This article describes an approach for maintaining a high-quality question test database, implemented in a real-world case study. The approach leverages answering history from past tests to identify and address problematic records, thus enhancing the overall quality of the database. Key criteria for selection include the percentage of incorrect answers, the proportion of choices for single distractors, and the time spent on each question. These criteria effectively highlight questions needing revision. The method demonstrated simplicity and efficacy in managing a large dataset from past examinations, ensuring continuous database quality. Almost all records flagged for revision were modified by reviewers, leading to significant improvements in the percentage of correct answers in subsequent tests. This indicates that the primary issue was often the question construction rather than test-taker knowledge. An automated module was developed using PHP and MySQL, integrated with frontend frameworks for enhanced usability. This module assists in identifying problematic questions, with final decisions made by reviewers. The study concludes that this tool is valuable for systematic improvements, maintaining the integrity and reliability of the assessment process.