Introduction

Most adults are insufficiently physically active [1, 2], i.e. below the recommended 150 minutes of moderate-to-vigorous intensity physical activity (MVPA) per week [3, 4]. Cycling for transport may be a time-efficient way to integrate physical activity (PA) into daily routines, possibly increasing total PA levels [5, 6]. As transport cycling commonly reaches moderate intensity [7], enhanced cycling could prevent non-communicable diseases and decrease mortality [8–10]. When accounting for inhaled air pollution and traffic accidents, the health benefits for individuals shifting from car to bike substantially outweigh the risks [11], with large societal benefits as well [12, 13]. In addition, replacing motorized modes of transport with everyday cycling has positive environmental effects, through a decline in greenhouse gas emissions [12, 14], noise and pollution [6].

Electric assisted bicycles (e-bikes) have emerged as a new mode of transportation, potentially health enhancing, which is attributable to overall increased levels of PA [15, 16]. Although entailing lower intensity than traditional bikes, e-bikes are found to attain the prescribed MVPA-standard in both active and inactive individuals, as well as in healthy adults and patient groups [17]. E-bikes enable speed maintenance with less effort, thus overcoming typical barriers to traditional pedal cycling like poor fitness, lengthy distances, hilly terrain, lack of time, and lack of end trip facilities (e.g. changing rooms and showers) [18]. Consequently, e-bike users tend to cycle more often and over longer distances [18–21].

Cargo bikes might provide benefits in terms of the transport of equipment and children due to substantial carrying capacity [22]. Because of the weight, it is likely that human-powered cargo bikes imply a higher PA intensity than traditional bikes, which in turn could generate additional health effects [23]. American data report a decline in car travel among cargo bike owners [22], illustrating the car-substitution potential. Although men are repeatedly reported to cycle more than women [24], recent Swiss data indicates that males and females find cargo bikes equally appealing [25]. Besides, if conditions and circumstances are favorable, the cargo bike has been shown to function as a substitute for the car for American mothers with access to one [26]. Nonetheless, cargo-bikes still belong to a niche-market, and few studies on them have been conducted.

In Norway, the prevalence of bike possession is approximately 75%, yet the annual cycling share is only 5%, with seasonal variations ranging from 1% in the winter to 8% in the summer [27]. Hence, it should be manageable to drive less and cycle more. In 2014, 59% of all trips between one and 2.9 kilometers (km) were conducted by car and 8% by bike, while the average distance of bike trips was 5.1 km [27]. However, for parents of young children, most factors influencing mode choice seem to facilitate car use, e.g. both parents being in paid work, organized leisure activities (for older siblings), increased perceived time pressure and enhanced car access [28]. Traveling to kindergarten by bike rather than by car could potentially teach children that alternative modes of transport exist, thereby encouraging future healthy and sustainable transportation and PA habits.

Understanding the essential determinants for the behavior or the activity of interest would increase the chances of developing effective intervention schemes and persistent behavior change. Travel has traditionally been considered externally motivated, e.g. as a means to move from one destination to another, but intrinsic motivation has gained increased attention [29]. According to the Self-determination Theory (SDT) [30], the degree to which the basic needs of autonomy, competence and relatedness are met would determine motivation on a continuum ranging from extrinsic motivation to intrinsic motivation [31]. Extrinsic motivation focuses on consequences that are separate from engaging in the specific activity, such as satisfying external pressures or achieving rewards, while intrinsic motivation is based on interest and engagement in the activity itself [31]. In order to sustain behaviors, intrinsic motivation should be strived for. Current knowledge supports the value of SDT in understanding and promoting exercise behavior [32], and higher levels of intrinsic motivation have been related to greater levels of active transport [33]. Moreover, the Theory of Planned Behavior (TPB) is commonly used to predict behaviors based on individuals’ intentions to perform the behavior of interest. In turn, these intensions are determined from attitudes towards the behavior, subjective norms, and perceived behavioral control [34]. Studies have linked active transport (cycling and walking) to these psychological constructs, i.e. attitudes, subjective norms, perceived behavioral control and intentions [35–37]. However, habit strength could moderate the impact of intentions on actual behavior, and has therefore been included in more extensive models assessing behaviors like travel mode choice [38]. Also, human behavior is commonly affected by environmental determinants, like accessibility [39]. Supporting this, access to bikes and accompanying equipment is suggested to be a key factor associated with transport cycling [40]. Correspondingly, a British study providing e-bike access for six to eight weeks was seen to influence employees’ travel behaviors and total PA level substantially, including in employees not usually engaged with PA, or feeling unable to use a conventional bike [41]. Bike share programs represents another aspect of accessibility, and a number of cities have now introduced e-bike sharing potentially encouraging new users to bike share [42].

