Bump outs (aka curb extensions and neckdowns)
One of the increasingly popular methods for “traffic calming” is the “bump out.” This is a short extension of a curb that balloons into the street at an intersection. A motorist approaching an intersection will therefore experience a narrowing of the traffic lanes.
These devices have several factors that might improve pedestrian safety. They reduce the amount of time that a pedestrian is at risk from traffic while crossing a street, because the distance to cross the street at the intersection is shorter. Cars executing turns, particularly right turns, go slower by an average of 2.6 mph. And visibility for both walkers and motorists is improved.
Pedestrian Activated Blinking Signals
Another pedestrian safety device is the “Pedestrian Activated Flashing Signal” (PAFS). These are placed on busy streets where walkers cross without the benefit of traffic signals or stop signs. The pedestrian activates the signal, which is a rectangular structure that alternates between two flashing yellow lights. This alerts drivers to slow down and yield to pedestrians.
The PAFS result in a 67 percent automobile “yield rate,” which is two percentage points better than marked crosswalks that do not have the flashing signals.
Distracted Driving
Distracted driving is defined as any activity that compromises attention to the road. This includes eating and drinking, adjustment of sound systems, talking, and, most prominently, cell phone usage–especially texting.
In 2023, in-car cell phone use contributed to 1.6 million automobile crashes, with 3,275 deaths and 327,819 injuries. These statistics are likely understated, as drivers are reluctant to admit to using a cell phone while driving.
Cell Phones and Pedestrians
Distracted walking involves using cell phones to talk, listen to music, text, etc. while walking. Pedestrians using cell phones are 46 percent more likely to be involved in a crash with a vehicle. Distracted walking was a factor in 2,500 pedestrian deaths in 2021.
Seat Belts
Seat belt (which includes shoulder straps) usage in the US was 91.2 percent in 2024.* This is estimated to save about 15,000 lives per year. 49 percent of passenger vehicle occupants killed in 2023 were unrestrained.
Speed Bumps
Speed bumps reduce automobile speeds by 10 to 15 mph. They reduce the number of high speed drivers from 14 to 1 percent, and result in a significant reduction in child pedestrian injuries.
Blind Spot Monitor Systems
Modern cars have signals in their outside rear view mirrors that warn drivers a car is in the driver’s blind spot. These have reduced lane change accidents by up to 23 percent.
Speed Limits
Speed limits are normally regulated by states, but in 1974 President Nixon signed into law a national speed limit of 55 mph. This was a response to high oil prices, and the regulation ended in 1995.
This law reduced the nation’s consumption of oil, but it also had a huge effect on highway mortality. In 1972, there were 56,278 traffic deaths nationally (26.9 deaths per 100,000 people). In 1993, 41,893 died in traffic accidents (16.3 deaths per 100,000).
Interestingly, traffic death rates have continued to fall over the last two decades, despite the loosening of speed limits. In 2023, there were 13.4 deaths per 100,00 people, despite a significant increase in population and drivers. This is undoubtedly due to ever safer cars, and one wonders what the effect of lower speed limits would have on this figure.
Effectiveness
The deployment of most of these traffic safety devices makes sense, but are they effective? Some of the interventions seem to be wishful thinking. There are many computer simulations that show reductions of pedestrian accidents with, say, pedestrian activated blinking systems, but there are no data based on real world observations. Given the rather low compliance rate with PAFS, I have often wondered whether these signals might actually be a hazard to walkers, as unwitting pedestrians may erroneously rely on the signal at peril to their safety. And the improvement in yield rates with PAFS is marginal at best.
Some interventions for safety are just plain obtuse. In Illinois, for instance, the use of hand held cell phones by drivers has been proscribed (with minimal enforcement, by the way). Researchers agree, however, that the safety threat is not in the manual manipulation of the phone, but rather in the distraction born of conversations.
There is, however, no disputing the effectiveness of reducing speed limits and using seat belts. These are the power interventions for automobile safety. Americans use seat belts in high percentages, but we have eschewed lowering speed limits in the name of speedier travel.
*Air bags are not a substitute for seatbelts; they are a supplement and are not to be used independent of seat belts.
I agree that physicians and public-health professionals should be involved in traffic safety. Motor vehicle crashes remain one of the leading causes of death for children in the United States, killing around thousand children every year and injuring tens of thousands more. At that scale, traffic safety is not just a niche transportation concern. It should be a core public-health problem.
