Theory-based flood disaster education enhances disaster preparedness among adolescents in Selangor, Malaysia: a cluster randomized controlled study Authors: Khairunnisa Makmon, Ahmad Zaid Fattah Azman, Rosliza Abdul Manaf, Mohd Rafee Baharudin Published: 2026 ABSTRACT AbstractFlood disasters are prevalent and severely impact humans and the environment, with adolescents being particularly vulnerable. Preparedness can mitigate this impact from disaster. This study evaluated a tailored disaster educational module known as ‘Get-Ready DEM’ based on the Extended Parallel Pathway Model (EPPM), to enhance flood disaster preparedness among adolescents in Selangor, Malaysia. A two-arm, single-blinded, cluster-randomised controlled trial involved 287 adolescents from eight secondary schools, divided into intervention (140 participants) and control (147 participants) groups. The intervention group received the ‘Get-Ready DEM’ once after the baseline assessment, while the control group was on the waitlist. Preparedness was measured using the Flood Disaster Preparedness Index (FDPI) scores at baseline, immediate post-intervention, and 3 months post-intervention. Data were analysed using a generalised estimation equation. Results showed a significant increase in FDPI scores for the intervention group immediately post-intervention (β = 2.17; 95% CI = 1.66, 2.66; p value < 0.001) and at the 3-month follow-up (β = 1.58; 95% CI = 0.99, 2.16; p value < 0.001) compared to the control group. These findings underscore the EPPM framework’s efficacy in designing effective flood disaster education. Implementing modules like ‘Get-Ready DEM’ can significantly enhance adolescents’ resilience and preparedness, suggesting a promising approach for disaster preparedness programs. Trial registration: This study was registered retrospectively under Thai Clinical Trial Registry TCTR20240426001 on 26th April 2024 and obtained ethical approval by the Ethics Committee of Human Research University Putra Malaysia (JKEUPM-2022-858) and the Ministry of Education, Malaysia (KPM.600-3/2/3-eras (15189). This study received funding from the University of Putra Malaysia through the university grant ‘Putra Berfokus’ (6302015-14001).Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-56110-7.Keywords: Extended parallel pathway model (EPPM), Flood disaster preparedness, Disaster preparedness, Disaster education, AdolescentSubject terms: Environmental social sciences, Natural hazards Preparedness is crucial for mitigating the severe impacts of floods, among the most destructive natural disasters globally. According to the Centre for Research on the Epidemiology of Disasters (CRED), in 2021, floods accounted for 52% of global disasters and 41% of disaster-related deaths in 20201. Despite being outside the Pacific Ring of Fire, Malaysia is highly susceptible to floods, pollution, and landslides, with floods representing 74% of natural disasters from 1998 to August 20182. From July 2012 to January 2019, Malaysia reported the highest number of individuals affected by floods in the ASEAN region3. Increasing frequency of flood disaster and it’s severity are worsening with climate change4. The late 2021 and early 2022 floods caused extensive damage across various sectors, with losses estimated at RM6.1 billion, or approximately 0.40% of Malaysia’s nominal GDP5. Aftermath of the flood disaster may have profound impact to adolescent either direct impacts such as injuries, displacement and loss of material or indirect impacts such as mental health issues, growth and development and educational disruption6–8. This may affect learning continuity, and contribute to poverty and affect their future opportunities6,9,10. Moreover, based on UNICEF Climate Change Risk Index (CCRI) report in 2021 which assess impact of climate change on children, ranked Malaysia at 61st countries where least performing in preparing with the effect of climate change for children11. There were 9 out of 10 youth experienced climate and environmental related effect. Therefore, this underscores the urgency of addressing flood risks, especially to adolescents12. Adolescents, particularly vulnerable due to their developmental stage, suffer disproportionately from floods, impacting their physical, mental, socioeconomic, and educational well-being13. Adolescents, defined by the World Health Organization as individuals aged 10–1914 and comprising nearly 11% of Malaysia’s population, are at a critical developmental juncture15. This phase presents both opportunities for capacity development and significant vulnerabilities16,17. Flood disaster preparedness in this contexts of this study encompasses of proactive measures taken such preparing emergency bags, identifying temporary shelter place, stockpiling food and water and designing evacuation plans either by individuals or families facing any potential flooding to minimize the impact of floods which occurs before, during and post flood disaster. Adolescents have significant role in flood disaster preparedness as they can enhanced their family preparedness and play active role as they are equip with knowledge and proactive measures11,18. Their involvement is crucial in promoting effective behaviours in disaster risk reduction, as outlined in the Sendai Framework for Disaster Risk Reduction 2015–203019. It aligns with the global sustainable development agenda, particularly the SDG 2030 education target 4.7, aiming to equip youth with skills to promote sustainable development and resilience against climate challenges, including Target 13.320. This flood disaster preparedness can be achieved through education, training, community engagement and risk