Trends and disparities in unintentional injury mortality among children under five in northern Iran: a 10-year population-based surveillance study (2013–2022) Authors: Roya Farokhi, Pouyan Ebrahimi, Mehdi Safarpour, Mohamed Ali jahani, Mehran Asadi-Aliabadi, Behnam Khodadoust, Leila Zoleikani Published: 2026 ABSTRACT AbstractAlthough unintentional injuries (UIs) are a well-recognized cause of early childhood mortality, regional evidence on their trends and disparities remains limited in Northern Iran, where subnational variation may obscure national progress. This study investigates the patterns and characteristics of unintentional injury deaths (UIDs) in this vulnerable population. We conducted a retrospective trend study using mortality data from the under-five death (U5D) surveillance system in Northern Iran, covering the years 2013 to 2022. Causes of death were classified according to ICD-10 codes. Temporal trends in UIDs were examined using Join point regression analysis, while Poisson regression was applied to estimate relative risks (RRs) and 95% confidence intervals (CIs) for sex- and residence-based disparities. Among 3390 U5Ds, 240 (7.08%) were attributable to UIs. The overall unintentional injury death (UIDs) rate was 0.81 per 1,000 live births, declining non-significantly from 0.93 in 2013 to 0.52 in 2022 (APC − 3.34%, p = 0.162). Transport accidents were the predominant cause (56.7%), followed by accidental drowning and submersion (10.8%) and other accidental threats to breathing (10.0%). Girls had a significantly lower risk of UIDs rate compared to boys (RR = 0.76; 95% CI: 0.58–0.98). No statistically significant difference was found between rural and urban areas (RR = 1.25; 95% CI: 0.97–1.61). Between 2013 and 2022, UIDs among children under five in Mazandaran Province showed a downward trend; however, the reduction was not statistically significant. During this period, boys consistently exhibited a significantly higher risk of injury-related mortality compared with girls. These patterns underscore persistent inequities in child injury outcomes and highlight the need for sustained, equity-focused preventive strategies, particularly those targeting high-risk groups and the leading mechanisms of injury, such as transport incidents, drowning, and suffocation.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-46126-4.Keywords: Unintentional injuries, Child mortality, Epidemiology, Cause of death, Risk factors, IranSubject terms: Diseases, Health care, Medical research, Risk factors Background Although much has been achieved to promote child health in numerous countries around the globe, under-five mortality rate (U5MR) is a highly important health system indicator, socioeconomic development, and equity. According to the United Nations Inter-agency Group for Child Mortality Estimation, the global U5MR dropped from 93 deaths per 1,000 live births in 1990 to 37 in 2022, yet over 5 million children under five still die each year, largely from preventable causes1. Among these, UIs have emerged as a persistent and underrecognized contributor, responsible for approximately 6–9% of under-five deaths (U5Ds) globally, depending on region2,3. The vast majority of these deaths occur in low- and middle-income countries (LMICs), where children are exposed to higher environmental hazards, limited injury prevention policies, and poor access to emergency care4. Common mechanisms of UIs include road traffic crashes, drowning, suffocation, poisoning, and falls, most of which are entirely preventable through community-based interventions, environmental modifications, and policy enforcement2,5. In spite of this, the prevention of injuries has been disproportionately underrepresented in child survival strategies in the world6. In Iran, substantial progress has been achieved in reducing child mortality, with the national U5MR declining from 57.6 per 1,000 live births in 1990 to approximately 12 in 2022, approaching the Sustainable Development Goal (SDG) target of 25 deaths per 1,000 live births6. Road traffic injuries, drowning, suffocation and burns have been the most common causes of childhood UIDs because of epidemiological evidence that has been indicated by various regional studies. In addition to these stable patterns of mechanistic associations, concerning socioecological gradient is observed, as gender of males, living in rural areas, and younger age are major factors contributing to high mortality risk7–12. These disparities emphasize the critical role of mortality data in identifying preventable causes of death, revealing hidden inequities, and informing region-specific child health strategies. Mortality statistics can be used to monitor health actions, the strain of disease, evaluate the output of health care systems and determine the achievement of national and global objectives including the SDGs13. Iran’s mortality surveillance architecture integrates two principal, complementary national systems: the Civil Registration and Vital Statistics (CRVS) system