Indian Journal of Public Health

: 2019  |  Volume : 63  |  Issue : 3  |  Page : 178--185

Community perception and risk reduction practices toward malaria and dengue: A mixed-method study in slums of Chetla, Kolkata

Debayan Podder1, Bobby Paul2, Aparajita Dasgupta3, Lina Bandyopadhyay4, Arkaprovo Pal1, Soumit Roy1,  
1 Junior Resident, Department of Preventive and Social Medicine, All India Institute of Hygiene and Public Health, Kolkata, West Bengal, India
2 Associate Professor and Head, Department of Preventive and Social Medicine, All India Institute of Hygiene and Public Health, Kolkata, West Bengal, India
3 Director Professor, Department of Preventive and Social Medicine, All India Institute of Hygiene and Public Health, Kolkata, West Bengal, India
4 Associate Professor, Department of Preventive and Social Medicine, All India Institute of Hygiene and Public Health, Kolkata, West Bengal, India

Correspondence Address:
Debayan Podder
Department of Preventive and Social Medicine, All India Institute of Hygiene and public Health, Kolkata, West Bengal


Background: Globally, vector-borne diseases account for 17% of infectious disease burden. In India, despite accelerated efforts, both malaria and dengue are major public health concern. Understanding local community perspectives is essential to strengthen ongoing program activities. Objectives: The study aimed to assess the community perceptions and risk reduction practices toward prevention and control of malaria and dengue at slums of Chetla in South Kolkata and to explore the perspectives of relevant local stakeholders in this regard. Methods: This cross-sectional study with mixed-method design was conducted from June to September 2018. Quantitative data were collected in 288 sampled households through face-to-face interview of respondents and environmental checklist, whereas, qualitative exploration was done with five key informant interviews. Quantitative data were analyzed using SPSS version 16 Software. Qualitative data were analyzed thematically using “cut and sort” processing technique. Results: In household survey, majority had unsatisfactory knowledge (68.4%), attitude (64.2%), and practices (62.8%) regarding diseases prevention/control, which was further substantiated by qualitative findings. The use of electric fans (79.5%), mosquito repellents (53.8%) were predominant practices; however, only a few used bed nets regularly (26.0%). Reportedly, “heat” and “poor ventilation” were key reasons behind reluctance to use bed nets. Multivariable regression analysis revealed that increasing age, proximity to a nearby canal, and unsatisfactory knowledge significantly predicted unfavorable attitude, whereas, respondents' gender and unfavorable attitude significantly predicted poor practices. Conclusions: Intensified social behavior change communication with active community participation is the need of the hour to prevent malaria and dengue occurrence and future outbreaks.

How to cite this article:
Podder D, Paul B, Dasgupta A, Bandyopadhyay L, Pal A, Roy S. Community perception and risk reduction practices toward malaria and dengue: A mixed-method study in slums of Chetla, Kolkata.Indian J Public Health 2019;63:178-185

How to cite this URL:
Podder D, Paul B, Dasgupta A, Bandyopadhyay L, Pal A, Roy S. Community perception and risk reduction practices toward malaria and dengue: A mixed-method study in slums of Chetla, Kolkata. Indian J Public Health [serial online] 2019 [cited 2020 Aug 5 ];63:178-185
Available from:

Full Text


Globally, vector-borne diseases (VBDs) account for 17% of infectious disease burden.[1] Among VBDs, malaria and dengue are of a public health concern due to their endemicity, outbreaks, and socioeconomic implications.[2] In 2017, nearly half of the world's population were at risk of malaria with an estimated 219 million cases,[3] whereas, Dengue, a neglected tropical disease is endemic in more than 100 countries with over two-fifths of world's population at risk.[4]

India, being the highest-burden country in South-East Asian region, showed a declining trend in malaria incidence in recent years[3] and is now poised to continue its stride toward achieving malaria elimination target by 2030.[5] However, dengue, endemic in all 29 states and 6 UTs in India,[6] witnessed a rapid increase in incidence over the past 50 years.[2] Despite ongoing efforts and surveillance activities, West Bengal reported recurrent outbreaks in recent past and had maximum reported cases among all states in 2016.[6]