communication. Educating this demographic about flood risks and appropriate responses is essential for their safety and fostering community resilience21,22. Disaster education can be acted as a tool for cultivating flood disaster preparedness among adolescents23,24. Both formal and informal disaster education are vital from an early age25. Advocates highlight the importance of initiating education and awareness efforts in childhood to lay a foundation for a culture of preparedness and resilience26,27. Although studies have shown that educating adolescents about disasters prepares them to handle such calamities, a systematic review of school-based disaster education programs in 27 countries revealed discrepancies between increased awareness and preparedness among adolescents28. It indicates a need for innovative strategies to enhance the effectiveness of school-based disaster education programs22,28. Furthermore, age28–30, gender29,31,32, personal disaster experience31,33,34, perceived risk or threat30,35,36, perceived efficacy37,38, disaster knowledge22,39, and socioeconomic status40 influence adolescent flood disaster preparedness. In Malaysia, research on disaster education among adolescents has been promising, yet there is a shortage of disaster education modules specifically targeting this group. This research crafted a flood disaster preparedness educational module for adolescents based on the Extended Parallel Process Model (EPPM), a prominent framework in health communication41,42. EPPM effectively combines cognitive (efficacy) and emotional (threat) elements to influence behavioural responses, aiming to shift adolescents’ perceptions of flood disasters and bolster their management skills. This is achieved by enhancing their awareness of risks and equipping them with the necessary tools to take proactive measures. The EPPM framework is especially pertinent for adolescents, who are at a crucial stage of developing independent decision-making and risk assessment skills43. By integrating the EPPM into educational strategies, the module educates about flood risks while simultaneously building the confidence required for taking preventive actions. This is vital as it meets the adolescents’ developmental needs for autonomy and competence, promoting a proactive rather than a reactive approach to disaster preparedness. This provides adolescents with the opportunity to make more informed and safer decisions while facing flood disaster. Furthermore, their increased awareness and preparedness may positively influence their family members, helping to mitigate the risks and severity of disasters, which often strike without warning. EPPM-based flood disaster educational module incorporates structured, skills-based components that offer realistic flood scenarios, outline potential impacts, and provide detailed steps for readiness, such as preparing emergency kits or planning evacuations. By ensuring that educational messages about flood risks are coupled with clear, achievable preparedness steps, the module does more than raise awareness, it actively fosters a resilient, proactive mindset in adolescents. EPPM is particularly suited for this purpose due to its proven effectiveness in previous studies on preparedness for various natural disasters, pandemics, and outbreaks44–47. Therefore, this study aims to assess the effect of designated theory based, flood disaster education module on flood disaster preparedness among adolescents in Selangor, Malaysia. Materials and methods Research objectives This study aims to develop, implement, and assess the effects of a flood disaster education module based on the extended parallel pathway model on flood disaster preparedness among secondary school students in Selangor, Malaysia. Study design Two-armed, parallel, single-blinded, cluster-randomised control trial. Secondary schools designated as temporary shelters constitute this study’s cluster. The clusters were divided into intervention and control groups. Both groups received follow-up during the trial. The intervention group received a ‘Get-Ready DEM’ flood disaster education program. The control group received no intervention and was put on the waiting list if the intervention group demonstrated a positive effect. Location of the study The research was carried out in Selangor, Malaysia. Selangor is one of fourteen Malaysian states on the west coast of the Peninsular. Selangor has 6.99 million people, the most populous state in Malaysia (21.6%), with approximately 674 people per square kilometres and 95.8% urban population48. The study was conducted in Selangor’s four flood-prone districts: Klang, Hulu Langat, Gombak, and Sepang49,50. Study population and respondent selection Randomization occurred at the cluster level, utilizing schools as the primary units. From a pool of 18 eligible secondary schools pre-selected based on study criteria and listed by the Selangor State Education Department, only ten were chosen through a random selection process. Of these, merely eight schools provided consent and were subsequently enrolled in the study. The selected schools were randomly allocated into the control and intervention groups using a 1:1 ratio. The allocation concealment was done using the subsequentially numbered, opaque, sealed envelope (SNOSE) technique. The eight schools were labelled by randomly assigning them to names (A, B, C, D, E, F, G and H). Then, it was concealed in eight identical black opaque envelopes size 2-inch by 2-inch, containing different colour-coded cards by the other third party. The white and red cards were labelled as control and intervention groups, respectively. Envelopes were opened only