To the best of our knowledge, no previous studies have tested the effects of providing free access to different bike types for a prolonged period with regard to the parents of young children. Therefore, the primary objective of the present study was to assess the feasibility and effect of an intervention providing participants with access to an e-bike including a trailer for child transportation, a cargo (longtail) bike, and a traditional bike including a trailer—each bike type for three months—on transportation habits (weekly frequency of cycling and driving to the workplace, the kindergarten and the grocery store) and total cycling (distance and time). The secondary objective was to assess possible intervention effects on intrinsic motivation for cycling, and psychological constructs and habit strength related to car use.

Methods

Study sample

A convenience sample was recruited, consisting of 36 parents (Fig 1) with children attending kindergarten and residing in Southern Norway. Recruitment was carried out via Facebook announcements and direct contact with kindergartens and selected organizations and businesses in Kristiansand municipality from May through August 2017. The main outcome was cycling amount (distance and time). Based on a standard deviation (SD) of 60 minutes (min) per week [45], a power of 0.80 and a significance level of 5%, a sample size of 16 subjects in the intervention group and 16 subjects in the control group would allow for detecting an increase in cycling time from 15 min to 75 min per week (i.e. half of weekly PA recommendations). This meant that we aimed to recruit participants that cycled less than 15 minutes per week initially. To account for a 10% drop-out, and to utilize the bikes optimally according to the study design, a total of 36 participants were enrolled by the project coordinator (HBB). Inclusion criteria were (i) being able to understand and read Norwegian, (ii) having one child born in year 2013, 2014 or 2015 attending kindergarten, (iii) being responsible for bringing to/picking up the child from kindergarten ≥5 times per week, or at least half of the time, (iv) residing 2–10 km from the workplace (i.e. excluding those working from home), (v) residing <3 km from the kindergarten and the grocery store, (vi) having car access, (vii) possessing a smartphone, (viii) being between 167–190 centimeters (cm) tall (due to the size of the accessible bikes), and (ix) having the opportunity to store the bikes indoors. Exclusion criteria were: (i) being physically active, i.e. meeting the PA recommendations, (ii) having cycled more than once weekly throughout the last twelve months to the workplace, kindergarten or grocery store, and (iii) suffering from severe cardiovascular diseases or upper respiratory tract diseases.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Participants flow.

The study was conducted in Kristiansand, Norway, from September 2017 to June 2018.

https://doi.org/10.1371/journal.pone.0219304.g001

Results

Baseline characteristics across intervention and control group participants are presented in Table 1, showing that groups were not significantly different at baseline.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics across intervention and control group participants.

https://doi.org/10.1371/journal.pone.0219304.t001

Travel behavior

At the nine-month follow-up significantly more participants from the intervention group (compared with the control group) were classified as cyclists to the workplace (n = 7 (38.9%) vs. n = 1 (5.9%), p = 0.04), yet not to the kindergarten (n = 6 (33.3%) vs. n = 2 (11.8%), p = 0.23) or to the grocery store (n = 2 (11.1%) vs. n = 0 (0%), p = 0.49). No differences in mode classification between groups were found for being a car user (Table 2). For the intervention group, a significant increase in cycling frequency from baseline to the nine-month follow-up was found for all destinations and seasons, except to the grocery store during the winter season (p = 0.16). Change in cycling frequency ranged from 0.1 (“grocery store, winter”) days per week to 2.0 (“work, autumn”) days per week (Fig 2). For frequency of driving a car, a significant decrease was observed for the travel to the workplace and the kindergarten for all seasons, yet not for the grocery store for any season (p = 0.15–0.49). Change in frequency of driving ranged from -0.2 (“grocery store, spring”) days per week to -1.7 (“work, autumn” and “kindergarten, autumn”) days per week (Fig 3). For the control group, no significant changes were observed for frequency of cycling or for driving.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Change in cycling frequency.

