Our findings from this study confer our hypothesis that given sufficient exposure public transportation in Toronto may pose a risk for noise-induced hearing loss. Both the bus and subway had louder mean Leq noise levels (79.8 +/− 4.0 dBA, 78.1 +/− 4.9 dBA) than streetcars, with subway platforms being significantly louder than in-vehicle subway noise (80.9 +/− 3.9 dBA vs 76.8 +/− 2.6 dBA). Furthermore, if we extrapolate the EPA recommended noise thresholds for an average Toronto commuter using public transportation (47 min), we would find that 9% of subway noise exposure and 12% of bus noise exposure exceeded the recommended 85 dBA threshold.
Our most important finding however may be the frequency of which peak noise levels measured in the public transport system exceeded recommended thresholds. Up to 20% of subway measurements had mean peak noises greater than 114 dBA, and up to 85% of bus platform measurements exceeded that threshold, with 54% greater than 120 dBA. Referring back to the EPA noise threshold guidelines, an exposure longer than four second for a 114 dBA noise exposure, and one second of 120 dBA may place the individual at greater risk of NIHL. Peak noise levels were louder in subway vehicle than platforms (Table 3), however, the loudest mean peak (Lmax) noise was found on the bus stop (120.4 +/− 5.0 dBA). Even if this exposure is measured in seconds, it is well known that impulse noise exposure and repeated trauma from noise exposures at this level may place an individual at greater risk of NIHL [32,33,34]. In fact, animal models suggest that impulse noise exposure may cause hair cell loss more rapidly, and greater hearing threshold shifts than continuous noise exposure [33, 34].
There have only been a few studies looking at dosimetry measurements of noise exposure from public transportation. Neitzel et al. 2009 similarly found that roughly 20% of their subway Leq measurements exceeded the threshold of 85 dBA, however, their mean Lmax noise measurements ranged from 88.0–90.5 dBA, with their loudest capture noise exposure being 102.1 dBA . This is several orders lower than the Lmax captured in our study of 128.1 dBA on a bus stop and 123.4 dBA on a subway platform (Table 3). Our measurements were closer to the measurements found on the Bay Area Rapid Transit system in the San Francisco area, with a mean Leq of 82 dBA, 22% of measurements exceeding the threshold of 85 dBA and majority of routes with over half their measurements with Lmax louder than 90 dBA . Measurements performed in Chicago, also demonstrated routes along the subway system where the noise exposure exceeded the 85 dBA threshold, attributing it to the effects of being in an underground tunnel . In all these transport systems, there is sufficient noise exposure to increase the riders’ risk to NIHL.
Indeed, to adapt and potentially mitigate the level of noise exposure from public transportation, the contributors to loud noise exposure deserve particular attention. Dinno et al. 2011 used a clustered regression analysis to identify train-specific conditions (velocity and flooring), and rail conditions (velocity and tunnels) that may contribute to levels of noise exposures . They found Leq measurements to increase linearly with average velocity by 0.52 dBA/km/h, with the effect tapering to a linear increase of 0.05 dBA/km/h above 53 km/h. Trains traveling through tunnels also increased the Leq by 5.1 dBA, with the type of flooring contributing a small effect to overall mean noise measurements.
Shah et al. 2016 studied the design of New York City subway platforms, finding that overall, curved stations trended louder than straight stations, with Leq noise levels reaching significantly louder intensities at the inbound end of the platform than outbound (89.7 dBA vs 78.7 dBA) . In our study, we found that stations built in the 1960–69 s, when majority of the Line 2 stations were built had louder peak noise levels, whereas the platform design, and location did not play a significant role. It is not known at this time why that decade resulted in subway designs with more intense peak noise exposure, as even older stations did not result in this finding. In addition to the overall layout of the station, there are engineering characteristics such as track curvature, train and rail age, use of vibration reduction methods, as well as environmental factors such as wall material and station size that can contribute to noise exposure while on a subway platform. Specific to train induced noise exposures, engineering studies have described three broad categories of noise: rolling noise, representing the vibration between wheel and rail surfaces; impact noise, representing any discontinuity between the wheel or rail surface; and wheel squeal, representing the friction between wheels sliding against sharp turns [35, 36]. As it may be difficult to address some of the noise derived from existing train paths (curved paths), other endeavours such as the implementation of rail friction modifiers, dampers, and sound barriers may be a more feasible solution [37, 38].