after it was time to allocate the intervention by research assistant. It should be noted that the allocation codes were masked from the investigator during this step. The intervention allocation was masked to the school and participants only. The sampling population consisted of adolescents between 13 and 15 years old who attended the selected secondary school during the study period. Study participants were recruited upon approval of the school principal. The recruitment of the participants was based on the selected class from Form 1 (13 years old) to Form 3 (15 years old) based on the selection by the random application of the computer generator. Power and sample size The sample size (N) for this study was calculated based on a comparison of mean differences between two groups using Lemeshow and Lwanga’s formula51 and cluster design52. The calculation based on prior study53 (µ1 = 65.0, µ2 = 50.10) with a mean difference (Δ) of 14.9 in disaster preparedness scores between intervention and control groups, assuming a pooled standard deviation (σ1) of 20.25, a two-sided significance level of α = 0.05, and a statistical power of 80%, fixed number of clusters (k = 4 per arm). Under individual randomization, the required sample per arm was estimated using the formula for comparing two means: Given randomisation occurred at the class (cluster) level with a fixed 4 clusters per arm, inflated this sample using the design effect (DE) = 1 + (m − 1)ρ, where m is the average cluster size and ρ is the intra-cluster correlation coefficient (ICC). Assuming an ICC of 0.01—supported by estimates from school-based cluster trials of adolescent health behaviours and small size of cluster54 which obtained design effects between 1.24 and 1.34 for expected class sizes of 25–35 students. After adjusting for this design effect and allowing 20% attrition, the required total sample size increased to approximately 90–98 participants (45–49 per arm). Hence, this study recruited a total of 287participants (140 in the intervention group and 147 in the control group), substantially exceeding the minimum requirement and thereby ensuring adequate power and robustness for detecting the pre-specified effect size. The flow of school and participant recruitment is summarised in Fig. S1 (Supplementary). Study instrument Flood disaster education module for intervention group The flood disaster education module, known as Get-Ready Disaster Education Module or with the acronym of ‘Get-Ready DEM’, was meticulously developed based on the Extended Parallel Process Model (EPPM), effectively integrating perceived threat and perceived efficacy-based information to create impactful content that able to influence the behaviour of individuals. The ‘Get-Ready DEM’ modules aim to serve as a platform for providing knowledge, awareness, and life skills to help individuals cope with disasters effectively and reduce their severity. The module comprised three comprehensive sessions was designed and developed following the objectives and learning outcomes of each specific session as showed in Table S1 (supplementary section). This module took approximately about 3 to 4 h as summarized below:-. Session 1 : general information on flooding This session aims to provide essential insights into the causes, effects, and dynamics of flooding. The content is designed to enhance participants’ understanding of how natural events can lead to severe flooding and the factors influencing flood risk and susceptibility. Additionally, it aims to increase awareness of local flood hazards and identify individual risk factors, particularly concerning adolescents. The session incorporated elements of perceived threat as a key component of the Extended Parallel Process Model (EPPM). It emphasized the severity and frequency of flood events, presenting detailed information on possible outcomes to enhance participants’ perceived susceptibility and severity. Consequently, this may improve participants’ awareness of their personal risk by presenting realistic flood scenarios, fostering a crucial perception of vulnerability and severity that motivates protective actions. Session 2 : flood disaster preparedness Through this session, participants learned risk identification, emergency kit preparation, early warning system recognition, and local communication understanding. This session utilizes best practices from disaster management authorities to propose strategies for safety before, during, and after a flood. The information presented was grounded in literature reviews from various sources55–62. The session’s information incorporated perceived effectiveness as a critical element of the Extended Parallel Process Model (EPPM). Focus was placed on response efficacy by informing participants about effective protective measures and preparedness strategies. The session included interactive activities, such as scavenger hunts and evacuation planning, to enhance self-efficacy. Therefore, participants may view themselves as competent in following safety protocols. Additionally, the activities, along with feedback and support, aim to bolster self-efficacy. Session 3: flood disaster risk communication and the emergency crisis This session underscores the significance of effective communication, knowledge sharing, and the use of credible sources for disaster-related information. The session’s material is derived from authoritative local organizations, particularly the National Agency of Disaster Management (NADMA) in Malaysia, concentrating on disaster risk communication. The aim is to augment capabilities for disaster risk reduction and resilience, empowering participants to serve as information