Knowledge, Attitude, and Prevention Practice on Dengue and Dengue Seroprevalence in a Dengue Hotspot in Sabah, East Malaysia

Background: Dengue, a major global public health concern, is growing more prevalent. Extensive measures have been undertaken to prevent and control dengue transmission in Malaysia including understanding and modication of human behaviour. This study was designed to assist in planning an effective health intervention for dengue by measuring residents’ knowledge, attitudes, and practice (KAP) in hotspot areas of Sabah, East Malaysia. Methods: A community-based and cross-sectional study was conducted at dengue hotspot area in Sabah involving 463 residents (aged 18 years and above). Information on the socio-demographic characteristics of the residents and their KAP towards dengue was collected using a structured pre-validated questionnaire developed by the researchers. In addition, residents were invited to participate in a dengue seroprevalence study using dengue Immunoglobulin M (IgM) and Immunoglobulin G (IgG) ELISA kit. Results: Most of the residents correctly identied common signs and symptoms of dengue such as fever (84.4%), chills (70.7%), and nausea or vomiting (56.0%). Nearly half of the residents (51.3%) had good knowledge, 66.0% had a positive attitude, and 63.7% had a positive practice for dengue prevention and control, respectively. The dengue seroprevalence study included 200 of the 463 residents; IgG and IgM seropositivity was 74.1% (n = 63) and 7.1% (n = 6) respectively. Based on logistic regression analysis, local and migrants differs considerably in terms of their knowledge and attitude toward dengue prevention and control, while gender and level of education were predictive of practice level (p<0.05) among all sociodemographic variables examined. Knowledge was correlated with attitude (p<0.0001) and practice (p< 0.0014). Discussion: To combat dengue epidemics in the hotspot area of Sabah, a proactive and behavioural changes among residents, are required. To effectively prevent and manage dengue transmission, a collaborative effort of government sectors, private sectors, housing residential committees, and community residents is essential. and strongly agree to negative comments received a score of one. The KAP level was calculated based on the total number of correct answers in each section. A resident would be labelled as "good" if he or she scored at least 80% in a category; if the score was less than 80%, the resident would be labelled as "poor." The KAP survey's scoring system and cut-off points were based on earlier research (17, 18, 19). Frequencies and percentages were used in the sociodemographic factors' descriptive statistics. The chi-square test or Fisher's exact test were used to examine the relationship between the independent variable and the KAP level (good or poor). The relationship between good KAP and demographic and socioeconomic characteristics was investigated using univariate logistic regression analysis. The statistical signicance level was set to 0.05. On the other hand, Spearmen's rank correlation was employed to establish the extent of correlation between KAP scores.


Background
In Malaysia, the annual accumulative dengue incidence continues to rise year on year; in 2018, there were 80,615 cases reported, which grew to 130,101 in 2019 (a 61.4% increase), indicating that the country's dengue incidence rate is currently at 390.4 per 100,000 people (1). Dengue is more common in areas undergoing rapid urbanization such as the Federal Territory of Kuala Lumpur, the state of Selangor, and the state of Johor in Malaysia (2,3). Interestingly, the state of Sabah in East Malaysia has consistently been ranked highly in the country for dengue cases (1). The annual dengue incidence rates in Sabah were spatially variable over the period of 2010-2016, with average annual incidence of 21 cases per 100,000 people (1). According to a recent study by Murphy, A et al (2020), dengue cases in Sabah primarily affected rural areas in the Western districts (Kuala Penyu, Nabawan, Tenom, and Kota Marudu), with a smaller number of more severe cases in the eastern districts (Sandakan and Tawau) (4).
According to the latest national statistics, as of mid-2015 Sabah has a population of approximately 3.7 million (5), making it the third most populous state in Malaysia. It is home to diverse ethnicities comprising mainly of indigenous communities of Kadazandusun (25%), Bajau (15%) and Murut (3%) (6). Over the years, Sabah has also experience increasing number of migrants, accounting for 27.7% (886,400) of Sabah population (7), with most migrants being of Indonesian (85%) and Filipino (15%) origin (8). Several reasons led to increase migrants in Sabah such as religions, political and economics advantageous (9). As a result of increasing population density, urbanization have extended outwards from towns and cities to the outskirts, resulting in enormous social and economic changes. For example, in the town of Telipok (urban expanse of Kota Kinabalu, the capital city of Sabah), urbanisation has resulted in the growth of low-cost homeownership, due to high demand from the low-income group. Water supply services are frequently disrupted in this town, which has an admixed population of both locals and migrants, as are irregular sewage systems for draining rainwater and ine cient garbage collection services. Unsurprisingly, these interconnected components linked to human migration, environmental changes, and urbanisation have facilitated the spread of dengue (10) across Sabah.
The fact that there are no effective antiviral medications available for dengue further complicates the situation. As a result, holistic prevention -which includes the eradication of breeding sites, environmental management, and good community engagement and cooperation, is the most effective method for dengue prevention and control. According to a study, sociodemographic variables, behaviour, and belief in the community are all closely linked to dengue prevention and control (11). Despite the fact that the health authorities in Sabah launched massive health outreach campaigns, the communities' attitudes and practices remained unchanged, as evidenced by high dengue infection rates. As a result, the current study measured community residents' KAP in hotspot areas of Sabah, East Malaysia, in order to gain a better understanding of population health and to stimulate public health formulations to develop an effective dengue health intervention.