Although most studies have focused their attention on subway transportation, we characterized the noise exposure while using other modes of public transportation including buses and streetcars. To our surprise, although in-vehicle bus measurements mean Leq noise levels were comparable to those previously reported in the New York mass transit system (78.1 +/− 4.9 dBA vs. 75.7 +/− 3.0 dBA), peak Lmax noise exposure were significantly more intense (120.4 +/− 5.0 dBA vs. 87.8 +/− 7.1 dBA).  Certainly, factors such as the distance between the bus stop and the bus play a role, however, with over 85% of bus stop noise level measurements exceeding threshold, more studies assessing the engineering characteristics are required. Recently, the importance of noise exposure within buses has been highlighted by a study demonstrating higher rates of hearing impairment and high blood pressure amongst bus drivers .
One of the strengths of this study, was the broad scope of commuting modalities studied. Noise exposure while driving with speeds up to 100 km/h had a Leq of 67.6 +/− 4.0 dBA with peak noise ranging from 109.6–122.2 dBA. Although no prior studies have reported measurements of in-vehicle noise while driving a closed automobile, a study comparing the difference in noise exposure of a top-open and top-closed convertible automobile also depicted the potential for excessive noise above a certain speed . Interestingly, when personal commuting was measured, biking exposed riders to a louder mean Leq noise level than walking or driving (81.8 +/− 3.4 dBA vs. 73.9 +/− 5.4 dBA, vs. 67.6 +/− 4.0 dBA). This also held true for mean peak noise exposures (Table 3). Although the sample size of this was low, and focused around the downtown core, a study mapping out the noise exposure of over 85 bicycling trips in Montreal supported our finding of the potential for significant noise exposure during morning peak traffic hours as well . In general, cyclist have shorter commute times than those using public transit or personal vehicles (Table 4), however, their exposure to louder peak noise also suggest they may benefit from hearing protection. Complicating this decision lies in the fact that hearing is essential for cycling road safety. Other strategies such as developing dedicated bike lanes in low-traffic areas should thus be considered.
Our findings add to the body of literature demonstrating potential sources of noise exposure while commuting. Criticism of these studies have revolved around the cross-sectional design which preclude causality. One study that has attempted to address this gap administered an extensive self-administered questionnaire to over 756 study participants in New York City, finding that at least approximately 32% of participants frequently experienced symptoms suggestive of a temporary threshold shift after using the mass transit system . They also found that two-thirds of their participants reported the use of MP3 players or stereos with an average use of 3.1 h, and that only 14% of participants wore hearing protection at least some of the time while using the mass transit system. When these factors, as well as others were added to their logistic regression model, the only significant predictor for a temporary threshold shift after riding was heavy transit use (OR = 2.9), and female gender (OR = 2.7). Overall, more studies characterizing the impact of concurrent use of MP3 players and lengthy transit times, as well as definitive audiometric evaluation of transit users would continue to clarify the relationship between transit noise exposure and hearing health.
Aside from the cross-sectional design, other limitations of our study include the lack of modeling of other potential factors that may contribute to noise exposure for personal transportation modalities, as well as buses, and streetcar. Although we chose the busiest routes for streetcar and bus modalities of transportation, the relative sample size may be relatively low and may not represent the entire sprawling Toronto transit system. Despite these limitations, these findings still illustrate that the potential noise exposure for Toronto commuters add to the risk for the development of NIHL, not to mention the other adverse health effects from excessive noise.