Impacts on practice

  • Recurrent stroke is very common among stroke survivors, with 50% of recurrent event occurred in less than two years. This may partly be a result of poor stroke knowledge and poor medication adherence among stroke survivors while improvement in risk factors management is also needed. Special attempts should be made for the vulnerable groups including those with advanced age, low level of education and stroke survivors with disability.

  • A large proportion of stroke survivors may not have adequate knowledge regarding stroke despite having a prior experience with stroke. As a result, all healthcare professional should make every attempt to constantly assess and improve such knowledge. Efforts should focus on reinforcement of patient’s understanding regarding the signs and symptoms of stroke, importance of early hospital arrival to ensure timely management of all stroke cases and importance of preventive treatment of stroke.

  • More than half of stroke survivors had suboptimal level of medication adherence. Pharmacists, as the drug experts, must play a role in assessing and improving medication adherence in this high risk population. Based on our study, stroke knowledge independently and strongly correlates with medication adherence. As a result, a comprehensive plan of patient education which covers both general stroke knowledge and medication therapy may be an effective way to improve medication adherence.

Introduction

Stroke has become a major public health burden both as the leading cause of death and disability in the developing countries [1]. In Indonesia, stroke has been ranked as the number one cause of death for more than 10 years [2]. Stroke also leads to the greatest number of Disability-Adjusted Life Years (DALYs) lost in 2017 [2]. Generally, there are a number of barriers to improve stroke care. However, poor stroke knowledge and low medication adherence for stroke prevention are of particular importance [3, 4]. Studies suggested that a large proportion of stroke survivors lack knowledge about the warning signs of stroke and the importance of rapidly seeking treatment [5, 6]. The scale of problem may be even larger among developing countries with less advanced health system and suboptimal health education among general public. Previous studies indicated that less than half of stroke survivors knew the purpose of an anti-platelet drug [7, 8]. In addition, previous studies reported that 31–65% of stroke survivors had poor self-reported medication adherence [8,9,10]. Poor adherence to preventive medications has been shown to increase the risk of recurrent stroke and mortality in stroke survivors [11]. Previous studies conducted in atrial fibrillation patients taking an oral anticoagulant suggested a positive relationship between stroke knowledge and medication adherence [12, 13]. Hence, attempts to improve stroke knowledge should be implemented to improve medication adherence. Despite the importance of stroke as a major public health burden for Indonesia, little data exists on the quality of stroke management along with patient aspects of care particularly stroke knowledge and medication adherence.

Aim of the study

This study aimed to assess the level of stroke knowledge and medication adherence to stroke preventive medications, along with the relationship between stroke knowledge and medication adherence. The availability of these data may be beneficial for health authorities in the country to improve stroke care quality.

Ethics approval

This study protocol was approved by Faculty of Dentistry/Faculty of Pharmacy Mahidol University Institutional Review Board (reference number: 0517.0319/274) and the Research Committee at Dr. Soetomo General Hospital (reference number: 1021/KEPK/III/2019) and Universitas Airlangga Hospital reference number:123/KEH/2019. Written informed consent was obtained from all study participants. Confidentiality and anonymity of all patients recruited into this study were maintained.

Method

Study design and participants

This study was a prospective, cross-sectional study among adults (aged ≥ 18 years) with recurrent stroke or transient ischemic attack (TIA) who had sufficient cognitive and communicative ability with Mini Mental State Examination (MMSE) score as no and mild cognitive impairment. The study was conducted at two ambulatory neurology clinics of the Dr. Soetomo Academic Medical Center Hospital (SAMC) and Universitas Airlangga Hospital (UAH) in Surabaya city, Indonesia, during April–September 2019. Patients who provided written informed consent were included into the study. Patients who had severe aphasia limiting comprehension, dementia or mental illness were excluded from this study. Data from the East Java’s Basic Health Research reported a stroke incidence of 16.4%, a sample size for this study was therefore estimated to be at least 210 patients.

Translation, validation and reliability test of study tools

Since there was no Indonesian version of the questionnaires to assess both stroke knowledge (Stroke Knowledge Test or SKT [14]) and medication adherence (Morisky Green Levine Adherence Scale or MGLS [15]), both questionnaires were translated from the original English version into Indonesian Bahasa version (Supplementary Material No. 1–2). Subsequently, a back translation was performed by two bilingual experts. The language experts compared the translated version against the original English version. After minor adjustments, a face validity test was performed in 20 stroke patients from 2 hospitals. Minor adjustments were done based on the comments of patients. Reliability tests were done on both Indonesian Bahasa version in 20 stroke patients. Cronbach’s alpha values for reliability and internal consistency were 0.69 and 0.65 for SKT and 0.81 and 0.61 for MGL indicating that both tools had acceptable reliability and consistency.

