Factors associated with road traffic injury severity among victims retrieved by pre-hospital emergency services in Rwanda

Study design

This study employed a retrospective cross-sectional design using secondary data from the registries of Emergency Medical Services (EMS), known as Service d’Aide Médicale d’Urgence (SAMU). It encompassed all patients attended by pre-hospital emergency ambulance services.

Study setting and data source

Rwanda, a landlocked country in East Africa, has a population of approximately 14 million, with a majority being young people. It is one of the most densely populated nations in Africa, covering an area of 26,338km².²⁶ The study was conducted within the Service d’Aide Médicale Urgente (SAMU), the national pre-hospital emergency care division of the Ministry of Health in Rwanda. SAMU operates in Kigali, capital city of Rwanda and provides emergency transportation and pre-hospital care for trauma victims, including those injured in RTIs. Patient information is systematically documented in a computerized registry developed in partnership with Virginia Commonwealth University, which captures demographic and clinical data relevant to emergency medical care.²⁷ This institution has a diverse team of healthcare professionals, including nurses and anesthetists, all of whom receive training in emergency care and first aid to ensure the delivery of high-quality services. Depending on the severity of the case, their main services also include patient transportation, pre-hospital care, first aid, and client transfer. The drivers, who are also trained in first aid, play an important role by providing safe, timely, and high-quality support during patient transport.²⁸

Participants, sampling, and eligibility criteria

We conducted this research using the SAMU pre-hospital registry to include all patients whose primary injury mechanism was a RTI between January 1 and December 31, 2020. Eligible cases included individuals injured as motor vehicle occupants, motorcyclists, bicyclists, pedestrians, or users of animal-drawn carts on roads. From 2,062 identified cases, 900 were excluded due to missing demographic or clinical data, resulting in a final analytic sample of 1,162 patients (56.5%). Inclusion criteria required that the registry listed RTI as the main cause of injury and that patient records contained complete demographic and clinical information; cases with injuries unrelated to RTIs or with incomplete data were excluded. Furthermore, a census sampling approach was used including all eligible cases within the study period to provide more comprehensive coverage of the target population and minimize sampling bias.

Study variables

The study variables in this study were categorized as dependent and independent. The dependent variable was injury severity, classified using the Injury Severity Score (ISS). This ISS was used as an anatomical scoring system that provide overall scores for patients with multiple injuries.²⁹ The ISS is derived from the Abbreviated Injury Scale (AIS) by selecting the three most severely injured body regions, squaring the AIS scores for these regions, and summing them to produce a total score ranging from 0 to 75.³⁰ In this research, injury severity was dichotomized as either severe (ISS ≥ 9) or non-severe (ISS <9).³¹ This cutoff was chosen based on established thresholds in trauma research, where an ISS of 9 or higher is commonly used to indicate clinically significant injuries requiring advanced medical attention.

The independent variables included a range of socio-demographic factors (e.g., age, gender, and health insurance status), clinical or health-related factors (e.g., primary complaints, trauma mechanism, time of collision), behavioral factors such as alcohol use, and environmental factors including distance from the accident site to the health facility. These variables were analyzed to explore their association with injury severity. The head injury or traumatic brain injury (TBI) measured using Glasgow Coma Scale (GCS) to indicate the level of consciousness. While this instrument show severity of brain injury, we categorised the patients who experienced head injury as severe (≤8) TBI or non-severe injury (>8). In addition, the classification of variables such as trauma mechanism and time of collision was guided by the structure of EMS records and prior studies conducted in sub-Saharan Africa.^11,32,33 These elements are already standardized within the EMS reporting system, ensuring consistency in data extraction and analysis. While previous studies did not categorize distance to health facilities, we created three distance-based categories to reflect access challenges. Distances over 5 km were considered significant due to their potential to delay care, aligning with national goals to improve timely access to emergency services.^34,35

Data collection and materials

Data was collected from SAMU’s electronic data registry between July 13 and November 30, 2022. SAMU staff collected demographic and clinical data using standardized client forms, which were later entered into an electronic registry. Data collection was supervised by SAMU’s pre-hospital team leader and overseen by the study’s primary author. Enumerators received two days of training on the data collection process to ensure accuracy. Data from the electronic registry was exported to the Statistical Package for the Social Sciences (SPSS) software, version 28, for statistical analysis.