Study design and setting
A community-based and cross-sectional study was conducted at Taman Telipok Ria and Taman Sri Rugading, which are two residential areas located at Telipok district, a hotspot dengue cases in Sabah (12). The residential apartments consist of ve-story walk-up apartment complexes with both low-and high-density units. The study utilized convenience sampling to collect data from residents' homes as well as a community clinic in the proximity of the housing area (Fig 1).

De nition of terms
Being of Malaysian nationality is based on Malaysia's citizenship constitution, which states that that someone is a Malaysian national if either one of his or her parents is a Malaysian or a permanent resident at the time of birth, or he or she was born in Malaysia and is not a citizen of any other country (13). Two staff nurses were assigned to properly check the residents' citizenship by questioning them about their family background. A dengue hotspot is de ned as an area where a dengue outbreak has been prevalent for more than 30 days (12).

Participation and sample size
Data was collected from April 2020 through February 2021, following written approval from the housing management bodies and district police o ces. Residents who were aged at least 18 years and above and those provided informed consent were solicited to participate in the study. Residents self-administered the questionnaire and were doublechecked to ensure that they had answered all of the questions. Based on G*Power3.1 software, with effect size = 0.3, α error probability = 0.05 and power (1-β error probability) = 0.95, a total of 134 residents were required to achieve the minimum recommended sample size (14).

KAP survey questionnaire
Information on the socio-demographic characteristics of the residents and their KAP towards dengue was collected using a structured questionnaire developed by the researchers. Prior to the actual data collection, a pilot study and exploratory factor analysis (EFA) study involving 30 and 468 residents respectively was conducted to validate the questionnaire written in native Sabah creole (15). There are three sections to the questionnaire. Section A consists of nine questions designed to test residents' general knowledge of dengue, including dengue transmission, vector biology, clinical symptoms, names of major dengue vectors, breeding habitats, active time of dengue vectors, and treatments for dengue (if any). Section B comprises eight items relating to residents' attitude towards dengue, such as importance of maintaining a clean and vector breeding free environment, vector control, and the role of community residents on vector control. Lastly, Section C consisted of 11 items relating to the residents' usual dengue prevention practices such as maintaining a clean environment, home gardening practices, covering of water storage tanks, and use of Abate ® , a larvicide, as vector control. The survey questionnaire is attached supplementary le.

Dengue seroprevalence test
In addition to the KAP survey, residents were invited to participate in a dengue seroprevalence study. Three millilitres of venous blood were collected in ethylenediaminetetraacetic acid (EDTA) blood tubes, maintained at room temperature, and delivered to the laboratory in four hours. The blood was then centrifuged for 4 minutes at 4,000 rpm, and the serum was aliquoted, coded, and kept at -80 °C for later use. According to the manufacturer's instructions, a dengue IgM and IgG ELISA kit (Standard Diagnostics Inc., Yongin-si, Gyeonggi-do, Republic of Korea) was used to determine the presence of anti-dengue immunoglobulins in the sera. The presence of anti-dengue immunoglobulins in the serum of each community resident was conducted in duplicate. Dengue IgM positivity indicates an acute dengue infection, whereas dengue IgG positivity indicates past dengue exposure (16).