Data collection

After obtaining consent from the patients, data were collected through interview, medical chart and hospital database where appropriate. Demographic data consisted of age, gender, educational status and status of smoking and alcohol use. Comorbidities were obtained from clinical diagnosis from the medical chart and/or the International Classification of Diseases 10th Revision (ICD-10) coding in electronic hospital database. Stroke related information consisted of types of stroke, risk factors of stroke, duration of stroke, level of locomotor disability, number and type of medications related to stroke prevention.

Assessment of stroke knowledge was performed using the Stroke Knowledge Test questionnaire which consists of 20 questions with multiple choices [14]. Medication adherence of patients was measured by Morisky Green Levine Medication Adherence Scale (MGLS) questionnaire which contains 4 questions with “yes” versus “no” answer for each item [15].

Data analysis

Descriptive analysis was used to analyze participant’s characteristics. Normally distributed continuous variables were described as mean ± standard deviation (SD), while categorical variables, as absolute and relative (%) frequencies. Kolmogorov–Smirnov test was used to test the normality of the data. SKT score ranges from 0 to 20. A score range of 0–10 indicates low level of stroke knowledge while a score range of 11–20 indicates high knowledge level. MGLS score ranges from 0 to 4. A score range of 0–1, 2–3 and 4 indicates low, medium and high level of adherence, respectively. Spearman’s rank correlation and Pearson’s Chi squared analysis were used to evaluate the correlation between knowledge of stroke and medication adherence. Multivariate analysis (binary logistic regression) was used to assess the relationship between SKT and MGLS scores. Odds ratio (OR) and 95% confidence interval (CI) were calculated with p < 0.05 considered as statistical significance. Subgroup analysis based on study site was also performed to ensure consistency of results. Data were analyzed using SPSS 20 software for Windows.

Results

Baseline characteristics

A total of 215 patients were recruited into this study. Mean age was 56.34 ± 8.69 years, 56.7% were male. The education level of the majority of patients (85.6%) was secondary education or less. (Table 1) The most common comorbidities were hypertension (84.2%), dyslipidemia (44.2%) and diabetes (27.9%). Approximately one-third were smokers while most patients did not consume alcohol. Most patients (96.3%) had a history of ischemic stroke with 13.9% with some levels of disability. The majority of the recurrent events were ischemic stroke (70.2%) and approximately half occurred within 2 years after the previous event (Table 1). Approximately one-third of patients took more than 5 drugs per day. The most commonly used medications were anti-thrombotic therapies (93.02%), anti-hypertensive drugs (94.88%), lipid lowering agents (39.07%) and anti-diabetic drugs (17.2%).

Table 1 Baseline characteristics of the study population
Full size table

Knowledge of stroke and medication adherence

The mean SKT score of the study population was 7.89 ± 3.38. One hundred and sixty-five patients (76.7%) had low level of stroke knowledge (Table 2). More than two-thirds of patients were able to identify that brain is an organ involved in stroke. Among 9 SKT items related to risk factors of stroke, patients were able to identify only 5 items as stroke risk factors. Patients performed poorly in recognizing stroke signs and symptoms and immediate response to stroke. In addition, most patients had poor understanding related to stroke treatment and importance of rehabilitation. Detail analysis of each item of SKT is provided in Supplementary Material No. 3. The mean MGLS score of the study population was 3.05 ± 1.11. There were 47.9%, 41.4% and 10.7% of patients who were classified as having high, medium and low medication adherence, respectively (Table 2).

Table 2 Stroke knowledge test (SKT) and Morisky-Green-Levine Adherence Scale (MGLS) scores of the study population
Full size table

Correlation of stroke knowledge level and medication adherence

Education level (Spearman’s rank correlation coefficient [rs]: 0.307, p = 0.001) and duration of stroke (rs: 0.128, p = 0.041) were found to significantly correlate with stroke knowledge level (Table 3). Age (rs: 0.169; p = 0.013) and disability (rs: 0.171; p = 0.012) were found to significantly correlate with medication adherence (Table 4). Results from the Pearson’s Chi squared analysis were similar to that of Spearman’s rank correlation analysis. Gender, education level, number of medication used, and duration of stroke did not show significant impact on the level of medication adherence. A binary logistic regression was performed to evaluate the relationship between level of stroke knowledge and medication adherence. Results showed that high stroke knowledge level was independently associated with high medication adherence (adjusted OR: 4.37, 95% CI 2.00–9.53; p < 0.0001) (Table 5). Subgroup analysis based on study sites were performed and results were consistent with the main analysis (Supplementary Material: No. 4–5).