Data analysis

Following data cleaning and the removal of incomplete variables, the analysis was conducted in two stages: descriptive and analytical. In the descriptive analysis, statistical parameters such as mean, standard deviation, frequency, and percentage were used to summarize the data. Cross-tabulations were performed to explore the distribution of injury severity across the independent variables. For the analytical phase, bivariate logistic regression was performed to calculate crude odds ratios (cORs) and assess associations between injury severity and each independent variable. Variables found to be statistically significant in the bivariate analysis were included in multivariable logistic regression models using a forward selection method. This approach allowed for the calculation of adjusted odds ratios (aOR) with 95% confidence intervals (CI) for each significant variable, identifying factors associated with severe injury. To address potential multicollinearity, the variance of inflation factor (VIF) was assessed for all independent variables. A VIF threshold of 5 was used; variables with VIF values exceeding this cutoff were considered to exhibit high multicollinearity and were subsequently reviewed or excluded from the model to ensure robustness.^36,37

Ethics

This research did not involve direct interaction with human participants or the collection of new data from individuals. Instead, it employed secondary data obtained from existing medical records and laboratory logbooks. Ethical approval for the study was obtained from the Institutional Review Board of the College of Medicine and Health Sciences, University of Rwanda (No: 233/CMHS IRB/2021). As the study relied solely on secondary data, informed consent from participants was not required. However, access to the dataset was granted after obtaining permission from the Emergency Medical Services (EMS) and data custodians who oversee records of patients. All data were anonymized prior to analysis to ensure confidentiality and privacy. All methods were conducted in accordance with the ethical standards outlined in the Declaration of Helsinki.³⁸

Descriptive analysis

The participants of this study had a mean age of 31.64 years (SD = 11.1), with the largest age group (38%) falling between 20 and 29 years, highlighting a high burden of RTIs among young adults. A substantial majority (78.4%) were male, and over half (55.9%) were uninsured at the time of injury. Most collisions occurred in the afternoon (2:00 pm–7:59 pm; 28.9%), followed closely by morning to early afternoon incidents (8:00 am–1:59 pm; 27.1%), indicating peak risk during active daytime hours. Alcohol use was reported in 24.5% of cases, while 54.9% of participants had no history of alcohol users, pointing to behavioral risk factors among a considerable segment of the population. Regarding clinical presentation, extremity injuries were the most common (46%), followed by TBI at 22.2%, indicating high prevalence of both types. On average, individuals traveled 15.8 kilometers (SD=12.45) to access health services, with 68.8% residing within a 0–20 km radius, suggesting moderate geographic access to care ( Table 1).

Prevalence of severe injury by demographic and clinical information

A total of 1162 victims were considered that the prevalence of severe injury was 14% (n=165), with males accounting for 82.4% of these cases ( Figure 1).

Figure 1. Prevalence of severe trauma.

The highest prevalence was observed among individuals aged 30–39 years (35.8%) and 20–29 years (35.2%). Individuals without health insurance had a significantly higher prevalence of severe injury (61.2%). RTIs occurring between 8:00am and 1:59pm had the highest prevalence of severe injury (38.2%). Extremity injuries were the most common severe complaint (40.6%), followed by traumatic brain injuries (35.8%). Those who were not alcohol users had a slightly higher prevalence of severe injury compared to those who used (35.2% vs 31.5%). Among the mechanisms of injury, car-to-motorcycle collisions had the highest prevalence of severe injury (31.5%) ( Table 2).

Bivariate and multivariate logistic regression analyses for the associated factors of severe injury

The findings from bivariate logistic regression analysis demonstrated that gender time of collision, primary complaints, trauma mechanism, alcohol use, and distance to the treating health facility were significant factors of severe injury among the pre-hospital emergency patients. All significant variables in bivariate logistic regressions were transported into multivariate logistic regression models. After adjusting for other variables, females had lower odds to experience severe injury compared to males (aOR=0.47, 95%CI: 0.26–0.79, p=0.032). In terms of trauma mechanism, we found that car-to-pedestrian collisions had more than twice the odds of severe injury (aOR=2.3, 95% CI: 1.25–4.1, p=0.007) compared to car-only accidents. Furthermore, car-to-motorcycle collisions posed an even greater risk, with nearly four times the odds of experiencing severe injury (aOR=3.88, 95% CI: 1.16–13.05, p=0.007) relative to car-only crashes. Alcohol use was also a strong predictor of severe injury, with alcohol users being more than three times as likely to experience severe injuries compared to non-users (aOR=3.37, 95% CI: 2.04–5.56, p<0.001). Regarding the timing of collisions, individuals involved in RTIs between 2:00 pm and 7:59 pm (aOR=0.52, 95% CI: 0.25–0.72, p=0.02) and 8:00 pm to 1:59 am (aOR=0.25, 95%CI: 0.04–0.81, p =0.012) were significantly less likely to experience severe injury compared to those involved in RTIs between 8:00 am and 1:59 pm. Additionally, the distance to healthcare facilities played a critical role in injury severity, with individuals traveling between 21 and 40 km after an RTI having nearly three times the odds of severe injury (aOR=2.91, 95%CI: 1.27–6.63, p=0.011), and those traveling over 40 km facing a similarly elevated risk (aOR=2.64, 95% CI: 1.11–6.25, p=0.028) compared to those within 0–20 km of a healthcare facility. These findings highlight the strong associations between severe injury and collision type, alcohol use, time of occurrence, and access to healthcare, underscoring the urgent need for targeted trauma prevention strategies to address these risk factors. In addition, the individuals who sustained extremity injuries (aOR=0.28, 95% CI: 0.15–0.52, p<.001), chest injuries (aOR=0.40, 95% CI: 0.23–0.71, p=.002), or other types of injuries (aOR=0.16, 95% CI: 0.05–0.57, p=.004) were significantly less likely to suffer from severe injury compared to those with TBI or head trauma ( Table 3).

