Measuring motorcycle helmet use is critical for countries to target enforcement and measure the impact of new laws for road safety for motorcyclists. Current methods to measure helmet use are time-consuming and costly and involve roadside observation and review of hospital records. This novel method, using street imagery and crowdsourcing internet marketplaces, has the ability to revolutionize how this life-saving intervention is measured by dramatically reducing time and cost.
Introduction: The majority of Thailand’s road traffic deaths occur on motorised two-wheeled or three-wheeled vehicles. Accurately measuring helmet use is important for the evaluation of new legislation and enforcement. Current methods for estimating helmet use involve roadside observation or surveillance of police and hospital records, both of which are time-consuming and costly. Our objective was to develop a novel method of estimating motorcycle helmet use.
Methods: Using Google Maps, 3000 intersections in Bangkok were selected at random. At each intersection, hyperlinks of four images 90° apart were extracted. These 12 000 images were processed in Amazon Mechanical Turk using crowdsourcing to identify images containing motorcycles. The remaining images were sorted manually to determine helmet use.
Results: After processing, 462 unique motorcycle drivers were analysed. The overall helmet wearing rate was 66.7 % (95% CI 62.6 % to 71.0 %). Taxi drivers had higher helmet use, 88.4% (95% CI 78.4% to 94.9%), compared with non-taxi drivers, 62.8% (95% CI 57.9% to 67.6%). Helmet use on non-residential roads, 85.2% (95% CI 78.1 % to 90.7%), was higher compared with residential roads, 58.5% (95% CI 52.8% to 64.1%). Using logistic regression, the odds of a taxi driver wearing a helmet compared with a non-taxi driver was significantly increased 1.490 (p<0.01). The odds of helmet use on non-residential roads as compared with residential roads was also increased at 1.389 (p<0.01). Conclusion: This novel method of estimating helmet use has produced results similar to traditional methods. Applying this technology can reduce time and monetary costs and could be used anywhere street imagery is used. Future directions include automating this process through machine learning.
What Makes Your Project Innovative?
This study presents a novel method of estimating motorcycle helmet use in Bangkok, Thailand. Using a combination of GSV
and Amazon’s Mechanical Turk, we were able to estimate motorcycle driver helmet use as well as obtain additional information on helmet use based on the type of road traveled, and whether the driver was a taxi driver. This method presents both a cost and time savings compared with traditional methods and could be applied to any region where street imagery is used. This is important considering the number of regions globally without helmet use data.
What is the current status of your innovation?
Recently we published our work in the top injury journal, BMJ Injury Prevention (see attached). We are now actively exploring methods to further reduce the time and cost of this methodology by replacing Amazon Mechanical Turk workers by an artificial intelligence system.
Collaborations & Partnerships
This project involved two public health researchers, a computer scientist and a biostatistician. The idea was developed after seeing the enormous problem of helmet use data under-reporting around the world, and actively working on projects where the methodology of helmet use counting was slow (years) and costly. Our small team of 4 individuals was able to develop this entirely new methodology with funding from Johns Hopkins University School of Public Health.
Users, Stakeholders & Beneficiaries
Beneficiaries of this innovation include every Ministry of Public Health around the world, especially in low and middle-income countries. This important data (helmet use) can now be extracted quickly and easily, a task that is too expensive and time-consuming for many countries to do. It can also be helpful to academics and NGOs working to improve road safety in any country where street view maps are available.
Results, Outcomes & Impacts
Using 12,000 images extracted from Google Street View from September 2001 to December 2016, and then processed by Amazon's Mechanical Turk, we analyzed a total of 462 motorcycle drivers. We were able to determine the overall motorcycle helmet use rate, the taxi driver helmet use rate, and the rates of helmet use by residential and non-residential roads. Overall, the study was successful and we were able to get an accurate helmet use estimate at a fraction of the cost and time compared to traditional methods.
Challenges and Failures
In general, the quality of the images is one significant limitation in Google Street View (GSV). In our study, it is unlikely that a helmet would have been completely missed, however, it is possible that certain types of hats may have been misclassified as helmets or vice versa. GSV is also limited by how frequently areas are updated in the database. Our dataset contained a large range of data from 2011 to 2016 and therefore point prevalence was not able to be estimated. Another limitation of this study is that we were unable to assess the quality of the helmets being worn which is important as substandard helmets can limit their effectiveness. The use of human workers with Mechanical Turk is also a limitation, demonstrated by the large number of false-positive images. This is potentially due to users not understanding the instructions properly as many of these images displayed bicycles. In future research, it would be helpful to do inter-rater reliability testing.
Conditions for Success
Now that we have demonstrated that this novel technology can work in a large city, this project would truly be a success if we could implement this in multiple countries. What we need is publicity and connections to public health officials in countries who would be willing to implement this strategy. This would be particularly useful in a country that currently has no helmet use reporting capabilities. This World Government Summit will allow us to achieve the global visibility we need to form partnerships and scale-up this innovation. It can be easily replicated in multiple countries.
This innovation has not yet been replicated in other countries, but it would be easy to do so. We focused on Bangkok, Thailand as a pilot project, and now that we have been successful, we can implement the project in any city that is mapped out using Google Street View. We would welcome partnerships with any government, public health officials and road safety NGOs to duplicate the project in other cities and countries.
It is critically important to work with people who have a completely different skill set than your own as they will be able to find solutions to problems you may think are too challenging. For example, for our innovation we collaborated with a computer scientist, who in just a few weeks was able to devise a completely automated process to extract thousands of Google street view images. This is not something we could have done on our own, or would have thought to be possible.
The three goals of the World Government Summit: Global Visibility, Platform for scale, and global movement, are precisely in line with what we need to move our innovative project from a research publication to a fully operational public health initiative at the national level.
- Diffusing Lessons - using what was learnt to inform other projects and understanding how the innovation can be applied in other ways
15 July 2021