| OHDSI Home | Forums | Wiki | Github |
Good afternoon!
Curious if anyone has modeled pain/pain score/pain scale/etc. attributes and values? We have some pretty messy, but somewhat standardized EHR data that we are going to put into the Observation table. If you’ve tackled this domain, please share your concept set and conventions.
Cheers,
Melanie
Hi,
I’m integrating emergency medical service patient care records into OMOP - they utilize a Pain Scale Score as part of typical trauma and other clinical condition protocols. The National EMS Information System (NEMSIS) reporting standard defines this pain score - usually 1-10 numerically. It can vary among EMS EMR systems and documentation practices.
My implementation saves this as a measurement under the more generic concept “Pain intensity rating scale” (4137083) unless I know the EMS system utilize the “Pain severity Wong-Baker FACES pain rating scale” (3035486) or other standard pain scale concepts. If the data is originating from a NEMSIS source, I then use a custom vocabulary of NEMSIS terms and enumerated concepts derived from the NEMSIS data dictionary if such concepts are not already in the OMOP vocabulary.
Hi, @mkwong
I’m also try to integrate Korean national emergency department information system (NEDIS) into OMOP-CDM.
My pilot project may be completed by March.
That is wonderful to hear. I am not familiar with EMS practices outside the US. In the US our EMS paramedics carry 12-lead defibrillators for ACS patients and other cardiovascular conditions. In addition to pain score - I have OMOP mapping for 3 common ECG algorithm criteria and measurements. Integration of device data into our research repository beyond just summary findings and impressions. Are you doing the same for your EMS system?
I would be interested in OMOP ecg mapping tools
What manufacturers are common in your EMS setting? I have draft ECG mapping for Philips 0B/0C, GE 12SL, and GLASGOW for computerized interpretations. For ECG measurements I have an expanded measurement set to cover 16 lead systems plus XYZ leads. A good portion of the ECG mapping (criteria and lead measurements) was mapped to standard OMOP concepts (SNOMED and LOINC). The others appear to be vendor specific and will end up in a “Custom” vocabulary is the way I’m handling it right now.
I’m having some of the vendors review the mapping for comment and suggestions.
Specific tools are specific to our projects and needs - Java based.
The ecg readers we are using are typically found attached to smart watches
or smart phones. Currently they can be kardia/alivecor or apple watch.
The ecg databases are downloaded from physionet and uci.
So you are capturing primarily single lead ECG data then. I’m not sure if my mapping would be useful in this use case as LOINC codes I map to refer to standard 12-lead measurements. What I refer to as “Global” measurements may be applicable - that is heart rate, QRS duration, PR-interval, QT and QTc measurements may apply in your situation.
If you have ECG algorithm needs - detectors, rhythm, etc - you are welcome to contact me directly.
Friends:
When and why did this debate turn from @MPhilofsky’s pain scores to EKGs? 
But yes, if you can make a proposal for EKG measurements we’d gladly bring it into the CDM Working Group or THEMIS.
Absolutely i would be interested
I am developing some algorithms for arrhythmia classifiers. So far the
kardia/alivecor only does Afib but i think there is need to do more
Hi Christian. The conversation went side-ways as I handle both pain scoring and ECG in the same emergency medicine context via my response. I am getting an EKG vocabulary proposal worked up to be submitted to the CDM Working Group is the plan. Stay tuned. Robert here is our resident OMOP expert helping me figure out the proper mapping mechanism before we submit our first draft.
We have some pretty messy, but somewhat standardized EHR data that we are going to put into the Observation table
Sorry for the silly question: Why this data should go in the observation table versus going to the measurement table.
Reading the documentation:
Measurements differ from Observations in that they require a
standardized test or some other activity to generate a quantitative or
qualitative result.
Then do pain score is not a standardized test ?
In a general way, I am not very confortable in the way to both choose how to split data from observation and measurement, and explain to statistician how they can easily state where is the data.
Thanks for any answer
It’s not a silly question. Most of the pain attributes seem to belong in the observation. Per the wiki: “The OBSERVATION table captures clinical facts about a Person obtained in the context of examination, questioning or a procedure”. However, I asked because I want community input on best practices.
The attributes we have are types of pain scales (NRS, CPOT, self reporting/ non-self reporting, yes/no), pain scores (generally 0-10, but also others), pain quality, location, duration, alleviating factors, aggravating factors, interventions, response to interventions, pain goal…
I see pain score as a mostly subjective report from the patient. Some pain scores are objective (ie. NIPS, FACES, CPOT), but most are reported from the patient. What do you think? Do you have pain data in your OMOP instance?