One of my simple requirements is a separation of patients at non-profit org based on the treatment they(patients) are taking like Allopathy, Ayurvedic, Homeopathy, Psychotheraphy, etc. So while registering new patient I can choose patient identifier (prefix) accordingly. For this I added new identifier sources, Is this ok to do? This solves my problem but just wanted to confirm if am utilizing it(Manage Patient Identifier Sources) correctly and if there could be any problem in future or any other best way you know of?
Usually Patient ID remains constant for a given person. Although the same patient can also be given multiple IDs.
Most often facilities prefer giving IDs based on Geographical Location / Region. But can also give it based on service/etc. You can choose to give an ID based on treatment type… but it might be possible that the same person is taking two different kinds of treatments for different conditions (like Diabetes and Arthritis,etc).
You can also possibly use a Patient Attribute to capture type-of-treatment, and use a generic Patient ID which is incremental/year-based.
Also - please note, in terms of privacy leak, coding treatment option in patient ID is not a good idea. I can see a person’s ID and say – oh they prefer Homeopathy over Allopathy. Or this person seems to be coming for Psychiatric treatment. People will not want such information to be decipherable from their IDs which are used in various places within a facility.
@gsluthra makes excellent points. The best patient identifiers are meaningless (e.g., randomly generated) medical record numbers. When you put meaning into the identifier, you eventually run into conflicts. For example, when identifiers contain geographic information, patients move around and end up getting assigned multiple identifiers (increasing chance for errors). When medical records (e.g., paper chart filing) depend on sequential identifiers, patients move around and either break the filing method or end up getting assigned multiple identifiers (increasing chance for errors).
Best practice is using meaningless identifiers with a check digit and using a paper filing method that doesn’t depend on sequential numbers (e.g., group records by the last digit or last few digits and then file sequentially within each group).
If you do end up categorizing identifiers, consider using a suffix instead of a prefix. It’s common for identifiers to be padded with zeros (0) to a common length. When a prefix is used, people pad after the prefix – e.g., “PTH123” becomes “PTH00000123” – which have different check digits. Most check digit algorithms are unaffected by leading zeroes, so “123PTH” can be padded with any number of zeroes and does doesn’t change the check digit.
Lastly, if you end up putting letters into your identifiers, best practice is to avoid any letters that can be confused with numbers (B, I, O, Q, S, and Z) .