Data management and statistical analysis
Both the KAP questionnaire and blood sample of the residents were coded with a unique identi cation number based on their residential address. Data analysis was performed using Statistical Package for Social Science version 23 (SPSS, Inc., Chicago, IL). A scoring system was utilised to evaluate KAP data. All correct and positive responses in the knowledge and practice sections received a score of one, while "don't know" and incorrect responses received a score of zero. The attitude section is scored on a Likert scale, with one being the lowest and ve being the highest. Strongly agree to positive statements and strongly disagree to negative statements obtained a score of ve, while strongly disagree to positive statements and strongly agree to negative comments received a score of one. The KAP level was calculated based on the total number of correct answers in each section. A resident would be labelled as "good" if he or she scored at least 80% in a category; if the score was less than 80%, the resident would be labelled as "poor." The KAP survey's scoring system and cut-off points were based on earlier research (17,18,19). Frequencies and percentages were used in the sociodemographic factors' descriptive statistics. The chi-square test or Fisher's exact test were used to examine the relationship between the independent variable and the KAP level (good or poor). The relationship between good KAP and demographic and socioeconomic characteristics was investigated using univariate logistic regression analysis. The statistical signi cance level was set to 0.05. On the other hand, Spearmen's rank correlation was employed to establish the extent of correlation between KAP scores.

Results
Social and demographic characteristics of study population A total of 463 neighbourhood residents agreed to participate out of 480 who were approached, giving a response rate of 96.6%. The sociodemographic characteristics of the residents are shown in Table 1. The mean age was 31 ± 9.467 years. Females made up the majority of the study population (88.1%), which could be explained by the fact that they were recruited at a community clinic within the vicinity of an apartment that caters mostly to pregnant women's followup.
Although both locals and migrants occupy the apartments, the majority are migrants as many of the apartment units are rented out to migrants. In terms of race, the majority were Bajau (37.6%), followed by Bugis (15.3%) and Suluk (13.3%). More than three-quarters of those surveyed had completed at least high school (74.1%). Only 28.1% of the total residents were employed, and 69.3% earned more than RM1,000 each month.
The ndings showed that 92.9 % of residents had no prior dengue infection. Among 463 residents, only 200 residents participated in the dengue seroprevalence study, out of which 74.1% (n = 63) and 7.1% (n = 6) tested seropositive for dengue IgG and IgM, respectively.
Knowledge about dengue prevention and control Table 2 highlights the residents' dengue knowledge, indicating a high degree of knowledge on dengue disease and transmission. Of the 86.8% of residents who answered the knowledge section, more than half correctly answered the statements (51.3% ±14.4). For example, more than half of the residents correctly identi ed fever (84.4%), chills (70.7%), and nausea or vomiting (56.0%) as signs and symptoms of dengue. Despite this, several residents misidenti ed coughing (19.3%) and swollen lymph nodes (13.9%) as signs and symptoms of dengue. Although only 30.5% of residents were aware that dengue is an infectious disease, the majority of the residents (94.7%) were aware that it is spread by mosquito bites, speci cally Aedes species (79.9%). Furthermore, the majority of residents were able to accurately answer questions about vector behaviour, namely vectors' active biting times (62.5%) and breeding preferences, such as discarded food containers (64.0%), open pools of water (94.8%), and exposed water reservoirs (69.9%). However, knowledge of the fact that there is no speci c treatment for dengue was lacking, with over half of the residents (60.8%) believing that there is. Despite this, the majority of residents were aware of the necessity to rest (57.6%) and drink abundant water (76.3%) for dengue recovery. Surprisingly, residents were aware of the potential of nes if mosquito larvae were discovered in residential property (87.5%).