Table 3 Correlation between baseline characteristics and level of stroke knowledge (N = 215)
Full size table
Table 4 Correlation between baseline characteristics and level of medication adherence (N = 215)
Full size table
Table 5 Binary logistic regression evaluating the relationship of baseline characteristics and stroke knowledge versus medication adherence (N = 215)
Full size table

Discussion

To the best of our knowledge, this is the first study evaluating the level of stroke knowledge and medication adherence along with their relationship among stroke survivors of Indonesia which is one of the countries that are most severely afflicted by stroke. Results from the Global Burden of Disease (GBD) study indicated that Indonesia ranked fourth in the number of stroke fatalities in the world and had the highest rate of increase in the stroke age-standardized rates during 1990–2016 [16]. As a result, despite a small sample size, findings from our study provides important information on the current status of stroke care of this country.

Data from the baseline characteristics were consistent with stroke in developing world where patients tended to be much younger compared to those from developed world [1]. Consistent with being a country with predominantly Muslim population, there was a very low rate of alcohol use [17]. However, one-third was active smokers. As a result, tobacco control campaign may be useful to reduce this modifiable risk factor along with better control of most common risk factors including hypertension, dyslipidemia and diabetes [18]. From our study, there was a discrepancy between the amount of comorbidities and medication use. This may potentially indicate suboptimal utilization of preventive pharmacotherapies such as statins to name a few. All these risk factors may contribute to a very high rate of stroke recurrence found in our study population. Half of these Indonesian stroke survivors suffered a recurrent event within 2 years while the reported annual rate of stroke recurrence in developed countries ranges from 4.9 – 8.7% due to the much better implementation of preventive strategies [19].

Our analysis on SKT indicates a number of areas of improvement on patient’s knowledge. Despite having a prior experience with stroke, stroke knowledge in our study population remains poor. Lack of understanding on signs and symptoms of stroke and failure to recognize the importance of early hospital arrival may contribute to the late hospital arrival which may limit patients’ eligibility for intravenous thrombolysis [20]. Even more troubling is the fact that these stroke survivors had inadequate knowledge on stroke treatment which may increase the risk of recurrent events due to non-compliance to treatment [17]. With late hospital arrival, high rates of morbidity, mortality and disability in this high-risk population are expected. This fact highlights an urgent need for a more rigorous effort to improve stroke knowledge among stroke survivors to lessen the burden on the country’s healthcare system.

Secondary prevention of stroke requires effective pharmacotherapy regimen along with optimal medication adherence [4, 18]. Findings from our study clearly indicates suboptimal adherence among stroke survivors. Less than half of our study population had high medication adherence level. Medication adherence is a result of a complex interplay of various factors such as patient-related factors (e.g., knowledge), socioeconomic factors, therapy-related factors, health system factors, and stroke-related factors [3, 4]. As a result, evaluating and tackling medication adherence issue is a unique challenge for each country. Our analysis indicated that stroke knowledge independently and strongly correlates with medication adherence. This is clearly a call to action that improving stroke knowledge and increasing medication adherence among stroke survivors are urgently needed.

There are several limitations in this study. First, due to the study design as a cross-sectional study, there is no longitudinal data to observe changing level of medication adherence through time. Second, causal relationship cannot be drawn with certainty from a cross-sectional study. Third, there is a possibility of bias from the subjectivity of adherence measurement, since it was based on self-administered questionnaire. Fourth, this study was performed in limited geographic area with data collection from only 2 hospitals in Surabaya city. As a result, generalizability of the study results is limited to the areas with similar social and health context and cannot be a representative of a vast country like Indonesia. More research on stroke is clearly needed to guide health policy and action against this important disease. Lastly, we acknowledged that the validity of MGLS is not perfect. However, among the self-reported medication adherence questionnaires available, MGLS is one of the most widely used and validated tools across broad range of patient population [21]. With only 4 items, it is very easy to use in clinical practice particularly in developing countries where overcrowded hospitals are common. Results from our validation suggested that this tool was acceptable to use in our study.

Conclusion

Despite having a prior experience with stroke, 3 out of 4 stroke survivors in Indonesia had low stroke knowledge. Lack of understanding on signs and symptoms of stroke, failure to recognize the importance of early hospital arrival and inadequate knowledge on stroke treatment are the three main areas of discrepancies in stroke knowledge. In addition, less than half of stroke survivors had optimal level of medication adherence. Stroke knowledge independently and strongly correlates with medication adherence. As a result, attempts should be made to increase stroke knowledge which may positively improve medication adherence among this high risk population. This, in the long run, will ultimately lead to a reduction in healthcare burden of this deleterious condition.