Sleepless on the road: Are mothers of infants with insomnia at risk for impaired driving?
Summary
Infant sleep problems have been associated with a myriad of adverse child and parent outcomes, yet whether these problems may pose a risk for parents on the road has received little research attention. This study sought to test whether mothers of infants with insomnia are at an elevated risk for vehicular crashes, by comparing their objectively measured driving performance with that of mothers of well-sleeping infants and with that of women without children. Fifty-four women from these three groups completed a simulated driving task. Outcome measures included standard deviation of lateral position, number of lane crossings, standard deviation of speed, average speed and maximum speed. Women additionally reported on their driving behaviour using the Driving Behaviour Questionnaire, and on sleep, sleepiness and insomnia symptoms using 7-day sleep diaries and questionnaires. Mothers of infants with insomnia demonstrated greater lane deviation (Wald = 9.53, p = 0.009), higher maximum speed (Wald = 6.10, p = 0.04) and poorer self-rated driving behaviour (Wald = 7.44, p = 0.02) compared with control groups. Analyses also indicated that driving performance in mothers of infants with insomnia tended to be poorer relative to control groups with the progression of time on task. While further research is needed to assess the scope of these effects, our findings suggest that parents, healthcare providers and policymakers should be aware of the potential consequences of infant sleep problems on road safety, and collaborate to establish strategies to mitigate these risks.
1 INTRODUCTION
Infant insomnia is a common concern for parents. Research indicates that 15%–30% of infants experience considerable and persistent difficulties initiating and maintaining sleep, prolonged nighttime wakefulness, or dependence on external regulation for sleep onset (American Academy of Sleep Medicine, 2014; Honaker & Meltzer, 2016; Williamson et al., 2019). Left untreated, these problems can have severe and wide-ranging consequences for the infant, related to cognitive, motor, emotional, behavioural and social functioning, as well as physical health (Sadeh et al., 2014; Spruyt, 2019). Notably, infant sleep problems also impact their caregivers. Given their dependency on parents early in development, young children's insomnia symptoms are usually associated with considerable sleep loss for parents (Richter et al., 2019), which may result in increased parental depression, anxiety, stress, poor physical health and reduced quality of life (Bayer et al., 2007; Liew & Aung, 2021; Sadeh et al., 2010). Another potentially life-threatening context that has received much less attention is the impact of infant insomnia on parents' ability to safely perform “next-day” applied tasks, such as driving a motor vehicle.
Driving while sleepy or fatigued poses a serious risk to road safety. According to data from the National Highway Traffic Safety Administration (NHTSA) in the USA, drowsy driving accounts for 7% of all vehicular crashes, and 16.5% of those that are fatal (Tefft, 2014). Notably, fatigue-related vehicle crashes tend to be more severe due to drivers' reduced situational awareness and potential micro-sleep episodes, which can result in lost control over the vehicle and impede timely corrective action (e.g. braking; Boufous & Williamson, 2009; Thygerson et al., 2011). Research examining the effects of drowsy driving has mainly focused on certain at-risk populations, such as individuals with sleep disorders (e.g. obstructive sleep apnea, narcolepsy; McCall & Watson, 2020; Rizzo et al., 2019) and night-shift-workers (Lee et al., 2016; Smith et al., 2020). However, parents of young children may also be a population at risk for drowsy driving due to the sleep disturbance they may experience during this developmental phase of their child's life. Indeed, high prevalence estimates of drowsy driving have been reported in this population, with 18% of parents of infants under 12 months old reporting driving while sleepy at least once per week (Armstrong et al., 2015; Macy et al., 2014; Malish et al., 2016). Retrospective vehicular crash data also indicate that women driving with an infant passenger are at elevated risk of fatigue-related vehicular crashes (Maasalo et al., 2017). Higher levels of chronic fatigue have been reported in women who are caregivers of infants aged 1–24 months old compared with non-caregivers, and this chronic fatigue is associated with a higher frequency of reported attentional driving errors (Sánchez-García et al., 2019). One study directly examining the link between poor sleep and driving patterns of new parents found that sleep disruption was more common in parents who reported near-miss or actual vehicular crashes (Malish et al., 2016). Poor sleep quality in this study was associated with a sixfold increase in reported near-miss motor vehicle crashes.