Attitudes of the residents toward dengue prevention and control
The attitude of residents towards dengue prevention is presented in Table 3. On average, the attitude score was 66.0% ± 9.0 (n = 427). More than two third of residents (76.5%) agreed that dengue is an illness that can be prevented. Furthermore, 93.0% of them agreed that mosquito larvae must be disposed of to avoid dengue spread. Less than half of the residents (45.0%) believed that killing a dengue-carrying mosquito was the only approach to control or prevent infection with 52.7% of the residents believed that "fogging" could fully prevent mosquito breeding. Furthermore, 32.2% of residents said that at least once a year, frequent supervision to control and monitor probable mosquito breeding sites should be conducted, and 57.6% of the residents agreed that nes may help with dengue control. Although the majority of respondents (91.2%) agreed that routine inspections of residential premises should be conducted, a large number of residents responded that health professionals and volunteers were not the only ones responsible for reducing mosquito breeding grounds (75.3%).

Practices of the residents toward dengue prevention and control
In terms of dengue prevention practices by the residents, the results are summarised in Table 4. The study showed that the average preventive practice score was 63.7% ± 21.1 (n = 435).According to the ndings, more than half of the residents (69.9%) utilised lids for any water containers at home and closed them after each usage or checked their stored water in containers on a frequent basis (71.4%). Abate ® , a larvicide that prevents vector-borne disease including dengue, was used by 28.7% of the residents, who were either given it by health inspectors or bought it themselves. More than a third (42.4%) of those who used ower vases checked for mosquito larvae on a regular basis and dumped stagnant water in the vase (49.9%) to avoid mosquito breeding. Furthermore, majority of residents had a habit of inspecting their home and property for a container or location that could provide a breeding ground for mosquitoes (81.7%) or removing containers that could contribute to water accumulation and stagnation (93.8%). Lastly, the majority of residents (97.5%) agreed that eliminating Aedes mosquitoes is a shared responsibility and that they are prepared to work with local authorities to execute dengue prevention measures at their residence (97.7%).
Factors associated with good dengue KAP Next, association analysis was performed between different factors and good KAP as shown in