Rapid unplanned urbanization with the growth of slums, poor sanitation, and other anthropogenic environmental changes has made communities more vulnerable. Furthermore, unavailability of effective vaccines, resistance to antimalarial drugs and insecticides, complexities of agent-host interaction, lack of community awareness has aggravated the situation.[7],[8]

It is well-known that community involvement and ownership are crucial for success and sustainability of VBD control program. Hence, understanding local community perspectives is essential to identify the gaps and design culturally appropriate, effective malaria and dengue prevention, and control strategies. The mixed-method approach is apt in this regard as it provides deeper insights into existing community perceptions and practices and thereby helps in identifying the attributes that influence the community in adopting healthy practices and responsive behaviors. With this background, the study aimed to assess the community perception and risk reduction practices toward prevention and control of malaria and dengue in slums of Chetla and to explore the perspectives of local stakeholders in this regard.

 Materials and Methods

Study design

This community-based cross-sectional study with mixed method design the following sequential explanatory approach was conducted from June to September 2018.

Study population

The household-level quantitative survey among residents (residing for >2 years) in selected slums was collected by interviewing any adult member (≥18 years) of sampled household, preferably the head. Qualitative data were collected from five-key informants (municipal field supervisor – K1; vector control in-charge – K2; secretary of a local club – K3; a medical officer – K4; and a public health nurse – K5). They were important stakeholders directly related to malaria/dengue prevention and control activities in the area.

Study area

It was conducted in densely populated slums of Chetla (22° 31′ N, 88° 20′ E) in urban field practice area of All India Institute of Hygiene and Public Health, Kolkata. It is situated in the southern part of Kolkata and falls in Borough IX of Kolkata Municipal Corporation - one of the high-risk boroughs considering urban high-risk criteria for mosquito-borne disease transmission, especially malaria.[9] The area had an estimated slum population of 36,190 (as per census 2011) with estimated households of 9601 as per available data with Urban Health Unit and Training Centre (UHU and TC), Chetla.

Sample size

A previous community-based cross-sectional study in Chetla, Kolkata, by Sahoo et al.,[10] it was reported that 97.9% of study participants had awareness that malaria is transmitted by mosquito followed by dengue (90.9%). The sample size was estimated based on the reported awareness on dengue (0.9) and considering confidence interval of 95% with Z1-α=Z0.95=1.96 and absolute error (L) of 5%. As multistage sampling was done, a design effect of two was used to calculate the sample size. Therefore, using the formula, (Z1-α)2 (P *Q/L2), where P = 0.9, Q = (1–0.9) = 0.1, estimated sample size was 288.

Sampling technique

The household was taken as one sampling unit. Multistage probability sampling method was used to select households. The entire slum population in Chetla was distributed across 15 major roads, out of which eight roads (primary enumeration areas) were chosen by simple random sampling (SRS) in the first stage. In the second stage, number of households that were selected from each primary enumeration area was determined using probability proportional to population size. Individual households in primary enumeration area were selected using SRS with replacement. If no one was available at a selected household for two consecutive days, next house was visited. Participants for qualitative data were purposively selected.

Data collection

In the 1st phase, quantitative data were collected by face-to-face interview of consenting participants, from June to August 2018, using predesigned pretested schedule and a checklist was used to assess the environmental characteristics at households. The content validity of the instrument was evaluated by public health experts. The schedule included items on sociodemographics, respondents' knowledge, attitude, and practices (KAPs) toward prevention and management of malaria and dengue. Geographical mapping of households with confirmed cases was done using mobile-based geographic information system application and meteorological data in the form of mean monthly temperature (maximum and minimum), mean monthly rainfalls were obtained.[11]

In the 2nd phase, qualitative exploration was done by key informant interviews (KIIs) with semi-structured interview guide. Interviews were conducted in selected facilities, and it lasted for around 30 min.


Institutional ethics committee permission was obtained prior to start of the study. Informed written consent was obtained before each interview, and all ethical principles were strictly adhered to throughout the process.

Data description and analysis

Each participant was assigned KAP score based on correct/appropriate responses. Maximum and minimum attainable score for knowledge (14 items/disease) was 28 and 0, respectively, for attitude and practice (10 items; based on 3-point Likert scale), it was 30 and 10, respectively. The scores were dichotomized into knowledge (≥19 - satisfactory/th percentile of attained score as cutoff.

