The use case is COPD and the CDS manages treatment and interaction prior to administration via formalised clinical guidelines. Here it is an example that I believe it may not worked for the indicator and labelling I commented on on my previous email:
Assume some Patient A has been diagnosed with severe COPD. Also, Patient A is a current smoker and is at high risk of Pneumonia. The COPD Gold guideline suggests 3 possible treatment paths: admin LAMA type drugs, admin LAMA + LABA combination drugs, admin LABA + ICS combination drugs. Also, there is a recommendation to suggest smoking cessation therapies to the patient and a warning that regular treatment with ICS increases risk of Pneumonia.
The CDS device provides a response with 3 suggestions or care plans:
1. admin LAMA, suggest smoking cessation therapies
2. admin LAMA + LABA, suggest smoking cessation therapies
3. admin LABA + ICS, suggest smoking cessation therapies, warn practitioner about the risks of regular treatment with ICS on a patient susceptible to Pneumonia.
Now, these are converted into 3 care plans with goals where the CDS card action group is composed of medicationRequest and ServiceRequest resources to be done together (eg, update COPD treatment, suggest smoking cessation therapies).
At the client end, the practitioner selects one of the care plans. That means there should be a means to create medication orders and check smoking cessation therapies by providing message boxes on the screen actionable by a simple button.
By following your initial suggestion, one card would hold 3 suggestions. However, only one of the suggestions comes with a warning. In this case, a clear solution is to have a card per suggestion instead. However, for each suggestion group, how can the CDS service preserve the format of CDS cards when specifying the actionable requests since this is defined at the suggestions level?