Mobile applications for road traffic health and safety in the mirror of the Haddon’s matrix

According to the World Health Organization report, 1.35 million people are killed due to traffic accidents, and millions are injured annually [1]. Moreover, road accidents have psychological consequences. Based on studies, road traffic accidents have been argued to be the leading cause of posttraumatic stress disorder (PTSD) [2]. As WHO reports on the global status of road safety, despite being among the leading causes of mortality, most traffic accidents are predictable and preventable. There is evidence of interventions that have been effective at preventing traffic accidents; countries that have successfully implemented these interventions have seen a considerable decrease in traffic accident deaths [3].

Considering the scope and pace of technological improvements, people have turned to technologies for solving their various problems [4]. Mobile phones are one of the most widely used technologies. The number of mobile subscribers was 5.2 billion in 2019. It is estimated to have about 5.8 billion unique mobile subscriptions worldwide by the end of 2025 [5].

Smartphone is a culprit for drivers’ distraction and road accidents [6‐10]. One study highlighted that tasks in the smartphones that take drivers' eyes off the road have a greater safety–critical risk than functionalities with no involvement of the eyes off the road [11]. However, another study has shown the conversation on hands-free (HF) and handheld (HH) phones can affect the driver reaction time (RT: as a response with a brake pedal to an event requiring a response in the traffic environment), RT detection, detection percentage, lateral position and speed in a similar way. This can be attributed to the fact that if the phone does not fit in the car automatically, interacting with the HF phone may require manipulating a device. Therefore the driver may require keeping his eyes off the road, which may increase the risk of an accident [12].

Despite being among the main culprit behind road traffic accidents, smartphone technology can facilitate road traffic management practices. Different capabilities of smartphones (such as lane detection, vehicle detection, and vehicle distance estimation) can turn them into potential accident prevention devices [13]. With the widespread use of mobile devices, many companies have recently developed apps to improve public service quality, people security, and safety [14]. Similarly, the development of smartphone technologies has driven the growth of mobile applications in the road traffic field. These apps are used for many transportation-related activities such as education, road traffic data collection, travel information, route planning, and navigation [15].

There are also traffic apps for promoting environmental safety, drivers/pedestrian safety, traffic alerts [16], and driving behavior feedbacking. The world has witnessed progress in these areas. For example, receiving traffic accident information via mobile apps has changed for recording driver behavior. According to recent studies, receiving information about driving behavior through traditional models such as questionnaires, Police report studies, and direct observation has been replaced with new approaches, including [17]:

Using smartphone sensors for recording the driving behavior data of 303 drivers at the road segment and junction level showed that if the average traffic volume per lane increases in the respective areas, the number of rough events in the road segments will increase. In addition, as the average occupancy increases in junctions, there is an increase in harsh accelerations, and as the average speed increases, more harsh decelerations occur [18].

As the number of road traffic apps increases, several issues may arise and influence their use. This increase makes identification of the apps difficult, making their classification inevitable. One way to help researchers and professionals to understand and analyze complex domains is to classify objects [16]. According to evidence [19], any taxonomy should have the following characteristics if it is to be beneficial: It must be brief, enough inclusive, comprehensive, and extendible.

A limited body of knowledge is available on the classification of apps relates to general mobile app classification methods such as hierarchical classification, two-dimensional classification, and three-dimensional classification [20]. There is no classification of apps in a particular domain based on their related scientific frameworks. Only two studies in the literature [21, 22] have presented the mobile app classification in traffic and tourism fields. According to one of the studies, road traffic apps [23] fall into the following three types:

Since each of the apps has a variety of features, characterizing them is crucial for appropriate use. It is essential to identify the whole spectrum of the apps' functionalities and enlist information and communication media used in mobile apps. The researchers characterized Information Feeding System (IFS) used by the apps. The aim was to enlist technologies of capturing information and types of communication media transferring information to end-users. Smartphones with a high-resolution camera, microphone, compass, accelerometer, 3-axis gyroscope, and GPS sensors can collect different data. They can capture traffic information to detect congestion, traffic signals, and rerouting traffic. These devices can also gather environmental information for monitoring road conditions to detect road anomalies and warn drivers of potholes and their location. Moreover, they can sense information about driving behavior to recognize aggressive or nonaggressive driving, drunk driving, lane departure, and assist eco-driving [21].

Despite the increasing trend of the apps on road traffic health and safety, there is no comprehensive analysis and classification of their features based on the traffic safety frameworks. This study aimed to explore the characteristics of available mobile apps on road traffic health and safety and classify them based on Haddon’s matrix as the most comprehensive framework in the area of injury prevention. As a mobile operating system, Android is widely used and has 72% of the mobile market share [23], compared with other operating systems.

