It is not clear to me what you mean with substudy. I would say that if for example you only looked at the results in the intervention arm, then it is probably not necessary. If you looked at secondary outcomes or process outcomes (usage etc) in both arms, then it is. Note that JMIR prefers to publish per-protocol reports of RCTs. Anything else is probably better suited for JMIR Res Protoc or sister journals as formative evaluation etc.