I have a question regarding patient reported history and how the time(s)
of that would be represented in the reference model.
1971-1992 patient married to X
2008-06-17 patient asked about social history, reports that he/she was
married to X 1971-1992.
What would be the HISTORY.origin and EVENT.time (or
INTERVAL_EVENT.width?) of this? Or would the timings of this be
archetyped? Is this an observation of the patients memory taking place
now or an observation of the past event of marriage?
I have a question regarding patient reported history and how the time(s)
of that would be represented in the reference model.
1971-1992 patient married to X
2008-06-17 patient asked about social history, reports that he/she was
married to X 1971-1992.
What would be the HISTORY.origin and EVENT.time (or
INTERVAL_EVENT.width?) of this? Or would the timings of this be
archetyped? Is this an observation of the patients memory taking place
now or an observation of the past event of marriage?
Daniel,
It depends on whether we want to record marriages as a computable fact.
There are two possibilities (at least):
1. as you say, use Observation, History, Interval_event etc to record
the marriage start, period and end. This would need an archetype.
2. record what the patient says using the 'patient story' archetype.
This just records what was said in a narrative form.
I have a question regarding patient reported history and how the time(s)
of that would be represented in the reference model.
1971-1992 patient married to X
2008-06-17 patient asked about social history, reports that he/she was
married to X 1971-1992.
What would be the HISTORY.origin and EVENT.time (or
INTERVAL_EVENT.width?) of this? Or would the timings of this be
archetyped? Is this an observation of the patients memory taking place
now or an observation of the past event of marriage?
Daniel,
It depends on whether we want to record marriages as a computable fact.
There are two possibilities (at least):
as you say, use Observation, History, Interval_event etc to record
the marriage start, period and end. This would need an archetype.
record what the patient says using the ‘patient story’ archetype.
This just records what was said in a narrative form.
Up to clinical people to decide what they want!
thomas beale
Dear Daniel and Thomas,
If you would like to study the correlation between length of marriage and another parameter, you would need alternative 1 for a computable fact, wouldn’t you? Then the narrative form isn’t enough.
It depends on whether we want to record marriages as a computable
fact.
There are two possibilities (at least):
1. as you say, use Observation, History, Interval_event etc to record
the marriage start, period and end. This would need an archetype.
2. record what the patient says using the 'patient story' archetype.
This just records what was said in a narrative form.
Up to clinical people to decide what they want!
- thomas beale
Dear Daniel and Thomas,
If you would like to study the correlation between length of marriage
and another parameter, you would need alternative 1 for a computable
fact, wouldn't you? Then the narrative form isn't enough.
- Rikard Lovstrom
excuse my jetlag, I wasn't clear ... I should have said that only option
1. is computable.
Tom,
let me rephrase my question a bit:
is the EVENT classes time attributes used to represent the time of the
observation or the time (related to?) of the thing observed. In examples
in the EHR IM I think it is the former and not the latter case, but its
not all clear to me.
A third possibility would be to archetype the times.
Not to contradict the technical discussion, I would argue that the recording of relationships are probably best done in the form of an evaluation archetype - as this is likely to be summary data and the dates are not likely to be due to repeated observations, rather recorded after the fact. I can see no advantages to using the observation class here which is designed for repeated recording of the same phenomena. It would be useful if we asked someone if they are married each time they came to see us - Are you still married?
The dates when information is recorded in a summary context can be more complex and just as computable. Thus if we were to have a relationship archetype it might have the date of meeting, of onset of sexual relationships, of cohabitation, of marriage/legal relationship, of separation, of divorce, of last contact. It could include the name and contact details and whether the person is the next of kin and exist in the Social History persistent composition.
The information classes in openEHR are just that - they are good for purpose. The observation class is really based on scientific observation (measurements, state and method) and is not suited to summary information of a complex concept. It is best for observations or measurements of a feature that vary with time and may be repeated. The names give an indication of when they are most useful, but it is the features of the class that determine which to use. It is important not to confuse this with Larry Weed’s SOAP, although they are often aligned.
