The principle of any retrospective on patient data-based investigation is searching the patients by problem or sign, but not by name. With a proper problem-encoded archival database, the data mining process would be easy. One would only need to input the request and obtain the proper data in a short time. Medical archives are frequently based on paper records only, with the patient name as the entry key. To find the proper record in such an archive, a detection strategy is needed. The process continues with collecting the usually enormous amount of papers, finding the appropriate records within them, and finally encoding and arranging them in a table. The whole process can be separated into patients, paper and data mining. Because of their slowness, these phases can be the most time-consuming part of a medical data-based investigation. The author describes his data mining experience.