CodeSystems for Fitness apps?

(Kev Mayfield) #1

I’ve a few devices around the house (and bike) which track my health and fitness. The information I get from my nokia (aka Withings) devices all seem to have SNOMED concepts (blood pressure, heart rate, weight, pulse wave velocity, body temp, etc)

Others like run/cycling app strava don’t seem to have any obvious candidates. From Strava I get Training Stress score, normalised power, (cycling) exercise energy (kJ). Also on Nokia (and Apple) has exercise measures: step count, (walking) exercise calories/energy.
This is a ‘pet’ project (so I can make my own codes and codeSystem) but any pointers to correct CodeSystems or relevant subject matter would be appreciated.

My main aim is to experiment with this data while refreshing my Angular skills. I’m intending to use existing open source sql database, system and api to record the data and build a angular app which visualises the combined data.

(Marcus Baw) #2

As far as I know, all these systems use arbitrary and supplier-dependent data models. I’m not aware of any standards being used. Apple HealthKit did start to set out some [specifications] although it’s very tied to Apple,
Swift and is an internal API (, but you might find some suppliers are adhering to those standards maybe.

I know a chap at NHS Digital who has done a fair bit of practical work on personal monitors such as BP, SaO2, and weighing scales for the MediPi project, so I’ll invite him to the discussion and see if he has any advice.

Anyone else with any suggestions?


(Kev Mayfield) #3


I’d had a look at Apple (and microsoft healthvault) a years back and they seemed to use a tables for each data type with date and value.
In the (sql) db model I’m using, I’d use one table with the date and value, plus a code to distinguish the entry plus a value units.
So similar to what’s seen here in JSON (but without the category):

(ian) #4

Hi Kevin,

You probably want to think in terms of LOINC for future ‘names’ of results as this is the clear line for now in the US but this might be useful to you also.


(Kev Mayfield) #5


I didn’t think to look at LOINC. Found a number of useful codes without looking in detail.

Some useful background info on openmheath :slight_smile:

(Kev Mayfield) #6

p.s. Patient 2 on the earlier link is from Nokia/Withings.

(Kev Mayfield) #7

Noticed Grahame Grieve provided some the answers here:

Which answers coding Observations. I’m considering also recording CarePlan.activity (for an implied Fitness/nutrition CarePlan) which can be linked to the Observations via an Encounter.
So the Encounter roughly equates to the fitness event, which would be activity in Strava or measure group (for 24 hour readings) in Nokia.
I’ve coded api connections to strava and nokia, so I’ll look at creating an example from real data (i.e. me).