Descriptive statistics was performed to assess unsatisfactory KAP. After assessing normality, nonparametric Spearman's correlation (rs) was performed between KAP score. Bivariate analysis evaluated the association between KAP level and confirmed cases of malaria/dengue. Logistic regression analysis identified predictors of unsatisfactory KAP. Significant (P ≤ 0.05) explanatory variables obtained from univariate analysis were entered into multivariable regression model to obtain adjusted odds. Statistical data were analyzed using the Statistical Package for the Social Sciences version 16 (SPSS Inc., released 2007, SPSS for Windows, version 16.0. SPSS Inc., Chicago, USA).

Qualitative data were first transcribed into Microsoft Word 2016. Sections in texts relevant to problems and objectives were identified, and data were coded thematically. Interview transcripts were read by all principal researchers to avoid bias. Thematic data analysis was done using cut and sort processing technique.[12],[13] QGIS 2.18.13 software (QGIS Development Team.2016. QGIS Geographic Information System. Open Source Geospatial Foundation Project. was used to illustrate the spatial distribution of confirmed cases of malaria/dengue in the surveyed households.

Operational definition

Overcrowding was assessed using household crowding index (HCI). It was defined as total number of co-residents per household, excluding the newborn/infant, divided by total number of rooms, excluding the kitchen, and bathrooms.[14],[15] HCI >2 was considered as overcrowding.

Malaria cases diagnosed by rapid diagnostic test kits and/or microscopy and dengue cases diagnosed by both Ns1 antigen and IgM-ELISA were considered as confirmed cases.


Background characteristics of surveyed households

Of 288 respondents, 56.6% were male and majority were Hindu (95.8%). Their mean age was 37.8 ± 13.6 years with range 55 (18–75); mean years of schooling were 6.1 ± 4.0 years with 31.6% of respondents having below primary education (below 5th standard). Most of the households (52.8%) belonged to Class IV (lower middle socioeconomic class) as per modified BG Prasad classification, January 2018)[16] with mean per capita income of INR 1895.2 ± 578.8. Only 29.2% of households were pucca, and rest were either kachha/semipucca. Overcrowding was present in 75.3% of surveyed households with average household size being 4.5 ± 1.4 persons.

Environmental characteristics were assessed using a structured checklist which revealed that 36.5% households situated in slums near Tolly Canal (also known as Adi Ganga, a small stream of Hooghly river) were more prone to waterlogging during tides. Although all households had access to improved water and sanitation facilities,[17],[18] majority (94%) were dependent on intermittent water supply through municipal standpipe and relied on shared sanitary latrine facility (91%) often leading to improper waste disposal and water storage. Indiscriminate waste disposal (62.5%) and the presence of potential vector breeding sites (68.3%) were observed in and around majority households. The absence of cross ventilation (86.40%) and inadequate natural lighting (88.20%) inside house, which might affect the indoor resting behavior of mosquitoes[19] was predominant across surveyed households.

Respondents' knowledge, attitudes, and practices toward malaria and dengue

All respondents were aware of the transmission agent of both malaria and dengue, but only 7.3% and 23.3% could identify the respective vectors. The most common symptoms reported for both were “high fever with nausea” (80.2%). Interestingly, none of the respondents were aware of the mosquito breeding sites inside the house. Knowledge score revealed that 197 (68.4%) respondents had inadequate knowledge with median knowledge score being 17 [Table 1]. Only, 117 respondents (40.6%) reported electronic media (television [TV]/radio), (36.2%) reported mobile campaigns and 23.2% reported print media (leaflets from health centers, posters, and hoardings) as primary sources of information on these diseases.{Table 1}

One hundred and eighty-five (64.2%) participants had unsatisfactory attitude. Majority (65.6%) perceived mosquitoes as a serious problem in the community, and 67.7% felt that they are at increased risk of getting malaria and dengue. Only 95 (34%) participants regarded waterlogging as a serious issue in context to mosquito breeding. Few (25.3%) were of view that responsibility of vector control activities lies solely on local municipal authority [Table 2].{Table 2}

Regarding risk-reduction practices, only 107 (37.2%) respondents had satisfactory practice scores. Inspection of water collection sites (at least weekly once) was reported by 43.8% of respondents whereas clearing them was reported by few (28.8%). Reportedly, the use of (biological/chemical) larvicide (32.3%) and door/window screens (7.0%) was seldom practiced. 53.5% reported not using bed nets regularly; however, the use of electric fans (79.5%) and mosquito repellents (53.8%) was predominant among surveyed households.

