Predicting the Intention to Adhere to Hypertensive Medication among Seniors: A Cross-Sectional Survey

Nonadherence to hypertension medication is a widespread problem, especially among senior citizens. Several approaches are needed to improve adherence and include identifying and prioritizing problems related to adherence. A formative study employing theory of planned behavior (TPB) has revealed some salient beliefs underlying the intention to adhere to hypertension medication among seniors. The correlation of these beliefs with adherence needs to be further investigated to identify which belief to be prioritized in a medication adherence strategy. This study aims to identify which belief had the strongest correlation with the intention to adhere to hypertension medication among seniors. A cross-sectional survey was conducted among seniors with hypertension in Yogyakarta. A questionnaire based on TPB was developed and used in this study. The questionnaire was distributed conveniently to three senior communities (offline) and through a Google Form link to five WhatsApp groups (online) in September – December 2023. Data were analyzed by using multiple linear regression analyses. A total of 49 and 48 seniors completed the offline and online questionnaires, respectively. The mean age of the participants was 66.9 ± 4.4 years. Subjective norms and perceived behavioral control, but not attitude, accounted significantly (p < 0.01) for 57.6% of the variance in adherence intention. Among beliefs, perceived behavioral control had the strongest correlation with intention, and frequent awareness of blood pressure level had the highest correlation with perceived behavioral control. Understanding which belief has the highest correlation with the intent to adhere to hypertension medication provides insight for developing strategies for improving adherence among seniors.


INTRODUCTION
Hypertension is the highest risk factor for death and disability in Indonesia (International Health Metrics Monitoring and Evaluation, 2019;Murray et al., 2020).The key to managing hypertension and preventing morbidity and mortality due to hypertension is to maintain blood pressure on target.Such a target is achieved through medication adherence (Cross AJ & George, 2020;MacLaughlin & Saseen, 2020).However, given that hypertension is often asymptomatic, nonadherence to hypertension treatment is common.Several studies have shown that nearly 50% of patients with hypertension have low adherence (Abegaz et al., 2017;Mahmood et al., 2021).Nonadherence is higher in senior groups than in other age groups and ranges from 55.9% to 62.7% of the studied cases (Hamza et al., 2019;Lo et al., 2016).
Education on hypertension is an important intervention in the management of hypertension.Educational interventions can be provided one-onone, in groups, alone, or in combination with other interventions (Ukoha-Kalu et al., 2023).They can be presented verbally, in printed form, or electronically (Ampofo et al., 2020).They can also be given in the form of reminder systems using text messages, voice calls (Arshed et al., 2023), and mobile applications for self-assisted educational intervention (Liu et al., 2020).The contents of Volume 35 Issue 1 (2024) educational material usually include the definition, detection, complication, risk factors, prevention and treatment of hypertension (Indonesian Society of Hypertension, 2023), dietary changes, physical activity promotion, and stress management (Chang et al., 2023).However, the quality of evidence supporting the improvement in medication adherence with educational interventions is low to moderate (Ampofo et al., 2020) and show only moderate effects on blood pressure control (Tam et al., 2020).
Numerous factors underlie nonadherence among seniors.They range from sociodemographic (age, race, gender, and education), medical (comorbidities and polypharmacy), psychological and behavioral (beliefs about and satisfaction with treatment), and economic (insurance type) (Al-Ruthia et al., 2017).Previous research has shown that beliefs about hypertension medication adherence among seniors include (1) those that medicines have adverse effects on the kidney and drugs will no longer be effective if consumed for a long time; (2) those that doctors, family, and neighbors, especially those with experience with herbal/alternative medicine, are not supportive of medication adherence; and (3) those about inhibiting factors, such as physical and cognitive decline, irregular meal times, tablets that must be cut before being taken, insurance that does not cover all drugs, changes in drug regimens, and drug packaging that is difficult to open (Widyakusuma et al., 2023).Theory of planned behavior (TPB) states that these beliefs produce attitudes, subjective norms, and perceived behavioral control (Ajzen, 2019) that are the predictors of intention to adhere to hypertension medication.
Designing an intervention that modifies the most powerful predictor, i.e., the specific beliefs that have the greatest influence on intention, would be effective and efficient to increase intention, which in turn improve behavior.While all salient beliefs from previous studies may be used as the target of an educational intervention strategy, the prioritization of the belief that is the most powerful predictor of the intention to adhere to medication needs to be further investigated.Therefore, this study aims to identify which belief has the greatest correlation with the intention to adhere to hypertension medication among seniors.