While these findings provide preliminary evidence for a link between parenting young children and impaired driving, they are limited by the subjective, retrospective and indirect assessment of driving performance. To the best of our knowledge, only one study has objectively assessed driving performance with new parents. Using a driving simulator, this study found no differences in measures of driver awareness or attention, reaction times, braking time, speed or distance in response to potential hazards at 2–3 weeks post-birth compared with 5–6 weeks post-birth (Harpham et al., 2020). However, a comparison group of childless participants was not included, and the chosen time frame may have limited the ability to detect differences (i.e. the lack of a prenatal assessment and focus solely on the first 6 weeks postpartum). Moreover, given that vast individual differences exist in infant sleep development, examining the parent population as a whole may not be as informative as focusing on what is likely the most vulnerable subpopulation—parents of infants with insomnia. In line with these notions, a recent systematic review highlighted the need to investigate the impact of infant sleep problems on parent driving (Sprajcer et al., 2022).
The aim of the present study was to test whether mothers of infants with insomnia are more likely to be at risk while driving, by comparing their objectively measured driving performance with that of mothers of well-sleeping infants and women without children. In line with previous studies (Armstrong et al., 2015; Maasalo et al., 2017; Sánchez-García et al., 2019), we chose to focus the present investigation on women. This decision was made because mothers are typically more likely to be the primary caregiver for their infant, which may result in greater sleep disruption, and therefore an increased degree of vulnerability to impaired driving (Horwitz et al., 2023). We hypothesized that mothers of infants with insomnia would show poorer driving performance compared with both control groups, based on both self-reports and objective driving assessment.
2 METHODS
2.1 Participants and procedures
Participants were 54 women from three distinct groups. (1) The clinical group (n = 20) included mothers of infants aged 6–23 months diagnosed with insomnia. Based on ICSD-3 criteria (American Academy of Sleep Medicine, 2014), infants whose sleep-onset latency was ≥ 30 min, wake after sleep onset (WASO) was ≥ 30 min, or had ≥ 2 awakenings per night on average were eligible to be included in this group. Additionally, these sleep difficulties had to be perceived as causing distress or impairment, and not be better explained by any medical, developmental or environmental condition, as reported by mothers in an initial telephone interview. (2) The non-clinical control group (n = 15) included mothers of infants aged 6–23 months who reportedly slept well and did not meet the criteria described above. (3) The childless control group (n = 19) included women who did not have children. This design was chosen based on the study undertaken by Sadeh et al. (2016), which controlled for both parenting experience (childless control group) and the presence of infant insomnia (non-clinical control group).
Inclusion criteria for women in all groups included possessing a valid driver's licence, being between 21 and 40 years old, and not undertaking night shift employment. Women in the clinical group and non-clinical control group were only included if they had an infant aged 6–23 months, with no diagnosed medical or neurodevelopmental disorder. Participants were recruited via advertisements posted on social media platforms, and in local general practitioner clinics and child daycare centres. Recruitment for the clinical group was also conducted at the university sleep clinic, with parents seeking treatment for their infants' insomnia invited to participate in the study.
Enquiring participants underwent a telephone interview, and those eligible were asked to complete a sleep diary for 1 week prior to attending the laboratory, reporting on both their own sleep, and—for women who were mothers—also reporting their infant's sleep. Participants were asked to attend the laboratory between 09:00 hours and 11:30 hours to minimize circadian rhythm effects on driving performance. They were instructed to abstain from consuming caffeinated foods/beverages for 12 hr prior to attending the laboratory, and consume at least 1 L of water (to assist with hydration) and their regular breakfast on the morning of the visit. The selection of the morning time frame was informed by several considerations. To begin with, driving performance simulations have often been undertaken during similar morning periods (Bartrim et al., 2020; Bragg et al., 2017). Moreover, a morning session facilitates adherence to the stipulated 12-hr caffeine abstinence prerequisite. Additionally, morning driving is a common occurrence among mothers of young children (Armstrong et al., 2015). Lastly, this choice aligns with the practical convenience of schedules of mothers with young children, thereby enhancing the feasibility of participation.
