A total of 410 participants were recruited randomly from the five main governorates of Lebanon between the months of February and August 2020. These include Beirut, Mount Lebanon, North, South and Bekaa. The regions of Mount Lebanon and Beirut share the highest number of inhabitants- around 75%- according to national registries.

In this cross-sectional study, participants were first given a brief explanation of the study, and consent was obtained upon filling a self-administered questionnaire. Due to the COVID-19 situation, it was modified to a google form survey in order to be able to proceed with data collection. According to the Epi info sample size calculations, providing a confidence level of 95%, a margin of error of 5%, and assuming 36% of the population have adequate KAP according to a study conducted in Saudi-Arabia (36), a total sample of 354 Lebanese adults were targeted.

Lebanese adults aged 18 years and over were included in the study. Participants under 18 or who do not hold a Lebanese nationality were excluded.

Review board approval was acquired on the 5th of October 2019 from the Lebanese University.

The instrument used was adopted from two validated and reliable stroke questionnaires namely the Stroke Knowledge Test (SKT) that has a validated Arabic version and the Stroke Action Test (STAT) which was translated into Arabic. The SKT and the STAT aimed at evaluating the participant's knowledge and practice toward stroke, respectively. The survey instrument was then piloted on 10 subjects prior to its finalization and distribution. The questionnaire's final version included 3 sections.

The first section was concerned with the respondents' socio-demographic characteristics, social factors and habits, family history, and medical profile. Registered data included age, gender, marital status, education level, social security insurance, occupation; whether they have ever known anyone who has ever had a stroke, where they get their health information. Presence of known stroke risk factors such as alcohol consumption, smoking, dyslipidemia, diabetes, hypertension, abnormal heart rhythm, previous cerebrovascular disease, previous MI, family history of cardiovascular or cerebrovascular disease.

The second part investigated the knowledge of stroke, risk factors and warning signs of stroke using the SKT.

The SKT is a 20-items questionnaire developed by Sullivan and Dunton (43). The cut off point for good knowledge level is 50% of the 20 items. The tool has demonstrated good internal consistency, test-retest reliability and construct validity (25, 43).

Results from psychometric investigations were established at Cronbach of 0.65 in the original questionnaire demonstrating its reliability and validity in stroke knowledge assessment (43). The instrument is used worldwide and has been translated into other languages including Arabic by Eshah (34) in her study entitled “Knowledge of Stroke and Cerebrovascular Risk Factors in Jordanian adults”. For the Arabic version, face and content validity were established with a Cronbach's alpha of 0.68 (34). A copy of the tool in both versions with the answers were obtained by directly contacting Dr. Sullivan and Pr. Eshah through email and permission to use the tool was granted. Minor adjustments were made to the Arabic version by changing certain terminologies as dialects differ in different Arabic countries (44).

Reliability of the instrument in our study was established with a Cronbach alpha of 0.877.

The last part of the survey explored appropriate practice toward stroke using the STAT. It is a 21 items questionnaire developed by Billings-Gagliardi and Mazor (45) in the study entitled “Development and Validation of the Stroke Action Test.” The objective of the study was to develop and assess a tool where the participant needs to link individual stroke and non-stroke symptoms with the most appropriate action. The reliability of the 28-item test was good as determined by a Cronbach's alpha of 0.83 (45).

A copy of the STAT tool was obtained by contacting Dr. Billings-Gagliardi directly by email and permission to use the tool granted. The instrument was translated to Arabic by two bilingual experts and back translation into English was performed by two additional bilingual experts. The language experts compared the original English instrument with the back translated Arabic instrument and edited to obtain the matched Arabic version.

Reliability of the instrument in our study was established with a Cronbach alpha of 0.923.

Statistical analyses were performed using Statistical Package for Social Science (SPSS) version 23 (IBM SPSS Software, Chicago, IL, USA). Descriptive statistics were calculated using mean and standard deviation for continuous measures, counts and percentages for categorical variables.

