Risk perception and health-related adaptive behaviour to climate change among high school students in Thiruvananthapuram district, Kerala, India: a cross-sectional study Authors: Arya Rajeev, Aswathy Sreedevi Published: 2025 ABSTRACT Abstract AbstractObjectivesTo determine the risk perception, health-related adaptive behaviours and associated factors related to climate change among high school students in Thiruvananthapuram district, Kerala, India.Study design and study settingsA cross-sectional study with multistage cluster sampling was conducted among high school students from Neyyatinkara Taluk in the Thiruvananthapuram district of Kerala, India. After identifying the taluk, 10 schools were chosen using probability proportionate to size to ensure adequate representation.Study participantsThe study was conducted among 600 high school students (mean age 14 years, SD 0.75) from Neyyatinkara Taluk in the Thiruvananthapuram district of Kerala.MethodsNeyyattinkara taluk was randomly selected from the six taluks in Thiruvananthapuram district. From each of the 10 selected schools, students from classes 8 to 10, section A, formed the study clusters, with cluster sizes ranging from 45 to 60 students. All students from classes 8 to 10 (section A) who were present on the day of the survey and had obtained informed consent from their parents or guardian were considered eligible to participate in the study. Risk perception and health-related adaptive behaviour scores for children were calculated using a pretested structured questionnaire with 8 and 17 questions, respectively. All questions were designed on a 5-point scale. For positively worded questions, scores ranged from 5 to 1 (strongly agree to strongly disagree), and for negatively worded questions, the scoring was reversed. Binary logistic regression analysis was used to determine the independent factors associated with risk perception and health-related adaptive behaviour.ResultsNearly three in four study participants (72.1%) were aware of the term climate change. The median risk perception score and health-related adaptive behaviour scores were 28 (IQR 26–30) and 52 (IQR 47–57), respectively. Study participants from urban areas had significantly better risk perception compared with rural counterparts (AOR 2.42; 95% CI 1.54 to 3.78). Similarly, children from above poverty line (APL) households demonstrated markedly higher risk perception than those from below poverty line households (AOR 28.77; 95% CI 16.84 to 45). Participation in a climate change awareness programme was also associated with higher risk perception (AOR 1.98; 95% CI 1.23 to 3.19). Positive health-related adaptive behaviour was more likely among children aged 14–16 years compared with those younger than 14 (AOR 1.92; 95% CI 1.3 to 2.84). Urban residence (AOR 20.72; 95% CI 5.04 to 85.17), higher paternal education (AOR 1.91; 95% CI 1.15 to 3.13) and APL household status (AOR 2.50; 95% CI 1.57 to 3.93) were also significantly associated with better adaptive behaviour.ConclusionsClimate change interventions and awareness programmes should prioritise rural, lower socioeconomic and younger populations and equip them with practical life skills for adaptive behaviour.Keywords: Climate Change, Health, PUBLIC HEALTH STRENGTHS AND LIMITATIONS OF THIS STUDY Assessment of climate change perception and health-related adaptation in a large, representative sample of high school students in Kerala, India. Assessment of climate change perception and health-related adaptation in a large, representative sample of high school students in Kerala, India. The study’s reliance on self-reported data introduces the possibility of bias. The study’s reliance on self-reported data introduces the possibility of bias. The studies were from different types of schools, such as government, aided and unaided. However, our study has not given adequate representation to each type of school where the students may be from different socioeconomic strata. Thus, the findings are not generalisable to each type of school. The studies were from different types of schools, such as government, aided and unaided. However, our study has not given adequate representation to each type of school where the students may be from different socioeconomic strata. Thus, the findings are not generalisable to each type of school. Limitations of cross-sectional studies also apply to this in that temporality cannot be assessed. Limitations of cross-sectional studies also apply to this in that temporality cannot be assessed. Introduction Over the past decade, approximately 83% of natural disasters have been attributed to events related to weather and climate.1 The increasing frequency and intensity of extreme weather events due to climate change pose significant threats to human health. A comprehensive approach involving emission reductions, adaptation measures, resilient healthcare systems and policies to protect vulnerable communities is necessary to address the health impact. Raising awareness, promoting sustainability and prioritising resilience-building efforts are essential for mitigating the threats posed by extreme weather events in a country like India. Climate change has intensified erratic and extreme rainfall in Kerala, India.2 Children are especially vulnerable due to their rapid development, immature adaptive capacity, greater exposure relative to body weight and prolonged lifespan risks.3 4 Evidence shows heightened morbidity among children during extreme weather events.5 6 Mitigating these impacts