“You were the first. “These words, spoken to me by Professor Peter Selby of the University of Leeds in Rome in 2016, almost stopped my heart, when in a flash I realised something about the importance of my research. The actual topic of our conversation was about cooperation between European cancer organisations, but then we got onto developing cancer care and his own research themes. They research topic sounded very familiar and I shyly suggested that he might like to read my 1989 article in the BMJ (British Medical Journal until 1988) “Do treatment protocols improve end results: a study on survival of patients with multiple myeloma in Finland”. Selby replied that of course he was familiar with my research and that I was one of the first to publish a scientific study on the impact of systematic treatment programmes on treatment outcomes in the general population. My research idea came about during a discussion with Ilmari Palva, Professor of Haematology. A clinical trial of myeloma treatment had been carried out in Finland, and only certain hospital districts had participated. The idea of our study was to compare the survival of all myeloma patients in different regions. In the districts where the clinical trial had been conducted, the treatment outcomes were better across the board than in those districts where it had not been conducted. Regional differences in survival are driven by many factors – including random variation. In the sub-publications of my thesis, I examined the role of random variation and some systematic factors in explaining regional variation in survival. Subsequently, regional survival rates have been boldly incorporated into the regular statistical reporting of the Cancer Registry. The data from the Cancer Registry can be used for a wide range of research on cancer risk factors and health care activities. However, it is important to be aware of the limitations of the data. For example, in Finland there is abundant information on differences in survival rates for different cancers among population groups, but the reasons for these differences are difficult to identify using registry data alone. It is possible to go much further by combining different types of registry data, but even here there are limitations. The study I described above, published in the BMJ, showed that all available information on health care activities and the research conducted within them can be used in an insightful way when conducting a registry study. And for the researcher, feedback on the relevance of their own publications may only come after more than 25 years.