MATERIALS AND METHODS
A cross-sectional survey was conducted among seniors with hypertension in Yogyakarta.A questionnaire based on TPB was developed and used in this study.

Instrument and its development
A questionnaire for determining the correlation of each belief with intention was constructed by following the TPB manual (Francis et al., 2004) with some modifications to facilitate administration to seniors.The questionnaire contained 24 questions and consisted of direct attitude, direct subjective norm, direct perceived behavior, indirect behavioral belief, indirect normative belief, indirect control belief, and total intention domains.On the basis of a previous study (Widyakusuma et al., 2023) and observations in the senior community, some salient beliefs were selected as indirect domains.The questionnaire was then tested for validity (corrected item-total correlation) and reliability (Cronbach's α), revised, and tested again for validity and reliability, as explained in the Supplementary Data.The first draft of the questionnaire was tested on 30 visitors of two community pharmacies in Sleman, Yogyakarta, during the first and second weeks of September 2023.After revision, the questionnaire was tested again on 15 online and 15 offline participants at the same community pharmacies during the third week of September 2023.All the participants in this validity and reliability test were chosen conveniently.After the items were verified to be valid and reliable, the final questionnaire was used as the research instrument.
The final questionnaire has two versions: offline and online.The offline version was a paperbased, researcher-assisted, or researcheradministered questionnaire.This questionnaire was written in Bahasa and Javanese with the help of two native speakers and printed in large font (Calibri 16).The online version was provided as a link to a Google Form to facilitate online sharing.Completing each version of the questionnaire took approximately 5-10 min.

Data collection and participants
The final questionnaire was distributed conveniently to three senior communities (offline) and five WhatsApp groups (online) during the last week of September 2023 until December 2023.Offline participants were recruited from three locations: Posyandu Lansia (an integrated healthservice post for the elderly) in Kotagede Yogyakarta, Yastroki (Indonesian Stroke Foundation) in Yogyakarta, and Apotek UGM Yogyakarta (a pharmacy retail owned by a university).
Online participants were invited from five WhatsApp groups: Pams Nordic Walk Community (a group of 80 seniors), the researcher's family members (a group of 53 people), SMP 2 Putri Muhammadiyah alumni '74 (a group of 75 seniors), STM Perindustrian alumni (a group of 50 seniors), and the researcher's neighbors in Kotagede, Yogyakarta (a group of 29 people).
The inclusion criteria were seniors (≥60 years of age) and diagnosed with hypertension or used antihypertensive medicine before data collection.The exclusion criteria were incomplete answers (for the offline questionnaire) and living outside Yogyakarta (for the online questionnaire).A minimum of 80 participants were needed (Francis et al., 2004).

Data analysis
This study, including its method of analysis, followed the manual for constructing a questionnaire on the basis of TPB (Francis et al., 2004).Eight direct questions applied a five-point Likert scale, and 16 indirect questions used a five-point Likert scale or a bipolar scale (−2 to 2).Direct domains were calculated by adding up each question score and calculating the mean value.Indirect domains were calculated by multiplying the scores of each correlated question, namely, questions that used a Likert scale with the results of the questions on a bipolar scale (behavioral beliefs × outcome evaluations, normative beliefs × motivation to comply, and control beliefs × influence of control beliefs).Multiple linear regression was performed by using SPSS 26 software to analyze correlations between direct domains and total intention (simultaneously), between each direct domain and total intention (individually), and between each indirect domain and its corresponding direct domain.This method of analysis has also been conducted in different studies to predict the intention of physicians to prescribe extemporaneous compounded dosage forms (Wiedyaningsih et al., 2016).