and the Ministry of Health and Medical Education’s (MOHME) Cause of Death Registry. Collectively, these systems represent Iran’s most authoritative and comprehensive data infrastructure for cause-specific mortality surveillance, with well-documented applications in injury epidemiology and pediatric population health monitoring14. The exploitation of such systems is especially important in such areas as Mazandaran with its valuable variety of environmental and infrastructural factors which can impact the pattern of injuries and health outcomes. Mazandaran Province, located in northern Iran along the southern coast of the Caspian Sea, is the most populous province among the three northern regions, with over 3 million inhabitants and a total area of approximately 23,833 km²15. While several studies have been run in other Iranian provinces such as Fars in the south7,11, Ardabil in the northwest12, and Isfahan9, the unique socio-ecological context of Mazandaran necessitates a separate, in-depth investigation. As the most populous northern province with a vast Caspian coastline, dense river networks, a distinct humid climate, and specific agricultural and tourism-related activities, the environmental hazards for children in Mazandaran are qualitatively different. Therefore, this study is not merely a replication of national or regional data, but an effort to generate context-specific evidence. By directly comparing our findings from Mazandaran with existing evidence from other regions, this study, a decade-long analysis, aims to highlight subnational variations in child injury patterns, thereby challenging the notion of a one-size-fits-all national prevention strategy and underscoring the critical need for province-level data to inform localized interventions. Methods Study Design and Setting The present retrospective trend analysis research used mortality data of U5D Surveillance System in Mazandaran Province, north of Iran, between January 2013 and December 2022, a population-based mortality surveillance initiative developed by the Iranian MOHME. The system was intended to record, examine, and assess all fatalities among children aged U5D years in the entire country. The program is run in Mazandaran Province under the oversight of the Deputy of Health Affairs at Mazandaran University of Medical Sciences and covers all county-level health networks, which makes it almost comprehensive in covering U5Ds. U5D system operates under a network of primary health-care centers (Health Houses), district health offices, civil registration, forensic medicine and hospitals. The routine health information system records each child death and investigates it with the help of standardized abstraction forms. These forms record demographic characteristics, environmental context, timing and location of death, pre-hospital and hospital care, and any delays in recognition or referral. Data are collected by trained community health workers or physicians and verified by district health officers for accuracy and completeness. Every case is subsequently reviewed by expert mortality committees at three levels, hospital, district, and provincial, to determine the underlying and contributing causes of death according to ICD-10 definitions. The review process systematically identifies modifiable risk factors related to health-system performance (such as quality of pediatric care, referral efficiency, or accessibility of emergency services) and socioeconomic or environmental determinants The system is periodically checked to guarantee data validity, such as cross-checking with the hospital records, civil registration data, and community reports, and regular completeness and timeliness audits by Iranian MOHME. Cases and records that are not resident are not analyzed and those whose key variables are implausible or missing. Results of such multi-level reviews are sent back to the provincial health system to inform specific interventions, reinforce referral networks, and avoid the reoccurrence of similar deaths. Where the death records were missing or ambiguous, further details were sought in the hospital records or through follow-up with the healthcare providers to fill in the missing details. In order to ensure the quality of the data, all the reports were firstly checked by the local health specialists and then, at the provincial level. Children that were not residents of the area (e.g., migrants or visitors) were not included in the analysis to guarantee the applicability of the results to the local population. The records containing invalid or unrealistic entries (e.g., age discrepancy or duplicate records) were thoroughly checked and, when needed, omitted of the dataset. Out of the initial 247 children under five, 4 non-resident children and 3 records with invalid or incomplete entries were dropped to give a final analytic dataset of 240 records to be used in all analyses. These quality assurance measures improved the internal validity and reliability of the data that was utilized in this