A reliability test was performed for each of the questionnaires used using Cronbach's alpha to determine internal consistency. Data was weighted according to age and gender. The assumptions of the statistical tests were checked. The Student t-test and ANOVA test were used to assess the association between each continuous independent variable (knowledge: SKT score, and practice scores) and the sociodemographic variables. Multivariable linear regression models were conducted to examine factors associated with the SKT and the STAT scores as dependent variables and taking all variables that showed a p < 0.25 (46) in the bivariate analysis as independent variables. A two-sided p < 0.05 was used to indicate statistical significance.

A total of 410 participants from the 5 districts of Lebanon completed the questionnaire. The respondents' socio-demographic characteristics, social factors and habits, family history and medical profile are summarized in Table 1 in Appendix A.

The study population included 281 women (68.5%) and 129 males (31.5%), with 77.6% aged between 18 and 44 and 22.4% aged 45 and above. The results showed that 48% of the subjects were married, 37.3% without any professional activity, 62% with social security insurance and 89% had completed a university degree. In addition, the majority of participants (48%) lived in Mount Lebanon and 20.2% in Beirut. 23.7 and 55.9% of the respondents were occasional smokers and alcohol consumers, respectively. Around 45% of the participants had a family member or a friend who suffered a stroke.

Of the 410 respondents, many modifiable stroke risk factors were observed. The most prevalent were diabetes (89%) and dyslipidemia (86.9%) followed by hypertension (63.8%) and atrial fibrillation (49.4%). Moreover, 29.7% declared getting their health information from television and 22.6% from a health or fitness center.

The mean SKT score of the participants was 9.16. Respondents were divided into 2 groups based on their SKT scores. 48.5% (n = 199) scored <50% and therefore showed a poor stroke-related knowledge level and 51.5% (211) a good knowledge level.

In addition, 61% of the sample was able to identify the pathological mechanism of stroke.

Moreover, of the 8 modifiable risk factors considered, the average score for knowledge was 3.49 of 8. 12.4% of participants were not able to identify 1 risk factor, around 20% were able to recognize 4 and approximately 43% were unable to enumerate more than 3. The most frequently recognized risk factors were physical inactivity (70.5%), obesity (69.3%), and hyperlipidemia (67.3%) followed by hypertension (52.9%).

In the bivariate analysis, being between 18 and 44 years old (p = 0.018), having a university degree compared with secondary education (p < 0.001), social security insurance (p = 0.02) and a medical history of diabetes mellitus (p = 0.07) were significantly associated with a higher SKT score. On the other hand, the mean SKT scores of participants living in Mount Lebanon compared with Beirut (p < 0.0001), who are occasional smokers vs. non-smokers (p = 0.015), occasional alcohol drinkers compared with non-drinkers (p = 0.035), who have a medical history of hyperlipidemia (p = 0.007) and who get their health information form the internet (p = 0.016) were significantly lower (Table 1).

In the multivariate linear regression, living in Mount Lebanon compared with Beirut (ß = −8.819, p < 0.001) and occasional smokers compared with non-smokers (ß = −13.099, p = 0.016) showed statistically significant lower mean SKT scores; whereas, university degree vs. secondary education (ß = 7.217, p = 0.006) and suffering from diabetes mellitus (ß = 12.719, p = 0.01) were associated with higher mean SKT scores (Table 2).

The mean overall STAT score was 41.3%. On average, for 36.8% of the stroke symptoms, respondents selected call 112. Moreover, of the 21 stroke symptoms considered, ~41.5% of respondents were unable to identify the most appropriate action to more than 6 stroke symptoms and around 14% responded correctly to all symptoms.

The stroke symptoms for which the fewest percentage of respondents would call 112 were “sudden loss of coordination” (0.7%), followed by “sudden confusion” (24.6%), “sudden dizziness” (25.1%) and “sudden severe headache with no known cause” (28%). On the other hand, call 112 was greatly chosen by the participants in situations where difficulty speaking presented together with sudden arm and face weakness. Moreover, 82.4% of the sample recognized the need to call 112 when the diagnosis of stroke was provided.