conduits for accurate community information. The session’s content is integrated with elements of perceived effectiveness, deemed essential within the framework of the Extended Parallel Process Model (EPPM). The intervention was implemented once following the initial baseline assessments of the participants. It was delivered by the similar proficient research assistant that act as trainer that has been trained to conduct the module. To facilitate active engagement and practical learning, various pedagogical tools were utilized. These tools included audio-visual aids such as use of short videos, PowerPoint lecture, interactive group activities, and educational games, all designed to enhance participants’ disaster awareness, knowledge, and preparedness skills. Furthermore, to aid in the retention of critical information, participants were provided with handouts in the form of stickers, along with models of emergency bags, first aid kits, and whistles. These stickers listed the essential items for an emergency bag and included vital emergency contact numbers, serving as a handy and accessible reference that reinforced the learning outcomes and preparedness for flood disasters. Additionally, each school received a model emergency bag and a banner detailing key flood disaster preparedness tips, which could be displayed in classrooms to promote and reinforce the message of flood disaster preparedness. The development of the ‘Get-Ready DEM’ modules was guided by a dynamic instructional design framework, ensuring a systematic and structured approach to creating educational content for the flood disaster63. The modules were crafted in Malay and English to accommodate the linguistic preferences of the participants, thereby enhancing accessibility and understanding. A comprehensive validation process was undertaken to ensure the relevance and appropriateness of the module content for the local context and the intended audience. It involved a panel of four expert committees consisting of two public health specialists, a disaster and risk management expert, and an educator. Prior to the full-scale implementation, a pilot study was conducted to assess the content validity and face validity of the modules. This preliminary phase was crucial in refining the modules based on expert feedback and ensuring their suitability for effectively conveying essential disaster preparedness education. Self-reported questionnaires This study used 24 items of the self-reported questionnaire that were mostly adapted from the pre-existing instrument and translated into the Malay language via forward and backward methods. A copy of the English version of the questionnaire is provided in the supplementary section. Flood disaster preparedness index score (14 items) In this study, flood disaster preparedness is referred to as the knowledge and capacities of individuals, communities, institutions, or governments to efficiently anticipate, effectively respond to, and recover from disasters. The instrument used to measure flood disaster preparedness encompasses 14 items adopted from the Disaster Preparedness Index score, demonstrating a high Cronbach alpha reliability of 0.91 in a prior study in Chennai involving 445 participants64. It evaluates the adoption of essential precautions or measures necessary to safeguard life and property during and after a major natural disaster. For this study, the items were adapted to reflect the local context, and the study population showed satisfactory internal consistency of Cronbach’s alpha 0.7, providing insights into the participants’ perspectives on flood disaster preparedness practices at both individual and family levels65. The items in the questionnaire cover three main factors: indoor safety measures, document preparation, and collective efficacy and escape plans. For instance, items A1, A2, and A3 focus on evacuation plans, while items A4 to A6 and A9 address collective efficacy. Document preparation is assessed through items A7, A8, and A14, and indoor safety measures are captured by items A10 to A13. Adaptations included modifications like replacing the item from ‘Acquired a fire extinguisher and learned how to operate’ to ‘Stored clean water supply for every household’ and from ‘Acquire an all-hazards alert radio’ to ‘Download applications or devices that help get related disaster information’ to suit local needs. A copy of the questionnaire is enclosed in the supplementary. The response format is binary (Yes/No), with ‘Yes’ responses scored as one point and ‘No’ or ‘Not Sure’ as zero, summing up to a total score of flood disaster preparedness range between 0 and 14. Score-based measurement preferred to measure flood disaster preparedness due to its sensitivity and quantitative precision. This method reveals subtle variations in preparedness levels that categorical classifications might miss. It allows for detailed statistical analyses, including regression and trend analyses, to examine variable correlations and intervention impacts. Scoring reduces the loss of information associated with categorical classifications, effectively capturing gradual improvements in preparedness, and increasing the efficiency of disaster preparedness evaluations. Detailed participant’s background questionnaires (10 items) This section asked about the participant’s sociodemographic background (age, gender, parent’s educational background and family monthly income), disaster experience and severity, and exposure to disaster education. This question was requested once during the baseline or preliminary stage. Data collection method The data collection point was taken face-to-face at three-time points for both