Chances and challenges of a long-term data repository in multiple sclerosis: 20th birthday of the German MS registry

In 2001, the German Multiple Sclerosis Society, facing lack of data, founded the German MS Registry (GMSR) as a long-term data repository for MS healthcare research. By the establishment of a network of participating neurological centres of different healthcare sectors across Germany, GMSR provides observational real-world data on long-term disease progression, sociodemographic factors, treatment and the healthcare status of people with MS. This paper aims to illustrate the framework of the GMSR. Structure, design and data quality processes as well as collaborations of the GMSR are presented. The registry’s dataset, status and results are discussed. As of 08 January 2021, 187 centres from different healthcare sectors participate in the GMSR. Following its infrastructure and dataset specification upgrades in 2014, more than 196,000 visits have been recorded relating to more than 33,000 persons with MS (PwMS). The GMSR enables monitoring of PwMS in Germany, supports scientific research projects, and collaborates with national and international MS data repositories and initiatives. With its recent pharmacovigilance extension, it aligns with EMA recommendations and helps to ensure early detection of therapy-related safety signals.

The GMSR's design. The GMSR established a network of participating centres across different healthcare sectors in Germany (see Fig. 1). Centres that were accredited 'MS Centre' , 'Specialised MS Centre' , or 'MS Rehabilitation Centre' by the German MS Society participate in the GMSR documentation. Certificates are awarded to university clinics, acute care clinics, rehabilitation clinics, MS outpatient clinics and resident neurologists complying with specific criteria. Compliance is re-evaluated every 2 years 17 . With this certificate, each centre commits itself to recruiting PwMS and collecting data for the GMSR. Physicians and other medical staff carry out patient enrolment and data collection for the registry during routine examinations. After the provision of written informed consent by adult PwMS (see Table 1), previous medical data can be collected, and prospective data collection starts and is carried out until the patient's death or withdrawal of informed consent.
Members of the medical staff enter data into a web-based electronic data capture (EDC) system. This system allows a single patient's data to be reported by multiple centres. To avoid duplicate patient entries, patientidentifying data (first name, last name, sex, date of birth) is entered into the registry at enrolment and used to generate a unique pseudonym. A trusted third party stores these pseudonyms and the corresponding identifying data separated from the registry data. To guarantee patient privacy, only de-identified medical data is stored in the registry's database and used for analysis. Data collection. In 2005, the registry started regular operations and has been continuously improved since then. At first, data was collected through the 'MSDS Klinik/Praxis' software, which had been locally installed in centres. Encoded datasets stored digitally were transmitted quarterly to the MSFP-gGmbH (MSFP) for quality checks and analysis. Since 2016, it is mandatory to document via the standardised web-based platform and device-independent EDC-system 'secuTrial' and documentation through MSDS has been discontinued. The web-based EDC-system is compliant with established tools and concepts of the Technology and Methods Platform for Networked Medical Research e.V. (TMF) and is certified for conducting clinical trials (GCP, GAMP5, FDA 21 CFR Part 11). PRO can be integrated to the system through patients uploading information themselves through an app or web browser. Data entry happens directly through interfaces of the EDC system.
Governance of the GMSR. The GMSR is operated by the MSFP, a non-profit organisation founded in 2001 as a subsidiary of the German Multiple Sclerosis Foundation to develop and manage the registry on behalf of the German Multiple Sclerosis Society. The governance of the GMSR includes a scientific advisory group of clinicians and representatives of MS centres as well as methodology experts (of medical statistics, epidemiology and medical informatics), that approves and oversees scientific projects and research activities 18

History of GMSR.
The history of the GMSR is described in Supplementary Document 1.

Dataset.
In 2014, the GMSR's dataset was revised to fulfil the increased requirement for harmonization and comparability of data of different data sources. As a result, the dataset is a combination of datasets of the German Multiple Sclerosis Society and the German Competence Network Multiple Sclerosis. The current minimal dataset (see Supplementary Table 1) contains information on disease onsets, diagnoses, MS disease courses 19,20 ,   21 . In 2019, the GMSR dataset was extended by a pharmacovigilance module to include documentation on medical history, body mass index, adverse events and pregnancies (see Table 2). This documentation has been rolled out to eligible centres. The GMSR dataset is expandable regarding new research questions and adapts to the specific requests of documenting parties using its dynamic web-based EDC system and modular structure. Physicians schedule frequency of visits according to patients' needs for medical care and routine follow-ups, which the registry recommends to be carried out at least on an annual basis. Relapses, adverse events, and pregnancies are documented as separate events.

