Recommendations for the use of propensity score methods in multiple sclerosis research

Abstract

Background: With many disease-modifying therapies currently approved for the management of multiple sclerosis, there is a growing need to evaluate the comparative effectiveness and safety of those therapies from real-world data sources. Propensity score methods have recently gained popularity in multiple sclerosis research to generate real-world evidence. Recent evidence suggests, however, that the conduct and reporting of propensity score analyses are often suboptimal in multiple sclerosis studies. Objectives: To provide practical guidance to clinicians and researchers on the use of propensity score methods within the context of multiple sclerosis research. Methods: We summarize recommendations on the use of propensity score matching and weighting based on the current methodological literature, and provide examples of good practice. Results: Step-by-step recommendations are presented, starting with covariate selection and propensity score estimation, followed by guidance on the assessment of covariate balance and implementation of propensity score matching and weighting. Finally, we focus on treatment effect estimation and sensitivity analyses. Conclusion: This comprehensive set of recommendations highlights key elements that require careful attention when using propensity score methods.

Oral Antibiotics Alone versus Oral Antibiotics Combined with Mechanical Bowel Preparation for Elective Colorectal Surgery: A Propensity Score-Matching Re-Analysis of the iCral 2 and 3 Prospective Cohorts

“…The strengths of this study are the large number of enrolled patients in a well-defined time-lapse study, representing a real-life snapshot of the surgical units performing colorectal resections in Italy, and its PSMA methodology. Following recommendations for the use of propensity score methods [ 72 , 73 ], a rigorous patients selection from the parent population and the reasoned inclusion of 20 conditioning variables were performed to limit data imbalances. Moreover, both a clear and restrictive balance algorithm, together with the evaluation of the treatment effects through an adjusted multiple regression model including the same 20 covariates used for matching, were used ( Figure 1 ).…”

Abdominal drainage after elective colorectal surgery: propensity score-matched retrospective analysis of an Italian cohort

“…The main strength of this large sample size study is that it followed rigorous guidelines for applying PSMA 29 , 45 , being based on the following items: rigorous patient selection from the parent population, performed upon explicit criteria: to limit data imbalance, several potential confounders related to the surgical procedure (delayed urgency, operations without any abdominal incision/trans-anal procedures) or exclusively impacting on a subgroup of patients (anastomosis located <10 cm from the anal verge, neo-adjuvant therapy, proximal protective stoma, administration of perioperative steroids, patients treated by dialysis) were excluded; a reasoned inclusion of 21 conditioning variables (covariates): hospital type, surgical unit type and centre volume to account for the potential imbalance of multicentre, clustered data; adherence to the ERAS pathway items to account for the potential imbalance of medical, anaesthetic and surgical perioperative management; resections for benign and malignant diseases, mini-invasive or open surgery, standard and non-standard procedures 24 , intracorporeal (anastomosis performed under visual control through the scope) or extracorporeal (anastomosis performed under direct visual control through an open access) anastomoses, stapled or handsewn anastomoses, end-to-end or different fashion anastomoses, and operation length, in relation to the imbalance of the surgical treatment; pre- and intrapostoperative blood transfusion(s) to account for transfusion-related morbidity rate 46 ; age, sex, ASA class, body mass index, diabetes, chronic renal failure, chronic liver disease, and Mini Nutritional Assessment–Short Form, to account for patient imbalance; evaluation of the treatment effect through an adjusted multiple regression model including the same 21 covariates used for matching 40 ; a clear, sheer and restrictive balance algorithm ( Fig. 1 ); a sensitivity analysis for unmeasured confounders.…”

COPP-MS: COrticosteroids during the Post-Partum in relapsing Multiple Sclerosis patients

“…This therapeutic decision should reflect the decision of the neurologist based on the risk to reach the outcome, for instance, the risk of relapse. Thus, PS was obtained from a multivariate logistic regression including variables statistically associated with each specific outcome (Cox model, level of significance: 20%) [27][28][29] For each outcome, following variables were considered potentially associated with: age at delivery, MS duration at delivery, ARR over the two years before pregnancy, number of relapse during pregnancy, EDSS at pregnancy onset, MRI within the two years before pregnancy, breastfeeding, DMD within the year before pregnancy and DMD during pregnancy. The association of each variable with each outcome were studied using Cox model (level of significance: 20%).…”