It is always interesting to read someone who is clearly thoughtful and well educated working through an area adjacent to, but distinct from, their primary expertise. I think this piece reflects that. Much of the reasoning is careful, but some of the skepticism around traffic safety measures rests on assumptions that don’t quite align with how risk is managed in complex systems, not merely complicated ones.
In your writing on Red Dye #3 and fluoride, you emphasize population-level effects, exposure, and the reality that public health rarely relies on single, perfectly measurable interventions. You also acknowledge that policy often has to move forward despite uncertainty, because waiting for clean attribution has real costs. That way of thinking maps closely to what safety science describes as the Swiss cheese model. No single intervention layer is sufficient, but multiple imperfect layers together meaningfully reduce harm.
Traffic safety works the same way. Measures like RRFBs are not meant to stand alone, and evaluating them in isolation misses their function. They are most effective with other measures in place with them.
It is also worth noting that their role is to change the interaction at a point of conflict by inserting a pause and a separation, keeping pedestrians out of harm’s way until vehicles slow or stop. Even if they had no measurable effect on aggregate yielding rates (thought they do depending on location installed (https://cmfclearinghouse.fhwa.dot.gov/detail.php?facid=9024)), they would still reduce risk by changing sequencing. The pedestrian waits on the sidewalk rather than stepping into the crosswalk and relying on the driver to comply after the fact, as Illinois law technically requires (Illinois Vehicle Code 625 ILCS 5/11-1002).
Speed is a fundamental risk factor, you correctly identify it as a core issue and that speed limits have historically been effective. Where the analysis needs updating is in how speed limits actually work. We now know that setting a number on a sign does very little by itself. Speed limits are most effective when reinforced by design choices that make the target speed self-enforcing, such as lane width, vertical deflection, and intersection geometry (pmc.ncbi.nlm.nih.gov/articles/PMC4989038/).
You also correctly identify that seat belts are a powerful intervention. Design and enforcement had impact on the uptake of buckling up, but one often overlooked reason they succeeded is the in-vehicle reminder chime, required in U.S. passenger cars beginning in the mid-1970s under FMVSS 208. Its importance is not that it educates or persuades, but that it provides immediate, unavoidable feedback at the moment of risk. The signal persists until the behavior changes, after which protection remains in place for the rest of the trip. From a systems perspective, that matters. It reduces reliance on memory, motivation, or enforcement and shifts the burden onto the environment itself.
That same logic applies to other effective safety measures. Speed humps, for example, do not depend on persuasion or perfect compliance. They physically constrain behavior. Their effect is repeatable, durable, and largely independent of driver intent. Public health is comfortable with this kind of intervention precisely because it reduces exposure directly, even when it is difficult to tie a specific prevented injury to a specific device.
There is little doubt that transportation planning has long suffered from two related failures. As Wes Marshall argues in Killed by a Traffic Engineer, traditional engineering practice prioritized abstract, idealized models of behavior over how people actually interact in real streets. Roads were designed around hypothetical “reasonable users,” even as evidence accumulated that real human behavior is variable, error-prone, and predictably imperfect.
At the same time, transportation was largely insulated from public-health thinking. Despite injury and death occurring at a scale comparable to major diseases, these outcomes were treated as unfortunate side effects of mobility rather than preventable population-level harms. Marshall’s point is not that engineers were malicious, but that systems optimized for throughput and theoretical efficiency systematically discounted human vulnerability.
Modern traffic safety measures emerge directly from correcting those two failures. Roads are increasingly designed for real behavior, with the assumption that people will make mistakes, and with the explicit goal of reducing harm when they do (https://nacto.org/program/reimagining-city-streets/safe-system-approach/). Seen through that lens, layered interventions like RRFBs, speed humps, speed management, and geometry changes look much less like speculative add-ons and much more like standard public-health risk management.
Taken together, these points suggest that modern traffic safety measures are best understood not as isolated interventions to be judged individually, but as parts of a layered system designed to anticipate human error and reduce the severity of its consequences. When viewed through that lens, many of the measures you question align closely with standard public-health risk management rather than standing apart from it.
For that reason, I genuinely think you would find “Killed by a Traffic Engineer” interesting and worth your time. It speaks directly to the tension between traditional engineering assumptions and the public-health framing you already apply so effectively elsewhere. Thanks for sharing your thoughts and for engaging seriously with the topic.
Thanks so much for the thoughtful comment. It was quite educational. I rarely get much pushback, even if it mild. Blog readers will benefit from your piece.