Spearman's correlation (rs) indicated significant positive correlation between knowledge-attitude (rs = 0.26, P < 0.001), knowledge-practice (rs = 0.18, P = 0.002), and attitude-practice (rs = 0.27, P < 0.001). Significant predictors of unsatisfactory knowledge in univariate analysis were respondents' age, sex, below primary education, occupation status, household type, overcrowding, social/public health advertisements, health worker's visit in the past 1 month, and history of malaria/dengue. Multivariable regression model deduced was of good fit (Hosmer–Lemeshow test: P > 0.05), and Nagelkerke R2 indicated that covariates in the model predicted 24.7% of the variance of unsatisfactory knowledge [Table 3].{Table 3}

Significant explanatory variables for unsatisfactory attitude among respondents' were age, household type, proximity to nearby canal, and unsatisfactory knowledge of respondents [Table 4].{Table 4}

Respondents' sex, education, household type, and unsatisfactory attitude significantly predicted poor practices among participants. After adjustment attenuated odds were obtained for all covariates in the model. However, no significant association was observed with socioeconomic status, overcrowding, health worker's visit to households, and viewing of IEC materials/advertisements [Table 4].

In the past 1 year (excluding 2 weeks before and during survey), 60 household respondents reported of having malaria (P. vivax/P. falciparum) and eight reported of dengue, whereas, during data collection, 75 (26.0%) households reported of having cases with acute fever, out of which, 46 were confirmed cases (44 malaria and two dengue cases). Only eight households with acute fever cases reportedly sought formal health care at a government facility. Spatial distribution of confirmed cases in surveyed households is shown in [Figure 1].{Figure 1}

Bivariate analysis (Pearson's Chi-square) revealed statistically significant association between confirmed cases and unsatisfactory knowledge (χ2 = 5.11, degrees of freedom [df] = 1, P = 0.024), attitude (χ2 = 4.68, df = 1, P = 0.03), and practice (χ2 = 7.25, df = 1, P = 0.007). This is suggestive of the fact that KAPs of a community might be related with disease outcomes especially in terms of malaria/dengue.

During the data collection period, the proportion of confirmed cases was highest in August (27.7%). However, it was observed that proportion of households with confirmed cases was lowest (8.3%) in July, when mean monthly temperature (maximum 32°C and minimum 26°C) was least and mean monthly rainfall (315.3 mm) was highest of all 3 months.

KIIs revealed two main themes – perceived barriers and strength toward disease prevention and control.

Perceived barriers were lack of community awareness, poor infrastructure in slums, proximity to Adi Ganga canal, unavailability of specific medicines for dengue, referral of dengue cases to higher centers and irregular logistics. Notable verbatim in this regard - K1 remarked, “…inspite of weekly house to house visits along with biweekly mobile awareness campaign most are not responsive in clearing stagnant water…when we try to clear they start arguing…”

K1, K2 were of consensus that despite continued efforts, community awareness, and responsiveness was low and it was “out of their habit” people engaged into unhealthy practices.

K3 mentioned, “…we are more concerned of dengue as it has affected many in our locality in recent years …and unlike Malaria there is no medicine to treat it…”

K3 felt that existing slum infrastructure is responsible for preponderance of malaria/dengue in the area, “…with fewer number of shared sanitary latrines, intermittent water supply and inadequate space for living, slums are difficult place to live in…also there is problem of water stagnation which is more in slums near Adi Ganga, and during rains situation worsens…”

On inquiring about bednets usage in the community, K3 mentioned “heat” and “poor ventilation” as key barriers, “…it feels hot sleeping inside most of us prefer to use fans over bednets…often we have to use coils and incense sticks…”

K4 mentioned, “…diagnosis of dengue is not done here…so when a probable case of dengue comes we have no option but to refer them to higher centres…often this leads to delay…”

Perceived sources of strength were biweekly mobile awareness campaign, active involvement of local political leaders in community mobilization, availability of Malaria Clinic at Urban Health Centre, municipal tie-up with a private hospital for dengue management, regular outreach sessions.

