In keeping with the other two answers, I would pursue a strict overview / edit mode information architecture with this type of application.
The overview layer should provide neatly ordered high-level drug type, delivery mode, and dosage information with emphasis on legibility and ease of finding. Present this as a deck of cards, as you're already doing. Some people are on complex drug regimes that can easily make for 15-20 item lists. My apologies by the way, I am merely inferring that this is the overall purpose of your application - it might be helpful to add user workflow information to minimise conjecture by your respondents.
If you stick to one of the Material recommended card designs with a recurrent pattern, you can optimise your overview mode UI detailing to facilitate passive consumption of the information, i.e. read above edit.
In order to offer your users ergonomically feasible edits, have them toggle easily from overview to edit mode, one drug at a time. The affordance to do so can be a single edit button or tap target in a recurrent location - this way each card needs only one interactive element, which solves your spacing problem. You might even map a double-tap-to-edit type interaction onto the entire card if you manage to build a micro-animation of sorts that zooms the card to fill the entire screen; something you should definitely user test before committing to.
In edit mode you can take advantage of the entire screen, which makes interaction patterns like...
- sliders or scrubbers
- numeral spinners (with floor and ceiling)
Material has UI pattern elements for all of these things.
A simple save call-to-action causes the app to consume any changes made, and reverts to overview mode, which is a highly learnable pattern. Besides handling the space layout problem, this slightly more compartmentalised interaction mode is also safer. This is a medical / pharmaceutical application, by the looks of it, so you need to minimise the opportunity for user error above all else - the last thing you want to see happen is someone accidentally overdosing, which could have legal as well as moral consequences. On that note, test your design with users thoroughly. Consider that with age and conditions like diabetes come vision and hand-to-eye coordination problems in some patient cohorts.
It is tempting to have users edit information in-situ, as you might do on a desktop application where space is less of a concern. However space conflicts - between high information density and ergonomically feasible interaction patterns - are an inherent challenge with mobile apps, which is why I would recommend working with depth of navigation instead.
Best of luck!