The desired output is memory inefficient, since there will be lot of NA values generated. If the further postprocessing doesn't require these NA values then it should be avoided.

Instead you can concatenate each patient data row wise and use a separate index column to identify and access a specific patient data.