Stream: implementers
Topic: rehab exercise results
Maurizio Macagno (Feb 20 2020 at 17:30):
Hello. New here. Hope the question is ok for the group.
We are implementing a remote rehab monitoring solution. The patient needs to performs rehab exercises for their knee, and we want to capture individual exercise rep results (e.g. flexion and extension angle). A typical "observation" should include:
exercise type
number of sets/repetitions in set
flexion/extension angles for each single repetition.
We could model a single exercise as an observation.
The flexion/extension angles series could be valueSampleData components, although semantically it's incorrect, but I cannot find other suitable ways.
Also not sure how to model the individual sets, or to track them as metrics.
Any feedback from the experts?
Lloyd McKenzie (Feb 20 2020 at 18:49):
Given that it's not a time series, agree that SampledData is probably not appropriate. There's some discussion about a new type for non-time-series data, but nothing specific to look at yet. The 'best' way right now is probably a separate Observation for each rep, and a 'hasMember' relationship to an Observation that represents each set (and perhaps hasMember relationships from those to an Observation that represents the overall session.
Maurizio Macagno (Feb 20 2020 at 21:24):
Lloyd McKenzie said:
Given that it's not a time series, agree that SampledData is probably not appropriate. There's some discussion about a new type for non-time-series data, but nothing specific to look at yet. The 'best' way right now is probably a separate Observation for each rep, and a 'hasMember' relationship to an Observation that represents each set (and perhaps hasMember relationships from those to an Observation that represents the overall session.
Thanks for your reply.
Semantically it may make sense, but we would have a proliferation of NxMxT (T exercises, N sets, M reps) observations for each day of therapy... seems overkill.
Lloyd McKenzie (Feb 20 2020 at 21:29):
Proliferations of Observations is pretty normal. Every separate assertion that would reasonably be queried or analyzed independently should be a separate Observation. A few days in the ICU could easily result in 10k+ Observations from the various continuous vitals monitors. If you wanted to look at the progress of extension angle over several weeks, you should be able to retrieve that without necessarily grabbing anything else. Typically we only use Observation.component for stuff that is almost always looked at and evaluated together (systolic/diastolic; parts of an APGAR) or pieces that don't have any meaning when looked at independently (e.g. susceptibility and antibiotic).
Last updated: Apr 12 2022 at 19:14 UTC
