Retrieving information from EHRs that are represented as XML documents is an important aspect for the users of this domain. Such retrieving may lead to some vague queries. There is an extensive need for designing an IR approach to process these kinds of queries and retrieve relevant XML documents. This paper proposes an IR-approach that decomposes the vague query into two sub-queries, Content-Only (CO), and Content and Structure (CAS). The proposed approach processes each subquery using one of the existing techniques and then combines the results, in order to answer the user's vague query. We expect that this approach will improve the retrieval precision of the retrieved documents.