Takes a table in the EAV style for a single patient and a single data item and transforms it into a tidy format
process_item(df, var_name, metadata, start_time, cadence, params)
a data table with column names as for the EAV style database tables employed in CC-HIC. The table must contain data only for one dataitem (e.g. have only a single unique `code_name`) and for a single episode (e.g. have only a single unique `episode_id`)
character string of length 1 containing the code_name to be processed
a table containing metadata. Can be found within the package in the `.variables` object
datetime of the episode start
numeric scalar indicating the table base resolution
a table containing exactly 3 columns called: code_name, short_name and coalesce_rows. The coalesce rows column should contain a summary function which is used when duplicate row level data is present.