Comprehensive evaluation of demographic, socio-economic and other associated risk factors affecting the occurrence of dengue incidence among Colombo and Kandy Districts of Sri Lanka: a cross-sectional study

Study design: selection of locations

Colombo District (6.70° to 6.98°N and 79.83° to 80.22°E), the commercial capital of Sri Lanka, accounted for approximately 18.42% (n = 34,274) of the dengue cases reported from Sri Lanka in 2017 and is the most high-risk area for dengue incidence in Sri Lanka [5]. Being located in the lowland of the country, Colombo is the most urbanized metropolitan area of the country, which hosts a highly variable multi-culture and multi-ethnic population of 2,309,809 within an area of 699 km² [15]. In contrast, Kandy District (6.93° to 7.50°N and 80.43° to 81.04°E), located in the central highlands (Fig. 1), extends over an area of 1940 km², covering a wide array of natural environmental features in contrast to Colombo. It is of major tourist interest due to its natural location, and places of historical and religious importance. At present, Kandy District is the third highest risk area for dengue transmission in the country, contributing to 7.73 % (n = 14,378) of the total dengue cases reported in 2017 in Sri Lanka [5].

In this context, the study focused on socio-economic risk characterization of highland and lowland populations of the country with notable variations in land use, demography, living standards and KAPs. The lowland areas included in this study were predominantly urban areas compared to the semi-urban and rural highland areas (except for the Kandy Municipal Council area) in the highland. From each district, five high-risk Medical Officer of Health (MOH) areas were considered for the socio-economic survey.

Selection of study population

An analytical cross-sectional survey was conducted from February to April 2017. The five most high-risk MOH areas that have reported the highest number of dengue cases during 2010–2016, were selected as the study areas in each district. A sample of 200 dengue impacted households were selected along with another 200 non-dengue reported households from a single MOH area (400 households from a single MOH area), on a systematic random basis following Krejcie & Morgan [16]. Cumulatively, 2000 dengue-impacted households and non-dengue households (with a ratio of 1:1) were used as the total sample from each district (400 × 5). During the calculation of sample size, it was assumed that the marginal error is 3.5% and population proportion is 0.5, while the actual population size of the whole Kandy and Colombo districts were taken as 1,369,899 and 2,309,809, respectively.

Households, from which the residents were not willing to cooperate in the study due to one or more reasons such as religious beliefs, absence of a household head or their opinion that it is not worthwhile participating in our survey, were not considered for the survey. On such occasions, the sample size was achieved by randomly selecting new households with consent to participate for the study.

Data collection

The demographic, socio-economic and living standard related factors, along with knowledge, attitudes and practices (KAPs) of the selected households were obtained by using an interviewer-administered questionnaire prepared in three local languages (Sinhala, English and Tamil). The selected household heads were interviewed by a group of trained interviewers. The household head was defined as the person who is perceived by the household members to be the primary decision maker in the family and the household was defined as individuals living together and taking meals from a common cooking facility [14]. In the absence of a household head, a responsible adult above 18 years, appointed by the family, was considered for the study.

The questionnaire covered the following areas: (i) demographic information (age and gender of the household head, number of family members, monthly income of the family and number of years residing within the relevant MOH); (ii) living standards (size of the homestead, vegetation coverage, temporary or permanent nature of the household, status of the household, accessibility conditions, surrounding land use types, number of rooms in the house, roofing, drinking water source and sanitary conditions); (iii) knowledge about dengue vectors, symptoms of dengue, vector ecology and transmission modes; (iv) attitude towards dengue and dengue management approaches; and (v) preventive practices against dengue (source separation of solid waste, waste disposal methods, maintaining a clean environment, home gardening practices and composting, covering of water storage tanks, cleaning and proper maintenance of roof gutters, use of mosquito repellents and insecticides etc.).

Data interpretation and statistical analysis

All collected data were double-checked and verified on the same day for completeness and consistency. The data were then entered into Microsoft Access® 2007 data sheets, while adhering to quality controlling procedures by trained personnel. The accuracy of data was routinely checked by cross tabulations and logical checks. Discrepant data were checked against original data forms and any mistakes were promptly corrected.