requires fostering climate awareness and building adaptive capacities in children.7 Risk perception on climate change refers to how individuals perceive and understand the potential risks and impacts associated with climate change. Children’s perception of climate change risks may differ from adults, evoking emotions like fear and anxiety, along with a sense of urgency and responsibility to protect the planet and their future. Adaptive behaviour towards climate change is any action that moderates the risk of climate change impacts. Health-related adaptive behaviours specifically address behaviours focused on maintaining and improving health, as well as preventing diseases, in response to the challenges posed by climate change.8 However, evidence pertaining to the awareness levels of children and their adoption of health-related adaptive behaviours in the context of climate change is limited. Our study endeavoured to evaluate the perception of risk and health-related adaptive behaviours among school children in response to climate change. Furthermore, we aimed to identify the factors associated with risk perception and health-related adaptive behaviours. Objectives Primary To determine the risk perception of climate change among the high school students in Thiruvananthapuram district. Secondary To determine the adaptive behaviour pertaining to climate change among the high school students in Thiruvananthapuram district and to determine the factors associated with risk perceptions and health-related adaptive behaviour among the high school students in Trivandrum district. Methods Study design A cross-sectional study was conducted among high school students from Neyyatinkara Taluk in the Thiruvananthapuram district of Kerala. Study setting The study was conducted in Government, Aided and Unaided schools of Thiruvananthapuram District, Kerala. The participants included students from classes 8 to 10. All students who were present on the day of the survey and had obtained informed consent from their parents or guardians were considered eligible to participate. Sample size estimation Using an anticipated 26.1% prevalence of risk perception of climate change in a previous study,8 with a 95% confidence level, 5% absolute precision and a design effect of 2 to account for cluster sampling design, the sample size was estimated to be 592. Sampling technique Thiruvananthapuram district in Kerala has 215 schools and is administratively divided into six taluks. We randomly selected Neyyattinkara taluk, which includes 46 government, aided and unaided schools, using the lottery method. A taluq is a subdivision of a district which consists of a group of several villages organised for revenue purposes.9 A multistage cluster sampling method with population proportionate to size (PPS) was used to select schools. Ten schools were thus chosen. From each selected school, students from classes 8 to 10, section A, formed the study clusters, with cluster sizes ranging from 45 to 60 students. The first available 60 students were included, approximately 20 each from the three classes, except in three schools where class 10 students were not permitted to participate, reducing the cluster size to 45. Necessary permissions and class lists were obtained from each school principal. Government schools are fully funded by the government. Aided schools are privately managed but government-funded and unaided schools are privately managed without government funding. The schools as a whole are representative of adolescents in this particular age group as the school enumeration is 100%. Thus, the inclusion criteria were students of class 8, 9 and 10 who had provided parental consent and assent. There were no exclusion criteria. Data were collected at a single point and the time period of study was between March and July 2023. Data collection After obtaining permission from the school authorities, the schools were visited in sequential order. Prior to data collection, informed consent forms were sent to parents through the class teacher, outlining the study’s purpose, procedures and voluntary nature. All the parents provided consent. On the day of data collection, assent was obtained from the available students of classes 8, 9 and 10, section A. The questionnaire had 8 and 17 items. It was pretested by administering it to 20 students in the same classes from another section before the start of the study, following which a few questions related to health-related adaptive behaviour were modified for greater clarity. The questionnaires were self-administered. However, the investigator, along with the class teacher, explained every question to ensure clarity and understanding. The completed questionnaires were collected from the students after a period of 45 min. A total of 603 students participated in the survey; however, 3 responses were found to be incomplete. Therefore, data from 600 students were included in the final analysis. Tools The study employed a pretested structured questionnaire,8 10 adapted from two previous studies. The tool included multiple sections including: Sociodemographic factors (age, sex, household type, socioeconomic status, place of residence and parental education), Knowledge of climate change: awareness (programme attendance, understanding of importance, causes, impacts and possible solutions), Sources of information, risk perception (participants’ awareness of risks posed by climate change, perceived barriers to mitigation, perceived benefits of behavioural changes and knowledge