Change (mean (SD), days/week) in cycling frequency from baseline to follow-up in the intervention group (n = 18), grouped according to destinations (work, kindergarten, grocery store) for the autumn, winter and spring seasons respectively. Follow-up measures were conducted after 3, 6 and 9 months, i.e. following the autumn, winter and spring season.

https://doi.org/10.1371/journal.pone.0219304.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Change in frequency of driving.

Change (mean (SD), days/week) in frequency of driving a car from baseline to follow-up in the intervention group (n = 18), grouped according to destinations (work, kindergarten, grocery store) for the autumn, winter and spring seasons respectively. Follow-up measures were conducted after 3, 6 and 9 months, i.e. following the autumn, winter and spring season.

https://doi.org/10.1371/journal.pone.0219304.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Mode classification across intervention and control group participants.

https://doi.org/10.1371/journal.pone.0219304.t002

Cycling within the intervention group

There were major individual differences in cycling among the participants in the intervention group, with cycling distance for the total trial ranging from 44 km to 1408 km, with a weekly median of 10.9 km (IQR 20.8 km). In total 6 (33.3%) participants cycled ≥75min per week throughout the entire trial period (42.9 min (65.3 min), median (IQR)). Further, for the trial period in total, the traditional bikes accrued 2798 km and 11 503 min of cycling, the longtail bikes 2750 km and 10 572 min, and the e-bikes 4164 km and 12 371 min, entailing an average speed of 14.6, 15.6 and 20.2 km per hour respectively. On the individual level, cycling distance within each season ranged from 25 km to 561 km (traditional bike) for the autumn, from 0 km to 401 km (e-bike) for the winter, and from 3 km to 504 km (e-bike) for the spring. Further, cycling time per participant ranged from 26 mins to 1906 mins (traditional bike), from 0 min to 1351 mins (e-bike), and from 18 mins to 2015 mins (longtail) within the autumn, winter and spring seasons respectively. E-bikes achieved the greatest cycling amount (distance and time) for the entire trial period, with the smallest sample variability (IQR). During the first cycling period, i.e. the autumn season, the traditional bikes achieved the greatest amount of cycling per week (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Cycling distance and time per week for the total trial and for each season across bike types.

https://doi.org/10.1371/journal.pone.0219304.t003

Discussion

The present trial, providing participants with access to different bike types for in total nine months, resulted in increased cycling and decreased car use in parents with children attending kindergarten, as well as higher intrinsic motivation for cycling. These findings agree with the socio-ecological framework [39], claiming that accessibility is one relevant environmental determinant for PA, including cycling for transport [40]. Accordingly, yet targeting e-bikes only, a Norwegian real-life controlled experiment providing 66 participants with e-bikes for two or four weeks, generated considerably more cycling [19]. Fyhri and Fernley reported the number of cycling trips to increase from 0.9 to 1.4 per day, distance cycled to increase from 4.8 km to 10.3 km, and cycling shares out of all transport to increase from 28% to 48% [19]. Likewise, Cairns et al. [41] found that after loaning 80 employees an e-bike for six to eight weeks, car mileage was reduced by 20%, 59% of the employees increased their overall PA, whilst 73% would cycle to work at least once a week, if they had an e-bike accessible [41]. Concerning cargo bikes specifically, American data has shown a decrease in car travel among cargo bike owners [22]. Also, if conditions facilitate it, American mothers accessing a cargo bike seem to consider it a realistic option for trips usually made by car [26]. Regarding bike share, representing another means to bike access, a pilot study testing a university-based e-bike share in North-America reported that new users were attracted to cycling [53]. Multi-city analyses of regular bike shares’ impact on car use and PA suggest that car use decreases, yet this is of limited magnitude [54]. Nevertheless, PA levels could increase [55] due to mode shifts following bike share usage.

