Potential Health Impact of Switching From Car to Public Transportation When Commuting to Work

Abstract

We assessed humidity-corrected particulate matter (PM_2.5) exposure and physical activity (using global positioning system monitors and diaries) among 18 people who commuted by car to Queens College, New York, New York, for 5 days, and then switched to commuting for the next 5 days via public transportation. The PM_2.5 differed little between car and public transportation commutes (1.41 μg/M³·min; P = .226). Commuting by public transportation rather than by car increased energy expenditure (+124 kcal/day; P < .001) equivalent to the loss of 1 pound of body fat per 6 weeks.

In 2007, the US population took an estimated 10.3 billion public transportation trips, a 32% increase compared with trips taken in 1995.¹ If sustained, this behavioral change may impact health favorably, by increasing physical activity.

Increased use of public transportation can potentially generate health benefits from the persistent aerobic physical activity that results from walking and climbing stairs when one is riding buses and trains, and from moving to, from, and within stations.^2–6 To determine the magnitude of such effects if car commuters switch to public transportation, we compared personal exposure to PM_2.5 and levels of physical activity between car and public transportation commutes to work.

METHODS

Between October 27, 2008, and May 29, 2009, 18 of 21 recruited participants continued commuting by car for 5 days and then switched to public transportation for another 5 days, all while carrying the Forerunner 305 (Garmin, Kansas City, KS) Global Positioning System (GPS) receiver and an AM510 SidePak (TSI, Shoreview, MN) aerosol monitor during the 10-day commute.

Eligibility criteria, equipment, measurements, and analysis had previously been described in detail.⁶ The only major change was the addition of an air drier jacket to the AM510 SidePak to prevent the artificial increase in particle detection resulting from the high humidity levels of New York City air.

To match the particulate matter of a diameter of 2.5 microns or smaller (PM_2.5), and GPS data, each volunteer was required to maintain a time–activity diary with preprinted, minute-by-minute time and activity columns.⁶ During car commutes, participants used a hands-free digital voice recorder (Model ICDP520, Sony, Los Angeles, CA), worn with a neck strap, to dictate diary information as they drove. All digital voice recordings (for the car days) or print copies (for the public transportation days) were completed.

We assessed commute-specific energy expenditures based on GPS tracking and diary entries.⁶ The GPS device failed to record waypoints for 7.3% of the segments, mostly at the beginning of the commute, while the GPS receiver was searching for a satellite connection. These missing waypoints were easily imputed on the basis of commutes with complete recordings. We used conventional metabolic equivalents (METs; 1 MET = 1 kcal/kg of body weight/hour) for various modes of activity.⁷ Travel by subway was assigned a MET of 2.0.⁶

Geometric means from log-transformed PM_2.5, 95% confidence intervals, and statistical tests were computed from the data recorded on a 1-minute interval time base, with variance corrected for the clustering of observations within participants, by using the SAS version 9.1 procedures Surveymeans and Surveyreg (SAS Institute, Cary, NC). The commute-specific energy expenditure was computed as described previously.⁶

RESULTS

On average, the 7 men and 11 women were aged 31 years, were 66 inches (167 cm) tall, and weighed 159.5 pounds (72.3 kg); two thirds of the participants were White. The group difference in PM_2.5 over the 5 commute days by car (5.60 μg/M³·min) and 5 commute days by public transportation (7.01 μg/M³·min) was weak (1.41 μg/M³·min; P = .226; Figure 1).

FIGURE 1.

Boxplots of the logarithm of exposure to particulate matter of a diameter of 2.5 microns or smaller (PM_2.5; μg/M³·min) for (a) morning work commutes, (b) evening work commutes, and (c) all day: New York, NY, October 27, 2008–May 29, 2009.

Note. Five-day averages of each commute are presented. Boxplots depict the 25th and 75th percentiles (edges of the box), the median (line within the box), and the 2.5th and 97.5th percentiles (whiskers) Circles indicate outliers.

The excess cumulative energy expenditure for public transportation commutes of 622 kcal (P < .001; Table 1) corresponds to 622 kcal / 5 days = 124 kcal/day, or 124 kcal/day × 30 days = 3720 kcal/month, and would amount to approximately 1 pound of body fat over 6 weeks, assuming 5 commutes per week and 3500 kcal per pound of fat.

TABLE 1.

Physical Activity Energy Expenditure, Duration, and Walked Distance for 5 Days of Car or Public Transportation Commute to Queens College Among 18 Participants: New York, NY October 27, 2008–May 29, 2009

Variables	Car	Public Transportation	Paired Difference	t; P
Cumulative energy expenditure, kcal				5.76; < .001
Mean	612	1234	622
SE	51	150
Median	598	1117
Total duration, min				4.44; <.001
Mean	321	545	224
SE	28	61
Median	284	520
Metabolic equivalents, (kcal/kg·hour)				3.84; .001
Mean	1.61	1.88	0.27
SE	0.04	0.07
Median	1.58	1.74
Energy expenditure per minute (kcal/min)				3.5; .003
Mean	1.94	2.24	0.3
SE	0.10	0.10
Median	1.96	2.17
Total distance walked, km				5.8; <.001
Mean	3.42	9.82	6.4
SE	0.50	0.96
Median	2.89	9.32
Cumulative number of steps				5.8; <.001
Mean	5682	16 093	10 411
SE	871	1517
Median	4739	15 919

DISCUSSION

This study has demonstrated the feasibility of having car drivers switch to public transportation when they commute to work. The extra energy spent by public transportation commuters in this sample could amount to substantial weight loss if the public transportation commute were sustained over 6 weeks. The humidity-corrected excess exposure to PM_2.5 for public transportation versus car commute was modest, not statistically significant, and unlikely to exceed the current recommended threshold of 15 μg/M³ per year.⁸

In the absence of a direct subway line to the Queens College campus, public transportation commuters are forced to use a combination of buses and subways, which makes public transportation commutes substantially longer (median = 104 min/day) than car commutes (median = 57 min/day). Also, the suburban location of Queens College—where, by Manhattan standards, traffic is fluid and parking spaces abound—resulted in shorter commute times by car than if the college had been located in Manhattan. Thus, the advantage of using Queens College commuters was a greater statistical power in detecting public transportation–car differences. On the other hand, it is probable that workplaces located in dense urban environments with rich public transportation service, slow traffic, and poor parking availability may produce longer car commute times and shorter public transportation commute times when compared with those for Queens College. In a denser urban environment, both PM_2.5 and energy expenditure differences may be weaker, but switching to public transportation may also be more sustainable and may provide long-term benefits.⁹

Because of budgetary constraints, our exposure assessment focused solely on PM_2.5, just 1 of the many air pollutants present in urban environments.^10–15 Noise-induced hearing loss among riders of bus lines and subways may also be a concern.¹⁶ Our assessment of energy expenditure could be improved by using a pedometer integrated to a GPS receiver, which would provide time-stamped stride.

Because there were much greater differences in energy expenditure than in exposure to air pollution, we conclude from this analysis that the physical activity benefits associated with using public transportation to commute to work probably outweigh the risks associated with the greater exposure to PM_2.5.