OHDSI Home | Forums | Wiki | Github

Missing age groups in propensity score covariate results

Hi all,

We ran propensity score analyses using the PLE package. In the covariate balance table, the sums of the percentages in the tabulated age groups equal to only ~67% for both the target and comparator group. The age groups 70-74 and 80-84 for both target and comparator are missing. Clinically it is unusual to have no one present in these age groups as well. Does anyone know how to fix it? Many thanks!

An example of data output is shown below

Hi @wallislau . Iā€™m sorry. I see the cause of the problem. It is this code that filters covariates if they have a zero value anywhere (in this case probably the standardized difference of means equals 0). The correct code should be something like this, where covariates are filtered if they have a zero value everywhere.

Unfortunately, this means that after fixing the code the exporting of covariates will need to be rerun at the site.

1 Like

Thank you so much @schuemie !