Co-administration patterns of interacting drugs vary depending on age, gender and healthcare system
Jon Sanchez-Valle, Rion Brattig Correia, Marta Camacho-Artacho, Alba Lepore, Mauro M. Mattos, Luis M. Rocha, Alfonso Valencia
Blumenau:
Blumenau, in Southern Brazil, is a city with a very high Human Development Index and a population of approximately 340,000 inhabitants.
Brazil has a universal public health-care system, and Blumenau possesses a city-wide Health Information System (HIS) with prescription and
dispensation information for its entire population. Drug dispensing data for 18 months (January 2014 - June 2015) were gathered from the Pronto HIS.
Electronic Health Records were anonymized at the source and only drug dispensation and demographic variables such as age, gender, heighborhood and
educational level were kept. In the study period, drugs were administered to 132,722 people.
Catalonia:
Catalonia is an autonomous community in the northeast of Spain with a population of 7,6 million people. The Spanish National Health Service is a
public healthcare system with copayment prescription charges that provides universal coverage, financed mainly by tax revenue. The Catalan Health
Institute manages primary health care teams that serve 74% of the population. The coded clinical information recorded in Electronic Health Records by its
274 primary health care teams is contained at the Catalan Health Institute's Information System for the Development of Research in Primary Care (SIDIAP) since 2006.
Diagnoses, drug prescriptions, and demographic variables such as age and gender from January 2008 to December 2018 were gathered.
In the study period, 5,5 million people were diagnosed with at least one disease or bought a drug.
Indianapolis:
Indianapolis is a city in the United States with a private healthcare system and a population of 876,682 people. The Indiana University
Health System provides information on drug dispensation and disease diagnosis using ICD10 codes from primary, secondary and tertiary. Electronic Health Records
from January 2017 to December 2018 were analyzed.
Strength of the association:
The strength of the association (edges' width) is calculated as the mean number of days/months two drugs known to interact
are given together vs. separated in the population.
Relative Risks (RR):
The Relative Risks for females are calculated as the percentage of females taking the drug-drug interaction divided
by the percentage of males taking the drug-drug interactions (we assign them positive values), while the RR for males
is calculated the other way around (and a negative value is assign). In the network, those interactions with a higher
risk associated to females are colored in red, while those at a higher risk for males are colored in blue.
Nodes' size and color:
The size of the node ranges from 0 to 1, and is calculated as the probability of the drug to be involved in an interaction. The
color is based on the drugs.com category they belong to.
Severity and description of the interaction:
The severity of the interaction is extracted from drugs.com, while the description of the interaction is extracted from DrugBank.
Significance of the interactions:
Fisher's exact tests are conducted to analyze the significance of the co-dispensation of each pair of drugs based on the number
of patients taking each of the drugs, the number of patients taking both drugs together and the total number of patients
in the given population at the selected age range.