Patient demographics form the core of the data for any medical institution. They allow for the identification of a patient and his categorization into categories for the purpose of statistical analysis. In my work I have found that there are more or less 5 different ways of interpreting the term “demographics.”

1. Date of birth, gender (Ref: Google Health)
2. Birth year, gender, country, postal code, ethnicity, blood type (Ref: Microsoft HealthVault: Personal Demographic Information, Basic Demographic Information)
3. (A or B) + Contact information (Name, Phone, Address)
4. C + Emergency contact information, family doctor, insurance provider data
5. (C or D) + Allergies, major diagnoses, major medical history

Now here is the problem: somewhere around item C, the definition of patient demographics starts to get polluted with items such as patient and emergency contact information and patient medical record data. There is a very good reason for this concept drift to have occurred in practice and I’ll break it down as to why it happened.

We can agree that date of birth and gender creates the most classic type of demographic. We can all remember reading news articles about key demographics such as males aged 55-65. Add in ethnicity and some geographical elements such as country and postal code and we have most of the classic of categories. Google Health and Microsoft HealthVault have been careful to adhere to this formal definition.

In the medical domain, however, there are additional elements that can be used to categorize patients, such as patients with a particular medical condition, allergy or past medical event, e.g. coronary bypass surgery in 1999. I would caution against including these as core demographics though. The reason for this is that they are specific enough that they can be called something else. Even Microsoft HealthVault felt the pressure and included blood type in one of its demographic “things” and this is probably because the term sometimes acts as a catchall and they didn’t know where else to put it.

The term patient demographics, therefore, should be used primarily for the same data items as marketing people use. Everything else has a different name like patient contact information, emergency contact information, family doctor, insurance provider information, etc.

A word of caution though, medical people will almost always include patient contact information when talking about patient demographics. There’s probably no traceable reason why this has happened, but if it helps, patient contact information (+ identifiers) does make it possible to divide the patient population into groups of 1 and this is needed to count how many people belong to the real demographic categories.