In the bivariate analysis, medical history of heart attack and atrial fibrillation were significantly associated with a higher mean STAT score (p = 0.026 and p = 0.002, respectively). On the other hand, getting health information from the internet was significantly linked to a lower mean STAT score (p < 0.001; Table 3).

In the multivariate analysis, the mean STAT scores of participants who get their information from the internet, a family member and church were statistically significantly lower (ß = −10.177, p < 0.001; ß = −9.069, p = 0.001; and ß = −36.975, p = 0.001, respectively). Also, no association was found between the SKT score and the STAT score (Table 4).

The current research showed that stroke-related knowledge in the Lebanese community is insufficient. Almost half of the respondents demonstrated a poor level of stroke knowledge which is similar to studies conducted in numerous other countries (26–29, 34, 35, 47–49).

Most of the participants (61%) were capable of identifying the pathological mechanism of stroke by answering that it is due to blockage of the blood supply to the brain. This revealed that Lebanese adults had some awareness about the development of the illness. The understanding of the disease development is important for them to be able to distinguish a stroke from a heart attack which tend to be confused by a large number of people (50–53). This result is in accordance with studies conducted in Malaysia (66.7%) (48) and in Ireland (60.3%) (47), however the percentage is higher than studies conducted in India (31%) (54) and a previous study conducted in Lebanon where 44.6% of the participants stated the pathological mechanism of the disease (55). An explanation could be that in the current study, closed-ended questions were used instead of open-ended questions employed in the previously mentioned studies. Also, it is noteworthy to mention that the Arabic translation of stroke is more descriptive and therefore facilitates the understanding of disease development.

Moreover, there was a lack of knowledge about controllable risk factors of stroke which is a finding that is consistent with many previous studies (26–29, 34, 47, 56–64). This finding may provide an indication into the cause for the higher stroke prevalence in Lebanon compared to other developing countries in the middle East.

The well-recognized risk factors in our research were physical inactivity (70.5%), obesity (69.3%), hyperlipidemia (67.3%), and hypertension (52.9%). Similarly, numerous studies have revealed these 4 as being the most identified risk factors of stroke (22, 59, 65). In contrast, in a recent study conducted in Ethiopia among hypertensive patients, hypertension was the least recognized risk factor (3.6%) (28).

Interestingly, although smoking is a well-known stroke risk factor, it was the least perceived risk factor in our study (9.8%). This finding contrasts with many other studies (22, 59, 65) including a previous study conducted in Lebanon where smoking was one of the most recognized risk factors (55). Furthermore, participants underestimated the most pertinent risk factors of stroke-diabetes and atrial fibrillation-despite their high prevalence among respondents in our research. Only a small percentage recognized that they were important stroke risk factors (37.8 and 17.6%, respectively). Numerous research has also shown poor identification of diabetes and atrial fibrillation as risk factors for stroke (22, 57, 65, 66). On the other hand, participants with diabetes were associated with a higher level of stroke-related knowledge which is a result confirmed by several other studies (22, 67).

Living in Beirut was associated with a significantly higher level of stroke-related knowledge than living in Mount Lebanon. This result needs to be confirmed as the number of participants in Mount Lebanon was larger than that of Beirut (197 and 83, respectively).

In addition, research has demonstrated that educated and younger participants have a higher level of stroke-related knowledge (26, 54–59, 68, 69). Our results ad to the literature by confirming this association. It is noteworthy to mention that in Lebanon; the proportion of the population which is literate is higher among the younger adult population which could explain the latter.

Also, our findings revealed that non-smokers have a higher stroke knowledge level. This finding correlates with other studies conducted in Jordan and in Ireland (34, 47). This may be because non-smokers are more conscious about their health which stops them from smoking hence leading to a pursuit of health-related information.

Our results revealed a poor correlation between stroke symptoms and appropriate action. This may be due to insufficient knowledge to effectively recognize stroke symptoms and hence act accordingly, which is in accordance with a former study conducted in Lebanon (55). A similar result was reported in numerous other studies. The findings showed that knowledge of stroke symptoms was significantly associated with the intent to call 112 (24, 50, 70). Moreover, a study conducted in India showed a significant association between poor knowledge of warning signs and delay in hospital arrival (20).