Upon arrival in the lab, participants provided signed informed consent, completed a series of questionnaires, and performed a 25-min simulated driving task. They were provided a gift voucher ($30 AUD) for participation. The study was approved by the Flinders University Social and Behavioural Research Ethics Committee (approval #7928).
2.2 Measures
2.2.1 Sleep diaries
Women completed 7-day sleep diaries during the week prior to their laboratory visit. Mothers in the clinical and non-clinical groups simultaneously recorded 7-day sleep diaries for their infants. Sleep diaries were used to further corroborate the group inclusion–exclusion criteria in relation to sleep (e.g. verify that no infants of women in the non-clinical group met criteria for insomnia). Sleep diaries have long been established as a reliable and valid technique for sleep assessment in both infants and adults (Carney et al., 2012; Sadeh, 2015). Derived metrics for this study included 24-hr sleep duration, nighttime sleep duration, number of nighttime awakenings, and WASO.
2.2.2 Questionnaires
Demographic and background characteristics
Participants completed a questionnaire, which included items regarding their age, marital status, employment status, average intake of daily caffeinated beverages (including coffee, tea, energy drinks, chocolate drinks and sodas) and number of children. Socioeconomic status was assessed using address postcodes, based on the Socioeconomic Indexes for Areas Score (Australian Bureau of Statistics, 2016), from 1 (lowest 20%) to 5 (highest 20%). Mothers were also asked to report their infant's age and sex, whether they currently breastfed, and if the infant usually slept in the same room as them.
Epworth Sleepiness Scale (ESS)
The ESS is a validated and broadly used measure of daytime sleepiness (Johns, 1991, 1992). This questionnaire consists of eight everyday situations (e.g. sitting and reading), in which participants are asked to rate their likelihood of dozing off on a four-point scale. Total scores range from 0 to 24, with higher values indicating greater sleep propensity.
Insomnia Severity Index (ISI)
The ISI is a seven-item measure of the perceived presence and severity of insomnia symptoms (Bastien et al., 2001). Items are rated on a five-point Likert scale, and subsequently summed to generate a total score, with higher scores indicating greater insomnia severity (range: 0–28). The psychometric properties of the ISI are well established, and it has been widely used in both clinical and non-clinical samples (Gagnon et al., 2013; Morin et al., 2011).
Driving Behaviour Questionnaire (DBQ)
Self-reported driving behaviour was assessed using the DBQ (Lajunen et al., 2004). This questionnaire includes 28 items that measure driving lapses (e.g. misreading traffic signs), errors (e.g. failing to notice pedestrians), ordinary violations (e.g. speeding) and aggressive violations (e.g. honking to indicate annoyance to another road user). Items are rated on a six-point Likert scale based on the frequency of engaging in these behaviours, from 1 (never) to 6 (nearly all the time). The total score is calculated by averaging individual items, with higher scores indexing a greater number of unsafe driving behaviours. The validity and reliability of the DBQ have been previously demonstrated (Zhao et al., 2012), and internal reliability in the present sample was high (alpha Cronbach = 0.85).
2.2.3 Driving simulation
Driving assessment was conducted on a desktop computer (SCANeR Studio Simulation Engine v1.6r85, OKTAL, Paris, France) equipped with peripheral devices for force-feedback steering and accelerator/brake pedals (Logitech G29, California, USA). Visual images were displayed on three 22-inch LCD monitors using a TripleHead2Go multi-display adaptor for 3840 × 1024 resolution (Matrox Graphics, Quebec, Canada), and set to provide a 100° front field of view. Auditory feedback was provided using an integrated stereo sound system (Logitech, California, USA).
