Study Design and Sample Size
Ethical approval
was obtained from The Scientific Research and Ethics Committee of the Medical Research Council (BMEÜ/2801-3/2022/EKU) in Hungary. Data collection took place in 2022 and 2023 among patients suffering from hypertension in Hungary who were taking some type of antihypertensive medication. The data collection was carried out by SZLEM Service L.P. in Hungary. During the study procedure, the participants signed an informed consent form containing all the details of the study. The researcher explained to the patients the aims of the study and assured them that participation was voluntary. No incentives were provided. The participants were asked using an online form, which took about 15 minutes to fill out. The answers are completely anonymous to all authors of this study. The original goal was to ask at least 1,000 adult hypertensive patients. However, at the end, 1,067 adults completed the questionnaire. The survey took into account representativeness in terms of age, sex, regions, and the size of the municipalities at a national level. The criteria for participating in this research was being 18 years of age and taking some form of antihypertensive medication.
Measures
The Beliefs About Medicines Questionnaire (BMQ) was used in this study, [13, 21] to assess both the positive and negative beliefs of the participants about medicines. [22]
The Hungarian translation of the BMQ-Specific was previously used to assess attitude towards cholesterol-lowering medication. [14] The original English version [13, 23] was translated into Hungarian by two independent translators. Once completed, the two translations were merged into a single translation and modified so that the questions focused on antihypertensive medications. After that, as part of the validation process, each translation was tested by ten Hungarian-speaking citizens taking antihypertensive medication, and the translations were revised based on their feedback. A third independent translator translated the questionnaire back into English. The back translation was deemed adequate by all members of the research team.
The BMQ consists of 18 questions that assess general beliefs about pharmacotherapy (BMQ-General), and perceptions about pharmacotherapy in more specific situations such as having a chronic illness (BMQ-Specific). The BMQ-General has 8 items [21] and is divided into two subscales. The General-Harm subscale assesses beliefs about the harmful effects of medications, and the General-Overuse subscale addresses the notion that physicians place too much trust in medications. [21] The BMQ-Specific consists of 10 items and is also divided into two subscales. The Specific-Concerns subscale assesses the likelihood of side effects as a result of taking the prescribed medication. The Specific-Needs subscale examines the patient's belief that he or she must personally adhere to the prescribed medication. [21] Each question of the BMQ has five possible responses ranging from 'strongly disagree' to 'strongly agree' (from 1 = strongly disagree to 5 = strongly agree). Depending on the distribution of the data, the scores of all subscales are determined by the mean or median of the items.
Higher scores on the General-Harm subscale and the General-Overuse subscale indicate a generally negative attitude toward drug therapy. Similarly, higher scores on the Specific-Concerns subscale indicate the belief that side effects of regular medication use can be harmful, and higher scores on the Specific-Needs subscale indicate that patients need to adhere to their medication regimen to maintain their health. [21]
The following socio-demographic data were also collected alongside the BMQ: age, sex, highest education level, county, size of municipality, marital status, self-perceived financial situation, self-perceived health status and being a healthcare worker.
Data Analysis
Categorical variables were expressed as frequency (%), while continuous variables were expressed as mean and standard deviation (SD). Cronbach's alpha was calculated to evaluate the reliability of Necessity, Concern, Harm and Overuse subscales. Internal reliability was considered acceptable if the alpha value was 0.70 or greater. [24]
As a preliminary analysis, patients were divided into groups based on their beliefs about medication; [8, 25] these groups were created by splitting the BMQ Necessity and Concerns scores at the median. Four categories were created using this method: "sceptical", "ambivalent", "indifferent" and "accepting". Respondents in the "indifferent" category are neither convinced of the need for nor concerned about antihypertensive drugs, while "ambivalent" about antihypertensive drugs means that the respondent agrees with the need for antihypertensive drugs but is also concerned about their possible side effects. “Accepting” antihypertensive drugs means that respondents accept the necessity of antihypertensive drugs but have low concerns about their possible adverse effects. Respondents who agreed with low necessity and high concern about antihypertensive drugs were categorised as "sceptical". They have doubts about their personal need and high concerns about taking antihypertensive drugs.
Robust regression analysis was performed to evaluate the association between sociodemographic variables and the four subscales. The results were presented as coefficients and the corresponding 95% confidence intervals (CI). Statistical significance was set at a p-value of less than 0.05. A confirmatory factor analysis was performed using the chi-square (X2), ratio of chi-square to degrees of freedom (X2/df), comparative fit index (CFI), Tucker-Lewis index (TLI), root mean square of standardized root mean square (SRMR) and root mean square error of approximation (RMSEA) methods. According to the literature, the recommended values for the X2/df ratio are less than 3, [26] the values for CFI and TLI are greater than 0.90, [27] SRMR is less than 0.08, and for RMSEA, lower values such as 0.08 [28] are considered acceptable and values less than 0.10 are only slightly acceptable. [29] When evaluating factor loadings, 0.4 was set as the minimum acceptable value. [30] The STATA v13 (Stata Corp LLC, College Station, TX, USA) software was used to analyse the data.