Name: Propensity score matching methodology
Uploaded: 2023-03-14T00:00:00.0000000
Duration: T00H03M02S
Description: Propensity score matching methodology

Name: Propensity score matching methodology

Description: Propensity score matching methodology

Thumbnail URL: https://cadmoremediastorage.blob.core.windows.net/0ff884ff-08a4-4586-b1ca-7adbe74c32f2/videoscrubberimages/Scrubber_1.jpg

Duration: T00H03M02S

Embed URL: https://stream.cadmore.media/player/0ff884ff-08a4-4586-b1ca-7adbe74c32f2

Content URL: https://cadmoreoriginalmedia.blob.core.windows.net/0ff884ff-08a4-4586-b1ca-7adbe74c32f2/PSM Medical Version.mp4?sv=2019-02-02&sr=c&sig=tX4kHq74UyDJGGl%2FZeGa29mm1iU%2B89dRk9gzGh9z7Lk%3D&st=2024-11-21T09%3A56%3A39Z&se=2024-11-21T12%3A01%3A39Z&sp=r

Upload Date: 2023-03-14T00:00:00.0000000

Transcript: Language: EN.
Segment:0 .

SPEAKER: Propensity score matching. Why and how it is used to evaluate effectiveness data from observational studies that do not have an internal comparator. Randomized controlled trials, or RCTs, randomly assign patients who meet the inclusion criteria to either a treatment or a control group, and are considered the gold standard for evaluating treatment efficacy. However, observational studies without an internal control group, such as product registries and open-label studies, are sometimes needed to assess the effectiveness of treatments, including those for rare diseases.
SPEAKER: These studies can enable long-term data collection in a heterogeneous population in order to evaluate the real-world safety and effectiveness of a drug. Propensity score matching is a statistical method for reducing bias, and estimating treatment effects when it is not possible to randomly assign patients to different groups. This well-established methodology aims to select a subgroup of patients from an external comparator group, such as a natural-history study who have similar predefined characteristics to those in the study population of interest, thus mimicking the randomization process in RCTs.
SPEAKER: By matching patients according to these criteria, confounding factors can be minimized so that outcomes can be more confidently attributed to the treatment, rather than any other differences between the groups. So how does propensity score matching work? Before patients can be matched, a propensity score based on key patient characteristic data or covariates must first be calculated for all patients in both the study of interest and the comparator group.
SPEAKER: The score is calculated using preselected covariates allowing a valid comparison between the two groups to be made. For example, the age of patients when they first experienced symptoms is one covariate that helps match patients with a similar predicted rate of disease progression. The propensity score is a collation of the information about all the covariates of interest into one number for each patient.
SPEAKER: Once propensity scores have been calculated, a one-to-one matching process is initiated whereby each patient in the study of interest is paired with the patient in the comparator group with the most similar score until patients are matched. The quality of propensity score matching can be confirmed by showing that significant differences in covariates between the groups before matching are eradicated after matching.
SPEAKER: To summarize, in observational studies, such as those for rare diseases, propensity score matching is used to mimic the randomization process of RCTs, enabling treatment effectiveness to be assessed.

Cadmore media player playing video Propensity score matching methodology

Video Player

Transcript

Segments

End of Video Player Control