Study strengths and limitations

This study has some notable strengths. It utilized a large registry dataset covering pre-hospital care in Kigali, which enhances statistical power and supports broader generalization of the findings. Additionally, this study addresses a less-researched topic in Rwanda, shedding light on a significant public health issue that necessitates the implementation of targeted strategies. The insights gained from this study can inform policy decisions and interventions aimed at reducing the burden of road traffic injuries and improving road safety measures.

Despite its strengths, this study has several limitations that warrant discussion. Firstly, the use of secondary data from registries restricted our ability to include certain key predictors, such as pre-hospital care details (e.g. care provided at the scene, mode of transport to the hospital, use of spinal immobilization or trauma protocols), socio-demographic factors (e.g., marital status, education, employment status, and type of residence), and hospital-related variables (such as length of hospital stay, surgical interventions). Additionally, safety-related variables, such as adherence to traffic regulations, data on safety device usage, seatbelt and helmet utilization, and the duration of possessing a driver’s license before the accident, were not documented, limiting our understanding of factors influencing injury severity. Secondly, environmental factors, including road type and accident location, were not captured, which may have influenced injury outcomes. Moreover, the classification of distance to the health facility was not based on established evidence, which might have introduced bias through possible over- or underestimation. Finally, the retrospective cross-sectional design of the study prevents the establishment of causal relationships between the identified predictors and injury severity.

Public health implications

The findings of this study highlight an urgent need for public health interventions to address road RTIs and their associated consequences. The high prevalence of severe injury among alcohol users reinforces the importance of targeted awareness campaigns focused on reducing alcohol and alcohol use, particularly among young male drivers, who represented a significant proportion of severe cases. Behavioral change programs tailored for this demographic could play a critical role in reducing high-risk behaviors on the road. Moreover, the elevated risk of severe injury in pedestrian and motorcycle-related collisions underscores the vulnerability of these road users. Policy interventions such as creating pedestrian-friendly infrastructure, implementing protected lanes for cyclists and motorcyclists, and enforcing speed reduction zones in high-traffic areas are essential to safeguard these populations. The study also found that injury severity was influenced by the time of day, suggesting a need for improved traffic regulation and law enforcement during high-risk periods, particularly in the early morning and evening hours. Finally, access to timely medical care remains a critical issue, especially for those living in rural or remote areas. Strengthening pre-hospital emergency systems and decentralizing trauma care services to underserved regions will be key to improving trauma outcomes and reducing preventable deaths. These findings highlight the importance of prioritizing high-risk patients within emergency medical services and developing targeted prevention and trauma care strategies to improve patient outcomes and reduce the overall burden of severe injuries. These implications call for an integrated public health approach that combines education, policy enforcement, and infrastructure development. By prioritizing these measures, policymakers and stakeholders can make significant strides in reducing the burden of RTIs and promoting safer road environments.

Future directions

To expand upon the findings of this study, future research should prioritize longitudinal and prospective designs to establish causal links between various risk factors and severe injury outcomes. A more comprehensive investigation of socio-demographic and environmental factors will offer deeper insights into their influence on road traffic injuries. Additionally, exploring healthcare-related aspects including hospital-specific factors could enhance understanding of their effects on injury severity, recovery, and disability.

Given the limited understanding of how prehospital emergency care influences trauma severity, the integration of innovative methodologies like participatory action research (PAR) could prove particularly valuable in preventing RTIs and their outcomes.⁵² By actively engaging community members, stakeholders, and policymakers throughout the research process, PAR fosters a collaborative exploration of local challenges and facilitates the co-creation of interventions aiming sustainable changes especially reducing RTIs. Moreover, employing mixed methods in future research can provide a holistic understanding of the complexities surrounding road traffic injuries. While quantitative data can reveal statistical trends in injury prevalence and associated factors, qualitative methods can uncover the lived experiences, perceptions, and behaviors of road users. This integrative approach allows for a nuanced analysis of underlying issues and barriers to road safety, informing more effective public health interventions. Lastly, collaborative efforts among public health officials, policymakers, and community stakeholders are essential for devising and implementing strategies to reduce road traffic injuries and promote safer road practices. By incorporating PAR and mixed methods, future research can not only guide policy development but also empower road users to actively participate in creating safer environments.