research. Variables and Definitions The dataset included key demographic and epidemiological variables for each child, including sex (boy or girl), cause of death, and place of residence (urban or rural). Only deaths resulting from UIs were included in the analysis. Causes of death were coded according to the International Classification of Diseases, 10th Revision (ICD-10), in accordance with the World Health Organization (WHO) guideline16,17. Unintentional injury deaths (UIDs) were defined using ICD-10 codes ranging from V01 to X59 and categorized into major external causes, including transport accidents (V01–V99), falls (W00–W19), exposure to inanimate mechanical forces (W20–W49), contact with heat and hot substances and exposure to smoke, fire, and flames (X00–X19), accidental poisoning and exposure to noxious substances (X40–X49), drowning and submersion (W65–W74), other accidental threats to breathing including suffocation (W75–W84), and other specified external causes (e.g., W87, X38, and X57). These classifications enabled systematic grouping and interpretation of injury mechanisms among U5Ds. Statistical analysis Descriptive statistics were used to summarize the frequency and distribution of UIDs by age group, sex, place of residence (urban/rural), year of death, and cause. Annual crude mortality rates were calculated by dividing the number of UIDs by the total number of live births for each corresponding year, expressed per 1,000 live births. Cause-specific mortality rates and proportional mortality (dividing the number of UIDs by the total UIDs for each corresponding year) distributions were also calculated to characterize leading mechanisms of injury. Temporal trends in UIDs from 2013 to 2022 were analyzed using Join point regression (Join point Regression Program, version 5.0.2; National Cancer Institute, USA). We selected the Join point model because it provides a powerful, data-driven approach for identifying the precise timing of significant changes in trends, which is superior to simple linear models that assume a constant rate of change over time. This method fits a series of connected straight lines on a logarithmic scale to the data and determines the optimal number of breakpoints (join points) where the trend changes significantly, using a Monte Carlo permutation test, with statistical significance set at p < 0.05. For each segment between join points, the model estimates an Annual Percent Change (APC) with corresponding 95% confidence intervals, which quantifies the rate of increase or decrease per year, providing a detailed understanding of mortality trends over specific time periods. This approach is particularly valuable in public health surveillance to objectively evaluate the impact of interventions, policy changes, or emerging hazards on mortality patterns over extended periods. The Joinpoint methodology is well-established as a gold standard for cancer trend analysis and has been widely validated and adopted for injury surveillance and other epidemiological studies18,19. To assess disparities in UIDs by sex and residential setting, Poisson regression models were fitted to estimate relative risks (RRs) and 95% confidence intervals. Poisson regression was selected as the most appropriate method for several theoretical and practical reasons pertinent to our data structure. Firstly, our outcome variable was a count of rare events (number of deaths) occurring within a fixed and known population denominator (annual live births). Poisson regression is specifically designed for such rate data, where the logarithm of the population denominator is included as an offset term, allowing for the direct modeling of mortality rates and providing unbiased estimates of incidence rate ratios (interpreted as Relative Risks). Secondly, for rare outcome events (typically considered those with an incidence below 10%), Poisson regression provides more reliable and interpretable estimates of the relative risk compared to logistic regression, which would estimate odds ratios that can substantially overestimate the relative risk when the outcome is not rare. This characteristic makes Poisson regression the preferred and widely recommended method in epidemiological studies of mortality rates, particularly for cause-specific child mortality, which is a relatively rare event. The use of Poisson regression for analyzing mortality data with denominator counts is strongly supported by the epidemiological and biostatistical literature20,21. All analyses were conducted using Stata version 14 (Stata Corp, College Station, TX, USA), with statistical significance defined as p < 0.05. Ethical approval and consent to participate This study was approved by the Ethics Committee of Mazandaran University of Medical Sciences, Iran (Ethical code IR.MAZUMS.REC.1403.351). Result During the 10-year study period (2013–2022), a total of 240 UIDs were recorded among children under five years of age in Mazandaran Province, accounting for 7.08% of all U5D2s. The proportion of all UIDs