementary Figure 3(a) and (b)], although a similar direction of effects was noticed for all variables included when compared to the main analysis, only treatment status reached statistical significance (PPMS: p = 0.002; SPMS: p < 0.001), with PwMS not being on treatment having a lower probability of being a PRO user. In both PPMS and SPMS, the odds of being a PRO user increased in the age range between 60 and 70 years, although age overall did not reach significance.

Patient-reported outcome users’ behavior

Of the 2,121 PRO users, 994 (46.9%) submitted at least 1 questionnaire after signing up, with 113 (5.3%) doing so without a visit yet (Figure 5).

Figure 5.

Distribution of PRO engagement among patients following enrollment (N = 2,121). Patients were categorized according to whether they completed at least one questionnaire and whether they had at least one clinical visit after PRO enrollment. Overall, 46.9% of patients completed ≥1 questionnaire, while 53.1% did not complete any. PRO: patient-reported outcome.

During the 3 years after PRO launch, a total of 4,373 questionnaires were submitted, with the consultation questionnaire being the most represented (27.7%, Figure 6).

Figure 6.

Distribution of all completed PRO questionnaires by type (N = 4,373). The figure shows the proportion of completed questionnaires across categories, including consultation, Fatigue Severity Scale (FSS), Multiple Sclerosis Impact Scale (MSIS-29), Multiple Sclerosis Walking Scale (MSWS-12), and smoking status questionnaires. PRO: patient-reported outcome.

Out of 881 PRO users who had submitted at least 1 questionnaire and had at least 1 visit after signing up, the vast majority (87%) provided at least 1 questionnaire at the first visit (Figure 7), with 35% submitting all 5 questionnaires. Adherence dropped at the following visits (between 70% and 82% of patients did not submit questionnaires).

Figure 7.

PRO adherence. The figure shows, among PRO users who had a planned clinical visit, the percentage of patients who submitted 0, 1, 2, 3, 4 or all the PRO questionnaires at the indicated visit. Number of patients at each scheduled visit, numbered after PRO enrollment, is reported at the top of each bar. The analysis was restricted to the first five visits after PRO enrollment due to sample size limitation. PRO: patient-reported outcome.

The consultation questionnaires were the most frequently submitted among patients who had submitted at least 1 questionnaire prior to each visit (ranging from 86% to 100%, Figure 8). Other questionnaires had lower submission rates but were generally above 50%. The median time between questionnaire submission and the next scheduled visit was ∼30 days at the first visit (Figure 9).

Figure 8.

Distribution of submitted questionnaires by visit number and questionnaire type after PRO enrollment. The analysis is restricted to patients who submitted at least 1 questionnaire at each indicated visit. Number of patients analyzed at each visit is reported below each visit label. PRO: patient-reported outcome.

Figure 9.

Box-plot of median time (in days) between questionnaire submission and next planned clinical visit, by questionnaire type and visit number after PRO enrollment. The analysis is restricted to patients who submitted at least 1 questionnaire at each indicated visit. PRO: patient-reported outcome.

Discussion

This study describes the Danish PRO users who signed up during the first 3 years. The results revealed that an online PRO platform appears applicable to patients regardless of age, EDSS score, disease phenotype, or sex, as we observed broad representation of these factors among PRO users. This is valuable knowledge, as one might think that elderly or disabled patients would be challenged by using an unsupervised online solution.

Despite a diverse group of patients benefitting from the PRO initiative, certain characteristics were more prevalent among PRO users. The most pronounced factor was geographical affiliation; Region Hovedstaden, which includes Denmark's largest MS clinic (The Danish MS Center), had the highest number of PRO users. Despite this factor not being significant in the multivariate analysis, we still find it relevant to address it, as a single main factor likely drove it. The PRO implementation was coordinated by the DMSR, which is connected to the Danish MS Center. This likely increased awareness of the PRO offer at The Danish MS Center, which contributed to the substantial between-clinic variability in the probability of being a PRO user (high ICC), despite attempts to mitigate this effect: First, each region had appointed local PRO ambassadors. Secondly, two nurses in eastern and western Denmark organized events to promote PRO soon after its launch. Nevertheless, the distributed tools (flyers, QR codes, and electronic reminders) and staff members (including student assistants) were likely used more frequently at The Danish MS Center.