William Haddon had worked for many years on road safety in the USA. In 1970, he presented Haddon’s Matrix to the world of injury prevention [24]. Researchers have used the Haddon matrix as a tool for developing ideas for safety promotion and prevention of various injuries or fatalities. The matrix is a framework table with three rows and four columns related to the public health concepts. This framework aims at changing the concepts of primary, secondary, and tertiary prevention [25]. Although Haddon initially presented his matrix for incident analysis at the individual level, this matrix has been used at the level of group analysis and even in methodological studies as a method of qualitative analysis. In the present study, the Haddon matrix is used as a research methodology to provide more practical study objectives and at the same time improve the level of study analysis and facilitate the output of the analysis for users in the field of traffic safety who are familiar with the Haddon matrix [26]. Each row frames the timing of incident respectively as pre-event, event, and post-event phase. While designing a mobile app for primary prevention, we can categorize it as an intervention under the pre-event phase. For example, the apps that provide road traffic rules education and apps that inform about the vehicle status before the trip belong to this category. Analyzing the event phase will be valuable when interested in secondary prevention measures. For example, apps for drowsiness and distraction management relate to this phase. Apps for driving/driver behavior feedbacking belong to the post-event phase of Haddon’s matrix.

A comprehensive review of the mobile applications for road traffic health and safety was carried out using qualitative content analysis. This study included all apps related to the field of road traffic. First, the researchers set eligibility criteria to retrieve the most relevant apps in the search phase. After retrieving the apps, the researchers (HA and LRK) screened them for selecting the appropriate apps. Then data was extracted about the apps and their features. In addition, their functionalities were analyzed. Finally, the apps were classified in general and in terms of Haddon’s matrix framework. Figure 1 illustrates these stages in detail.

In this study, we extracted the features of traffic and transportation apps. If they had health and safety attributes, the researchers categorized them using Haddon’s matrix. According to the app features, we created four categories: Road Traffic Health & Safety (RTHS), Road Traffic Training (RTT), Road Traffic Navigation (RTN), and Other Road Traffic Apps (ORT). Out of the four categories, two groups (RTHS and RTT) had health and safety features. Therefore, we divided them based on Haddon’s matrix. We considered Haddon’s matrix factors as the human, vehicle, physical, and social environment. Then we categorized the apps based on these factors. We also modified the phases of Haddon’s matrix to the pre-driving, driving, and post-driving stages. We finally classified the apps based on these phases. The supplementary file [see Additional file 1] includes details of data extracted from the mobile apps in terms of 4 factors and three stages of Haddon’s matrix.

In this study, we first explored the characteristics of available mobile apps on road traffic health and safety and then classified them based on Haddon’s matrix as the most comprehensive framework in the area of injury prevention. Haddon's efforts aimed at using systems theory to explain ways to reduce the frequency and severity of traffic accidents [27]. We could not consider the Haddon framework as a systems theory approach. However, it reflects the importance of three levels of prevention and the importance of working with all elements of the system, not just the road user, to identify causes as well as preventive actions [28].

By classifying the apps in terms of Haddon’s matrix, we were able to find out which categories of apps can help users before, during, and after driving. Also, we found that each of the apps is related to which Haddon’s matrix factors (human, vehicle, social and physical environment) toward reducing the errors of that factor in driving. For instance, if a driver is drowsy during driving, the drowsiness management app can warn him/her at the right time before getting late. As another example, if there is a problem in the driving route (physical agent error), the apps of real-time traffic information/alerting subgroups can help the driver and prevent possible accidents.

Haddon’s matrix provided us a relevant framework for structured analyses of the mobile apps in the domain of Road Traffic Health & Safety. We classified the apps in terms of 3 different factors of Haddon’s matrix (including human, vehicle/equipment, and environmental factors) and three phases of the pre-accident, accident, and post-accident. When we applied Haddon’s matrix factors to different subgroups, we found that the human factor is the most widely considered factor for digital intervention in mobile applications across all the phases, before, during, and after driving.

According to the results, a few apps have mentioned that they have been developed or supported by traffic and transportation organizations, and most of them were individuals or software companies. However, there were relevant organizations among the developers including, the Abu Dhabi Department of Transport (DOT), Keeping Roads Safe Technologies Inc., Cambridge Mobile Telematics, American Geriatrics Society, National Highway Traffic Safety Administration, Wyoming Department of Transportation, Department of Transport, and Main Roads Queensland. Supporting such organizations may reflect a recognition and perceived importance of mobile apps in road traffic health and safety. Since there is no evidence, we could not judge the quality of apps based on their designers or sponsors. However, it appears that the more we could have an engagement of road traffic safety experts and the more we could use the evidence for designing the apps, we may end up with higher quality mobile apps. Such an approach may decrease the individuals' concern regarding the safety of the apps. During the evaluation of the mobile apps, one may ask if there is any road traffic safety organization and evidence behind the developed app.