Data quality.
To constantly ensure high data quality, the GMSR established a wide range of quality control measures. The web-based data entry system ensures that centres have the latest versions of questionnaires. Each questionnaire is checked for consistency of dates, of value ranges and for plausibility. Branching logic (asking only relevant questions based on previous answers) ensures that data is only entered for applicable fields. Furthermore, cross-reference-checks compare data between different questions and forms. Longitudinal checks aim at preventing implausible data changes over time. Additionally, the registry uses R-Syntax for automated quality control to find discrepancies and implausibilities. Participating centres receive findings through query management system within the EDC-System. All data changes are recorded in audit trails. The database only accepts records with the completed minimal dataset. Participating centres are provided with training materials and tutorials. Yearly benchmarking reports include visualised comparative data, based on the centre's population versus other centres' populations of the same type and individual feedback on data quality. Centres and MSFP can execute ad-hoc statistics and reports as well as customised database searches at any time.
Data analysis. The GMSR conducts retrospective analysis for a variety of MS-related research questions.
These include demographic studies, measurements of care availability, causal inferences but also pharmacovigilance studies and epidemiological assessments 6,22-25 . Most analyses are performed using R 26 . Before data is analysed, it is checked for duplicates, completeness, plausibility and consistency. The GMSR's data is accessible at Status quo. As of January 2021, 187 centres of all healthcare sectors throughout Germany participate in the GMSR, and the new registry database includes documentation of more than 196,000 consultations of more than 33,000 PwMS. Figure 1 shows the geographic distribution of participating centres across Germany. 5% of registered patients were reported by multiple centres. Recent estimations based on the claims data of participating centres put the number of PwMS treated annually by the participating centres at approximately 90,000. Thus, covering about one-third of the German MS-population (estimated at 240,000-250,000). At time of data extraction, the yearly documentation for 2020 included 26,491 visits of more than 13,000 patients, with an average of two consultations per patient per year (see Fig. 2). 3530 patients joined the registry in 2020. Table 3 shows the number of visits in which specific (sets of) variables are present.
In total, more than 33,000 PwMS were included in the MS Registry since 2014. Baseline data separated by disease course as well as the current distribution of EDSS scores of patients are presented in Supplementary Table S2 and in Fig. 3. In 2020, 4% of visits were documented by rehabilitation centres, 43% by MS centres and 53% by specialised MS centres. By January 2021, 75% of queries on visits data of 2020 were answered and closed. The average time for a final response to a query was 4 (± 7) months (median 1 month) and 81% of all queries were answered.
The GMSR baseline demographic data is comparable to recent publications based on German statutory health claims data regarding female to male ratios, most affected age groups and geographical distribution. The age    Table 4 compares baseline GMSR patient data with data from the registries of Italy, Denmark, France, and Sweden 28-32 .

Funding. The establishment and maintenance of the GMSR is funded by the German Multiple Sclerosis
Foundation and the German Multiple Sclerosis Society. Additionally, the MSFP has received financial support from Biogen, Bristol Myers Squibb, Merck Serono, Novartis, Roche and Sanofi for the specific purposes of operating the GMSR. Principles of the cooperation between MSFP and pharmaceutical companies are: (1) Transparency: each stakeholder gets the same information at the same time and the type and scope of sponsorship is made public, (2) Financing: to avoid an increased influence of a single stakeholder or a financial dependence on one funder, stakeholders contribute equal amounts, regardless of the company's market share, (3) Information exchange: stakeholders meet at least once a year to discuss results, ideas for scientific evaluations and further joint activities, (4) Scientific independence: the MSFP has the right to publish registry results independently.

Registry-based studies and sub-cohorts.
Besides the operation of the MS registry, the MSFP develops and provides the EDC system for MS research projects that can (partly) be linked to registry data (see Supplementary Table 2) 33 .

