Study population
This report is based on cross-sectional data from the Mannheim Industrial Cohort Studies (MICS). Employees (n = 4881; 41.1 ± 11.5 years; 21.7 % female) of an industrial company in Southern Germany participated in a voluntary health risk assessment during working hours between September 2009 and May 2011. Of these, 3805 participants (41.1 ± 11.3 years, range 16–64 years, 21.1 % female) presented a full data set. The study was approved by the Medical Ethic Committee of the Medical Faculty Mannheim of Heidelberg University (approval number 2010-296E-MA). Written informed consent was given by each participant.
Measurements
All data collection were conducted by an external agent (HealthVision Ltd, Berlingen, Switzerland). A comprehensive online health questionnaire included the question about minutes commuting to work one way. According to previous studies, commuting to work was grouped into four categories: 0–19.9, 20–44.9, 45–59.9, ≥60 min [1, 5]. Modes of commuting were unknown but likely by private car given travel patterns in Germany, especially in rural areas [3]. Sociodemographics (age, gender, marital status, single earner status, shiftwork), current smoking status, absenteeism and presenteeism (days per year) were assessed by self-report, as well as characteristics of well-being such as the 6-items mental and 6-items physical health subscales of the short form health survey (SF-12) (0–100, higher values = better health) [13], 5-items Cohen’s perceived stress scale (5–25, higher values = higher stress) [14], 6-items Maastricht Vital Exhaustion Questionnaire (6–30, higher values = higher exhaustion) [15], and 4-items Jenkins sleep scale (4–24, higher values = worse sleep quality) [16].
Clinical measurements assessed body mass index (kg/m2), waist circumference (centimeter), heart rate variability (root mean square of successive differences in milliseconds), diastolic and systolic blood pressure (mmHg). Fasting blood samples were collected between 7 and 9 a.m., immediately transported to a laboratory (Synlab, Augsburg Germany) and included following biomarkers: serum C-reactive protein, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glycosylated hemoglobin, and fasting plasma glucose.
Long term heart rate was recorded and beat-to-beat intervals were determined as the interval between two successive R-spikes and analyzed by researchers at the Center for Neuropsychological Research (University of Trier, Germany). Heart rate variability measures were calculated by usual means [17] as described elsewhere [18].
Statistical analyses
We present descriptive, univariate analysis. Differences between the four commuting categories were determined by Student’s t test for continuous variables and by analysis of variance (ANOVA) for categorical variables. Bivariate associations were calculated using Pearson correlation. Linear regression models tested the association of commuting (independent variable) with multiple health variables (dependent variable), adjusting for age, gender, marital status, and shiftwork. Skewed variables were transformed according to the ladder of power to better approximate a normal distribution [19]. We used Stata 12.1 MP (College Station, TX: StataCorp LP) for all statistical analysis.