Interestingly, our results showed that 82.4% participants would call 112 when provided with the diagnosis of stroke which is a finding reported in numerous other studies (66, 71–73). However, this result contrasts with Khalil and Lahoud's study where only 57.7% of the respondents would call 112 when stroke was suspected (54).

Moreover, in our study, only 36.8% of participants chose call 112 in response to specific warning signs. A comparable finding was reported in studies conducted in Australia and the US (24, 45).

In addition, our results revealed that the symptoms for which most participants would call 112 are difficulty speaking presented together with sudden arm and face weakness which is a consistent finding with other studies (45, 47). On the other hand, sudden dizziness was one of the symptoms with the lowest intent of respondents to call 112 which is in accordance to another study conducted in Ireland (47).

Furthermore, our findings showed that getting health information from the internet was negatively linked to adequate response to stroke. Tonsaker (74) highlighted that online health information-being difficult to regulate-can be misleading and hence patients might misinterpret it and make important health decisions that conflicts with appropriate medical practices.

It is important to note that there was no significant association between knowledge of stroke and the decision to call an ambulance in our study which is in agreement with a study conducted in Ireland (47). However, this differs from Lackland's et al. study which revealed a decline in stroke mortality in the US that was attributed to the implementation of interventions and programs resulting in better knowledge of stroke risk factors, hence improving stroke practice (38–42).

Also, our findings showed a positive relationship between having a medical history of atrial fibrillation and heart attack which is a finding that differs from a study conducted in Ireland where no association was found with the presence of any diagnosis of cardiovascular diseases (47) and another research in the US where a positive relationship was seen in participants with no prior history of heart attack (45).

Previous studies have shown that the achievement of primary preventive strategies and early treatment succeeding a stroke is strongly linked to the community's knowledge and practices of stroke (25, 35, 38–41, 56, 75). Sufficient stroke knowledge and perception could lead to rapid and accurate stroke identification, therefore timely management. Hence, public knowledge and the right practices regarding stroke are required, with aims to lessen the burden of stroke.

Our results revealed poor stroke-related knowledge and a poor relationship between stroke symptoms and intent to call 112. These results are disturbing bearing in mind the increasing occurrence of stroke in Lebanon. Therefore, these results warrant the need to develop and implement national community awareness campaigns. The latter can be designed through institutions, universities, medical care facilities such as doctor's clinics and pharmacies. The use of various media methods is also fundamental in order to disseminate the information and raise public knowledge. In addition, the development of health promotion agendas is of outmost importance in order to aid people embrace healthier lifestyles and subsequently reduce the risk of stroke. The provision of health education focusing on stroke risk factors, warning signs and appropriate action is crucial in order to reduce stroke incidence, morbidity and mortality.

Additionally, Lebanon's current economic and financial crisis is resulting in poor available infrastructure, facilities and qualified personnel despite ranking 23rd in the world for healthcare efficiency in 2019 (76). This adds to the importance of raising stroke awareness in the general public as the lack of amenities, services and equipment is alarmingly rising in Lebanon today.

Our study has faced numerous limitations. First, the sample in our study is not representative of all of the Lebanese people as our respondents filled a google form survey which would exclude nationals with no access or not familiar with technological advances leading to selection bias. Moreover, it is important to mention that the scales used in the current research are not validated in Lebanon yet. In addition, the questionnaire was designed with close-ended questions which can overestimate stroke knowledge of respondents. Furthermore, all of the respondents' characteristics and answers were self-reported, therefore the degree of true answers or confirming participant's statements is not possible in the current study which may have caused information bias. The current research is the first study aimed to assess stroke knowledge and practice in adults 18 years and above residing in all five Lebanese districts. Previous research only targeted older adults living in Beirut (55). In addition, the weighting of the data according to age and gender allowed to adjust the sample representation in our study.