Discussion
This is the rst study to describe dengue KAP in a mixed population of locals and migrants in Sabah, Malaysia, that has consistently contributed to high dengue cases. To the best of our knowledge, no research has been done in Malaysia to investigate the link between dengue KAP and migrants. Overall, more than half of the residents were found to have adequate knowledge of dengue, an acceptable attitude toward dengue prevention, and a satisfactory practice of dengue prevention. Furthermore, this study reveals that knowledge and attitude are associated with nationality independently; however, among resident in the community dengue prevention practices are more related to gender and level of education. Both attitudes and practices are substantially correlated with community residents' knowledge of dengue.
KAP has been assessed in a variety of research studies in Malaysia, but only two studies have examined KAP relating to dengue in both hotspot and non-hotspot areas -nding that knowledge is in uenced differently by race, marital status, monthly income, employment (18) and educational level (18, 20). On average, the percentage score for KAP was higher than previous studies in hotspot areas in Malaysia (18) (51.3% versus 50.7%, 66.0% versus 53.2% and 63.7% versus 50.2%, respectively). This disparity could be explained by a variety of factors, including a person's lack of exposure to a dengue awareness campaign (18), or in the case of this study, reluctance to cooperate due to fact that most of the residents are migrants (70.4%): they might feel additional pressure, a loss of a sense of community, disinterest, or concern as some of them do not possess valid documents to stay in Sabah. It is worth noting that unreliable or improper data collection methodologies or questionnaire use could also contribute to the differences in ndings of a study (21). It has been demonstrated that measurement properties are population-speci c and vary based on the questionnaire's setting and context (22). A questionnaire designed for one demographic may not be appropriate to another in a different country or region with differing socio-cultural, healthcare, economic, and political circumstances, resulting in discrepancies in the ndings. Prior to the actual data collection, a pilot study and EFA were conducted, and the resulting KAP items were kept in the nal version of the survey questionnaires, indicating a reliable and valid questionnaire for the Sabah population (15).
As this study included 88.1% females, the majority of whom were pregnant (86.2%), it will provide insight into the level of knowledge on dengue among pregnant women in Sabah. Several reviews of case reports have found that dengue infection before and even during pregnancy can result in negative effects for the foetus due to vertical antibody transmission (23). This is worrying as 74.1% of the 85 pregnant women who participated in the study's seroprevalence testing had previously been unwittingly exposed to dengue. When these two factors are combined, pregnant women may inadvertently become infected with dengue, putting this susceptible group, as well as the foetus, at risk of complications during pregnancy as a result of unnoticed dengue. A previous study in Malaysia (24) (26) suggest that increased length of stay in a country and availability of dengue information in the local language (27) may lead to greater dengue knowledge among migrants. This highlights the urgent need for effective steps to eliminate obstacles to acquiring dengue information for migrants, such as increasing healthcare provider training regarding migrants and implementing health literacy education programmes focusing on migrants.
Community residents residing at these hotspot areas have some noteworthy aws in understanding as observed in the context of our study. For example, more than a third of residents had the misconception that dengue is an illness that cannot be prevented. This misconception must be dispelled since people will not engage in healthy behaviour unless they value the behaviour's outcome (28). Out of 444 residents, slightly more than half (50.2%) had the erroneous belief that chemical fogging by the health personnel is adequate to reduce dengue transmission, which is comparable to the results reported by Kamel et al. (2017) (29) and Selvarajoo, S et al (2020) (18) (47% and 49.6%, respectively). Paradoxically, fogging is always carried out after cases are reported in Malaysia, and studies have indicated that chemical fogging does not prevent the breeding of mosquitoes completely (12). According to the logistic analysis, residents with a higher level of education had a positive attitude toward dengue prevention, which is consistent with ndings from previous studies in Malaysia (30) and Colombia (31). This could be because people with a higher level of education are more involved in the community: they are more likely to lead meetings, educate other family members and/or neighbours, and more frequently believe they have the ability to take control of their actions, with their suggestions accepted and a sense of responsibility. This shows that a person's level of education may be connected to their ability to integrate and coordinate efforts to limit dengue transmission. However, education alone does not have a signi cant impact on attitudes. This could be due to societal differences, such as those found in this mixed-nationality community, which prohibit the populace from participating in the national strategy for dengue prevention and control.
The current study also reveals some intriguing ndings on community residents' dengue-related activities. Although the majority of residents agreed that eradicating Aedes mosquitoes is a shared responsibility and stated their willingness to work with local health authorities to conduct dengue-preventive initiatives at their homes, surprisingly, a third of these residents stated that health inspectors aren't required to inspect residential properties. This phenomenon may be explained by the fact that this residential area is primarily occupied by migrants who are frightened of being discovered working without a permit by government o cials. Another possible explanation is that these migrants usually originated from con ict-affected areas, had deprived sociodemographic and poor educational backgrounds, and strong beliefs in traditional health. These factors could act as a barrier to health communication, as low health literacy has been associated with adverse health outcomes across many domains and contexts (32). An interesting approach suggested by Gilder, M.E et al (2019) (33) dealing with migrant's worker communities on the Myanmar-Thailand border showed the importance of public health posters that need piloting prior to implementation due to the fact that unanticipated misperceptions are common among migrants. The study further emphasized that verbal communication remains a key method of sharing messages with individuals whose health literacy is low; strengthening the educational system and audiovisual messaging are critical approaches to improve health outcomes that might be applicable among the migrants.
