I assume you are asking whether you can support a statement like X% of patients with condition Y have characteristic Z. Patient support program enrollment data would not meet the standard level requirements for epidemiological statements about the disease (e.g. demographics). A published and peer-reviewed epidemiological study which is relevant to the Canadian market could be considered as could an authoritative source such as Canadian guidelines.
Another important consideration, in addition to the evidentiary issue, is whether discussion of that particular sub-population would be acceptable in the branded piece (e.g. is the subpopulation discussed in the TMA?)