The simulated driving scenario was adapted from that used in previous research (Bartrim et al., 2020; Bragg et al., 2017). Briefly, participants completed a 2-min familiarization drive, followed by a 25-min test drive. Data from the familiarization component as well as that collected in the first 2 min of the test drive (to allow the participant time to accelerate to the designated speed limit) were excluded from analyses. The 25-min scenario was chosen given that previous research demonstrated its sufficiency for detecting impairment under conditions of fatigue or drowsiness (Bartrim et al., 2020). Moreover, research has shown that short drives are highly common in parent populations (Koppel et al., 2011). The simulated scenario consisted of driving along a two-lane dual carriageway road resembling a highway, with long straight stretches and gentle curves. Elevation deviations were embedded throughout, requiring the driver to accelerate when driving uphill and braking when driving downhill to maintain a designated adherence speed. Infrequent oncoming traffic was present, but programmed to exhibit non-conflicting behaviour (i.e. required no interaction or response from the driver). Participants were instructed to comply with normal road regulations for Australia, drive in the centre of the left-hand lane, and adhere to a 100 km hr−1 speed limit throughout the drive.
Data were automatically recorded by the simulator's software at a rate of 20 Hz. Primary outcome measures included the lateral control variables: standard deviation of lane position (SDLP) and number of lane crossings; as well as the longitudinal control variable: standard deviation of speed (SDSP). Secondary outcomes included average speed and maximum speed. These metrics were chosen based on sensitivity to sleep disruption demonstrated in previous research (De Valck & Cluydts, 2001; Reyner & Horne, 2000). Previous investigations using monotonous driving scenarios have also demonstrated that driving performance tends to deteriorate with time on task (Verster & Roth, 2011). In light of this, data were also subsequently divided into five epochs (approx. 5 min), with driving metrics calculated for each.
2.3 Data analysis
Data processing was performed in SPSS v26 (IBM, USA), and data visualization conducted using the ggplot package in R v4.2.2, using RStudio v2022.07.1 + 554. Sleep diary data were missing for two infants (one from the clinical group and one from the non-clinical group) and six women (three from the clinical group, one from the non-clinical group, and two from the childless control group). One mother from the clinical group experienced simulator sickness and was only able to complete < 10 min of the driving task. As a result, driving performance data collected from this participant were excluded from analysis. Sample characteristics were compared between groups using ANOVAs for continuous variables, and chi-square tests for categorical variables. Generalized estimating equations (GEE) were used to model sleep diary data, as this approach allows for estimation of parameters using all available nights nested within participant (Hardin & Hilbe, 2002). Given the nature of distributions, negative binomial models were computed for the number of nighttime awakenings, and linear models were computed for duration of sleep and nighttime wakefulness. GEE analyses were also used to assess differences between groups in simulated and self-reported driving metrics. For simulated driving, models contained main effects for group (clinical, non-clinical, and childless control), epoch (1–5), and their interaction term. The latter was used to inspect time on task effects. Significant effects were interpreted using post hoc pairwise comparisons, with significance accepted when p < 0.05.
3 RESULTS
Participant characteristics are presented in Table 1. Women were mostly of middle socioeconomic status, with no differences between groups. Groups did not differ in intake of daily caffeinated beverages, nor did infant age, infant sex and number of children differ between the clinical and non-clinical groups. Significant differences were found in women's age (F2,51 = 9.92, p < 0.001), with women in the clinical and non-clinical groups older than those in the childless control group. Similarly, marital status (χ2 = 25.29, p < 0.001) and employment status (χ2 = 34.69, p < 0.001) significantly differed as a function of group, with mothers more likely to be married or in a domestic partnership and employed part-time or on maternity leave, compared with childless controls.