K1 mentioned, “…active involvement of our local councilor in awareness campaigns has helped us in community mobilizationand opened the door in accessing elite households and now to our surprise we are encountering more mosquito breeding sites in those places…”

Regarding service availability, K2 stated, “…Malaria drugs and test kits are available at health center throughout but we have tie-ups with three local nursing homes for Dengue ELISA test…all these are provided free of cost, supervisorsare there who coordinates with these nursing homes…”

K4 mentioned, “…separate malaria clinic runs at our health center providing both diagnostics and treatment facility…and often counseling is done to promote use of mosquito nets…”

K5 remarked, “…awareness campaigns are regularly conducted during outreach and field visits…leaflets and booklets are distributedto promote healthy practices towards prevention and control of mosquito-borne diseases…”


Key revelations were community perceptions, and risk-reduction practices were unsatisfactory among the participants. Respondents' education played an important role in determining ones' knowledge and practice. Notably, “health workers” visit to household in the past 1 month' and visibility of “IEC materials/advertisements” on malaria/dengue had a significant association with respondents' knowledge. Increasing age, proximity to nearby canal and knowledge had significance influence on participants' attitude, whereas respondents' gender, education, and attitude were significantly associated with respondents' risk-reduction practices. The study also revealed a significant association with confirmed cases and participants' knowledge, attitude, and practices.

Environmental characteristics in/around surveyed households were not dissimilar from any other slums. Waterlogging was a serious problem in slums situated near Adi Ganga. All households had access to improved water and sanitation facilities (as per the WHO criteria) and these were higher than the reported global average of 91% and 68%, respectively,[17] although majority (94%) were dependent on intermittent municipal water supply and used shared sanitary latrine facility (91%) leading to improper waste disposal and water storage.

About 75.3% surveyed households had overcrowding, which was higher than reported overcrowding in a previous study in the same setting by Maharana et al.[20] Majority (53.8%) households reported regular use of mosquito coils/repellant as a protective measure. This was in concordance with two previous studies in India,[20],[21] where 62.5% and 65% households were found using insecticide repellant which potentially contributed to indoor air pollution.

All respondents were aware of the transmission agent of both malaria and dengue, whereas, in a previous study, at the same setting, it was 97.9% and 90.9%, respectively.[10] This was indicative of increased awareness among the slum dwellers. Kohli et al.[22] in Delhi also demonstrated a significant association of knowledge on mosquito-borne diseases with education level of participants. About 58.6% urban participants responded TV as main source of information, whereas in our study, it was only (40.6%). In contrast to our findings, a study by Das et al.[23] in Odisha, reported that although net was perceived as an effective tool for protection from mosquito bites, its reported use was low and utilization was determined by seasonality, affordability, whereas, in our study, bed net acceptability and perceived effectiveness as low. The use of electric fans was most reported practice. The result was consistent with that of a similar study[24] which reported significant positive correlation among knowledge-attitude, knowledge-practice, and attitude-practice scores. However, the degree of correlation was weak in our study.

This mixed-method study dealt with at-risk slum population is strength in itself. Qualitative findings helped in a deeper understanding of community beliefs and practices about disease transmission. Due to cross-sectional design temporal association could not be established. Cumulative environmental data collected could not be associated. Potential sources of bias were recall and social desirability bias (especially in attitude domain). Efforts were made to minimize them using a short recall period, pretested survey instrument, and ensuring anonymity of participants.


Despite ongoing awareness campaigns on malaria/dengue prevention and control, community perception, and practices were suboptimal. Improving public health infrastructure in slums (construction of sanitary latrines and continuous water supply), household-level distribution of biological larvicide may help in preventing vector breeding. Targeted approach is required for mosquito net promotion and increasing utilization of public health services. The research highlighted the need for community capacity building through intensified social behavioral change communication. Future researches directed toward the development of community-based intervention model in similar settings, addressing local contextual factors, may help in reducing disease occurrence and outbreaks.