intervention and control groups. Prior to the intervention, baseline data (T0) were collected. Second, a similar questionnaire was used for the immediate post-intervention evaluation (T1) after the disaster education intervention, except the sociodemographic section was skipped. Post-intervention assessment (T2) was done three months following the disaster education intervention using the same self-reported questionnaire as T1. It is estimated that each survey took approximately 15–20 min to complete. Data analysis Data were analysed with SPSS 27.0. Data normality distribution were assessed using visual and statistical methods. Continuous variable median, interquartile range, and mean rank were reported. Independent T-test parametric tests was used to determine any significant difference in baseline scores for the flood disaster preparedness score in the control and intervention groups. For categorical data (gender, parent’s education, and monthly income), frequency and percentages were applied. The Chi-square test determined the significant difference in baseline categorical variable distribution between the intervention and control groups. Accordingly, alpha level for statistical significance (p-values) equal to or below 0.05 were deemed to reflect statistically significant results, which is the same two-sided alpha (0.05) used in the sample size and power calculations. This threshold, conventionally employed in empirical research, implies a 5% risk of committing a Type I error—erroneously rejecting the null hypothesis. Adopting this standard minimises the probability that the observed effects are due to random variation. The Generalized Estimation Equation (GEE) is preferred for flood disaster preparedness analysis due to its effectiveness in managing longitudinal or clustered data. It effectively addresses within-group correlations and repeated measures, providing improved estimation of population-level effects66. Besides that, the significant covariates that may influenced the outcome can be controlled by the GEE. Its adaptability for diverse outcome variables and resilience against correlation structure misspecification render GEE optimal for thorough and statistically robust preparedness score evaluation. Moreover, a sensitivity analysis using a generalized linear mixed model (GLMM) was conducted to account for clustering at the school level and to validate the robustness of the primary findings. The model included fixed effects for group, time, and covariates, with a random intercept for clusters to capture between-school variability. Ethical consideration Approval was obtained by the Ethics Committee of Human Research University Putra Malaysia (JKEUPM-2022-858) and the Ministry of Education, Malaysia (KPM.600-3/2/3-eras (15189). The study retrospectively was registered under the Thailand Clinical Trial Registry: TCTR 2024040426001). All methods were performed in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki and the institutional and Ministry of Education requirements for research involving minors. No intervention-related harms, adverse events, or unintended consequences were identified or reported among participants during the study. This cluster randomised controlled trial was designed, conducted, and reported in accordance with the CONSORT 2025 statement. Participants in the study were selected based on voluntary participation, and no personal identifiable information was gathered, and confidentiality of all data was strictly maintained. Participants and their parents or guardians provided informed consent before the study. Participants have the option to withdraw from the study at any point without consequence. Inclusion criteria for schools and participants The selected schools were the public government secondary schools designated as temporary shelters in the districts of Klang, Gombak, Hulu Langat, or Sepang. These schools are coeducational and enrols students aged 13 to 15 years. Participants needed to be Malaysian citizens attending one of the selected schools, willing to participate in the study, and have parental or guardian consent. Additionally, participants were required to be proficient in either Malay or English. Exclusion criteria for schools and participants Any school and the participant who wished to withdraw from the study at any point in time. Expected benefits of the study This study is expected to significantly enhance the participants’ knowledge of flood disaster preparedness through a targeted educational module, thereby contributing to the scientific foundation in public health and augmenting the existing literature. By integrating into national disaster preparedness programs, the study aims to bolster the capabilities of key agencies such as the National Disaster Management Agency of Malaysia (NADMA) and the Ministry of Health Malaysia (MOH) in equipping communities to handle flood disasters better. Additionally, aligning with Malaysia’s commitment to the Sendai Framework 2015–2030, the research supports global efforts to strengthen disaster resilience. Focused specifically on adolescent education regarding flood disaster preparedness, this initiative not only empowers a vulnerable demographic but also cultivates a resilient community capable of effectively withstanding and recovering from catastrophic events. Start and end dates of the study and estimated duration The data collection phase of the study was conducted between May 2023 to January 2024. Results Sociodemographic background of participants The study included 287 participants with a median age of 14 (± 1 year). Female participants constituted 54.7% of the sample, while males accounted for 45.3%. Most