Chi-square test of independence was used for the statistical comparison of significance among study populations in terms of epidemiological, demographic, socio-economic and KAP factors [17, 18]. The proportions of demographic, epidemiological and socio-economic characteristics of the study populations were subjected to a cluster analysis (based on Euclidean distance) [19] followed by an analysis of similarities (ANOSIM) (i.e. a non-parametric analogue of MANOVA) [20], after square-root transformation. Furthermore, multi-dimensional scaling (MDS) analysis in Plymouth Routines in Multivariate Ecological Research version 6 (PRIMER 6) was utilized for the visual representation and comparison of the study samples in terms of overall socio-economic characterization [19].

Demographic and socio-economic factors

Living standards of the community

Knowledge, attitudes and practices on dengue

Practices related to dengue control

Overall characterization of study populations

Although study groups were associated with two distinct geographical and urbanization levels, the four study groups clustered into two major clusters at a Euclidean distance of 40 as dengue-impacted (test) and dengue non-impacted (control) groups based on the overall demographic, KAPs, socio-economic and living standard related characteristics (Fig. 2). The analysis of similarities (ANOSIM) further confirmed the formation of the above clusters by yielding a Global R value of 0.99 (P = 0.033). The multidimensional scaling (MDS) plot also resulted in the same observation (Fig. 3).

Demographic factors

The importance of demographic and socio-economic aspects in dengue control programmes, especially during disease epidemic episodes, has been progressively recognized by many countries [21–23]. In both dengue-positive groups of Colombo and Kandy districts, males belonging to the 15–35 age category who usually spend a notable time in public environments (school, tuition classes, work places and bus stands etc.) mainly for occupation or education purposes, have been recognized as the most susceptible category for dengue. Therefore, they are more exposed to the contact with Aedes vectors, which may be responsible for their high disease susceptibility [24].

Human movement, especially between urban/semi-urban and rural environments, has been recognized to significantly contribute to the increased transmission of dengue, confirming the above deduction [25, 26]. Findings of Udayanga et al. [14] and Nadeeka et al. [27] have also reported a similar trend regarding the susceptibility of younger age groups, while several studies conducted in other countries such as Brazil [28], Puerto Rico [29], Singapore [30] and Thailand [31], have reported a reverse trend. The presence of a high number of family members was a contributing factor for low transmission of dengue, especially within the dengue-free populations. According to Alobuia et al. [32], the existence of a high number of family members in a household has been recognized as imposing high responsibility on parents or guardians, thereby encouraging them to maintain a clean and safe environment to prevent their family members becoming infective.

Families with higher income levels were having lower rates of dengue incidence, as their economic strength and education level would enable them to take effective preventive actions against dengue [33]. However, it was interesting to note that a study conducted in urban areas of Thailand, reported a higher dengue incidence rate among people with secondary and higher degrees of education than with a lower level of education [31]. High dengue prevalence rates among residents with a relatively shorter residence time in a neighborhood was another special feature highlighted by the current study, which had already been reported from several other studies [31].

House conditions and infrastructural characteristics

Dominant characteristics of the living environment, especially the degree of urbanization, house density and surrounding land use practices, often influence the vulnerability to dengue outbreaks [14, 23, 27, 31]. As suggested by Alobuia et al. [32], occupants of moderate or smaller households have a relatively lower possibility of contacting dengue, since they tend to initiate necessary prevention and protective practices while maintaining a cleaner environment, when compared to the families living in relatively larger households. The number of dengue positive households with more than 1 floor (< 5 floors) were relatively higher than that of the control group in both districts of Colombo and Kandy. Hence, the apartment buildings could be a possible risk factor that support the transmission of dengue due to social negligence.

The relative coverage of grass and bushes (maintained as horticultural vegetation), were higher in households of dengue patients, which may provide ideal resting habitats for Ae. albopictus that remain as the secondary vector of dengue [34–36]. The high prevalence of built-up environment (urban environment) and marshy land has been found to be a critical risk factor associated with the incidence of dengue outbreaks in many countries including Malaysia [37, 38], Thailand [34] and Sri Lanka [14]. Such land use types may provide ideal breeding and resting grounds for Ae. aegypti, the primary vector for dengue with a high preference for urban settings. In particular, swamps and marshy land that may hold a shallow layer of standing freshwater provide ideal breeding grounds for Aedes mosquitoes, raising the risk associated with such surrounding land uses [34, 39]. Furthermore, majority of abandoned lands in two study areas were noted to be misused by the community, especially for improper disposal of solid waste, resulting in increased dengue vector populations [33].

Concrete, asbestos and metal sheets were the most preferred methods of roofing in all the studied communities. Such roofing materials, especially concrete and blocked roof gutters, have been identified as potential risk factors associated with dengue, especially during the rainy season by many studies conducted throughout the world [40–43]. However, the majority of the dengue recorded households claimed to remove roof gutters after an incidence of dengue, indicating the effectiveness of awareness programmes conducted within the high risk areas by different entities.

