The four primary objectives of this descriptive study were to: 1) design a quality-measurement instrument for institutional cancer registries (ICRs), 2) evaluate the existing ICRs in Colombia with the designed instrument, 3) categorize the different registries according to their quality and prioritize efforts that will efficiently promote better registries with the limited resources available, and 4) determine the institution with the greatest likelihood of successfully establishing Colombia's second population-based cancer registry. In 1990 the National Cancer Institute of Colombia developed 13 institution-based cancer registries in different Colombian cities in order to promote the collection of data from a large group of cancer diagnostic and treatment centers. During the first half of 1997, this evaluation reviewed 12 registries; one of the original 13 no longer existed. All of the Colombian institutions (hospitals) that maintain institution-based cancer registries were included in the study. At each institution, a brief survey was administered to the hospital director, the registry coordinator, and the registrar (data manager). Researchers investigated the institutions by looking at six domains that are in standard use internationally. Within each domain, questions were developed and selected through the Delphi method. Each domain and each question were assigned weights through a consensus process. In most cases, two interviewers went to each site to collect the information. The university hospitals in Cali, Pereira, and Medellín had substantially higher scores, reflecting a good level of performance. Four of the 12 institutions had almost no cancer registry work going on. Five of the 12 hospital directors considered that the information provided by the cancer registries influenced their administrative decisions. Three of the registries had patient survival data. Four of the institutions allocated specific resources to operate their cancer registries; in the other 8 hospitals there was no clear budget allocation. Seven of the hospital directors could not identify five or more objectives of a cancer registry. Data management was usually poor and resources insufficient at most of the institutions. In summary, the cancer registry system in Colombia varies greatly from institution to institution. A few of the hospitals do a good job while others have neglected the registries. The high, identical total scores for Pereira and Medellín suggest they would be good locations to establish new population-based cancer registries similar to the existing one in Cali. However, the overall characteristics in Pereira may provide a more appropriate environment for the second registry, with Medellín as an alternative.