participants came from families with educational backgrounds up to the secondary school level (69%) and a monthly family income of less than RM4850 (64.8%). Notably, 68.1% of participants reported no prior experience with flood disasters. Conversely, 31.9% had experienced flood disasters, with 33% describing their experiences as moderately severe. Additionally, 60.2% of respondents indicated that their family or friends had encountered flood disasters, with 41.2% characterizing these experiences as moderately severe. Significant proportions of participants 238 (82.9%) had not received any disaster education before the study, while 49(17.1%) had some form of disaster education. Detailed sociodemographic and disaster experience data are presented in Table S2 in the supplementary section. Participation rate The study achieved high response rates, with 97.2% immediately post-intervention and 96.5% at the three-month follow-up. Missing data were analysed using Little’s MCAR test, indicating that the missingness was completely at random (p-value > 0.05). Given that only 4.9% of the data was missing, primarily from the sociodemographic section, no imputation was performed as it was unlikely to impact the outcomes significantly. Data analyses were conducted using the Generalized Estimating Equation (GEE) model on non-missing observations. Baseline study of flood disaster preparedness score At the baseline of the study, the flood disaster preparedness (FDP) scores, as assessed by the Independent T-test, revealed no statistically significant difference between the control group (Mean (SD) = 7.442 (2.95)) and the intervention group (Mean (SD) = 7.40 (2.42)). The t-value was 0.133 with a p value of 0.894, indicating a mean difference of only 0.042 (95% CI − 0.59, 0.67). However, significant differences were identified at baseline between the intervention and control groups concerning age (p value < 0.001), perceived efficacy (p value = 0.002), perceived threat (p value = 0.018), individual disaster experience (p value < 0.001) and disaster severity (p value = 0.005), and family or friends’ disaster experiences (p value = 0.019) and disaster severity (p value = 0.023). Effect of ‘Get-Ready DEM’ on flood disaster preparedness score on intervention group after adjusting the significant covariates The study employed a robust multivariable analysis using the Generalized Estimating Equation (GEE) to assess the impact of the ‘Get-Ready DEM’ module, after accounting for significant covariates such as age, perceived efficacy, perceived threat, individual and family or friend’s disaster experience as detailed in Table s3 (supplementary). The findings indicated that the intervention group experienced a significant increase in the flood disaster preparedness scores by 0.80 units (β = 0.80, 95% CI 0.11,1.49; p value = 0.02) compared with the control group. There were no significant changes in FDP scores based on the timelines (p value > 0.05). Further interaction analysis between the timeline and trial groups revealed marked improvements in the intervention group’s FDP scores when compared with the control group. Specifically, there was a significant increase of 2.17 units in the intervention group’s scores immediately post-intervention (β = 2.17; 95% CI = 1.66, 2.66; p value < 0.001), and a 1.58 unit increase at the 3-month follow-up (β = 1.58; 95% CI = 0.99, 2.16; p value < 0.001), compared to baseline. Figure S2 at the supplementary section illustrates the total mean score of flood disaster preparedness score between trial groups at three different time points. Sensitivity analysis A sensitivity analysis (Wild-Bootstrapping) using a 3-level linear mixed-effects model (lmer) was conducted to account for clustering at the school level and validate the robustness of the primary findings. To obtain robust p values given the small number of clusters (N = 8), a CR2 small-sample correction was applied. The variance component for the random effect at the school level was not statistically significant (Z = 0.863, p value = 0.388), indicating minimal between-school variability. However, after adjusting for the small cluster size, the group-by-time interaction was no longer statistically significant (p value = 0.26 for time 1 and p value = 0.43 for time 2). While the positive coefficient for the interaction (Estimate = 2.099) indicates the intervention effect was in the expected direction, the high uncertainty and large standard error resulting from the small number of clusters prevented statistical significance. This suggests that while the primary GEE model showed a promising signal, the study is ultimately underpowered to detect a statistically significant effect once these necessary corrections are applied. As part of the sensitivity analysis, a subgroup analysis using GLMM was conducted to examine whether prior disaster education exposure influenced outcomes within the intervention group. The analysis tested for an interaction effect between disaster education exposure and group assignment on Flood Disaster Preparedness (FDP) scores. Results indicated no statistically significant interaction (β = 1.07; 95% CI − 0.22, 2.36; p value = 0.104) between intervention group and exposure to disaster education. This suggesting that the observed intervention effect was consistent regardless of students’ prior exposure to disaster education. Discussion The study demonstrated that the ‘Get-Ready DEM’ module significantly enhances FDP scores among participants in the intervention group. The observed improvement in flood preparedness scores, both immediately and at 3 months, indicates the module’s long-term effectiveness. The results imply