We have many patients in our source data who have more than one race and primary care provider. I know that PCP is not a required field, so we can leave this null (not preferable, but it will do), but how do I handle more than one race?
Thanks!
You may want to touch base with the folks from THEMIS Focus Group 3 for the multiple races, and Focus Group 4 for the Providers. They are debating this hotly right now.
If you want my opinion:
- People can’t have multiple races. So, you need to find a heuristic to pick one, or drop it if you cannot.
- People usually have one PCP at a time, but can have more than one over time. Pick the last one, if you can.
While I can not offer a solution to the question posed, I will offer a contrary perspective on the issue of “multiple races”.
As race is relevant in interpreting metrics or symptoms, in clinical care or epidemiologic research, knowing that a subject is Native American and African American parents would be more useful than seeing only a “mixed race” concept or “picking one” lineage, for example, when assessing creatinine or whether to consider sickle cell anemia…
By the way, I cringe at thought of coding gender in the future. On the OKCupid.com (dating) web site, where one might expect to choose one from “Straight”, “Gay”, and “Bisexual” registrants are asked to pick “up to five” amongst “Straight”, “Gay”, Bisexual", “Asexual”, “Demisexual”, “Heteroflexible”, “Homoflexible”, “Lesbian”, “Pansexual”, “Queer”, “Questioning”, and “Sapiosexual”.
Best,
Gerry
I totally understand and agree. But the reason people have more than one race is sloppy data capture in most cases. They check some box on an ugly sheet that is the 50th xerox copy of the orginal on a clip board. Plus, human race is a very weak concept to begin with. And most difference in healthcare outcomes are not due to those few true genetic differences, but due to socio-economic reasons that correlate with the perceived race.
So, not sure what to do about all that.
@Christian_Reich,
I agree that sloppy data collection accounts for a portion of cases. It may well be, as you suggest, the predominant reason. However, my impression, which I concede is not a validated instrument, suggests that 1) incidence rate of inter-racial relationships resulting in offspring has been increasing substantially over recent decades and 2) referring to oneself as multi-racial rather than choosing one, is becoming more common.
Looking at recent news-
In the US, in recent times, a person with one parent of African ancestry and another of European ancestry, would, typically, be referred to as “Black”, unless their complexion was so light that they might “pass for white”, in which case they might choose that label.
Recently, the actress Meghan Markle of the US and Prince Harry of the UK announced their engagement to wed. Ms. Markle’s parents are of disparate race. In the press accounts with which my eyes have been bombarded, it seems that she is more frequently referred to as being “bi-racial” than “black”.
Conventions are evolving. How OMOP should address this, I don’t know.
I suppose that we could have a one : many relationship from person to a new table RACE. But that would seem way more additional overhead than would be reasonable, and would then raise additional questions such as: should a person’s list of races be ordered? Do we have parallel options for people who refer to themselves as multi-racial and for those who are explicit in how they self-categorize?
My point exactly. What kind of data is that? Genetically, you have no idea whether or not the Harry/Meghan princesses/princes (or my own daughter, for that matter) have the sickle cell gene or not. It’s useless. I wouldn’t waste time on something that is 90% social and 10% biologically accurate.
Gents,
Will you guys be able to attend the next THEMIS working group? We can talk a bit more about this if you guys have the time this Thursday. The meeting is this coming Thursday at 2 PM EST/. @Tom_Galia Can you please provide the meeting info.