We have read with interest the article by Bernal et al.1 published in Revista Española de Cardiología and would like to make several comments.

First, we would like to congratulate the authors on their study and on the research topic chosen. In the era of big data, new opportunities to use large databases have greatly enhanced prospects for research into health care outcomes. The study by Bernal et al. is a clear example of the usefulness of the minimum basic data set (MBDS) for research into care outcomes for acute coronary syndrome and the possibility of linking MBDS information with data from other disease-specific clinical registries for this condition, such as the DIOCLES (Descripción de la Cardiopatía Isquémica en el Territorio Español [Description of Ischemic Heart Disease in the Spanish Territory]) registry.

The authors point out that the main limitation of their study was the percentage of matches between the 2 registries that could not be resolved. According to the authors, the linkage procedure applied in DIOCLES was adversely affected by using the variable of age, rather than date of birth, as well as quality issues related to admission and discharge dates. Quality issues can also arise with coding of the principal diagnosis at discharge in the MBDS. For instance, a study was conducted at the 9 general hospitals in the Health Service of Murcia, using an MBDS for the first half of 2012 and the second half of 2013 based on a principal diagnosis at discharge of ST-segment elevation acute coronary syndrome (STE-ACS) (International Classification of Diseases, ninth revision [ICD-9] code 410.X1, except 410.71). In that study, 29.1% of 898 cases initially coded as STE-ACS were actually found to be inaccurate during a review of the events by expert cardiologists. Ultimately, non–ST-elevation acute coronary syndrome [NSTE-ACS] was diagnosed in 87.7% of the cases excluded and other conditions in the rest.2

Because STE-ACS and NSTE-ACS differ in terms of therapeutic approach, mortality, complications, and rehospitalization rates, this quality issue could cause problems when the MBDS database is used to analyze care outcomes in STE-ACS.

Furthermore, this issue may have been recently heightened by the switch from ICD-9 to ICD-10 for clinical data coding. Another study performed in a regional health service concluded that the information it collected in 2017 in an MBDS using ICD-10 codes may be useful for understanding certain general aspects related to health care and service quality, in comparison with previous years. However, it would not be useful for analyzing trends regarding frequency of patient consultations, monitoring of the management of specific medical conditions, or identifying cases for research projects.3

Quality issues have always been a problem with MBDS4 and should be considered when this kind of information system is used in clinical research.