| Clinical group | Non-clinical group | Childless controls | F/χ2 (p) | |
|---|---|---|---|---|
| Women | ||||
| Age (years) | 33.7 (1.0) | 33.3 (1.1) | 27.9 (1.0) | 9.92 (<0.001)a |
| Marital status, n (%) | 25.29 (<0.001) | |||
| Domestic partnership or married | 16 (80%) | 10 (67%) | 2 (11%) | |
| In a relationship | 0 (0%) | 0 (0%) | 4 (21%) | |
| Never married | 4 (20%) | 5 (33%) | 13 (68%) | |
| Employment status, n (%) | 34.69 (<0.001) | |||
| Full-time | 1 (5%) | 2 (13%) | 9 (47%) | |
| Part-time | 11 (55%) | 7 (47%) | 7 (37%) | |
| On maternity leave | 3 (15%) | 4 (27%) | 0 (0%) | |
| Unemployed | 3 (15%) | 2 (13%) | 0 (0%) | |
| Student | 2 (10%) | 0 (0%) | 3 (16%) | |
| Socioeconomic status (ISRAD) | 3.55 (0.31) | 3.07 (0.36) | 3.53 (0.32) | 0.61 (0.55) |
| Average caffeine consumption (cups per day) | 2.47 (0.28) | 1.96 (0.34) | 1.97 (0.29) | 0.98 (0.38) |
| Daytime sleepiness (ESS) | 9.95 (0.88) | 5.20 (1.02) | 4.47 (0.90) | 10.95 (<0.001)b |
| Insomnia symptoms (ISI) | 12.40 (0.99) | 5.67 (1.15) | 4.24 (1.02) | 18.50 (<0.001)b |
| Infants | ||||
| Age (months) | 11.6 (1.1) | 13.6 (1.2) | – | 1.50 (0.23) |
| Females, n (%) | 10 (50%) | 6 (40%) | – | 0.34 (0.56) |
| Number of children | 1.5 (0.2) | 1.6 (0.2) | – | 0.06 (0.81) |
| Room sharing with parents, n (%) | 15 (75%) | 2 (13%) | – | 13.05 (<0.001) |
| Breastfeeding, n (%) | 15 (75%) | 3 (20%) | – | 10.38 (0.001) |
- Note: Data are given as mean (standard error) unless otherwise indicated. Bolded font represents significant effects.
- Abbreviations: ESS, Epworth Sleepiness Scale; ISI, Insomnia Severity Index; ISRAD, Index of Relative Socio-economic Advantage and Disadvantage.
- a Significant differences: Clinical group, Non-clinical group > Childless controls.
- b Significant differences: Clinical group > Non-clinical group, Childless controls.
The GEE analyses of sleep diary data revealed expected patterns, with significantly shorter and more fragmented sleep in both women and infants in the clinical group compared with the control groups, thus corroborating group allocation. Detailed descriptive and inferential statistics are provided in Table 2. Correspondingly, women in the clinical group reported significantly higher daytime sleepiness (F2,51 = 10.95, p < 0.001) and more severe insomnia symptoms (F2,51 = 18.50, p < 0.001) compared with mothers from the non-clinical group and childless controls. Mothers in the clinical group were also more likely to report room sharing (χ2 = 13.05, p < 0.001) and breastfeeding (χ2 = 10.38, p = 0.001) compared with mothers in the non-clinical group.
| Clinical group | Non-clinical group | Childless controls | Wald (p) | |
|---|---|---|---|---|
| Women | ||||
| 24-hr sleep duration (hr) | 6.3 (0.3) | 8.2 (0.2) | 7.9 (0.1) | 23.64 (<0.001)a |
| Nighttime sleep duration (hr) | 6.2 (0.3) | 8.0 (0.2) | 7.9 (0.1) | 24.71 (<0.001)a |
| Number of nighttime awakenings | 1.9 (0.3) | 0.9 (0.3) | 0. 6 (0.2) | 16.11 (<0.001)b |
| WASO (min) | 98.9 (15.5) | 35.2 (9.2) | 9.0 (3.0) | 36.96 (<0.001)c |
| Infants | ||||
| 24-hr sleep duration (hr) | 10.0 (0.5) | 11.2 (0.6) | – | 2.24 (0.13) |
| Nighttime sleep duration (hr) | 8.3 (0.4) | 10.0 (0.3) | – | 10.32 (0.001) |
| Number of nighttime awakenings | 2.6 (0.4) | 1.3 (0.4) | – | 6.24 (0.01) |
| WASO (min) | 125.8 (18.3) | 31.8 (7.6) | – | 22.63 (<0.001) |
- Note: Data are given as mean (standard error). Bolded font represents significant effects.
- Abbreviations: WASO, wake after sleep onset.
- a Significant differences: Clinical group < Non-clinical group, Childless controls.