of harmful consequences). Health-related adaptive behaviour included practices to cope with adverse effects of climate change such as transportation methods, clothing choices in extreme climates, personal hygiene, and related behaviours. Health hazards included air pollution, water pollution, flood, waste management and extreme weather events. The questionnaire included 8 items assessing risk perception and 17 items evaluating adaptive behaviour (online supplemental file 1). All questions were designed using a 5-point scale. For the positively worded questions, responses of 1–5 (strongly agree, agree, neither agree nor disagree, disagree and strongly disagree), respectively, were given. For the negatively worded questions, scores were assigned in reverse. The total scores were 40 for risk perception and 85 for health-related adaptive behaviours, respectively. A lower score indicates an elevated understanding of risks and the practice of beneficial health-related adaptive behaviour in relation to climate change. Statistical analysis Data were entered into a Microsoft Excel sheet (V.1808) and analysed using statistical package for the social sciences, SPSS (V.21). About 10 of the entries each day were double-checked by a colleague to minimise errors. Data were anonymised. The categorical variables such as gender, socioeconomic status, type of household, source of information, awareness programme and parental education of both father and mother were summarised using frequency and percentages. The continuous variables, such as age, risk perception and health-related adaptive behaviour, were summarised using median and IQR as they did not follow normality. Based on the median score for risk perception and health-related adaptive behaviour, it was categorised as high (below median) and low (above median), respectively. The sociodemographic factors and risk perception were analysed using the χ² test, and an unadjusted OR with 95% CI was calculated. Binary logistic regression analysis was done to identify independent determinants including variables with p<0.2, and an adjusted OR (AOR) with 95% CI was calculated. A p<0.05 was considered statistically significant. Results The age of the participants ranged from 12 to 16 years, with a mean age of 14 years, SD 0.75. Among the 600 participants, 413 (68.9%) were between 12 and 14 years of age, and more than half of the participants, 329 (54.8%), were girls (table 1). The vast majority, 521 (86.8%) of the study participants resided in urban areas. A higher proportion of the mothers (401, 66.8%) than fathers (341, 56.8%) had university-level education. More than three-fourths of participants (455, 75.9%) belonged to the above poverty line (APL) category. The majority of the study participants (362, 60.3%) were from private (unaided and aided) schools. Table 1Sociodemographic characteristics of study participants (N=600)Variablen%Age (in years) 12–1341368.9 14–1618731.1Gender Male27145.1 Female32954.9Residence Urban52186.8 Rural7913.2Level of education-father Up to higher secondary25943.1 University and above34156.9Level of education-mother Up to higher secondary19933.1 University and above40166.9Socioeconomic status APL45575.9 BPL14524.1Type of households Nuclear40166.9 Joint19933.1Type of schools Government23839.7 Others36260.3Aware of climate change Yes43372.1 No16727.9Attended an awareness programme Yes33956.5 No26143.5Importance of climate change Important38764.5 Not important21335.5Source of information Media40166.9 Friends19933.1APL, above poverty line; BPL, below poverty line. Sociodemographic characteristics of study participants (N=600) APL, above poverty line; BPL, below poverty line. Close to three-fourths of the participants (433 72.1%) were aware/had heard of the term climate change. Furthermore, 339 (56.5%) of the participants reported that they had attended an awareness programme on climate change organised in schools, facilitated by teachers and excise officers (table 1). Among all the study participants, 387 (64.5%) reported that climate change is an important issue. The majority of participants had heard of climate change (401, 66.9%) from media sources such as audio-visual and print media, and the remaining 199 participants (33.1%) reported the source of information as their friends. The median scores for risk perception and health-related adaptive behaviour in response to climate change were 28 (26, 30) and 52 (47, 57), respectively. We used binary logistic regression analysis to find the independent factors of risk perception and health-related adaptive behaviour of climate change among the study participants (tables2 3). Independent factors of risk perception of climate change (table 2) were participants from urban areas. They were found to have 2.42 (95% CI 1.54 to 3.78) times higher risk perception of climate change compared with rural residents. The risk perception of participants belonging to the APL category was 28.77 (95% CI 16.84 to 45.82) times higher than those belonging to the below poverty line (BPL) category. Participants who reported attending an awareness programme were found to have 1.98 (95% CI 1.23 to 3.19) times higher risk perception compared with their counterparts. The model also included variables such as age and parental (mother’s and father’s) education. Table 2Factors associated with risk perception on climate change: results of binary logistic regression analysis (N=600)Variables (n)Risk perceptionAdjusted OR*P valueHighLow(95% CI)N (%)N (%)Age (in years) 14–16158 (84.5)29 (15.5)1.07 (0.64 to 1.82)0.762 <14245 (59.3)168 (40.7)Ref–Place of residence Urban410 (78.0)111 (22.0)2.42 (1.54 to 3.78)<0.001 Rural24 (27.8)55 (72.2)Ref–Parental education-father University259 (76.1)82 (23.9)1.57 (0.9 to 2.75)0.112 Till higher secondary171 (66.3)88 (33.7)Ref–Parental education-mother University294 (73.3)107 (26.7)1.37 (0.94 to 1.99)0.1 Till higher secondary135 (67.5)64 (32.5)Ref–Socioeconomic status APL397 (87.3)58 (12.7)28.77 (16.84 to 45.82)<0.001 BPL35 (24.2)110 (75.8)Ref–Attended an awareness programme Yes266 (78.5)73 (21.5)1.98 (1.23 to 3.19)0.01 No163 (62.5)98 (37.5)Ref–*Variables that had a p<0.2 in the bivariate analysis were included in the model.APL, above poverty line; BPL, below poverty line. Factors associated with risk perception on climate change: results of binary logistic regression analysis (N=600) Variables that had a p<0.2 in the bivariate analysis were included in the model. APL, above poverty line; BPL, below poverty line. Table 3Factors associated with health-related adaptive behaviour on climate change: results of binary logistic regression analysis (N=600)Variables (n)Health-related adaptive behaviourAdjusted OR*P valueHighLow(95% CI)N (%)N (%)Age (in years) 14–16118 (63.1)69 (36.9)1.92 (1.30 to 2.84)0.001 <14217 (52.5)196 (47.5)Ref–Place of residence Urban430 (82.5)91 (17.5)20.72 (5.04 to 85.17)<0.001 Rural21 (26.5)58 (73.5)Ref–Parental education-father University264 (77.4)77 (22.6)1.91 (1.15 to 3.13)0.011 Till higher secondary177 (68.3)82 (31.7)Ref–Parental education-mother University352 (87.8)49 (12.2)1.05 (0.71 to 1.55)0.8 Till higher secondary95 (47.7)104 (52.3)Ref–Socioeconomic status APL358 (78.6)97 (21.4)2.5 (1.57 to 3.93)<0.001 BPL89 (61.4)56 (38.6)Ref–Attended an awareness programme No199 (76.3)62 (23.7)1.12 (0.70 to 1.80)0.69 Yes247 (72.8)92 (27.2)Ref–*Variables that had a p<0.2 in the bivariate analysis were included in the model.APL, above poverty line; BPL, below poverty line. Factors associated with health-related adaptive behaviour on climate change: results of binary logistic regression analysis (N=600) Variables that had a p<0.2 in the bivariate analysis were included in the model. APL, above poverty line; BPL, below poverty line. Binary logistic regression analysis was performed to identify independent predictors of health-related adaptive behaviour towards climate change (table 3). Participants aged 14–16 years had significantly higher adaptive behaviour, being two times more likely to adopt health-related adaptive practices compared with those aged 12–13 years (AOR 1.92; 95% CI 1.30 to 2.84). Parental education was also a significant predictor: participants whose fathers had a university-level education or higher were 1.91 times more likely to demonstrate adaptive behaviour than those whose fathers had completed up to higher secondary education (AOR 1.91; 95% CI 1.15 to 3.13). In addition, socioeconomic status and urban residence played a role, with participants from the APL category showing 2.50 times higher adaptive behaviour compared with those from the BPL category (AOR 2.50; 95% CI 1.57 to 3.93) and participants from urban areas showing higher adaptive behaviour compared to rural (AOR 20.72; 95% CI 5.04 to 85.17). The model also included variables such as maternal education and participation in awareness programmes, but these were not found to be statistically significant. Discussion About three-fourths of the participants (433, 72.1%) were aware of the term climate change. Higher risk perception of climate change was found among participants from urban areas, children from APL category households and those who reported attending an awareness programme. Correspondingly, the findings highlight that older age, urban residence, higher paternal education and better socioeconomic status are key determinants of adaptive behaviour towards climate change. Urban residents exhibited higher risk perception of climate change and greater health-related adaptive behaviour compared with rural residents. This may be attributed to higher education levels and better access to diverse information sources, including the internet, news outlets, social media and healthcare providers.11 12 Concepts of climate adaptation from literature were also less in rural and aboriginal populations.13 Limited access to such resources in rural areas contributes to lower awareness and adaptive practices. Supporting evidence from other settings indicates that access to information, socioeconomic factors and belief in climate change strongly influence adaptive behaviour, as seen among farmers in the Mekong Delta of Vietnam14 and urban populations in Pune, India.15 Risk perception and health-related adaptive behaviour was higher among participants from the APL category than those from the BPL category, likely due to greater access to information, proactive health-seeking behaviour, economic stability, education, awareness and social support. These factors may contribute to a better understanding of climate change-related risks among APL individuals. Evidence from the developing countries supports this pattern, showing that higher-income groups exhibit greater health-seeking behaviour and invest more resources in managing health risks compared with economically disadvantaged groups.16 A study in Punjab province, Pakistan, demonstrated that at the individual and household levels, adequate awareness contributed significantly to mitigation and perceptions of risks associated with climatic extremes.17 Reinforcing these findings in India, the adaptive