declined from 7.57% in 2013 to 3.81% in 2022, marking a relative reduction of approximately 49.7%. Overall, the distribution by residence was relatively balanced, with 52.1% of deaths occurring in urban areas and 47.9% in rural settings. Although the overall distribution of deaths was nearly balanced between urban and rural areas, notable annual fluctuations were evident; for instance, urban areas accounted for 69.6% of deaths in 2021, while rural areas had a higher share (63.6%) in 2022. In terms of gender, boys consistently experienced a greater burden, constituting 58.7% of UID cases versus 41.3% among girls, equivalent to a boy-to-girl ratio of 1.4:1 (Fig. 1, Table S1). Annual Distribution of Unintentional Injury Deaths among Children Under Five by Sex and Residence. The leading causes of UIDs were transport accidents (56.7%), accidental drowning and submersion (10.8%), and other accidental threats to breathing (10.0%), which together accounted for over 77% of all UIDs. While deaths from transport injuries declined from 56.3% in 2013 to 36.4% in 2022, they remained the dominant cause across all years. Less common causes included falls (2.9%), poisoning (7.1%), exposure to mechanical forces (5.0%), and contact with heat and hot substances (4.2%) (Fig. 2, Table S1). Building on these patterns, population-level metrics further illustrate the substantial burden attributed to the major cause of UIs. Transport-related accidents accounted for an overall mortality rate of 0.46 per 1,000 live births across the decade, contributing 4.01% of all U5Ds. Accidental drowning and submersion contributed a mortality rate of 0.09 per 1,000 live births (0.77% of all deaths), while other accidental threats to breathing showed a comparable magnitude, with a cumulative rate of 0.08 per 1,000 and a proportional contribution of 0.71%. Accidental poisoning represented 7.1% of all injury deaths but showed notable temporal variation. Pronounced peaks occurred in 2017 (14.8% of that year’s injury deaths) and 2019 (25.0%). The overall decline in mortality rates was largely driven by a marked reduction in transport-related fatalities, which decreased by approximately 77.8% in absolute numbers between 2013 and 2022 (Fig. 3, Table S2). Distribution of Unintentional Injury Deaths among Children Under Five by Cause ,2013–2022. Mortality rate (‰) and proportion (%) of injury-related deaths by cause among children under five years,2013–2022. Temporal Trends The best-fitting Join point regression model for the 2013–2022 period contained no significant join points (p > 0.05 for all potential change points), indicating that a single, steady trend line best fits the mortality data. Consequently, the overall trend is accurately summarized by the APC for the entire period, which was − 3.34% (95% CI: -9.41 to + 1.76; p = 0.162), describing a non-significant decreasing trend. The apparent numerical fluctuations in specific sub-periods (e.g., 2015–2019, 2019–2021) are visually present in Fig. 4 but are not supported as statistically distinct trends by the formal analysis (Fig. 4). Trends in UIDs among children under five years, 2013–2022. Gender and Residence-Based Differences Overall, compared to boys, girls had a significantly lower U5MR (RR = 0.76, 95% CI: (0.58 to 0.98; P = 0.032). while accidental poisoning was significantly more common in boys (0.10‰ vs. 0.01‰; RR = 0.14; 95% CI: 0.03–0.63; P = 0.01) (Table 1). Table 1Sex-specific Causes of Unintentional Injury Deaths among children under five years,2013–2022.Causes n GirlN = 142,056Rate Per 1,000 (95%CI) n BoyN = 152,694Rate Per 1,000 (95%CI)P- valueRR (Girls vs. Boys; 95% CI)Transport accident660.46 (0.36 to 0.59)700.46 (0.36 to 0.58)0.9381.01 (0.72 to 1.41)Falls20.01 (< 0.001 to 0.05)50.03 (0.01 to 0.07)0.310.43 (0.08 to 2.21)Exposure to inanimate mechanical forces30.02 (< 0.001 to 0.06)90.06 (0.03 to 0.11)0.1240.36 (0.10 to 1.32)Accidental drowning and submersion110.08 (0.04 to 0.14)150.10 (0.06 to 0.16)0.5490.79 (0.36 to 1.72)Other accidental threats to breathing100.07 (0.03 to 0.13)140.09 (0.05 to 0.15)0.5230.77 (0.34 to 1.72)Contact with heat and hot substances, and Exposure to smoke, fire, and flames30.02 (< 0.001 to 0.06)70.05 (0.01 to 0.09)0.2610.46 (0.12 to 1.78)Accidental poisoning by and exposure to noxious substances20.01 (< 0.001 to 0.05)150.10 (0.06 to 0.16)0.010.14 (0.03 to 0.63)Other20.01 (< 0.001 to 0.05)60.04 (0.001 to 0.09)0.2090.36 (0.07 to 1.78)Total990.70 (0.57 to 0.85)1410.92 (0.77 to 1.09)0.0320.76 (0.58 to 0.98)RR: Relative risk. Sex-specific Causes of Unintentional Injury Deaths among children under five years,2013–2022. RR: Relative risk. Although not statistically significant, children residing in rural areas had a higher overall UIDs rate than those in urban areas (0.92 vs. 0.74 per 1,000 live births; RR 1.25, 95% CI 0.97–1.61, p = 0.083). Except for exposure to inanimate mechanical forces, which was slightly higher in urban areas, rural children had higher mortality rates for other specific causes of unintentional