Since traffic behaviors are more culturally relevant, they should be considered as a strategic and ongoing goal. Moreover, every possible solution should be used for improving traffic safety. One of the technology-based solutions could be the road traffic apps developed, approved, and recommended by the relevant agencies or authorities. It appears imperative to have organizations in place for examining, validating, and accrediting the apps produced, especially if their developers have some claims about road traffic health and safety, and accident prevention.

We also looked at the communication media used in mobile apps. Communication media that was observable among apps under the distraction management subcategory ranged from the auto-answer to Block cell phone, Silence cell phone, and Screen magnifying. Managing communication media in multi-functional apps appears critical since different tasks such as drowsiness management, distraction management, real-time traffic alerting, driving behavior feedback or other features may require different communication media.

The other issue that matters in app development is the information feed. If credible centers feed information into the mobile application, their validity will be more likely. These centers may include and are not limited to traffic control centers, police, road network administration, emergency medical centers, and insurance companies. It appears the reliable information sources can lead to the effective management of road traffic issues and give confidence to the app users. Evidence shows that some of the apps developed in the field of traffic health and safety (ranging from environmental detection [29, 30], to drowsiness management [31, 32], medical supporting [33, 34], real-time road condition [35, 36], and driving behavior [37‐39]) provide information feeds using mobile sensors.

Smartphones have the potential to prevent possible accidents. Researchers have developed a mobile app that prevents accidents by monitoring, analyzing driving behavior, and advising driver based on unsafe driving behaviors. They used smartphone cameras and internal sensors (accelerometer, gyroscope, GPS, and microphone) to monitor driving behavior [40]. It is important to note that many mobile apps with the functionality of driving safety have been developed. However, there is limited empirical evidence on whether these apps are effective in promoting road safety or not. Furthermore, the evaluation of these apps through clinical trials seems to be inevitable. The field evaluators should assess the effectiveness, efficiency, and impact of these apps on health and traffic safety for the most popular apps in the real world or simulated environment. Moreover, the validity of these apps should be evaluated. Because they are in road traffic safety and their poor performance may cause irreversible injury or fatal results.

Authorities and legislators should regulate the use of road traffic health and safety apps in terms of their approval, encouragement, or prohibition. They should take a position on the validity and use of these apps. These authorities can also provide some domain standards and red lines for the development of road traffic apps. It appears inevitable to define a clearance process and dedicated clearance body for approval of these apps similar to what we have as FDA clearance for medical apps. This can contribute to the safety and accuracy of the apps. Therefore, the user may use them with confidence and trust.

Our study was limited to the apps developed for android platforms and did not include apps developed for other platforms. In most cases, the retrieved apps on Google Play have an iOS version. Despite the fact that both operating systems have sufficient popularity worldwide, the penetration of the Android operating system is more than iOS. Another limitation of our study is that we did not analyze the user comments for the apps included in this study. The results from such analysis could identify the issues related to the usage of apps in the real world. It could also shed light on the unmet needs of the user and reveal their recommendations for consideration in future app updates or developments.

Future studies should examine and evaluate traffic apps on iOS and other existing operating systems. Moreover, it is also recommended to do original research on existing apps and compare their safety and performance of their features. It is also ideal to evaluate the user-friendly of traffic apps with evidence of secondary data in a simulated environment or real world. Future studies should define a suitable and comprehensive final model for the development and evaluation of traffic apps. Such a model can be used as a reference for developing apps that are more compatible with road traffic safety goals and have full features.

Finally, the following users may use the results from this study:

Haddon’s matrix proved useful in categorizing existing mobile apps of road traffic health and safety. It means the mobile apps can be designed and used as an intervention on different factors (including human, vehicle, physical and social environments) and phases (before, during, and after the accident) the Haddon’s matrix. Therefore, mobile apps can be considered as intervention tools in road traffic health and safety. Developers of mobile apps may think systematically about the comprehensive contents for mobile apps in road traffic health and safety mobile considering multiple factors presented in the Haddon matrix. Therefore, this could end up with more meaningful and effective mobile apps in this domain. Since road traffic accidents and injuries are multi-factorial issues, they require multi-factorial solutions.