E-bikes achieved the greatest cycling amounts for the trial period in total in the present study, which corresponds with previous studies reporting e-bike users to cycle longer distances, and more often [18–21]. It is likely that increased cycling with an e-bike is explained by its tendency to reduce typical barriers for cycling such as long distances, hilly terrain, time use and the need to shower afterwards [19, 21, 56]. Also, e-bikes could be speed competitive with both public transport [57] and private cars [58]. Besides, seasonal variations that were previously reported to decrease commuter cycling with regular bikes significantly [59], may in fact become less problematic when assistance from an electric motor is provided [57]. The heavy weight of an e-bike, combined with the power, could offer more traction in winter conditions, hence it being easier to ride during all seasons [60]. Nevertheless, this would apply to a greater extent for experienced cyclists than for newcomers.

Even though cycling increased and car use decreased in the intervention group, there were major inter-subject differences. In Norway substantial seasonal variation is normal due to the climate, yet the winter season in the present trial was even more icy and snowy than usual in Southern Norway, making cycling more challenging. For subjects not used to cycling in particular, this likely affected cycling negatively, and may be one reason for the large individual differences. As suggested by the SDT [30], intrinsic motivation would increase the chances of behaviors being performed, and further that they would be sustained. In turn, increased compliance with the basic needs of autonomy, competence and relatedness would result in a move on the motivational continuum from more controlled regulations to autonomous regulation, with intrinsic motivation as the desired endpoint [31]. Following this rationale, it could be the case that those cycling even during the tough winter months experienced (unexpected) mastery, which in turn facilitated the basic need of competence, and further increased intrinsic motivation. Supporting this, the intervention group increased their intrinsic motivation for cycling throughout the trial period. However, we also found that scoring on “interest/enjoyment”, “introjected regulation” and “intrinsic regulation” for cycling increased in the intervention group and remained stable or decreased in the control group. The accompaniment of introjected regulation (being one controlled form of regulation) with intrinsic regulation may be somewhat surprising from a strict SDT perspective. Nevertheless, this finding agrees with previous work demonstrating that introjected regulation can follow intrinsic regulation, and further associate positively with exercise [61], total PA and active transportation [33]. It is noteworthy, though, that the amount of cycling was higher during the first period (autumn) than during the latter period (spring), with the greatest amount found for the traditional bikes. These findings may express fatigue and decreased motivation in some participants, as one might have expected the opposite trend in cycling based on season and weather conditions [62], as well as previous findings showing that e-bike users cycle more than those using a traditional bike [18].

For the present sample, switching from car to bike seemed the most challenging for the travel to the grocery store. Carrying capacity probably represents one key factor, together with combining more activities into one journey, as combined travels are found to be conducted by bike to a lesser extent than commuting trips separately [57]. The bike trailer complementing the traditional bike and the e-bike would allow for extra items to be carried if bringing one child, yet to a limited extent if bringing two children. The bags following the longtails could carry substantial loads, yet the feet protection on the child seat somewhat limited the carrying capacity of the bags. In addition, solid loads combined with child transportation entails considerable weight. Hence, for those not cycling regularly, nor being classified as physically active (due to inclusion criteria), the weight load may be too high. In turn, this could reduce the utility of the human-powered longtails.

Conclusion

Access to different bike types for families with children in kindergarten resulted in overall increased cycling and decreased car use. Substantial individual differences in cycling likely express the importance of the motivational aspects, in addition to accessibility. For subjects who are not well-trained, e-bikes with trailers or electric assisted longtails may be more feasible and sustainable for transporting goods and children by bike than traditional bikes with trailers and human-powered longtails. Thus, providing parents with children in kindergarten with access to e-bikes might increase cycling for some, also during the winter season. Higher levels of intrinsic motivation for cycling may contribute to increased active travel, also in the long term.

Supporting information

Acknowledgments

The authors would like to thank the participants in the present study.