injuries. A similar non-significant trend was observed for other major causes of UIDs (Table 2). Table 2Unintentional injury death rate among children under five years by residence, 2013–2022.Causes n RuralN = 124,889Rate Per 1,000 (95%CI) n UrbanN = 169,861Rate Per 1,000 (95%CI)P-valueRR (Rural vs. Urban; 95% CI)Transport accident640.51 (0.39 to 0.65)720.42 (0.33 to 0.53)0.2691.21 (0.86 to 1.69)Falls30.02 (< 0.001 to 0.07)40.02 (< 0.001 to 0.06)0.9791.02 (0.23 to 4.56)Exposure to inanimate mechanical forces40.03 (< 0.001 to 0.08)80.04 (0.02 to 0.09)0.5290.68 (0.21 to 2.26)Accidental drowning and submersion140.11 (0.06 to 0.19)120.07 (0.03 to 0.12)0.2411.59 (0.73 to 3.43)Other accidental threats to breathing110.08 (0.04 to 0.16)130.08 (0.04 to 0.13)0.7321.15 (0.52 to 2.57)Contact with heat and hot substances, and Exposure to smoke, fire, and flames50.04 (0.01 to 0.09)50.03 (0.001 to 0.07)0.6271.36 (0.39 to 4.70)Accidental poisoning by and exposure to noxious substances90.07 (0.03 to 0.13)80.04 (0.02 to 0.09)0.3811.53 (0.59 to 3.97)Other50.04 (0.01 to 0.09)30.01 (< 0.001 to 0.05)0.2622.67 (0.54 to 9.46)Total1150.92 (0.76 to 1.11)1250.74 (0.61 to 0.88)0.0831.25 (0.97 to 1.61)RR: Relative risk. Unintentional injury death rate among children under five years by residence, 2013–2022. RR: Relative risk. Discussion This population-based study provides a comprehensive assessment of UIDs among children under five in Mazandaran Province over a 10-year period. The overall burden of UIDs showed a non-statistically significant declining trend, falling from 7.57% in 2013 to 3.81% in 2022. Despite this decline, UIs remained a major cause of preventable child mortality, accounting for nearly 7.08% of all deaths in this age group. Transport-related incidents emerged as the leading cause of death throughout the decade, followed by drowning and threats to breathing, together responsible for over 77.5% of UIDs. Furthermore, the data revealed significant disparities in UIDs by sex: boys faced a statistically significantly higher risk than girls. Additionally, although the difference was not statistically significant, children in rural areas consistently showed elevated mortality rates across most injury mechanisms compared with those in urban areas. The observed decline in UIDs in Mazandaran Province from 2013 to 2022, although not statistically significant, should be interpreted within the context of concurrent child safety initiatives. Two major national programs were systematically introduced during this time using the primary healthcare infrastructure of Iran, the National Program to Prevent Accidents and Child Safety, which focused on home safety inspections, parental education on the dangers of accidents, and community sensitization; and the Child Mortality Surveillance System, which created standardized procedures to investigate and review child deaths to identify risk factors that could be modified, according to national guidelines by the Iranian MOHME. Although our study design does not allow causal inference, the progressive mortality decrease can be in part due to the cumulative effect of these concerted prevention and surveillance measures. The longitudinal studies conducted in the future, including formal policy evaluation elements, will be justified to quantify the effect of such a program on the population level in terms of injury-related outcomes. Our study demonstrated a declining trend in UIDs among children under five in Mazandaran Province from 2013 to 2022, with an APC of − 3.34%, though this reduction did not reach statistical significance. Despite temporary increases, particularly from 2019 to 2021, the overall rate demonstrated a consistent downward trajectory, with a total decrease of 44% over the study period, a trend that aligns with global patterns, albeit with substantial regional variations. The same negative development has been observed in southern Iran, where UID among children declined by a significant margin between 2004 and 2019, with an APC of -5.2% in boys and − 4.3% in girls7. These data are in line with our results, but the extent of the decrease in the southern area is somewhat stronger. Likewise, in the Sichuan province of western China, UID mortality among children under five years old dropped by 3.8 to 1.7 per 1,000 live births in the period between 2009 and 2017, which is an annual rate of 8.8%17. The national-level analysis in China from 2010 to 2020 also supports this trend, reporting an average annual decline of 3.7% in under-five UID mortality22, closely mirroring the APC observed in our study. However, unlike our setting, where declines were occasionally interrupted, China exhibited a more consistent trajectory. Global studies, such as the GBD-based analysis by Liu et al., indicate a marked global decrease in non-transport UID deaths, but highlight persistent inequalities, especially in LMICs. Although absolute mortality has declined, the relative inequality index remained static, suggesting uneven progress23. More recent