RESULTS AND DISCUSSION
In total, 97 participants completed the online (n = 48) and offline (n = 49) questionnaires.The number of online participants shows that many seniors can use technology (Table I).However, in general, the online participants are slightly younger, tended to be male, have higher education, and have had hypertension for longer than the offline participants.These characteristic differences are unsurprising because seniors who access the Internet and are familiar with smartphones are usually younger, more affluent, and highly educated than those who do not (Smith, 2014).
The variables of the TPB model significantly explained 57.6% of the variance (R 2 ) in this study (subjective norm, β = 0.291, p < 0.01 and perceived behavioral control, β=0.494, p<0.01 but not attitude, β = 0.095, p = 0.293).The simultaneous correlation coefficient (R) of the three predictors of intention, individual correlation of each predictor with intention, and correlation of each belief with the corresponding predictor (attitude, subjective norms, and perceived behavioral control) (Figure 1).The analysis showed that the strongest predictor of the intent to adhere to medication is perceived behavioral control (R = 0.698).The belief that most facilitates perceived behavioral control is frequent awareness of blood pressure level (R = 0.705), followed by the availability of help from family (R = 0.605) and whether the participant experiences hypertension symptoms or not (R = 0.597).Significant differences in education and years of hypertension between online and offline participants were observed.Subgroup analysis was conducted between the online and offline groups to determine the effect of these differences.In the online group (Figure 2), subjective norms served as the strongest predictor of the intention of adherence (R = 0.718), whereas in the offline group (Figure 3), perceived behavioral control served as the strongest predictor of the intention of adherence (R = 0.753).In the online group, subjective norms, but not attitude and perceived behavioral control (R 2 = 0.539; subjective norm, β = 0.563, p < 0.01; attitude, β = 0.164, p = 0.228; perceived behavioral control, β = 0.073, p = 0.710), explained the variance in intention significantly.In the offline group, perceived behavioral control, but not attitude and subjective norms (R 2 = 0.667; perceived behavioral control, β = 0.635, p < 0.01; attitude, β = 0.115, p = 0.379; subjective norm, β = 0.244, p = 0.069), explained the variance in intention significantly.
TPB has been widely used to explain and predict health-related behaviors, such as physical activity intention and behavior in school-aged children (Foley et al., 2008) and oral health behavior in adults (Rajeh, 2022).TPB can also be utilized to predict the decision of physicians to prescribe extemporaneous compounds in Indonesia (Wiedyaningsih et al., 2016).Moreover, TPB has frequently been applied to predict medication adherence in patients with hypertension, coronary heart disease, and schizophrenia or medication adherence in general (Ho & Lee, 2014;Javaran et al., 2020;Kopelowicz et al., 2015;Suh, 2021).
However, the strongest predictors of intention to adhere varied in each study.Even in this work, the results show discrepancies between subgroups.This inconsistency may be due to differences in educational level and hypertension duration.Consistent with those of other studies that predicted intent to adhere to hypertension medication, the aggregate results of this study show that the best predictor is perceived behavioral control (Ho & Lee, 2014;Javaran et al., 2020).Our finding that perceived behavioral control is the strongest predictor confirms that high self-efficacy is associated with good medication adherence (Rajeh, 2022).However, the results of other studies revealed that the best predictor of intention is subjective norms (Kopelowicz et al., 2015) or attitude and subjective norms (He et al., 2022).Another study also found that attitudes and perceived behavioral control significantly predicted the intention of hypertension medication adherence, whereas subjective norms did not reach significance in predicting intentions (Bane et al., 2006).
This inconsistency may be due to variations in the nature of the disease, treatment, and sociodemographic characteristics among the study subjects.For example, a study by Kopelowicz included patients with schizophrenia aged 18-50 years.Given the central role of family in Mexican-American culture, subjective norms would be the single best predictor of medication adherence (Kopelowicz et al., 2015).Meanwhile, our study included participants who were older than those included in other works.
TPB acknowledges the role of background factors.While TPB explains that a person's intention to engage in a behavior can be estimated from attitude, subjective norms, and perceived behavioral control, background factors are believed to have an indirect effect on intentions and behavior by affecting behavioral, normative, and/or control beliefs.These background factors include personality traits, intelligence, demographic characteristics, life values, and other such variables (Ajzen, 2020).