Almost all participants of the study heavily depended upon pipe-borne water, which is often associated with water storage tanks. It was noted that the relative percentage of partially covered or uncovered water storage facilities in both dengue patient groups was higher than the control groups; these facilities may provide stable breeding grounds for Aedes vectors increasing the possibility of dengue outbreaks [44, 45]. Furthermore, water storage and retention time period were relatively higher in dry periods in the country, favoring storage of excess water to cater to the day-to-day requirements, due to the limited supply of water. Such dry conditions may also be an indirect driving factor for the breeding of vectors as suggested by a recent study conducted in Australia [46]. Regarding toilet facilities, the majority of dengue patient households had separated toilets, mostly in outdoor settings. Such toilets may often contain water storage tanks or basins (especially in semi-urban or rural localities and in public toilets), which also could act as ideal breeding grounds for dengue vectors [45].

Knowledge, attitudes and practices on dengue

High levels of knowledge on transmission, symptoms, patient care and prevention of vector breeding, have found to lower the risk of dengue incidence among communities from all over the world [13, 14, 47]. All the participants were aware of the fact that dengue is transmitted by the bites of infected mosquitoes and a large majority was even capable of naming the primary and secondary vectors of dengue. A study from Laos has also reported a similar situation whereby about 93% of the participants knew the name of the specific vector of dengue [48]. Furthermore, a notable fraction had the knowledge on the presence of four serotypes of dengue, even though some (especially the dengue patients from Colombo) were not aware of the fact that a single person is vulnerable to dengue more than once. The moderate to high literacy level in Colombo and Kandy (approximately 96.3% for Sri Lanka in 2015), and the continuous awareness programmes may be the contributing factors for maintaining higher levels of knowledge on dengue transmission [14]. A study conducted in Jamaica [47] also reported similar findings, while an opposite trend has been reported in Pakistan [48] and Nepal [13] whereby the respondents were only aware of the transmission by mosquitoes without any in-depth knowledge of the transmission.

Regarding the symptoms of dengue, all participants were capable of identifying lasting fever as the typical symptom along with headache (> 70%). However, a relatively higher percentage of dengue patients were capable of stating joint pains, rash, muscular pain and nausea/vomiting as other possible symptoms. A lower proportion identified other symptoms such as retro-orbital pain, abdominal pain and itching sensation. In general, a considerable amount of respondents were able to correctly identify typical symptoms of dengue, unlike several other studies conducted in India [49], Thailand [50], Laos [48], Nepal [13] and Jamaica [47]. However, dengue patient groups had a relatively higher level of knowledge on symptoms of dengue, than the participants of the control groups with no personal experiences of dengue in their households [13, 47]. The relatively acceptable level of knowledge on dengue symptoms may also be the outcome of awareness activities conducted within the relevant study areas. Such notable levels of knowledge on the symptoms may minimize the chance of patients confusing them with other general causes of fever such as influenza, typhoid, etc. and thereby enabling them to receive the required patient care from the health sector of Sri Lanka [47].

Aedes mosquitoes are known day-biters who prefer to engage in blood-feeding mainly during several hours after dawn and before dusk [13]. Unlike a study conducted in Jamaica, where only about 3% of respondents were aware of this biting behavior of dengue vectors [47], a notable fraction of Sri Lankan respondents was aware of the day-biting preference. However, the awareness level among dengue patient groups were low (especially in Colombo with only 48.9%), that may have caused their high vulnerability to dengue, since adequate preventive measures are not followed by them during the daytime. The situation may be more serious since they have not yet gathered the correct and important information about the disease even after the infection. In addition, the majority of the respondents were familiar with the basic facts of vector breeding such as the preference of vectors to breed in clean standing water and potential breeding sites in the household and premises that enable them to considerably reduce vector breeding at their premises [13]. Surprisingly, a majority of the control groups were familiar with the basic morphology of dengue vectors (presence of white spots on the legs and black-colored body), while only a lower fraction among the test groups were familiar with the above morphological features. A similar moderate level of knowledge on vector morphology has been reported from another study conducted in Saudi Arabia [51]. However, a previous study conducted in Kandy has reported a relatively lower level of awareness on the vector morphology [14]. Moderate to high levels of awareness on dengue was found among dengue patient and non-dengue patient groups in the present study. The efficacy of intensive awareness programmes conducted by different parties (government, NGO and other community based organizations) in improving the knowledge of community on different aspects of dengue could be the reason for the current knowledge level.

