Stream: implementers
Topic: tattoo
David Hay (Jul 17 2016 at 21:45):
What's the best way to represent the tattoos that a person may have? My initial thought was to have an Observation for each one, but I'm wondering if an extension to Patient might be a simpler approach...
Grahame Grieve (Jul 17 2016 at 21:58):
you want photos with that?
David Hay (Jul 17 2016 at 22:42):
eventually...
Grahame Grieve (Jul 17 2016 at 23:13):
sounds like an administative observation to me
Brian Postlethwaite (Jul 17 2016 at 23:51):
Yes, not an extension on the Patient, especially if you also want the photo's too.
John Moehrke (Jul 18 2016 at 14:15):
I understood this was a use of Patient.photo? Is it really an Observation? Is the Tattoo being considered medically relevant, or as an identity confirmation method?
Lloyd McKenzie (Jul 18 2016 at 15:54):
Patient.photo is for a head-shot to use in recognition. It's not for specific images of particular tattoos. (I'm quite curious what the use-case is to capture images of a patient's tattoos in a healthcare setting . . .)
Reuben Bijl (Jul 19 2016 at 10:20):
I have been thinking about other contexts for using the patient data structure and want to represent tattoos for identity confirmation. So yes, it would be for specific images of particular tattoos.
Lloyd McKenzie (Jul 19 2016 at 13:42):
I've heard of doing that for law enforcement. Not something I'm familiar with in the healthcare space.
David Hay (Jul 19 2016 at 17:44):
Patients in a prison...
Andy Stechishin (Jul 19 2016 at 17:46):
Identification of a corpse (patient identity is unknown and pre-deceased)
Andy Stechishin (Jul 19 2016 at 17:47):
In this instance a picture of the face may not provide useful information
Bret H (Jul 19 2016 at 18:32):
HSPC has (or is developing) a FHIR profile for skin assessment. Would tattoo fit in there? but not for identification purposes
Grahame Grieve (Jul 19 2016 at 19:16):
generally, there's an overlap between healthcare and law enforcement in lots of jurisdictions
Grahame Grieve (Jul 19 2016 at 19:18):
@Brian Postlethwaite - going back to an earlier comment - there's no extension for an Administration Observation. Shoudn't we define one?
Brian Postlethwaite (Jul 19 2016 at 20:21):
I don't know how it would/should be coded it, but a tattoo I would have expected in the observation, as they are timed, and may change over time, just like a mole. (and could be removed too).
Which is really why I don't lean toward the extension on patient, or photo on patient.
(That and would need some form of tag on the photo to indicate which to use, and would be a lot of baggage to drg around with the patient resource, than a low res photo to assit in identification)
Brian Postlethwaite (Jul 19 2016 at 20:21):
I'd leave how to code the observation up to the PC folk to define. (well not me anyway ;)
Grahame Grieve (Jul 19 2016 at 20:42):
well, I was asking about administrative observsations generally. we've certainly seen lots of them over the years, and there's a number of the v2 segments. So wouldn't it make sense for PA to define a standard extension on Patient where you can refer to an observation made about the patient? Of course these are observations precisely because these are changeable values....
Lloyd McKenzie (Jul 19 2016 at 22:12):
It would also be reasonable for there to be a distinct "category" for administration observations too (as a sibling to vitals, lab, etc.
John Moehrke (Jul 19 2016 at 22:39):
I think we should be clear on the intended use of the tattoo picture. If it is just for Identification, then it is best in Patient.photo. If it is something being treated, or affecting treatment; then it logically becomes an observation. Both use-cases seem reasonable, but using Observation for both could be very combersome for identity purposes. For example the person that needs access to the photo might not be the kind of person you want to give access to Observations. (going back to the Jail, but applicable everywhere too).
Lloyd McKenzie (Jul 19 2016 at 22:45):
I was surprised to find that Patient.photo was a 0..* as the primary use-case - showing a headshot when retrieving a list of possible matches requires only 1 - and would need to choose from among the list to display only one.
Robert Horn (Jul 19 2016 at 22:58):
By the way, If doing this for treatment purposes, there are photo image quality and calibration issues that can't be ignored. FDA, ICC, DICOM, and others have image objects and calibration requirements covered. That's a different use case.
Richard Townley-O'Neill (Jul 19 2016 at 23:47):
Tattoo removal in cosmetic surgery would be health care. Before and after photos could be useful.
Brian Postlethwaite (Jul 20 2016 at 00:22):
Just an FYI on the photo as implemented in one of our production systems (not FHIR) we only have a single photo, and it is always reduced in size to be contained within a 300x300pixel box (which ensures size is more sensible, as had some 14MB images which slowed things down somewhat)
Last updated: Apr 12 2022 at 19:14 UTC
