| Titre : | Analyzing Recurrent Events in Multiple Sclerosis: A Review of Statistical Models with Application to the MSOAC database |
| Auteurs : | David Herman ; Ecole des hautes études en santé publique (EHESP) (Rennes, FRA) |
| Type de document : | Mémoire |
| Année de publication : | 2024 |
| Description : | 53p. / fig., tabl. |
| Langues: | Anglais |
| Classement : | MPH/ (Mémoires MPH à partir de 2024) |
| Mots-clés : | Sclérose plaque ; Modèle statistique ; Revue de littérature ; Handicap ; Evolution ; Pathologie |
| Résumé : |
Introduction: Patients with multiple sclerosis (MS) are susceptible to experience recurrent events of disability progression and relapses. Many studies still focus on analyzing MS events with traditional methods such as Cox Proportional Hazard, Poisson, and logistic regression that either ignore subsequent events or fail to account overdispersion and dependency between events. Therefore, considering recurrent event methods may improve treatment development. Objective: To conduct a literature review to identify recurrent event methods in the context of MS and apply them on the MS Outcome Assessments Consortium (MSOAC) to provide recommendations for MS research.
Methods: A literature review was conducted to identify main methods, which were then summarized based on their classification and main characteristics. These methods were applied to the MSOAC database to evaluate the effect of the disease course on the number of confirmed disability progression (CDP) and the Annualized Relapse Rate (ARR). Results: A total of 54 articles were included in the literature review, identifying 9 main recurrent event models. The most documented were the Andersen-Gill, Prentice Williams and Peterson and Frailty models. Marginal models may be recommended in experimental studies over conditional approaches, while event-specific models are accurate for estimating overall or event-specific effects in patients with one or more events. Random effect models are suited for studies with patient heterogeneity. In the MSOAC database, recurrent events have provided more precise estimates than traditional methods. Common and event-specific estimates for CDP and ARR were consistent across models. Conclusion: This study provides methodological guidance for health researchers to select and implement appropriate methods in recurrent event analyses. The model choice may vary depending on the research study and different factors. Researchers should prioritize recurrent event methods in their statistical plans to avoid information loss and improve the precision of estimated effects. |
| Diplôme : | Master MPH of public health |
| Plan de classement simplifié : | Master of Public Health - master international de Santé Publique (MPH) |
| En ligne : | https://documentation.ehesp.fr/memoires/2024/mph/david_herman.pdf |
Documents numériques (1)
 FullText URL |