Most of participants in all four study groups knew that they were residing within high dengue risk areas. However, the percentage of dengue patients with the above understanding was relatively lower in both districts. More than three quarters of all four study groups desired to further improve their knowledge on numerous aspects of dengue, such as general transmission, symptoms and patient care of dengue, along with the control of vector breeding and vector contact. It is interesting to note that a considerable fraction is also willing to focus on solid waste management (SWM), which remains a key factor in dengue epidemic incidence. Many studies have highlighted the role played by solid waste management practices in governing dengue outbreaks [33, 52]. Therefore, the current desire of the community to further increase their knowledge is a positive indicator symbolizing that the general public is ready to take part in the process of dengue control through the reduction of vector breeding habitats.

The satisfaction of the community on the adequacy of steps taken by the government stakeholders in managing dengue was very poor (except for the control group of Colombo with 45.5% satisfaction). The limitations in the role played by field-based staff, who are responsible for the coordination and initiation of vector management activities at the ground level, may be a potential factor for such dissatisfaction [14]. Furthermore, the poor coordination between government entities and other stakeholders (including NGOs, the private sector and the community), outdated vector management strategies and local political conflicts, could also be listed as potential reasons for the low efficacy of government vector controlling activities.

Even though the government was named as the responsible party for the management of dengue, a considerable portion (approximately one quarter) understands that the community itself has a role to play in managing the dengue outbreaks. The highly positive attitudes on community-based vector management among study groups (especially in the control groups), bear evidence for the fact that local communities have also ascertained the potential risk of dengue and are ready to join hands with VCE to fight against it. The recent severe outbreaks of dengue might be the motivation factor for current attitudinal change of the people, which made them realize about their responsibilities in vector management. However, the current finding might also be partially influenced by the respondents trying to appear responsible in front of a stranger by providing socially desirable responses, without expressing their true self. Similar difficulties have also been reported in several other studies [13, 47].

Only a limited number of participants were practicing composting or home gardening, in all study populations, with a greater number contributing from the control groups. Even though government and private sector based agricultural entities have introduced numerous cultivation and composting techniques along with crops requiring less space and water, the restrictions in time and space were mentioned as the main reasons against not practicing composting or gardening. The present findings also agree with a previous study conducted in Kandy and the limited interest raises a question on whether the community is actually moving towards environmentally friendly lifestyles and vector management or not [14]. Collection of waste by the Municipality or Urban Council remained as the major waste disposal method followed by open burning and disposal into a garbage pit. As emphasized by Gubler & Clark [52], properly planned urbanization and waste disposal services are key infrastructural features that minimize the incidence of dengue outbreaks. Unfortunately, the irregular nature of the collection date, rejection of certain items, use of no alarming sound to inform the residents that they are visiting the area and not reaching certain roads or houses, were found as major weaknesses of the waste collection service, (especially among dengue impacted households), that may have clearly contributed to the elevated risk of dengue within the study areas [33, 52, 53]. Therefore, implementation of a proper functioning system for waste collection is recommended, which is more user-friendly and caters for the requirements of the community to ensure proper waste management in the study areas, thereby assisting the management of dengue vector breeding.

Regarding practices, the use of mosquito coils and creation of smoke were more common preventive measures practiced against mosquito–biting, followed by the use of nets and closing/covering of windows. Another study conducted in Jamaica [47], has reported that approximately 80% of study participants were not using any effective preventive methods such as mosquito screening and bednets, due to higher costs of implementation. Conversely, Sri Lankan communities were rationally utilizing available resources to minimize and prevent bites of vector mosquitoes at their households, by incorporating some of the traditional methods. Use of coconut husks, cashew shells and dry leaves of plants to create smoke to avoid mosquito bites were highly practiced in many households (especially in Kandy), due to their low economic costs and high efficacy with limited side-effects. Other studies conducted in Pakistan [54] and Mexico [55] have also found the same tendency of using mosquito coils and screens as methods of mosquito bite prevention and have reported their successful contributions in minimizing the severity of dengue outbreaks.