- b Significant differences: Clinical group > Non-clinical group, Childless controls.
- c Significant differences: Clinical group > Non-clinical group > Childless controls.
For the driving performance parameters, a significant group difference was found for SDLP (Wald = 9.53, p = 0.009), with greater lane deviation (i.e. lane swerving) observed in the clinical group compared with both the non-clinical (Mdifference = 0.07 m, p = 0.002, Hedges' g = 0.95) and childless control groups (Mdifference = 0.06 m, p = 0.02, Hedges' g = 0.83; Table 3). Post hoc comparisons revealed no significant differences between women in the non-clinical and childless control groups. Similarly, a significant group main effect was found for maximum speed (Wald = 6.10, p = 0.04), with women in the clinical group having significantly higher maximum speed than those in the childless control group (Mdifference = 2.75 km hr−1, p = 0.008, Hedges' g = 0.80). No significant group effects were found for the number of lane crossings, SDSP or average speed. However, a significant main effect for group was found for self-reported driving (Wald = 7.44, p = 0.02), with higher total DBQ scores recorded in the clinical group compared with both the non-clinical (Mdifference = 0.35, p = 0.003, Hedges' g = 0.87) and childless control groups (Mdifference = 0.23, p = 0.03, Hedges' g = 0.57).
| Clinical group | Non-clinical group | Childless controls | Group effect Wald (p) | Group-by-epoch effect Wald (p) | |
|---|---|---|---|---|---|
| SDLP (m) | 0.34 (0.02) | 0.27 (0.01) | 0.28 (0.01) | 9.53 (0.009)a | 4.10 (0.85) |
| Number of lane crossings | 13.52 (1.48) | 11.47 (1.30) | 12.63 (1.48) | 1.11 (0.57) | 4.64 (0.79) |
| SDSP (km hr−1) | 3.35 (0.21) | 2.97 (0.38) | 2.89 (0.25) | 2.18 (0.34) | 17.89 (0.02) |
| Average speed (km hr−1) | 96.18 (0.48) | 96.04 (0.36) | 95.09 (0.44) | 3.62 (0.16) | 3.29 (0.92) |
| Maximum speed (km hr−1) | 109.82 (1.23) | 107.76 (1.51) | 105.85 (1.04) | 6.10 (0.04)b | 18.37 (0.02) |
| DBQ | 1.83 (0.09) | 1.51 (0.08) | 1.63 (0.06) | 7.44 (0.02)a | – |
- Note: Data are given as mean (standard error). Bolded font represents significant effects.
- Abbreviations: DBQ, Driving Behaviour Questionnaire; SDLP, standard deviation of lane position; SDSP, standard deviation of speed.
- a Significant differences: Clinical group > Non-clinical group, Childless controls.
- b Significant differences: Clinical group > Childless controls.
When time on task effects were analysed, significant group-by-epoch interaction terms were found for both SDSP (Wald = 17.89, p = 0.02) and maximum speed (Wald = 18.37, p = 0.02). Post hoc analyses revealed that for SDSP, there were no significant group differences in epochs 1–3. Significant group differences emerged in epochs 4 and 5, with women in the clinical group demonstrating greater speed deviation compared with women in the non-clinical group in epoch 4 (Mdifference = 1.06 km hr−1, p = 0.002) and compared with childless controls in epoch 5 (Mdifference = 1.04 km hr−1, p = 0.009; Figure 1). For maximum speed, significant differences between groups were also not apparent from the start of the drive, but emerged in latter epochs. In epochs 2 and 4, higher maximal speed values were detected in the clinical group compared with both the non-clinical (Mdifference = 2.92 km hr−1, p = 0.03; Mdifference = 3.56 km hr−1, p = 0.002, respectively) and childless control groups (Mdifference = 4.32 km hr−1, p = 0.002; Mdifference = 2.76 km hr−1, p = 0.05, respectively). In epoch 5, women in the clinical group had significantly higher maximal speed than women in the childless control group (Mdifference = 2.89 km hr−1, p = 0.009). Group-by-epoch interaction terms were not significant for other driving metrics, and are depicted in Supplementary Figure S1.
