“Data are just summaries of thousands of stories—tell a few of those stories to help make the data meaningful.” – Dan Heath
One morning Charlotte turns on the radio and hears that fatal collisions have risen by 30%. Later, she reads an online story claiming the fatal collision rate has dropped 20%. Confused? That’s understandable. So, this blog explores the importance of road data and its role in making communities safe.
Perhaps the news story Charlotte heard reported local or provincial data while the online news release shared national data. Or perhaps one story compared changes between 2022 versus 2021 while the other compared changes over a decade. It could also be that one story compared the number of collisions while the other examined fatal collision rates per 100,000 population. This blog explains how data are collected and reported, who relies on these data, and how the privacy of victims and their families is protected.
The journey from data collection to reporting
When it comes to data collection, interpretation and use, the devil is in the details as the saying goes. Collision data are an essential source of information to help measure road safety issues, track trends, and guide educational and enforcement campaigns in communities across Canada. As Research Associate & Data Collection (a.k.a. Steve), my role at TIRF is to collect and analyze these data. Each year I travel to several provinces, visiting coroner’s or medical examiner’s offices to collect it. But this is only the beginning of my work. One could say what comes next is where the rubber meets the road.
After data are collected, they are cleaned to remove impossible values (e.g., a driver age recorded as 212 instead of 21). Then comes the analysis to explore topics of interest including age, sex, prevalence, and the frequency of various risks among other factors. Results are applied to specific road safety topics from alcohol- and drug-impaired driving to distracted driving, to fatigue and speeding. These findings enable TIRF to answer important questions such as:
- Do more young drivers die in fatal crashes than drivers aged 40 and older?
- Are more people killed on the road as a result of impaired drivers or distracted drivers?
- Which road users pose the most risk?
- Do more crashes involve passenger vehicles or commercial vehicles?
The final step is to organize these data to create fact sheets and reports which present data in meaningful and relevant ways. This can help the public understand risk factors and better protect themselves, as well as enable decision-makers to make informed choices about road safety policies and programs.
More importantly, data are used to shape safe communities. Sharing findings with decision-makers helps to guide municipal policies and traffic enforcement strategies. It also help to identify at-risk populations, crash types, or crash locations which can enable engineers to design safe roads with protective elements to reduce crashes.
Data are an incredibly powerful tool in the proverbial Road Safety Toolkit. It’s pretty hard to solve a problem if you don’t know what it is; or can’t prove to others the problem exists. But we all know that with great power comes great responsibility (thanks, Stan Lee). That means data must be accurate, timely, precise and complete, and interpreted with consideration of the context of those data. Not only are these data essential to tell us if problems are getting worse or better, but also to tell us if the investment of resources to implement solutions is working. For instance, the effective use of speed cameras requires that cameras be placed in locations where crashes occur, or where speed measurement devices tell us many drivers are exceeding the speed limit on a particular road.
Why is TIRF’s National Fatality Data often cited & requested?
Each year, TIRF receives numerous requests for aggregated and anonymized [for more on this, see Data protection a bit further on in this post] collision data, particularly fatal collisions. This data source is considered the gold standard in research because it is the most complete and accurate, making it highly relevant to diverse stakeholders working to improve road safety. Organizations that frequently rely upon or request our data include:
- Governments (national, provincial, local)
- Police services
- Public health agencies
- Media (print, TV, radio, online)
- Universities (for research purposes)
Representatives of these organizations requesting data may want to investigate national or provincial/territorial trends in road safety. They may plan to develop a road safety campaign for individuals most at risk, or they may want to conduct a research study on a particular road safety problem (e.g., young drivers, males, motorcyclists, victims dying in crashes in July). Requests may also specify certain years, demographic characteristics, or types of vehicles.
And, then there are people like Charlotte who are just listening to the news and trying to figure out whether cycling is safe in her community, or where crashes are most likely to occur so she can make safe choices when choosing a route to her destination.
From the early days of data collection to now
As the person who has managed TIRF’s National Fatality Database since the 1990 data year, I am directly or indirectly involved in answering these queries. Presently, much of the data we collect are provided electronically through secure channels. However, when I was a rookie, we sometimes received large magnetic tapes containing collision data. We would deliver these tapes to a location with a large, sterile room full of mainframe computers that could read these tapes, often resembling the set of a James Bond movie (think Sean Connery/Roger Moore era). I would hurry back to the office where the data were sent to us via telephone, hooked up to our old dot matrix printer. Spoiler alert…I am, in fact, older than Google.
