Publications are often used to estimate the biomedical scientific production of an institution or country. Through its “affiliation” filter PubMed is a useful open-access tool for this purpose. In 2013 this filter was modified, beginning to include all authors. To estimate the impact of this modification we carried out the same search strategy before and after the change took place. We evaluated publications in PubMed from the 2010–2012 and 2014–2016 periods. Using the “affiliation” filter, we identified publications from three countries from different continents, with similar income level and number of scientific publications: Argentina, Thailand, and South Africa. We compared the slopes of both periods for each country and for the total references in PubMed. Regarding the number of publications, Argentina, Thailand and South Africa showed a significantly steeper slope in the 2014–2016 period versus 2010–2012 period, while the slopes of the total number of publications in PubMed showed no difference between periods. The inclusion of all authors in affiliation indexing by PubMed significantly impacts the evaluation of the scientific production of countries, reducing under-registration.