measures taken by farmers were primarily influenced by socioeconomic variables such as the farmer’s age, gender, household size, education level, income and farm size.18 In this study, participants who reported attending an awareness programme were found to have higher risk perceptions compared with their counterparts. In contrast, a study in Germany found that a better understanding of the effects of climate change was associated with lower perceived levels of hazard or risk.19 Participants whose fathers had a university-level education or higher demonstrated greater adaptive behaviour, likely reflecting higher self-efficacy and a stronger tendency to respond proactively to climate change.20 Similarly, participants aged 14–16 years showed higher adaptive behaviour than those aged 12–13 years, which may be influenced by differential access to information. Evidence from China indicates that risk perception accounts for a significant proportion (26.1%) of the variance in health-related adaptive behaviour,7 a pattern supported by studies in multiple countries.21 22 However, a study in North American cities found that even with abundant access to adaptive information, only half of respondents had modified their behaviour.23 Another study from the USA further highlighted a value–action gap: while personal experience and perceived harm increase concern, most Americans rarely discuss climate change, limiting broader societal engagement and action.24 The causes and consequences of climate change remain inequitably distributed.25 It disproportionately affects the Global South,26 where nearly 85% of children live.27 Although the current generation of adults has largely caused this crisis, today’s children and young people will bear the consequences.28 Youth play a critical role in combating climate change. A study from a western state of India reported that 54.5% of respondents believed young people could make a major contribution in addressing this challenge.29 Youth are quick learners and adaptable to changing circumstances. The 2003 tsunami and the 2018 Kerala floods vividly demonstrated their capacity to mobilise communities and lead action. During the Kerala floods, volunteers from the National Service Scheme and National Cadet Corps were at the forefront, managing shelters, supporting rescue operations, distributing relief materials and coordinating community responses.30 Youth-led social media initiatives further amplified these efforts by raising funds and locating missing persons. Collectively, these experiences underscore the vital role of youth in promoting awareness, education, innovation and community resilience in the face of climate-related crises. Children affected by floods showed a significant positive correlation between resilience and coping, drawing strength mainly from camp activities and family support.31 Another study from Kerala found that adolescent volunteering fostered a stronger sense of community togetherness during crisis and created greater opportunities for life-skill development. However, the reasons for volunteering varied significantly according to gender, marital status, socioeconomic status and health perception.32 Therefore, sensitising adolescents and youth and enabling them to take practical roles during extreme weather events could be truly transformative. Limitations The limitations of this study include its reliance on self-reported data, which may be subject to recall bias as participants might not accurately remember or report past behaviours and experiences. The findings may not be generalisable across government, aided and unaided schools, as the study did not provide adequate representation of each school type, where students may come from diverse socioeconomic backgrounds. Furthermore, the cross-sectional design limits the ability to establish causal relationships, as data were collected at a single point in time, preventing assessment of temporal sequences between variables. Climate change interventions and awareness programmes should prioritise rural, lower socioeconomic and younger populations and equip them with practical life skills for adaptive behaviour. Furthermore, region-specific studies among adolescents and young people would help capture variations and support the development of contextually relevant climate and health education programmes not only in India but across the world. Supplementary material Acknowledgements The investigators thankfully acknowledge all the school principals, teachers and students who participated in the study. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Prepublication history and additional supplemental material for this paper are available online. To view these files, please visit the journal online (https://doi.org/10.1136/bmjopen-2025-103996). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Consent obtained from parent(s)/guardian(s). Ethics approval: The data collection process commenced after obtaining approval from the review board and the ethical committee of Amrita School of Medicine (ECASM-AIMS-2023-106), dated 24 February 2023. Before the study, written informed consent was obtained from parents, and assent was obtained from the students. Students were free to opt out of the survey at any point in time. Patient and public involvement: Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details. Data availability statement Data are available on reasonable request. References Supplementary Materials Data Availability Statement Data are available on reasonable request.