Hello FHIR experts!
I am building a system that transmits medical device data using FHIR. For waveform data, I am using SampledData as shown in this EKG example.
However, many medical device signals (including EKG) will often contain NaN for a value. Its a common issue when the sensor does not produce a valid value. So sometimes the NaNs will be scattered throughout the array. For example, you might see an array like this:
2041 2043 NaN 2047 2060 2062 2051 2023 NaN 2027 2034 2033 2040 2047 2047 2053 2058 NaN NaN 2063 2061 2052
This page explains how to deal with NaNs from medical devices. But it doesn’t explain what do when those NaNs are scattered through a waveform.
My guess is that each waveform must be split into pieces so each individual NaN can be extracted and labelled with “data-absent-reason”. If that is true, it will be really tedious!
Any thoughts on how to handle NaNs in a SampledData array?