The creation of medical reports requires considerable time and effort from clinicians, often resulting in documents with unstructured text intended for human use but with limited usability for machines. An alternative approach is to maintain electronic health records (EHR) as structured data, which are usable by machine and automate generation of report for humans. RES-Q is healthcare registry which collects structured healthcare data relevant to the hospitalization of stroke patients with over 800 000 records globally. Recently, RES-Q introduced a basic deterministic (non-AI) discharge report generator based on existing case records, with support for seven languages. However, this system lacks advanced linguistic features such as gender and number agreement. The primary objective of this work is to address these limitations and improve the quality and applicability of the generated reports in clinical settings. Given the importance of reliability in the generated healthcare reports, testing should be part of this work to ensure the robustness of the solution.