Moreover, the differences in the determinants that predict antihypertensive medication adherence are determined by the theory or construct used.TPB is not the only health psychology theory used to predict adherence to medications (Holmes et al., 2014).The health belief model and theory of reasoned action prevail among other theories.Some variables in the health belief model differ from those in TPB.The variables include perceived susceptibility; cues to action; self-efficacy; and perceived barriers (Yue et al., 2015), high susceptibility, severity, and benefit (Kamran et al., 2014) and were significantly Volume 35 Issue 1 (2024) associated with antihypertensive medication adherence (Yue et al., 2015).However, a metaanalysis has confirmed the effectiveness of TPBbased interventions (Steinmetz et al., 2016), so this study used TPB constructs.
The results of this study affect the practice of pharmacists in the way they communicate with seniors.Providing new information to improve seniors' behavioral beliefs alone may be ineffective in improving medication adherence among seniors given that the behavioral beliefs that constitute attitude are not important predictors of adherence intention.Optimizing the role of social support systems for seniors may work well to improve control and normative beliefs.However, this situation does not mean that pharmacists should avoid patient education.A study conducted at an integrated health-service post for the elderly in a rural village in Yogyakarta showed that a considerable percentage of participants still had poor hypertension knowledge (Nazihah & Rahmawati, 2022).Therefore, patient education must always be provided while improving effective two-way communication between pharmacists and senior patients with their family or other social supports.
Providing new information will not immediately change the behavioral, normative, and control beliefs that seniors already possess.Efforts can be made to provide new information such that seniors have new beliefs without necessarily changing their existing beliefs.For example, although the belief that hypertension medication is harmful for the kidney may not disappear, patients can be introduced to the new belief that the benefits of adhering to a medication regimen will be greater than the risks to the kidney.Discussions and deliberations can also play a critical role in changing normative beliefs and lifting pluralistic ignorance (Bicchieri & Mercier, 2014).For example, during pharmacist-patient interactions, senior patients and their families may be invited to ignore peers who encourage switching medications to alternative medicines without the supervision of health professionals.Pharmacists may encourage family, health cadres in the community, and other available social support systems to provide frequent access to blood pressure measurements and thus improve control beliefs.Integrated healthservice posts for the elderly (Posyandu-Lansia) in Indonesia usually provide routine blood pressure measurement for hypertension screening.Thus, Posyandu-Lansia should be optimized by improving the participation of seniors, emphasizing the role of health cadres, simplifying the care pathway, and facilitating access to hypertension treatment (Gibran & Rahmawati, 2020;Rahmawati & Bajorek, 2015).
The strength of this study lies in its participant selection method.Given that its participants were recruited from senior communities (not bound to any health facility), those who participated may be nonadherent to medication and can present various beliefs.However, it also has limitations in terms of the difference in educational level and years of hypertension between the offline and online participants.Half of the participants were online participants with significantly higher education and longer experience of hypertension than offline participants.This discrepancy might introduce selection bias to the results.

CONCLUSION
This study supports the notion that perceived behavioral control has the highest correlation with the intention of medication adherence among hypertensive seniors.Understanding control beliefs that constitute perceived behavioral control provides insight for developing strategies to improve adherence among seniors. (2023).

Figure 1 .
Figure 1.Correlation of beliefs with the intent to adhere to hypertension medication among seniors (N = 97).

Figure 2 .
Figure 2. Subgroup analysis of the correlation of beliefs with intention among online participants (N = 48).

Figure 3 .
Figure 3. Subgroup analysis of the correlation of beliefs with intention among offline participants (N = 49).

Table I .
Participant Characteristics Data are presented as mean ± SD or number (%)."Low-middle education" refers to participants who attended elementary, junior, or high school or did not attend school."High education" refers to participants who have a diploma or bachelor's, master's, or doctorate degree.*p < 0.01.