**7. Results**

#### *7.1. Sample Demographic Characteristics*

The total sample was made up of 864 prescriptions issued to 820 individuals. Five per cent of those prescriptions involved tetracyclines, 61% involved thiazides, and 46% angiotensin II receptor blockers. Some of these prescriptions consisted of a combination of a thiazide and an angiotensin II inhibitor.

Sociodemographic characteristics are shown in Table 2. The proportion of women was 57%. The majority of the sample was non-Hispanic white (70%) and had secondary-level education (57%). Regarding civil status, 66% of the sample was defined as married or with a partner. Some 88% had been born in the U.S. and a higher percentage (97%) had American citizenship. The main BMI category was type II or III obesity (32%). Only 30% of the sample did moderate-intensity sport.


**Table 2.** Sociodemographic characteristics of the sample (n = 820).

\* The following variables contain one missing value: civil status, citizenship and country of birth.

#### *7.2. Prevalence of Potential Interactions*

Table 1 shows the prevalence of potential interactions among the sample. Forty-four percent of the people using tetracyclines were consuming calcium and 26% and 37% of them were consuming magnesium and zinc, respectively. Among the users of thiazides, 54% and 52% used calcium and vitamin D, respectively. Finally, 26% of consumers of antagonist II receptor blockers presented a potential interaction due to the concomitant use of potassium. Overall, 49% of the participants were at risk of at least one of the studied interactions.

#### *7.3. Factors Associated with Potential Interactions*

Table 3 shows the factors associated with su ffering a potential interaction between medications and dietary supplements. According to the adjusted analysis, age and educational level were strongly associated with the probability of a potential interaction. Also, compared to the non-Hispanic white population, the non-Hispanic black population presented a lower probability of a potential interaction. The remaining variables (sex, civil status, citizenship, country of birth, BMI category, and physical activity) showed no statistically significant association with the risk of a potential interaction.

**Table 3.** Factors associated with potential interactions based on the multivariate weighted logistic regression model \$.


\$ Potential interactions include the interaction of tetracycines with calcium, magnesium or zinc, thiazides with vitamin D and Angiotensin II receptor blockers with potassium; \* Few values to be considered. CI = Confidence interval; OR = Odds ratio.

Older people had a higher risk of using a prescription medication and a dietary product with a potential interaction e ffect (OR = 1.02 (95% CI 1.01, 1.03)) per year, i.e., OR is 1.22 in patients 10 years older). People with a higher educational level (secondary or university) showed a higher risk of using a dietary product with a potential interaction with one prescription medicine (OR = 2.0 (95%CI 1.18; 3.23) and OR = 1.6 (95% CI 1.01; 2.63), respectively).
