Performing a simple count, how to count particular entries as 2 and others as 1

[CM1]

Dear R Community,

I have a dataset of drugs and need to count how many each individual has been prescribed using:

drugs_filtered >%>
distinct(prochi, approved_name) >%>
count(approved_name, sort = TRUE -> table_1

BUT some of the drugs (defined as approved_name variable) are combination drugs and for a small number of drugs I need to count them twice. For example, co-codamol. Can I write a caveat into this code so that these combination drugs count = 2 and ll other drugs count = 1?

Many thanks,

Clare

[fredoxvii]

You could create new data. One new column splits the drug name by '-', another new column counts the number of elements or drug names in the first column. Then you can sum the second column and get the total number of drugs used.

[CM1]

Thanks very much. Would be difficult to do as many of the combination drugs don't have a "-" so would have to work through manually?

There are probably 20 drugs that are combination - is there a way to add this to the existing code and ask R to count these specific names with a value or 2 and default for all the others as 1?

Many thanks

[CM1]

Unless I can create a new variable for all data frame called drugs_filtered where the default is 0 (ie there is no second drug component) and then I write code to include any of the 20 or so drugs where there are two drugs? Not quite sure how I would do that - would it be mutate() function?

For example, do I want to ask R to mutate a new column and only populate the new variable approved_name_combination for certain drugs (where the original variable approved_name == "CO-CODAMOL")?

mutate(approved_name_combination... )? how to reference the approved_name variable and the caveat to include only some of the drugs?

Many thanks,

Clare

[fredoxvii]

Without looking at data, it would be difficult to propose a solution. Can you run an n_distinct on the approved_name column to see how many combinations there are?

[CM1]

Thanks - yes I can but then we are aggregating the figures for each individual and performing logistic regression according to individual characteristics so it needs to be individual level counts.

I am really new to R so making a new column is really difficult! Is it something like

mutate(approved_name_combination = approved_name IF *it happens to be called co-codamol), just don't know how to do the IF thing? Presumably then for drugs not called co-codamol then the new variable would have a value of 0?

Many thanks again, really appreciate your help.

Clare