provincial-level studies, such as in Ardabil, confirm ongoing challenges and the need for localized prevention efforts, especially targeting environmental factors12. Collectively, these comparisons indicate that the downward trend in our area is in line with more national and international initiatives. Nevertheless, the reversals that have been witnessed since 2019 to 2021 could be the result of structural weaknesses. These are injury prevention programs disruption, rural infrastructure lapses, and external forces like economic strain or reallocation of the health system. To guarantee a lasting improvement in the UID reduction, therefore, long-term and resilient strategies are needed to withstand such fluctuations in the context. The patterns of UIDs of 2019–2021 indicate the possibility of redistribution of risks during the COVID-19 pandemic, showing three major epidemiological patterns that match the pandemic disruptions reported. First, injuries associated with transport showed a temporary decrease and a rebound, which was related to the limitation of mobility, and the consequent release of it. This trend is reflected by the evidence in Iran that demonstrates a considerable decrease in healthcare use in the case of strict lockdowns. Second, there was an increase in home-based injuries such as drowning and suffocation, which probably represents a longer period of home confinement and changed supervision patterns. The results correspond to the Iranian research on the change in the patterns of pediatric trauma during the pandemic24,25. Third, the trend of poisoning deaths was decreasing, which means that there is a specific weakness of the mechanism to pandemic-related factors. This differential impact across injury categories suggests the pandemic may have functioned as an effect modifier, simultaneously reducing environmental exposures while intensifying domestic risks. The convergence of our findings with both regional Iranian evidence and global reports26,27. Strengthens the plausibility that pandemic-related environmental and health system changes contributed to the observed injury patterns. While our data cannot establish causality, these findings highlight how public health emergencies can rapidly reshape injury epidemiology through changes in exposure patterns and supervision dynamics. Beyond the temporal patterns observed in our data, Transport accidents were the leading cause of UIDs among children under five in Mazandaran Province, followed by drowning and threats to breathing. Together, these three causes accounted for more than three-quarters of all UIDs in this age group, consistent with global injury profiles in young children. Transport accident injuries remain a major contributor to childhood injury deaths worldwide, particularly in LMICs where enforcement of traffic safety laws, pedestrian infrastructure, and child restraint use are limited28. In Iran, several provincial studies have identified Transport accidents as the dominant mechanism of fatal injury in children, including in southern Iran and Ardabil7,12, mirroring the pattern we observed in Mazandaran. However, fundamental differences emerge in secondary causes and methodological approaches. In sharp contrast to Southern Iran, where burns represent the second leading cause, our coastal region demonstrates drowning as the predominant secondary cause (10.8%), reflecting profoundly different environmental risk profiles. Furthermore, unlike the Ardabil study, which focused exclusively on traffic accidents, our comprehensive approach captures the full spectrum of injury mechanisms, revealing drowning and suffocation as critical concerns in Mazandaran. The high burden of transport accident deaths in Mazandaran may be influenced by urban sprawl, inadequate pedestrian safety measures, and the frequent use of motorcycles for family transport in peri-urban and rural areas. Drowning ranked second among the causes in our study. This is especially concerning in a coastal province like Mazandaran, where children are frequently exposed to water bodies such as rivers, canals, and the Caspian Sea. Environmental proximity to unguarded water sources, limited parental supervision, and insufficient public awareness contribute significantly to this pattern. Local evidence from northern Iran has long highlighted the significance of drowning as a leading cause of preventable mortality in coastal communities29,30. This finding is reinforced by recent regional analyses indicating that drowning remains the second leading cause of unintentional injury deaths among children across the Eastern Mediterranean region31. The persistence of this pattern, despite notable progress in other areas of child survival, underscores the limited impact of current preventive strategies and the urgent need for sustained, community-based interventions. Globally, the WHO estimates that approximately 73,000 children under five years of age died from drowning in 2021, accounting for nearly one-quarter of all drowning deaths worldwide32. This