Medical researchers that work at the National Institutes of Health have a preponderant role in the scientific production of Mexico. This article focuses their behavior during the last 12 years, within the National System of Investigators (SNI). According to the citation and publication indices, the medical researchers are of greater productivity and quality. The number of investigators of the Institutes that have obtained their entrance to the SNI has grown in 129.8%, during the period of study and has been constant in each one of the Institutes of Health, with the exception of the INPer, that besides to have smaller number of researchers within the SNI, shows a tendency to low. The investigators level 1 have stayed between 10 to 12 per year in most of the Institutes, with exception of Nutrition, Cardiology and Public Health that have managed to maintain effective to more than 20 investigators of this level during the last 5 years; in the case of Nutrition over 30. These three Institutes also have maintained the greater number of investigators level 2 and 3; it emphasizes Nutrition that is counted with the average of publications by researcher in the institutes is of 4.9 +/- 3 papers. The researchers of Nutrition, Pediatría, the INPer and Neurology contribute with more than 4 publications by each year. Nevertheless, when we only included the publications of groups III-V, only the researchers of Nutrition publish annually 6 to 7 high quality research paper. All the professionals of the Institutes that have an appointment of Investigator in Medical Sciences do not belong to the SNI. In the case of the Institute of Medical Sciences and Nutrition, of 131 workers with this appointment (Julio 2003), only 47 (35.9%) belong to the NSI. When analyzing the participation of the professionals that do not belong to the SNI in the mentioned published papers from 1999 to 2002 it was found that of 484 publications of group III, in 108 (22%) nonmembers of the SNI and without appointment of investigator participated in these publications; of 85 of group IV, in 16 (19%) and of 52 of group V, in 12 (23%). Of the analysis of this information, strategies directed to particular groups can be undertaken, to mainly increase the scientific productivity in the National Institutes of Health and, to improve their quality and its scientific impact.