We are grateful to the officer-in-charge and staffs of UHU and TC, Chetla for their constant support and cooperation during this study. We express our heartfelt thanks to Dr. Kaushik Sarkar, public health expert, for his valuable opinion and Mr. Rajib Chakraborty for his support on GIS data presentation. Last of all, we extend our heartfelt gratitude to all the participants who have voluntary participated in this study.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1World Health Organization. Vector-Borne Diseases. World Health Organization; 2017. Available from: [Last accessed on 2019 Jan 02].
2World Health Organization. Dengue and Malaria Impacting Socioeconomic Growth: World Health Organization. World Health Organization; 2014. Available from: [Last accessed on 2019 Jan 02].
3World Health Organization. World Malaria Report 2018. World Health Organization. [Last accessed on 2019 Jan 04].
4World Health Organization. Neglected Tropical Diseases. Dengue. World Health Organization. Available from: [Last accessed on 2019 Jan 02].
5World Health Organization. National Framework for Malaria Elimination in India. 2016-2030. World Health Organization; 2016. Available from: [Last accessed on 2019 Jan 08].
6Disease Control Programme. Annual Report of Department of Health and Family Welfare 2016-17. Available from: [Last accessed on 2019 Jan 02].
7Ganeshkumar P, Murhekar MV, Poornima V, Saravanakumar V, Sukumaran K, Anandaselvasankar A, et al. Dengue infection in India: A systematic review and meta-analysis. PLoS Negl Trop Dis 2018;12:e0006618.
8Hati AK, Bhattacharjee I, Mukherjee H, Bandyopadhayay B, Bandyopadhyay D, De R, et al. Concurrent dengue and malaria in an area in Kolkata. Asian Pac J Trop Med 2012;5:315-7.
9Sen S, Mukhopadhyay B. Situation of malaria in Kolkata municipal corporation area: A Secondary data analysis report 2011. Indian J Community Med 2014;39:114-7.
10Sahoo SK, Taraphdar P, Mallick AK, Dasgupta A, Preeti PS, Biswas D. How aware are we regarding vector borne diseases? A community based study in a slum of Kolkata, India. Int J Res Med Sci 2017;5:2629-35.
11Customized Rainfall Information System. Available from: [Last accessed on 2018 Sep 21].
12Stewart M, Makwarimba E, Barnfather A, Letourneau N, Neufeld A. Researching reducing health disparities: Mixed-methods approaches. Soc Sci Med 2008;66:1406-17.
13Kohlbacher F. The use of qualitative content analysis in case study research. Forum Qual 2006;7:1-30.
14Melki IS, Beydoun HA, Khogali M, Tamim H, Yunis KA; National Collaborative Perinatal Neonatal Network (NCPNN). Household crowding index: A correlate of socioeconomic status and inter-pregnancy spacing in an urban setting. J Epidemiol Community Health 2004;58:476-80.
15Crowding Index. Available from: [Last accessed on 2018 Sep 21].
16Pandey VK, Aggarwal P, Kakkar R. Modified BG Prasad's socio-economic classification-2018: The need of an update in the present scenario. Indian J Community Health 2018;30:82-4.
17World Health Organization/United Nations International Children's Emergency Fund Joint Monitoring Programme; 2015. Available from: [Last accessed on 2018 Sep 21].
18World Health Organization/United Nations International Children's Emergency Fund Joint Monitoring Programme; 2017. Available from:[Last accessed on 2018 Sep 21].
19Paaijmans KP, Thomas MB. The influence of mosquito resting behaviour and associated microclimate for malaria risk. Malar J 2011;10:183.
20Maharana SP, Paul B, Garg S, Dasgupta A, Bandyopadhyay L. Exposure to indoor air pollution and its perceived impact on health of women and their children: A household survey in a slum of Kolkata, India. Indian J Public Health 2018;62:182-7.
21Nanjesh KS, Hedge R, Badiger S, Kiran KG. A study of mosquito borne diseases awareness, attitude and practices among the rural population in Karnataka, India. Int J Community Med Public Health 2017;4:4178-81.
22Kohli C, Kumar R, Meena GS, Singh MM, Ingle GK. A study on knowledge and preventive Practices about mosquito borne diseases in Delhi. MAMC J Med Sci 2015;1:16-9.
23Das A, Das Gupta RK, Friedman J, Pradhan MM, Mohapatra CC, Sandhibigraha D, et al. Community perceptions on malaria and care-seeking practices in endemic Indian settings: Policy implications for the malaria control programme. Malar J 2013;12:39.
24Dhimal M, Aryal KK, Dhimal ML, Gautam I, Singh SP, Bhusal CL, et al. Knowledge, attitude and practice regarding dengue fever among the healthy population of highland and lowland communities in central Nepal. PLoS One 2014;9:e102028.