that disaster educational interventions for adolescents can greatly bolster their disaster risk awareness, knowledge, and self-efficacy, particularly regarding floods. is consistent with similar studies conducted in Malaysia67–69, Indonesia70,71, Iran53, Pakistan72, Vietnam73, and Nepal74, all affirming the positive impact of disaster education on adolescent preparedness. Moreover, the findings correspond with previous Malaysian studies where Health Belief Model (HBM) based educational strategies yielded favourable outcomes. A quasi-experimental study on young students (9–10 years old) indicated that HBM-based flood education effectively improved perceived benefits, perceived severity and perceived susceptibility and enhancing knowledge retention(F (38,137) = t 2.517, p value < 0.001) post 1 month intervention67. Likewise, the game-based learning approach, Master of Flood Preparedness (MOFP) a cross sectional study, showed increased percentages of flood awareness and preparedness among Malaysian youth (13–25 years old) from 30.8% (37) to 51.7% (62)75. These interactive educational methods likely contribute to observed enhancements in awareness and behaviour in this study and others. Moreover, the EPPM offers a comprehensive framework for disaster education by integrating cognitive and emotional responses to perceived threats. Unlike the Health Belief Model (HBM), which primarily addresses cognitive factors, EPPM combines perceived threat with perceived efficacy, effectively motivating protective behaviors76. This integration is particularly beneficial for adolescents, as it aligns with their developmental stage and enhances their capacity to influence family and community preparedness. By addressing both the emotional and cognitive aspects of behaviour change, EPPM fosters proactive disaster preparedness among adolescents. This study further illustrates that majority participants (82.9%) in both the intervention and control groups had no prior exposure to flood disasters education and possessed similar baseline scores in flood disaster preparedness, despite the control group reporting a higher proportion of past flood-disaster experience. This common starting point underscores the significant improvement in flood disaster preparedness scores observed in the intervention group, which can be attributed to the effectiveness of the ‘Get-Ready DEM’ modules in enhancing disaster-related knowledge and awareness. These modules not only educate but also possibly influence participants’ attitudes and behaviours towards flood disaster preparedness. The cluster-randomized methodology demonstrates the broad applicability of the ‘Get-Ready DEM’ modules, effectively preparing adolescents with or without prior flood experiences. This underscores the module’s value in fostering proactive disaster education and resilience among diverse and vulnerable groups in Malaysia. The notable improvement in preparedness scores can be attributed to the comprehensive design of the ‘Get-Ready DEM’ module, which integrates theory-based (EPPM) with interactive and practical learning approaches22. These include game-based activities and audio-video lectures, which have proven effective in enhancing understanding and retention of disaster preparedness concepts69,71,77,78. Such interactive strategies have been shown to keep participants engaged and facilitate the translation of learned behaviours into actual preparedness measures22,71. However, the study also observed a decline in preparedness scores at the 3-month follow-up, suggesting that a single educational intervention may not be sufficient for long-term preparedness74. This finding indicates the need for continuous education, periodic drills, and regular reinforcement to maintain and enhance disaster preparedness behaviours79. Besides, this study underscores the necessity of implementing a booster intervention or continued flood disaster education, to sustain and enhance flood disaster preparedness. This should be conducted at a minimum interval of 3 months or on a quarterly basis, depending on its feasibility and practicality. Implementing visual aids like banners and models of emergency kits within educational settings also played a crucial role in reinforcing learning and promoting a culture of preparedness71,80. Furthermore, the study highlighted several covariates, such as age, perceived efficacy, and personal and family disaster experiences that significantly influenced the outcomes. Older adolescents often display a better understanding and are more responsive to disaster preparedness measures due to advanced cognitive and emotional development16,81. Conversely, younger adolescents may require more tailored educational interventions that cater to their developmental stage82. Additionally, personal experience with disasters significantly impacted flood disaster preparedness actions, as those who have previously experienced floods are more likely to engage in proactive measures81,83. The severity of past experiences also enhances the perception of risk, prompting more robust preparedness efforts34. This study addresses a documented gap in disaster preparedness education for adolescents, who are often underrepresented in disaster risk reduction (DRR) initiatives despite their vulnerability. By leveraging the Extended Parallel Process Model (EPPM), the research integrates cognitive and emotional determinants of preparedness behaviour into a tailored and innovative intervention. The findings affirm the effectiveness of EPPM-based flood disaster education programs in enhancing adolescent preparedness, highlighting the importance of incorporating such targeted educational strategies