Our staff use TIRF’s National Fatality Database to provide data to those who may be interested in some quick numbers. For example, a journalist may be working on a story about ATV fatalities and looking for the number of ATV riders killed annually. In addition, this data set is used by TIRF to produce reports, peer-reviewed articles, and fact sheets on a whole host of road safety issues.
Our database is unique because it is the only database in Canada containing both police-reported collision data and data from coroners’/chief medical examiners’ offices. Police-reported collision data provide information about the date, time, and location of collisions as well as the types of vehicles involved, use or non-use of safety equipment, and contributing factors like weather or road conditions. Data from the coroner or chief medical examiner provide TIRF with contextual details such as when the person died, if they were positive for alcohol or drugs, and the type of injury causing death (e.g., head injury, spinal injury, submerged in water).
To this day, some coroner/medical examiner data collection requires on-site visits to their offices while other offices send data electronically. The timeliness of preparing final datasets may take a year or two after a collision, at least for access to third-party researchers. I feel dutybound to report that, unlike the CSI effect shown on TV and in movies, data are not magically available within a one- or two-hour time frame.
And for the most part, the facilities we visit are not staffed by an army of people in scrubs. Many of these offices are in shared spaces staffed by a handful of professionals. It also bears mentioning that while road safety is paramount to us and our colleagues/clients, the staff at a coroner’s/medical examiner’s office deal with a wide range of issues and requests. I have met researchers on-site collecting data about homicides, suicides, overdoses, and boating mishaps.
Similarly, transportation agencies or provinces with public insurers that provide us with police-reported collision data may be engaged in undertaking their own analyses to guide policies and programs, as well as fulfilling research requests from those interested in photo radar, licence suspensions, or claims data.
Data protection
So, before we get any further into data collection, it’s important to emphasize that privacy protections for victims and respect for their families are paramount throughout the entire data collection and reporting process.
Our means of storing and securing data and longstanding working relationships with transportation agencies and coroners/medical examiners has ensured that TIRF has historical access to motor vehicle fatality data, and that data provided by other parties are safeguarded against privacy violations. TIRF often signs research agreements with these agencies specifying data protection protocols and the handling of data. There are also some agencies who request that TIRF obtain research ethics approval. Furthermore, some offices we visit require us to sign a non-disclosure agreement to protect the confidentiality of deceased persons and the circumstances of their death
Most importantly, data are anonymized once they are cleaned meaning personal information about victims (name, date of birth, address, driver licence number) are stripped from the data set. Such variables are never included in TIRF’s database files.
Secondly, data are reported in a manner that protects the identity of individuals, particularly when very small numbers are reported. If, for example, there are seven fatally injured drivers aged 16-19 but only one aged 20-25 in Province A in 2020, we combine these two age categories into a new one (16-25). Similarly, if a table is produced showing the number of fatally injured pedestrians by jurisdiction and there are only two in Province B and three in Province G, we will create a combined line in the table for Provinces B and G.
In addition, our reports and fact sheets may contain data about types of vehicles (e.g., automobiles, large trucks) but the categories are non-specific (i.e., vehicle make and model aren’t included). We would never want Charlotte to read a fact sheet and conclude, “I knew it! Uncle Roscoe was three sheets to the wind.”
To summarize, TIRF’s expertise working with and safeguarding these data for more than 60 years, our relationship with these agencies providing the raw data, and the care we take to protect the privacy of individuals whose information is never included in our dataset and reports has allowed us and other stakeholders to utilize these valuable sources of information to make roads safe.
The Sherlock Holmes approach
Typically, those who are identified as fatalities in the police-reported data also appear on the list of individuals provided by the coroner/chief medical examiner. As one might expect, sometimes a victim only appears in one data source but not the other. This is where TIRF researchers do some data-sleuthing. Reasons that a motor vehicle fatality will show up in police-reported data source but not coroner data may include:
A person crashes in one province but dies in another. Those injured in serious collisions are transported to the nearest hospital. Or they may be treated at one hospital and subsequently transported to a second hospital. In our National Fatality Database, there are examples of persons who crashed in southeastern BC but died in Alberta, those who crashed in Northwestern Ontario and died in Manitoba, or those who crashed in PEI but died in Nova Scotia.