alignment between global and regional patterns highlights both the preventable nature of these deaths and the critical importance of strengthening localized prevention initiatives in high-risk settings such as northern Iran. Breathing threats were in the third place in our results. These injuries are usually unreported and mostly happen in infants and toddlers, which are very susceptible to these injuries because of their developmental behaviors like the act of mouthing and poor motor coordination. The international research proves that suffocation is among the leading causes of fatal injuries in children younger than one year, which is often caused by unsafe sleeping conditions, food aspiration, or ingestion of small objects33,34. These causes are likely to be a result of a combination of environmental, behavioral, and systemic factors, as the prevalence of these causes in Mazandaran is significant. The distribution is caused by frequent access to open water sources, low levels of literacy about injuries at the community level, and insufficient investment in child safety interventions by the population. We found significant differences in UIDs of children below five years according to gender. In line with the past trends in the world and country, boys recorded higher levels of UIDs than girls9,35,36. As an example, the gender-specific social norms and activity patterns were found to disproportionately impact boys in rural and suburban China, who were disproportionately injured due to drowning and traffic accidents than girls17. This can be partly explained by an increased level of physical activity, an increased risk-taking behavior, and the difference in parental supervision between the genders as has been reported in various epidemiological settings35,37. Furthermore, boys are allowed to roam more freely in their environment and often play alone more frequently38,39. Children living in rural regions were found to have a greater risk of fatal UIs as compared to those living in urban areas, and this has always been the case in both developing and developed nations. As an example, in Sichuan Province, China, the risk of UIM was almost twice in the rural regions compared to the urban regions17. The same can be said in Canada, where children in socioeconomically deprived and geographically isolated regions experienced greater UIM rates, particularly due to transportation-related injuries40. However, while our findings directionally align with this global pattern (RR = 1.25), the difference did not reach statistical significance (95% CI: 0.97–1.61, p = 0.083). This attenuation of the typical urban–rural disparity may be explained by a combination of methodological and contextual factors specific to Mazandaran Province. The limited number of fatal events inherently reduced statistical power. Additionally, the province’s unique geographical context, characterized by closely integrated urban and rural areas along a narrow coastal plain, likely minimized disparities in risk exposure and emergency access. In Mazandaran, the distinction between rural and urban areas is often administrative rather than functional, due to the high proximity and intermingling of settlements. Furthermore, Iran’s robust primary healthcare network, particularly the dense distribution of rural health houses, may have mitigated traditional rural disadvantages by providing equitable service coverage, community-based prevention programs, and timely initial care. Thus, the non-significant trend observed here likely reflects not only statistical constraints but also the mitigating effects of Mazandaran’s geographical integration, administrative blending of rural and urban zones, and strengthened rural healthcare infrastructure. Nonetheless, the persistent (albeit non-significant) rural excess underscores the need for continued, geographically tailored prevention efforts. These include enhanced drowning-prevention programs in agricultural and coastal areas, as well as sustained investment in rural road safety and emergency services, to address residual vulnerabilities and further reduce child UIDs. Although this study has strengths such as longitudinal design and population-based surveillance, there are various limitations that it has to be noted. First, the use of registry information can create bias of underreporting or misclassification. This is especially true to the types of injuries like suffocation or drowning which can be wrongly reported because of insufficient autopsy or forensic examination in rural areas. Although the U5D surveillance system is based on a standardized national protocol in both urban and rural locations, we cannot completely exclude the contextual biases. In rural locations that are remote or topographically difficult, the distance to health centers, possible delays in reporting of cases, and differences in the workload and experience of community health workers may influence the timeliness and completeness of case documentation than more accessible urban locations. Second, the