Active 5 months, 3 weeks ago
Data Usability Workgroup Group
Reduce the Impact of Duplicates
Tagged: Reduce the Impact of Duplicates
- This topic has 3 replies, 1 voice, and was last updated 4 months, 3 weeks ago by Andrea Pitkus.
June 9, 2021 at 7:21 pm #33468Hera AshrafKeymaster
When clinical data is exchanged between multiple systems duplicate information is a frequent occurrence. Commonly this is the result of receiving the same information from more than one external organization. Unidentified duplicate information takes clinician time to filter and reconcile and can make it harder to find the most up to date information about a patient.
Reply to this post with your answers to these questions:
1. Where do you see the most significant problems with duplicate data — Problems, Meds, allergies or labs? Other data types, e.g., Immunizations, social or other historical elements?
2. Balancing reduction in duplicates with risk of information loss and patient safety concerns can be a challenge — would you prefer automation to remove specific duplicate data types altogether or collapsing them together and showing number of instances (e.g. Diabetes mellitus Type 2 (10 instances)?
3. Do you see duplicate information as a universal problem or variable from one organization to another?
4. Are there specific data types or scenarios in which safety concern is the highest when considering automated de-duplication?
June 24, 2021 at 4:16 pm #33687Andrea Pitkus, PhD, MLS(ASCP)CM
A clear operational definition of what constitutes duplicate data is needed as there were various definitions on the calls. Where a down or upstream system eliminates clinical data, errors of omission can occur and patient harm may result. Regulatory requirements may impact actions too. There is not a one size fits all responses as it depends on several factors (the type of clinical data, etc.)
How does a provider know a result/data element is truly a duplicate versus needed by another provider for a different reason?
June 30, 2021 at 10:28 pm #34591Riki Merrick
As Andrea states we need a defintion of duplicates.
If you mean the exact same result of a lab test for example (performed by 1 lab, but reported by 2 orgranizations – maybe the EHR to the patient and the lab to the patient, or the lab and EHR to another porvider in the care team) we need to have the same identifiers to be sure it is truely the same (and currently the results themselves don’t have identifiers, so you would need to use a common specimen id / accession number of the related sample as well as the performing lab and related dates etc to be able to identify it.
One important factor in any attempt to de-duplicate is that you MUST have the assigning authority for all identifiers, else you will have no luck being able to use them.
July 8, 2021 at 5:53 pm #36071Andrea Pitkus
Adding to what Riki posted….
Time is another factor which varies by lab test/scenario. Some genetic tests may only be done one in a lifetime. So what time period is utilized for assessing.
Also a hgb may be performed pre transfusion and another post transfusion and they may appear to be duplicate, but truly impact clinical decision making as to how much the transfusion raised the hgb level.
For medications, does it depend on the same drug with different dosing, ways it’s dispenses, etc. (IV vs oral pill)? Again over which time frame? Is it tied to a chronic problem like Parkinson’s and will be needed over a lifetime or is it a short term treatment such as antibiotics for an infection?
Hopefully more clinicians will jump in with more use cases/scenarios. I would urge extreme caution in deduplicating automatically as patient harm may result if the process is not clinically validated/implementable in all systems where data is utilized (up/down stream), etc.