Sampling and Procedures

This cross-sectional analysis was based on the Social, Economic, and Environmental Factor (SEEF) Study, which was conducted as a follow-up sub-study of the SAX Institute's 45 and Up Study [12], a large cohort study of adult residents aged 45 years and older in New South Wales, Australia (2006–2008; n = 267,153). Participants were randomly sampled from the Medicare Australia database, which includes all residents in the state of New South Wales. In 2010, the first 100,000 participants of 45 and Up were invited by mail and 60,404 participated in SEEF by completing the consent form and the questionnaire and mailed them back in prepaid envelopes to the study coordinating center (response rate of 60.4%). The 45 and Up Study was granted ethical approval by the University of New South Wales Human Research Ethics Committee (HREC 05035/HREC 10186) and the SEEF Study by the University of Sydney Human Research Ethics Committee (ref no. 10-2009/12187).

Measurement

Driving was measured using a single-item question “About how many hours in each 24 hour day do you usually spend driving?”

Health Behaviors

Smoking risk was defined as being a current regular smoker. Alcohol risk was defined as consuming more than 14 servings of alcohol per week. Dietary risk was defined as not meeting the Australian dietary guidelines for fruit and vegetable intake (2 servings of fruit + 5 servings of vegetables per day) [13]. Physical activity risk was defined as engaging in less than 150 minutes/week of moderate-to-vigorous intensity activity based on the Active Australia Survey [14]. Excessive sitting was defined as 8 hours or more of sitting in a 24 hour-day based on findings from 45 and Up [15]. Insufficient sleep was defined as sleeping less than 7 hours/day based on a meta-analysis on sleep duration and mortality [16].

Health Outcomes

Obesity was measured as body mass index (BMI)≥30 based on self-reported height and weight. Previous research among a random sub-sample of 45 and Up showed excellent agreement between BMI categories from self-reported and measured data [17]. Overall self-rated health was measured using a single-item question (“In general, how do you rate your overall health?”) from the Medical Outcomes Study 12-Item Short-Form Health Survey (SF-12). Quality of life was assessed using a parallel question “In general, how do you rate your quality of life?” Psychological distress was measured using Kessler-10, which has been validated in the Australian population [18]. Respondents' scores of 22 and above were coded as “high/very high psychological distress”, which correlates with a clinical diagnosis of depression and anxiety disorders needing interventions [18], [19]. Time stress was measured using a question previously used by the Australian Bureau of Statistics [20] “How often do you feel rushed or pressed for time” and those who answered “always” or “often” were coded as “stressed for time.” Social functioning was measured using one item from SF-12 “How much time during the past four weeks have your physical health or emotional problems interfered with your social activities?” Those who reported “all/most of the time” were coded as “poor social functioning.”

Data Analysis

Analyses were restricted to those with complete data, aged 75 and below, and who reported any driving (n = 36,430∼37,570 for different outcomes). Descriptive analyses of all variables were conducted to compare across participants with different amounts of driving using chi-square tests and ANOVA. Logistic regression analysis was conducted for each health behavior/outcome variable adjusted for age, sex, marital status, employment status, educational attainment, the number of motor vehicles per person in the household, rural/urban residence, and area-level socioeconomic disadvantage (measured by the Index of Relative Socio-Economic Disadvantage). Sex by driving and education by driving interactions were tested in all models for potential effect modification.

Driving and Health Behaviors

Based on unadjusted descriptive statistics (Table 1), there was a consistent trend that higher prevalence rates of smoking, excessive alcohol use, insufficient fruit and vegetable consumption, physical inactivity, and insufficient sleep were observed among those with higher driving time. When adjusted for socio-demographic characteristics, there was a positive linear trend for smoking, physical inactivity, and insufficient sleep across dose categories of driving (p<0.001), and for these three health behavior outcomes statistically significant associations occurred from the level of 31-60 min/day of driving (Table 2). Surprisingly, those who reported driving for more than 120 min/day were less likely to sit for more than 8 hours per day (OR = 0.88, p = 0.03) and there was an overall inverse association between driving time and excessive sitting (p = 0.042). No driving time by sex/education interaction was significant at p<0.05.

Download:

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

Table 2. Adjusted odds ratios^a for the associations between daily driving time and health behaviors and outcomes among driving adults aged 47–75 in New South Wales, Australia, 2010.

https://doi.org/10.1371/journal.pone.0094602.t002

Driving and Health Outcomes

Based on unadjusted analyses, those who spent more time driving were more likely to report obesity, fair/poor quality of life, high/very high psychological distress, time stress, and having physical health or emotional problems that interfered with social functioning (Table 1). When adjusted for socio-demographic characteristics, there was a positive linear trend for all health outcomes across dose categories of driving (p<0.001) and the significance occurred from 31–60 min/day for obesity, from 61–120 min/day for time stress and social functioning, and at more than 120 min/day for self-rated health, quality of life, and psychological distress (Table 2).

