The IIT-Delhi study on Delhi bicycle crashes presents a problem that extends beyond the number of collisions recorded by the police. Its central finding is that only 2.8% of bicycle crashes identified through the study were reported to police. The gap suggests that official records capture only a small part of the risks faced by people cycling in the capital, while also showing how the identity of the striking vehicle changes depending on which data source is used.
The study, titled “Characteristics and Incidence Rates of Self-Reported Bicycle Crashes in Delhi: A Cohort Study”, examined crash experiences through self-reported data collected from cyclists in Delhi. Researchers from IIT-Delhi’s Transportation Research and Injury Prevention Centre conducted on-road intercept surveys involving 844 cyclists at 27 locations last year. Of those surveyed, 742 cyclists provided crash and exposure information during at least one follow-up and were included in a prospective cohort followed for an average of 4.8 months.
The resulting evidence points to a road-safety issue that is difficult to understand through police records alone. Among 218 self-reported crashes, motorcycles were the striking vehicle in half of the cases. Cars accounted for 25%, while electric three-wheeled rickshaws were involved in 11%. Only 2% of the self-reported crashes were single-bicycle incidents, including loss-of-control events, mechanical failures and collisions with objects or potholes.
This distribution differs substantially from the profile of police-reported crashes in the study. Police records contained a higher proportion of crashes involving trucks and a lower proportion involving motorcycles. The contrast does not necessarily indicate that one dataset is incorrect. Instead, it shows that different reporting systems may be documenting different parts of the cycling-risk environment.
Police records are more likely to include crashes that become formal cases, particularly those involving serious injury, medical treatment or larger vehicles. Less severe collisions involving motorcycles may be settled informally, ignored or not reported at all. The study’s finding that just 2.8% of bicycle crashes were reported to police makes that limitation particularly significant for urban transport planning.
The researchers linked the prominence of motorcycles to their widespread presence in Indian cities and their use of kerbside lanes, where cyclists often travel, especially during congestion. This places cycling and motorised two-wheelers in direct competition for limited road space. The finding also complicates a road-safety approach that focuses mainly on high-severity collisions recorded by enforcement agencies.
The study does not establish that every unreported crash has the same cause or severity. It does, however, provide a clearer picture of the types of encounters that may disappear from official statistics. Half of the self-reported crashes involved motorcycles, while the police-reported cases had a greater share involving trucks. A database based largely on police complaints could therefore understate frequent interactions between cyclists and motorcycles while giving greater visibility to crashes involving heavier vehicles.
The injury data adds another layer to the findings. Among self-reported crashes that resulted in injury, 78% involved an injury and 67% of those injured sought medical treatment. Knee and lower-leg injuries made up 60.4% of reported injuries. Wrist, hand and finger injuries accounted for 27.8%, while head injuries represented 14.8%.
These patterns show that cycling crashes do not need to be fatal or involve a large vehicle to impose a meaningful health burden. Injuries to the legs, knees, hands and wrists can affect mobility, work and daily travel even when a crash does not enter the formal police system. The available material does not provide a detailed breakdown of lost work, household costs or long-term disability, but it does show that medical treatment followed a substantial share of self-reported injuries.
The study also reports a crash rate of 28.6 injury-causing crashes per million kilometres cycled. It states that this rate was 60 to 200 times higher than rates reported in Denmark, Germany and the Netherlands, countries described in the study as having extensive cycling infrastructure. The comparison is presented as evidence supporting the role of infrastructure improvements in reducing cyclist injuries.
The study’s international comparison must be read alongside its other findings. The researchers noted that single-bicycle crashes accounted for 40% to 70% of bicycle crashes in findings from Australia, Europe and the United States, whereas such crashes made up only 2% of the Delhi self-reported cases. They attributed this difference partly to lower cycling speeds in Delhi.
That explanation matters because it shows that cycling risk is shaped by local conditions rather than by a universal pattern. In Delhi, the dominant risks identified in the study are connected to interactions with other road users, particularly motorcycles. In other settings, falls or loss-of-control incidents may form a much larger share of the recorded crash burden. Infrastructure and safety interventions therefore need to be informed by the actual mix of conflicts present on local streets.
The study’s method is important to its interpretation. Researchers did not rely solely on records generated after a crash. They recruited cyclists through on-road intercept surveys and collected information through follow-up reporting. This allowed them to capture incidents that may never have reached the police. At the same time, the study covered cyclists surveyed at 27 locations and followed participating cyclists for an average of 4.8 months, so its findings describe the observed cohort and survey locations rather than every cycling trip in Delhi.
That distinction does not weaken the central reporting finding. It clarifies what the research contributes: a structured account of crash experiences that would otherwise remain largely invisible in administrative data. The study provides both an estimate of under-reporting and a basis for comparing the profile of self-reported crashes with police-reported incidents.
For urban governance, the institutional question is how such evidence should enter road-safety planning. Police data remains important for identifying serious crashes, enforcement needs and locations where formal complaints are concentrated. But the IIT-Delhi findings indicate that it cannot be treated as a complete measure of cycling risk. Self-reported surveys, medical information and other forms of local evidence may be necessary to identify frequent but less formally documented conflicts.
The policy issue is also tied to the allocation of road space. The researchers specifically identify kerbside lanes as spaces used by cyclists and motorcycles, particularly during congestion. The study does not prescribe a particular street design or enforcement programme, but its findings identify the interaction between these road users as a context-specific safety concern. Any assessment of cycling conditions that excludes this interaction would leave out the most common striking vehicle identified in the self-reported data.
The evidence also challenges the assumption that the absence of police cases indicates an absence of serious urban risk. If most bicycle crashes are not reported, a decline or low level of police-recorded incidents cannot by itself demonstrate that cycling conditions are safe. Conversely, the study does not provide enough evidence to calculate the full number of unreported crashes across Delhi or to determine whether reporting rates are identical across neighbourhoods, cyclist groups or road types.
What the study confirms is narrower but consequential. Delhi’s official crash picture differs from the experience reported directly by cyclists. Motorcycles dominate the self-reported striking-vehicle data, while trucks are more prominent in police records. Injuries are common among the reported crashes, and the estimated injury-crash rate is substantially higher than the comparison rates cited for countries with extensive cycling infrastructure.
The larger urban question is whether Delhi’s road-safety systems are measuring the risks experienced by all road users or mainly the incidents that become formal cases. The IIT-Delhi research suggests that cycling policy requires both administrative records and direct accounts from cyclists. Until those sources are read together, the city’s understanding of bicycle crashes will remain incomplete. Further attention will need to focus on how crash reporting can be strengthened and how cycling infrastructure responds to the motorcycle–bicycle conflicts identified by the study.