dichotomous urban-rural dichotomy can conceal heterogeneity in these groups. The peri-urban population, such as that of the peri-urban regions, is usually confronted with the same infrastructural problems as the rural population though termed urban. Also, the injury categorization was not detailed enough on situations and particular processes (e.g., what kind of vehicles were involved in traffic deaths or what was the source of water in drowning cases), which restricted us in making fine-grained sub-analyses. As an example, the overall group of transport accidents does not allow distinguishing between the deaths of pedestrians, cyclists, or vehicle occupants, and each of them would need different prevention measures. Third, even though the study was conducted over a decade, there were certain changes in the data, especially between 2019 and 2021, which might have been caused by external factors, including the COVID-19 pandemic. Nevertheless, the registry system was not designed to make systemic recording of contextual variables (e.g., health service disruptions related to the pandemic), which meant that it was challenging to assign these changes to a particular cause. Fourth, the socioeconomic factors that are known to influence the risk of injuries, including household income, parental education, and housing conditions, were not available, which prevented the use of stratified analyses that would have enhanced the knowledge of the patterns of vulnerability. Additionally, the behavioral aspects of the epidemiology like supervision of caregivers, safety equipment (e.g., helmets or car seats) and community awareness were not quantified, which restrained the behavioral aspect of the study. Lastly, the study is a regional study which is limited to one province and therefore the results may not be completely applicable to other regions in Iran or other countries with different demographics, environment and healthcare systems. Building on these limitations, future research should aim to incorporate more comprehensive data collection systems that capture behavioral, socioeconomic, and environmental variables alongside medical outcomes. Qualitative studies exploring caregiver perceptions, community safety practices, and systemic gaps in prevention could complement quantitative trends and provide deeper insights. Moreover, spatial mapping and geostatistical modeling could help identify local “hot spots” of risk and guide targeted interventions. Finally, collaboration between the health, education, and transportation sectors is essential for developing integrated child injury prevention frameworks at both the regional and national levels. Conclusion The UIDs of children below five years of age in Mazandaran Province decreased non-significantly during the 20,132,022 period, yet it is still a major cause of death that can be prevented. Accidental threats to breathing, such as transport accidents, drowning, etc., demand mechanism-specific interventions. Some priority measures are road engineering (speed bumps, pedestrian crossings), physical barriers surrounding open water sources in high-risk areas, community-based programs of Safety Champion with culturally adapted parental education on supervision, road safety, and safe sleep habits, economic incentives on child restraints, and enhanced inter-sectoral collaboration between health, transportation, and education sectors. These evidence-based strategies can considerably decrease these preventable deaths when implemented in a coordinated manner with assistance of province-level monitoring indicators. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors would like to express their sincere appreciation to the staff of the Deputy of Health at Mazandaran University of Medical Sciences for their invaluable support and collaboration throughout the study. Their efforts in data collection, validation, and management of the surveillance system were crucial to the successful completion of this research. Abbreviations Under-five mortality rate Sustainable Development Goal Unintentional injuries low- and middle-income countries Unintentional injury deaths Annual percentage change Weighted Bayesian Information Criterion Confidence interval Civil Registration and Vital Statistics Relative risk World Health Organization Author contributions R.F., M.S., B.K., L.Z. and M.A.A. were responsible for data collection and the overall methodology. P.E. and M.S. contributed to the initial manuscript drafting, while M.A.J., P.E., M.S., and R.F. were involved in editing the manuscript. All authors contributed to the writing or supervision of the manuscript and gave their approval for the final version. Funding This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Data availability The data that support the findings of this study are available from the corresponding author, uponreasonable request. 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 Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author, uponreasonable request.