within broader DRR initiatives to foster resilience among this vulnerable demographic. This approach is consistent with findings from other regions, reinforcing the broader literature’s perspective that structured disaster education not only raises awareness but also arms adolescents with the crucial skills and knowledge necessary for effective response to disasters22,28,79,84. The ‘Get-Ready DEM’ program, developed through systematic integration of effective communication theories and the strategic delivery of information via interactive practical sessions and audio-visual media, has effectively engaged participants. Situated within the school environment as the base for the program, it has significantly enhanced the learning process and successfully achieved the study’s objectives28. Therefore, school-based disaster education programs incorporating comprehensive disaster risk reduction (DRR) activities foster a safety culture and proactive disaster management among students85,86. These programs are particularly effective in environments with a strong synergy between community and school initiatives, underscoring the importance of collaborative strategies that include educators, community leaders, and policymakers87. Schools play a pivotal role in preparing a generation that is aware of natural disaster risks and adept at mitigating these risks39,73. This can be achieved by integrating DRR education into standard curricula and complementing it with community-based preparedness strategies, ensuring active participation of adolescents and providing a sustainable avenue for delivering disaster education. This study has several limitations. First, relying on self-reported knowledge and awareness introduces potential information bias and may not accurately reflect actual behaviours. Second, the short assessment period, without an intervening flood event, limits the evaluation of true proactive preparedness. Finally, the small number of clusters, which was constrained by the logistical realities of implementing a school-based intervention within a strict timeframe, may restrict the generalizability of these findings to regions with different socio-economic contexts or flood vulnerabilities. To mitigate this, random allocation and adjustment for the cluster design effect were employed to preserve the trial’s internal validity. Future research should broaden the study’s scope to include diverse geographical areas and cultural backgrounds to enhance generalizability and understanding of flood disaster preparedness. Investigating dissemination mechanisms for disaster risk reduction education and the roles of educators, volunteers, and guardians are crucial, along with fostering community involvement to build resilient partnerships. Furthermore, addressing the psychological impact of flood disasters on adolescents is essential, with future studies focusing on timely assessments and interventions to reduce anxiety, promote effective coping strategies, and enhance social support to improve disaster preparedness. Conclusion This study highlights the effectiveness of the ‘Get-Ready DEM’ module in enhancing flood disaster preparedness among adolescents, demonstrating the importance of tailored educational interventions. The findings, supported by rigorous methodology and control of covariates, align with research affirming the value of well-designed educational programs in building resilience. These results underscore the critical role of disaster education as a cornerstone of sustainable disaster risk reduction strategies, with potential applicability across disaster-prone regions and diverse populations. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors extend their profound gratitude to the Ministry of Education and the participating schools for their invaluable cooperation and support, which were pivotal in facilitating the execution of this study. Additionally, heartfelt thanks are directed to the Universiti Putra Malaysia (UPM) for the financial support provided through the grant (6302015-14001) that ensured the seamless progression of this research. This collaboration and funding were essential in achieving the study’s objectives and are deeply appreciated. Author contributions Conception and design: K.M, AZFAAcquisition, analysis, or interpretation of data: K.MDrafting the work or revising: K.M, AZFA, RAM, MRBFinal approval of the manuscript: KM, AZFA. Funding This work was supported by Universiti Putra Malaysia (UPM) through the Putra Berfokus university grant (6302015-14001), which provided essential financial resources for the conduct of this study. The funding facilitated the smooth implementation of all research activities, from data collection to analysis and dissemination of findings. Data availability The datasets generated during and/or analysed during the current study are not publicly available as it require permission from the Ministry of Education for data sharing. The original questionnaire, educational materials, and datasets used or analysed during the current study are available from the corresponding author only at a reasonable request to protect the participants’ privacy, as the dataset has some identifying variables. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated during and/or analysed during the current study are not publicly available as it require permission from the Ministry of Education for data sharing. The original questionnaire, educational materials, and datasets used or analysed during the current study are available from the corresponding author only at a reasonable request to protect the participants’ privacy, as the dataset has some identifying variables.