Observational Health Data Sciences and Informatics

• Requester: Robert Miller
• Discussion: here.

In the context of adding GIS capabilities to the CDM:

The location_id in the Person table is only capable of storing a single residence for an individual. We want to be able to know where a person's residence was for a given date. In another system, this problem is solved by introducing a Residency table, which acts as a relational entity between Person and Location.

There are several things to consider here:

• A person can change locations over a period of time
• A location can have multiple residents over a period of time
• A person can reside at multiple locations during the same period (e.g. vacation home)

The first two issues I believe a Residency table could solve. The third, if we are concerned about it, could be resolved by adding a “primary residence” boolean to the table. Note in the table below: both person and location IDs do not need to be unique in the residency table.

It was suggested that, since a large portion of implementations would only have a single location per person, we leave the location_id in the person table and have it refer to the most current residence.

The other necessary component for GIS capabilities would be to add latitude and longitude coordinates to the location table, though that may need to be discussed elsewhere.

• What zip code did person x live in when they were diagnosed with y?
• Are people who live closer to x more likely to y?
• How do care sites compare in terms of average travel time? Does it matter?
• Are there hot spots for specific conditions?