Discussion
The GMSR is a unique and reliable long-term data repository for real-world data on epidemiology and health of PwMS in Germany. It allows long-term monitoring of the disease course, analysis of different treatments and evaluation of the comparative efficacy, effectiveness and safety of an increasing number of DMT, which cannot be addressed as much by randomised clinical trials (RCT) 10,24,34 . The GMSR holds data on all approved DMT instead of focusing on specific drugs 24 . In RCT, inclusion criteria often do not fit the effect mechanism of the active substance, for instance, patients with low disease activity or long disease duration are included in studies on high anti-inflammatory effect substances 7 . Their high costs and intensity result in their rather short duration and insufficiency concerning real-world effectiveness in the long-term 1,7 . Therefore, registry data is important for  www.nature.com/scientificreports/ patients, physicians, industry, regulatory authorities and health insurance system to provide support to treatment decisions and the development of health promotion measures 5,10 . The increasing number of DMTs leads to individual treatment decisions, which consider disease related variables like symptoms or clinical findings, but also patient preferences. The evaluation of the severity and significance of specific patient symptoms and their consequences for disease management are often different from that of physicians 35 . Consequently, PRO will become more important 3,7 . The technical infrastructure of the GMSR allows implementation of PRO measurement via eCRFs that can be completed directly by patients and is directly linked to medical data 36 .
The creation of an added value within the GMSR for health professionals to document disease progression and treatment history directly through the EDC system, like automatically generated discharge letters and medical reports, could encourage them to participate in the MS registry and is constantly pursued by the GMSR. A technical interface to the centre's primary information system could have a similar effect due to the reduced documentation effort. It could furthermore establish a documentation system for several different prospective studies and reduce time and costs expenditure in the future. Despite its positive effects, the reuse and integration of electronic healthcare records for research purposes is still in its early stages 37 .
To increase follow-up rates, it is necessary to provide an incentive for participating centres to document visits regularly. Possible incentives which are offered by the GMSR to achieve this goal are financial compensation, requirements analysis at annual user meetings and guidance documents for follow-up processes.
Over the years, the GMSR's findings have benefited MS science. The monitoring of PwMS through the registry allowed for the first-time estimates concerning its prevalence, disease types, symptoms and degree of symptomatic treatment. Furthermore, the registry enables register based RCTs and General Data Protection Regulation compliant patient recruitment.
Recently, evaluations of registry data on symptom frequency and symptomatic treatment patterns showed that fatigue is one of the most common symptoms. This has only been analysed in smaller cohorts 38 but is of high importance as fatigue impacts patients well-being and employment 30 . Through GMSR's participation in international projects, insights and comparison of data of PwMS from other countries and healthcare systems, such as employment statuses 23 , time until the initiation of a DMT 22 and validation of secondary progressive MS classification methods 39 were gained.
However, the GMSR suffers from certain biases and limitations. The registry is not population-based. To date, there is no legal obligation to report MS cases in Germany. The introduction of an obligation would contribute to completeness of the registry's data and nationwide coverage. PwMS are treated in all areas of care (e.g. general practioners, urologist etc.) and are therefore not always seen regularly in a specialised neurologic unit, thus a legal obligation like in cancer would ensure the completeness. As the registry focuses on neurological care providers of special experience in the treatment of PwMS, the proportion of clinically diagnosed RRMS cases is higher than presented in the German statutory claims data. Our estimations show that since its revision in 2014, the registry has captured more than 33,000 PwMS resulting in about one-third of the PwMS seen in the participating centres. If data from the prior registry data collection, commenced at the turn of the century, is included, the coverage of the GMSR reaches more than 73,000 PwMS. On annual basis the participating specialised centres treat roundabout one-third of the prevalent German MS-population. Especially PwMS suffering from progressive disease courses tend to seek neurological expertise less often, possibly explaining their underrepresentation in the GMSR. However, by including PwMS from specialised rehabilitation centres, the GMSR still captures a higher percentage of the progressive patients than if it would just recruit from outpatient facilities 40 .
Recent research on the conversion of RRMS into SPMS within the GMSR and other registries showed that significant amounts of patients might be clinically labelled as RRMS while still receiving DMTs although objective algorithms would judge the disease course differently 27 .
With legal barriers in place that make the inclusion of under-age PwMS quite challenging and specific paediatric data collections already in place it was decided in the conception phase of the GMSR to allow registry entry only to adult PwMS. Paediatric MS patients are often treated in different settings than adult MS-patients. However, the informed consent allows the collection of retrospective data from patients and paediatric onset MS patients are included after their 18th birthday in the GMSR 41,42 . With no public funding for maintaining a long-term operation of a data source like the GMSR, funding remains a challenge that cannot be solely covered by the patient society. The GMSR has therefore started to involve the industry for funding. The extension of the dataset by the pharmacovigilance module was an important step for the use of registry data in non-interventional PASS, which was promoted by the EMA 43 . This may reduce costs, increase transparency, and create confidence in the value of pharmacotherapy of MS. www.nature.com/scientificreports/ T.F., J.H., K.H., T.P., C.W., F.P. and U.K.Z. critically revised the completed final draft for important intellectual content and contributed to the revision. A.S. substantially contributed to the conception, to the study design and to the data interpretation. He supervised this study and critically revised the completed final draft for important intellectual content.

Funding
The German MS Registry of the German MS Society was initiated and funded by the German MS Foundation and the German MS Society in 2001. It is operated by a not-for-profit company, the MSFP. In 2021, Biogen, Celgene (BMS), Merck, Novartis Roche and Sanofi are participating in the multi-stakeholder funding approach to support the registry's operation and to allow the collection and reporting of data required as part of the EMA-minimal data set. Industry funding does not result in restrictions to publishing data, nor do the funders have access to the raw data or have any influence over the scientific conduct of the registry.