The initiative to implement PRO originated from regulatory authorities and the Danish MS Society. As a result, all MS clinics should be equally motivated to encourage their patients to engage with the PRO platform. Furthermore, the DMSR distributed flyers containing patient-relevant information about PRO to all MS clinics for hospital staff to hand out. The Danish MS Society also shared written materials and videos across various platforms, including social media, while the DMSR provided information on its website. As a result, this information campaign should have reached every part of the country, and given the high heterogeneity of Danish society, an equal impact could have been expected. However, the steering committee's location has undoubtedly influenced the impact of the information campaign at The Danish MS Center (and thereby Region Hovedstaden), overshadowing other potential geographical factors, although our study did not investigate this aspect further.

Our implementation process has highlighted significant barriers to nationwide implementations of clinical tools, such as PRO. These barriers are (at least part of) the explanation for the relatively low enrollment during the first 3 years, with only 13% of Danish PwMS (2,653 [16.3%] if including the 532 PwMS completing questionnaires in trials).

The assumed impact of the steering committee's location emphasizes that successful implementation requires persistence and ongoing engagement from both staff and patients. This claim is further supported by the adherence analysis, which confirmed the commonly suspected decrease in adherence over time despite a stable proportion of participants after the drop between the first and second visits.

Introducing PRO to patients and physicians—and convincing them of its benefits for disease-monitoring and patient involvement—proved more challenging than anticipated, given PRO's ability to uncover the symptoms that matter most to the patient. ¹⁸ Patients must invest time and effort to complete PRO questionnaires before consultations, while physicians are required to interpret new types of data without additional time or resources. ¹⁹ Thus, PRO must be integrated into existing workflows without increasing the workload—a goal that is often challenging to achieve in the early stages of implementation. However, experiences from other fields, for instance, among cancer patients, have shown increasing satisfaction over time as participants become more familiar with the PRO solution. ²⁰

Patient feedback highlights the importance of acknowledging their responses, a finding also reported in the literature. ¹² Motivation quickly declines if PRO is perceived as a parallel activity with no impact, which is why physicians unfamiliar with PRO may unintentionally undermine patient engagement by skipping responses due to time constraints. This can lead to patient discouragement and withdrawal, which could contribute to the drop in adherence among PRO users after their first visit. Moreover, patients may be less inclined to participate in the first place if a physician appears uninformed or indifferent when asked about PRO.

MS management has advanced significantly with the introduction of high-efficacy DMTs that effectively suppress relapses and reduce MRI activity. However, increasing disability is not only a result of relapse activity, but in addition, progression independent of relapse activity (PIRA) is contributing, and most current therapies have a limited effect on slowing progression. In this evolving treatment landscape, PROs offer valuable insights by capturing subtle, gradual changes that may not be apparent in routine clinical or radiological assessments. ²¹ For instance, fatigue—one of the most common and debilitating symptoms in MS—is inherently subjective and can only be reliably quantified through patient self-reporting.^21,22 Thus, integrating PROs into clinical practice as standardized routine questions can help reveal subtle changes over time, which is particularly important when aiming to delay progression—an emerging goal for future treatment strategies. ²³ For the same reasons, PROs are considered a key component in composite outcomes that measure treatment effects on PIRA. ²¹ Given the growing recognition of PIRA, PROs may serve as a complementary tool to capture subtle, patient-perceived progression that may not be detected by conventional clinical or MRI measures.

As the number of PwMS continues to rise, ²⁴ the demand for efficient, scalable monitoring strategies becomes increasingly urgent, particularly given limited clinical time and resources. Therefore, optimization is not only beneficial but essential for a sustainable future healthcare system. One promising development is the Danish data collection platform, which was significantly enhanced in 2024 by the introduction of an automated text generator. This feature enables physicians to reuse DMSR data, including PRO responses, to draft an initial clinical note during consultations. As a result, PRO responses are integrated directly into the patient record, reducing documentation time and potentially improving physicians’ acceptance and integration of PROs into their workflow.

Conclusion

This study demonstrates that a heterogeneous group of patients can engage with an online PRO solution. However, regional disparities and relatively low uptake indicate the presence of structural barriers to implementation. To achieve long-term success, it is crucial to sustain staff commitment, recognize patient input, and integrate the PRO solution into clinical workflows. As the population of PwMS continues to grow and resources remain limited, optimizing the use of PRO will be vital for ensuring both clinical relevance and operational sustainability. Further qualitative research is needed to contextualize these findings and provide deeper insights into user perspectives and implementation experiences.