Although community clean-ups have been conducted by Jawatankuasa Penduduk Perumahan (Housing Residential Committee), some apartments are rented out, posing a barrier to successful dengue prevention programme uptake by the new occupants as reported previously (34). Residents who retain water in containers owing to frequent water supply disruptions in these apartments should be encouraged to properly close the container or not leave standing water in containers for an extended length of time, as this can contribute to mosquito breeding sites. Furthermore, the Housing Residential Committee should be aware of vacant units in these apartments, which may result in the establishment of mosquito breeding sites (e.g. toilet bowels). On the other hand, ine cient garbage collection such as dustbin containers that are not functioning but left at the garbage area, infrequent garbage collection resulting in garbage remaining on the dustbin container for too long, dustbin containers with covers but left open after garbage collection resulting in the collection of water in solid waste and leading to an increase in dengue breeding sites (35). However, while enforcing monetary nes ("compounds") for failing to comply with breeding laws can help manage dengue, this technique may have certain downsides. The owners of these apartments, may be government o cials in some cases, preventing them from being sanctioned as reported previously (36).
One of the factors thought to have a role in dengue transmission is dense populations (11). The presence of IgG seropositivity among various community residents in this study could be explained by the link between rich mosquito breeding areas (38) and high population density with high mobility. Not surprisingly, people who live in high-rise residential structures are more likely to be IgG seropositive than those who live in single or terraced houses, (11) who also experience more frequent fogging associated with mosquito reproduction. The current study, however, was unable to evaluate the link between seroprevalence and dengue because the sample size of residents with IgM seropositivity was insu cient.
The mean KAP scores were shown to be fairly correlated using the self-reported questionnaires employed in this study. In the study area, when residents' knowledge increased, their attitudes and/or practices improved as well, which is similar to other previous studies (19,39). It is also possible that some people exhibit favourable behaviours initially, but subsequently lose interest when their efforts are not rewarded by their peers. Seeing results such as a drop in Aedes population or dengue cases may encourage them to keep doing what they're doing (40). The reluctance to have good preventive practices could also be related to the 'reactive approaches' they used or received, as they are only noti ed when dengue cases appear and then take appropriate action. Furthermore, despite different attempts and activities to provide information and guidance to the public, a lackadaisical attitude within the resident's community may have resulted in poor dengue preventive practises. Continuous motivation, such as telling residents when dengue-positive mosquitoes are identi ed and reinforcing their capacity to successfully carry out source reduction, has been reported to improve people's attitudes and actions (41). Recently, the Sabah state government implemented a community-centered strategy to empower and mobilise the community through the implementation of the Communication for Behavioral Impact (COMBI), in which volunteers selected from among the community would lead prevention activities such as gotong-royong and conduct inspections and eliminate mosquito breeding grounds inside and outside residential premises (42). It is relatively easy to modify knowledge and practice -but inherited and poor habits, on the other hand, are di cult to break with simple knowledge exchange, requiring multiple treatments, processes, and time. Health education programmes must take a more personal and practical approach to achieve this. Health o cials should perform house-to-house inspections not just for Aedes surveillance but also to convey information and educate residents. One strategy to achieve this is to make changes to curriculums and provide continuing professional development to healthcare workers in order to promote cultural competency and, as a result, help to improve local and migrant attitudes around dengue fever (43). Although there was no association between attitude and practice in the current investigation, a Vietnamese study found that boosting attitude could improve dengue prevention practice at the same time (44). One explanation is that the population in the present study is mixed, with people from various sociocultural backgrounds.
There are certain limitations to this study that should be addressed. Firstly, because the majority of the participating residents in this study are pregnant women, the ndings cannot be applied to the entire population of the hotspot area. The most important sample attribute that may or may not be representative of Sabah in this study is ethnicity. We acknowledge the fact that the study's ethnicity distribution is not representative of the multi-ethnic Sabah population, with each ethnic group having distinct cultures and beliefs that may impact their KAP.

Conclusion
This study revealed knowledge regarding dengue symptoms, transmission and vector control measures to be generally high among residents at a dengue hotspot: Taman Telipok Ria and Taman Sri Rugading, Telipok Sabah. A signi cant difference was also observed between local Sabahan and migrants in terms of their KAP. Despite indications of a link between knowledge and attitude or knowledge and practice, some knowledge about dengue and prevention does not lead to a positive attitude or practice. Conventional health education methods should be adjusted to stimulate social mobilisation among the admixed population. Bottom-up techniques combined with a multi-disciplinary approach should be used to in uence people's attitudes and behaviours, particularly among migrants, about their role in combating dengue fever and eliminating vector breeding grounds. Clearly, a collaborate efforts of government sectors, private sectors, housing residential committee and community residents is required.

Consent for publication
The authors give consent for publication Availability of data and materials All data generated or analysed during this study are included in this published article [and its supplementary information les].

Competing interests
The authors declare that they have no con ict of interest