A person dies several days after the crash. Data from the coroner/medical examiner may reveal that a person in a crash that occurred on the 18th of the month but died on the 27th. Police-reported collision data may merely show that the person was seriously injured. Similarly, the coroner/medical examiner may not receive notification that a person injured in a collision died several days or weeks later.
In fact, sometimes a person can die in a different calendar year than the year in which they were involved in a collision. For example, Mr. X was injured in a crash in December 2010 but died in January 2011. The police will list Mr. X as a 2010 fatality, but the coroner/medical examiner will list Mr. X as a 2011 fatality.
In all of these types of incidents, TIRF’s access to both police-reported collision data and coroner/medical examiner data minimizes the possibility that cases will be unmatched or either omitted from or duplicated in the Fatality Database.
Data sources are contradictory. On very rare occasions, there may be discrepancies in the deceased person’s age, sex, date of crash and/or date of death, or position within the vehicle. On even rarer occasions, in a collision with several casualties, the wrong person is listed as deceased.
Whenever difficult cases must be reconciled, we have been very fortunate to receive assistance from our contacts who either provide police-reported collision data or coroner/medical examiner data.
A few years ago, there was a fatality that appeared in fatal collision data for two provinces. This person was involved in a collision on a roadway between two provinces. Fortunately, there was a detailed narrative description of exactly where the collision occurred, and we corrected our database to ensure this person was not counted twice.
So, how does data analysis create safe roads?
As mentioned earlier, one of these reports is the Alcohol and Drug-Crash Problem Report sponsored by the Canadian Council of Motor Transport Administrators. A section of the report is dedicated to data for Canada as a whole. In addition, there are sections for each Canadian province and territory, and these contain descriptions of the number of persons dying in alcohol-related crashes, the number of drivers killed who were drinking and/or positive for drugs, the number of drivers involved in alcohol-related serious injury collisions, and historical trends in alcohol and drug-involved fatal collisions. The reader can either look at national numbers or just their own jurisdiction.
TIRF’s National Fatality Database enables time series analysis which involves an analysis of data to examine points of data at consistent intervals (e.g., by year, month). Annual or monthly analyses of fatality data can be used to look at trends for certain age groups, vehicle types, or seasonal variations. For example, several provinces were provided with tables derived from TIRF’s Fatality Database showing the number of young drivers killed per month before and after graduated driver licensing programs were implemented. More recently, two areas of interest in road safety have been 1) the legalization of the recreational use of cannabis in Canada and 2) the effects of the COVID-19 pandemic on motor vehicle fatalities. These two topics will require further investigation to identify any lasting trends.
Road safety issues have become more diverse and complex in the last two decades. Thirty years ago, requests were related to the role of alcohol and use/non-use of safety equipment in fatal crashes. Today, requests span distracted driving, speeding, fatigue, daylight saving time, wildlife-vehicle collisions, and active road users (cyclists, pedestrians, operators of e-scooters and/or e-bikes). With this in mind, the National Fatality Database has evolved. More variables and values have been added to ensure TIRF can study and track new areas of concern.
So, the next time Charlotte hears a news story about trends in road deaths, she can better understand how the data shared are relevant to her community. It’s also important for Canadians to realize, that while data are a bunch of numbers, each statistic represents a person with family, friends and a community.
Every choice people make on the road can help ensure everyone gets home safely, and people don’t become a statistic in a preventable collision.
All TIRF’s National Fatality Database fact sheets, available in English and French, can be easily downloaded from our website project page: https://tirf.ca/projects/the-national-fatality-database/
#MySafeRoadHome authors: Steve Brown, TIRF Research Associate & Data Collection manages TIRF’s National Fatality Database and has co-authored many publications on road safety topics including distracted driving, alcohol and drug-impaired driving, commercial vehicles, and vulnerable road users. Robyn Robertson, TIRF President & CEO is a criminologist with 20 years of experience in road safety research, and the author of TIRF’s knowledge translation model. Robyn authored TIRF’s knowledge translation model and is well-versed in implementation strategies and operational practices across several sectors. Karen Bowman, Director, Communications & Programs, has a background in writing and program development. Since 2010, she has led TIRF’s award-winning Drop It And Drive® program which has been delivered to over 60,000 youth and workers across North America.
 |
 |
 |
 |