Covering water containers/water storage tanks and eliminating potential breeding sites of vector mosquitoes were recognized as the most common practices of the community. A number of studies have documented such preventive practices to be practiced by different communities all over the world [13, 14, 47, 48, 54]. However, proper disposal of solid waste, organizing “Shramadana” (clean-up programmes) to clean the surrounding and clearing bushes and other vegetation were practiced more by the healthy populations. According to studies conducted in Pakistan [54] and Thailand [50], implementation of clean-up programmes often is highly effective in controlling dengue transmission, especially if organized either prior to or at the beginning of the rainy season. Such practices often ensure the unity and social responsibility of the community residing at the ground level, which may be further improved and converted into community-based vector management strategies with the guidance and support from other government and private sector stakeholders. Several studies have highlighted the importance of driving the common public towards elimination of vector breeding sites at the household level via raising awareness as a successful solution for managing the transmission of dengue [13, 47, 50]. Therefore, regardless of the moderate to high knowledge levels of the community on various aspects of dengue, the preventive practices and attitudes requires further improvement to ensure a dengue-free environment, as highlighted by several similar studies [47, 56].

Based on the above, it is ostensible that the people residing in dengue-free households have a relatively higher degree of knowledge and more helpful attitudes toward dengue along with more preventive practices to ensure minimum levels of dengue vector breeding and human-vector contact. Comparatively, the patient groups of both districts had limitations in their awareness, social status, attitudes on dengue and practices against dengue. In particular, the knowledge on dengue symptoms, patient care, vector biology and behavior should be further improved. Therefore, the relevant VCE should design their awareness programmes to cater to the above requirements of the community and effective knowledge transmission methods should be followed to address the limitations in knowledge on different aspects of dengue. Improving knowledge on symptoms and treatment methods of dengue within the community will drastically reduce their reliance on traditional remedies and self-medication, while driving them towards immediate hospitalization. This would be immensely helpful for the VCE and other health staff in Sri Lanka, not only for the management of patients but also to minimize patient-based transmission of the virus to other vectors and thereby to humans. Regarding attitudes, the public should be made aware that the responsibility of managing dengue epidemics at the local level should be equally borne by the government based VCE, private stake holders and also by the general public to ensure efficacy of any intervention actions.

Furthermore, VCE and other government staff should work alongside the community to win the trust of the general public regarding the adequacy of their services provided in managing dengue, which was found to be poor. Routine inspections of the households for dengue vector breeding sites, organizing cleaning and awareness programmes (particularly before onset of rainy season) and facilitating local vector management activities organized by the community or other organizations are key steps to be followed by the VCE, to motivate the public towards community-based management of dengue [57]. In addition, local administrative bodies such as Municipal Councils and “Pradeshiya Sabha” (village councils), should provide key infrastructure facilities such as properly planned urbanization and waste disposal services, taking necessary steps to avoid weaknesses such as the irregular nature of the collection date, rejection of certain items, use of no alarming sound to inform the residents that they are visiting the area and not reaching certain roads or houses during waste collection. This would also motivate the public in practicing source separation of solid waste and proper disposal of solid waste, minimizing the potential of vector breeding. In addition to the chemical-based controlling of vector breeding, the VCE should encourage the public to move towards traditional methods of mosquito-human contact reduction such as use of coconut husks, cashew shells and dry leaves of plants to create smoke at dawn and just before dusk.

The findings of the study highlight the high susceptibility of males belonging to the 15–35 age category, who spend much of the day outside. The VCE should therefore identify that the risk of human-vector contact may arise from public environments such as schools, tuition classes, working places and bus stands etc., which are poorly managed in terms of vector breeding reduction. As solutions, the VCE could establish strong connections with the other government administrative entities (such as Municipal Councils and regional educational offices), the private sector and also with religious leaders of the locality to coordinate vector controlling activities and ensure vector-free environments in public places.

The recent outbreaks of dengue that have occurred since the beginning of 2017, simulated different parties such as Government VCE, NGOs and other private stakeholders to focus more on dengue while communicating the risk to the common public. This could be the reason for the elevated levels of knowledge, attitudes and preventive practices of the common public on different aspects of dengue. However, a number of studies including the WHO and CDC have recommended community-based vector management as the potential solution for the management of dengue, rather than relying upon conventional methods of chemical based control [11–13, 33, 47, 57]. Therefore, the VCE of Sri Lanka should focus more on bridging the gap in knowledge and attitudes on dengue among the general public, motivating them to work with other stakeholders to ensure personal and community-wise protection from dengue. As such, it is essential to design community educational campaigns to educate residents on different aspects related to dengue, while emphasizing the responsibility of the community in vector management, to ensure community-based controlling of dengue within the country [13, 14, 33, 57]