Regulatory Open Forum

Hello, 

I work for an SaMD organization in the obstetric space.  Our software is prescription use for hospitals and use by trained clinical personnel only.  Our current clearance includes some calculations and aggregation of data to provide alerts and color coding when certain thresholds are met/exceeded. We are exploring some requests from the end users to add functionality which calculates certain parameters (already obtained with our device) but provide an output of "sepsis" or "hypertension".  These calculations are published by NICHD (National Institute of Child Health and Human Development) and are done manually in the field currently, we are not looking to modify or change the calculation in any way, only provide an automated method to ease the workflow of our end users.  We would validate the calculations; however, it is unclear whether or not this would change our intended use to be "diagnostic" as we are not intending to diagnose, but the calculation would provide the clinician with a defined score for a diagnosis.

Thank you for your consideration.

Hi Anon.

First let me preface the below with the fact that I am not in this space and do not have direct experience in this space in my prior life.  However, I think my logic here will still be reasonable and potentially on track (although possibly others will be able to either say it more clearly or might know of some of the pitfalls that I myself would not at this point).

That said, the logic here in my mind is whether or not those "scores" or "calculations" that are routinely done in the field are the basis for the diagnosis in the field.  If that information is what is actually used clinically to make the diagnosis of "sepsis" or "hypertension" then yes, I think you end up being a diagnostic.  Even though you are only providing the calculation result I believe that in essence, if the calculation result is the key diagnostic value (or even a significant value) used for the purposes of the diagnosis, then your system providing that information rises to the level of being a diagnostic.

If on the other hand the value is more "consultative" meaning that the clinician will use other key evidence to make a final diagnosis in the field then you might still fall under the "measuring" definition and might not be considered diagnostic.  

Unfortunately as in most areas of Regulatory the ultimate answer is "it depends" because you as the manufacturer first of all know your device and its workings the most robustly.  And second you should know the medical space where your device is being used and therefore would know best whether or not your product actually meets the standard requirements for the product to be considered a diagnostic device in that space.

Hello, 

I work for an SaMD organization in the obstetric space.  Our software is prescription use for hospitals and use by trained clinical personnel only.  Our current clearance includes some calculations and aggregation of data to provide alerts and color coding when certain thresholds are met/exceeded. We are exploring some requests from the end users to add functionality which calculates certain parameters (already obtained with our device) but provide an output of "sepsis" or "hypertension".  These calculations are published by NICHD (National Institute of Child Health and Human Development) and are done manually in the field currently, we are not looking to modify or change the calculation in any way, only provide an automated method to ease the workflow of our end users.  We would validate the calculations; however, it is unclear whether or not this would change our intended use to be "diagnostic" as we are not intending to diagnose, but the calculation would provide the clinician with a defined score for a diagnosis.

Thank you for your consideration.

Hello Anon,

Without seeing the indications for use statement, the Instructions for Use (IFU), the current calculation parameters, and how the outputs are being used would only be able to provide some potential advice.  The first thought would be yes, it would be considered diagnostic if those features/functionality are added. If the information is being provided to the healthcare professional to make a clinical decision such as if the patient has sepsis, then it most likely would be considered diagnostic.  A good reference to review is the FDA's guidance on clinical decision making which talks through aspects of your scenario and other examples https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-decision-support-software. 

Hi Anon.

First let me preface the below with the fact that I am not in this space and do not have direct experience in this space in my prior life.  However, I think my logic here will still be reasonable and potentially on track (although possibly others will be able to either say it more clearly or might know of some of the pitfalls that I myself would not at this point).

That said, the logic here in my mind is whether or not those "scores" or "calculations" that are routinely done in the field are the basis for the diagnosis in the field.  If that information is what is actually used clinically to make the diagnosis of "sepsis" or "hypertension" then yes, I think you end up being a diagnostic.  Even though you are only providing the calculation result I believe that in essence, if the calculation result is the key diagnostic value (or even a significant value) used for the purposes of the diagnosis, then your system providing that information rises to the level of being a diagnostic.

If on the other hand the value is more "consultative" meaning that the clinician will use other key evidence to make a final diagnosis in the field then you might still fall under the "measuring" definition and might not be considered diagnostic.  

Unfortunately as in most areas of Regulatory the ultimate answer is "it depends" because you as the manufacturer first of all know your device and its workings the most robustly.  And second you should know the medical space where your device is being used and therefore would know best whether or not your product actually meets the standard requirements for the product to be considered a diagnostic device in that space.

Hello, 

I work for an SaMD organization in the obstetric space.  Our software is prescription use for hospitals and use by trained clinical personnel only.  Our current clearance includes some calculations and aggregation of data to provide alerts and color coding when certain thresholds are met/exceeded. We are exploring some requests from the end users to add functionality which calculates certain parameters (already obtained with our device) but provide an output of "sepsis" or "hypertension".  These calculations are published by NICHD (National Institute of Child Health and Human Development) and are done manually in the field currently, we are not looking to modify or change the calculation in any way, only provide an automated method to ease the workflow of our end users.  We would validate the calculations; however, it is unclear whether or not this would change our intended use to be "diagnostic" as we are not intending to diagnose, but the calculation would provide the clinician with a defined score for a diagnosis.

Thank you for your consideration.