For all health outcomes, there was no significant effect modification by sex or educational attainment at p<0.05. No driving time by sex/education interaction was significant at p<0.05.

Driving and Health Behaviors

Our findings highlight the clustering of poor health-related behaviors with driving. Associations with insufficient amount of sleep, physical inactivity, and regular smoking were significant and showed clear dose-response patterns. While other studies have highlighted the association of long commutes with sleep deprivation [5], [10] and physical inactivity [5], to our knowledge this is the first study to demonstrate that long driving time in general is associated with smoking. Time trade-offs are likely to be the most plausible explanation for compromised sleep and lack of physical activity, as an American time-use study indicated that more than 20% of the time of a 120-minute commute was taken from physical activity [5]. The mechanism by which driving induced smoking is less clear; it could be either a reaction to driving-induced stress, idle “time to kill”, or the “coupling” of behavioral rituals (e.g., driving) and the sensory/physiological reinforcements of smoking through second-order conditioning [22]. In this case, it is possible that each episode of entering the car “for a drive” or “sitting and driving” becomes a cue that evokes strong craving for smoking [23].

The lack of association with alcohol is not surprising given the risk associated with driving under the influence of alcohol. The lack of association with dietary behavior should be interpreted with caution because the measures only concerned fruit and vegetable intake, not other food items such as coffee, sweets, and sugar flavored beverages. Future studies should particularly include measures regarding fast food intake, dining out, and the consumption of energy-dense foods as previous studies have demonstrated that long commutes were associated with less time spent in food preparation [5], [7]. The slightly inverse association of driving time with excessive sitting is unexpected, as driving (or sitting for transport) contributes to overall sitting. This may be explained by other domains of sedentary behavior, such as occupational and recreational sitting, which may be disproportionately more common among those who spend less time driving. Regardless, our finding also implies that it may be oversimplified to assume that driving harms health mainly through prolonged sitting [23].

Driving and Health Outcomes

This study found a strong relationship between driving time and self-reported health and well-being, such as poor self-rated health, poor quality of life and high/very high psychological distress. The mechanisms that may moderate or mediate the associations between driving and mental health outcomes are likely to be multidimensional and complex. The associations between driving and health may be mediated by “travel impedance” [24], [25], such as exposure to stressful driving conditions, including adverse road environments, dealing with traffic congestion causing delays, and behaviors of other road users. Consequently, one would expect that the mental workload required for driving would have some measurable impact on health and well-being. Another possible mechanism relates to physical and leisure-time activities. Given the well-established relationship between physical activity and physical and mental health [26], it is plausible that reduced time for health-related activities as a result of more time spent driving will have consequences on health and well-being. The trade-off between driving and spending time with one's social network [27] is another important factor since limited social interactions have been shown to adversely impact health outcomes, such as mental health and mortality [28], [29]. Having supportive social interactions may provide a buffer against stressful driving experiences. Empirical work would be necessary to confirm these possible mechanisms.

In this study, the strength of the associations of driving with time stress and health-related social functioning are substantial, observed from an hour and more of driving per day. Whether time pressure leads to more driving or spending more time driving increases the odds of feeling stressed for time, the finding highlights the potential role of time stress in the relationship between driving time, health, and well-being. Finally, the link between driving time and obesity observed in this study is consistent with previous studies [9], [30], [31] and reaffirms driving as a potentially important risk factor for cardio-metabolic health.

Strengths and limitations

Key strengths of the study included a large population-based sample and the assessment of comprehensive behavioral and health outcomes using validated measures. The consistent patterns of associations across health behaviors and health outcomes enhance the validity of findings. The question about driving focuses exclusively on driving, rather than commuting, which not only includes other modes of transportation (i.e, train, bus, taxi) but also precludes driving for other purposes such as doing errands, chauffeuring children, and driving for recreational purposes. This is important because in Australia work commutes only accounted for 27% of the total kilometers traveled by passenger vehicles [32], suggesting that a large proportion of driving time is not accounted for in commuting based research. However, such a comprehensive measure has its merits, but also limitations, depending on the potential underlying mechanism through which driving affects health. For example, driving impedance is relevant to driving for all purposes while the lack of social interactions may be more relevant to commuting and less so to chauffeuring children or recreational driving. A cross-sectional study design is a limitation of the study. Although it is less plausible that unhealthy lifestyles are antecedents to driving behavior, it is possible that those with poor health depend more on driving because they have limited physical capacity for active travel (e.g., walking or cycling). Typical for large population-based studies, another limitation is that all measures were based on self-report data, which are subject to biases. The study could also be improved by including more detailed measures about driving, such as the duration of each trip (bout), to pinpoint risk associated with one episode of prolonged driving or cumulative driving time over multiple trips. Furthermore, future studies should include other modes of transport, such as public transportation, so that one can isolate effects of driving from overall daily travel time.