Best practices for Medical development


Follow the general "Best Practises" document

The document Best Practises describes the known good methods of utilizing Movesense sensors. This document builds on those, and emphasises the special things to pay attention when developing medical solutions utilizing Movesense Medical sensors.

Keep an eye on the Medical software updates

We update our medical software as specified by our software maintenance and feedback processes. If a bug in the medical release of the software comes to our attention, we will notify our medical customers about it as soon as possible via email and/or other means. For any bugs, the "Known bugs" part of the CHANGES.md in software release repository will be updated accordingly. Make sure that you can receive our emails and that they are not lost in the "Spam" -folder.

Always validate the incoming data

In the medical context, we recommend that you always validate the data that your application receives from the Movesense MD sensor – or any other data source that your application might be using. There is always a possibility that some data might be lost or is invalid, for example when operating in the extremes of the Bluetooth LE operating range. Validation is an important part the verification confirming that the data received from its source is unaltered.

From the two possible undesired outcomes, i.e. data lost or invalid data, the latter one is more serious one, but both could cause misdiagnoses, and therefore harm to the patient, if not accounted for. It is therefore very important to implement data validation in your application, so that your analysis algorithm is always protected from an erroneous input.

A few examples when data loss or invalid data might occur:

1. Marginal BLE connection, for example when operating in the extreme range of the BLE connection.

Follow data time stamps to confirm data integrity. If needed, implement a checksum calculation to complement BLE protocol level checks and prepare to duplicate data packets. Implement a “heart beat” signal, if your end application uses long time interval between actual payload data packets.

2. Bad bluetooth connectivity by mobile device, can cause inability to connect, random disconnects and lost data.

Follow data time stamps to confirm data integrity. If needed, insert a checksum calculation to complement BLE protocol level checks.

3. Possible anomalies in your own code

Thorough on-devoce testing with clean and purposefully bad data. A good habit is to add a separate testing interface where you can insert freeform data into your data processing pipeline.

4. Possible anomalies in the Movesense code: Movesense sensor framework or MDS mobile library

Follow data packet time stamps to confirm data integrity. Make sure the data is within expected range. Make sure your contact details are up to date, so that you are able to receive news, updates and other official information from Movesense. Movesense will inform about possible SW anomalies with emails and via the official documentation page available at https://www.movesense.com/docs