4 DISCUSSION
The primary aim of this study was to investigate whether mothers of infants with insomnia are more vulnerable to impaired driving, by objectively assessing their driving performance against mothers of well-sleeping infants and childless women. Overall, our findings lend support to this hypothesis. Mothers of infants with insomnia exhibited significantly greater lane deviation (i.e. SDLP) compared with both control groups across the entire duration of the simulated drive task. This increased tendency to swerve or drift can elevate the chances of collisions with other vehicles or roadside objects (Verster & Roth, 2011). Notably, the magnitude of deviation observed in the clinical group compared with control groups (Hedge's g = 0.83, 0.95) was similar to that documented in drivers with a blood alcohol level of 0.08% and 0.11% compared with sober drivers (Hedge's g = 0.83 and 0.98, respectively) in a previous study utilizing a similar driving scenario (Mets et al., 2011). In most western industrialized countries, the legal blood alcohol level limit is 0.05% or 0.08%, as driving at such conditions is associated with dramatic increases in vehicular crash risk (World Health Organization, 2015). Therefore, not only did mothers of infants with insomnia perform significantly worse than controls in terms of lateral vehicular control, but they also performed at a level that exceeds legal limits in most western countries.
Significant differences between groups in maximum speed were also detected, with mothers in the clinical group demonstrating a higher maximum speed than childless women. The average maximal speed in the clinical group was approximately 110 km hr−1, exceeding the designated scenario speed limit by 10%, while the childless control group recorded an average maximum speed of approximately 106 km hr−1. Speeding is considered a serious driving-related risk behaviour, as it increases both the probability of being involved in a car crash and crash severity (Elvik, 2005). The augmented propensity for speeding in combination with the greater tendency to experience lane weaving further highlight the potential road safety risk for mothers of infants with insomnia (Aarts & Van Schagen, 2006). These findings align with previous evidence for the association between inadequate sleep and impaired driving among parents of young children (Maasalo et al., 2017; Malish et al., 2016; Sánchez-García et al., 2019). Our study extends previous evidence by demonstrating these links using objective assessment of driving performance, and by highlighting the potential risk for road safety specifically associated with having an infant with insomnia, rather than simply having an infant.
Our findings also suggest that mothers of infants with insomnia may be aware that their driving is compromised. Self-reported driving behaviour differed significantly between groups, with mothers in the clinical group reporting poorer self-perceived driving behaviour (e.g. more errors, lapses and/or violations) compared with the two control groups. The concurrence between subjective and objective assessments of driving reinforces the validity of differences detected between groups. Furthermore, this awareness may be critical when considering risk management countermeasures, such as direct education and targeted awareness campaigns (Sprajcer et al., 2022). Mothers of infants with insomnia may be more open to implement countermeasures such as daytime napping, use of engineering controls (e.g. in-vehicle fatigue detection technology) or alternate transportation given that they acknowledge their heightened driving-related hazard (Sikander & Anwar, 2018).
Analysis of simulated driving measures by epoch yielded partial evidence of deterioration in driving performance with time on task in the clinical group relative to control groups. While no group differences were apparent in speed deviation in the first three epochs, these emerged in the latter two epochs, with mothers of infants with insomnia demonstrating greater deviation than controls. Correspondingly, maximum speed was significantly higher in the clinical group only in epochs 2, 4 and 5. These results are consistent with prior research suggesting that driving performance may deteriorate over time when driving on monotonous roadways, although previous studies documented such deterioration in lateral position measures, rather than speed measures (Marando et al., 2022; Zeller et al., 2020). While further work is warranted to replicate this study's findings, the results suggest that lateral control may be hindered in mothers of infants with insomnia even during brief drives, yet speed control may only begin to deteriorate after 5–15 min of driving.
Building on previous research, this study focused on the potential driving risks faced by female drivers, specifically mothers of infants with insomnia, as opposed to fathers. Mothers of infants with insomnia may be particularly prone to impaired driving for several reasons. First, maternal caregivers typically bear most of the responsibility for infant care, including feeding and nighttime care, particularly when breastfeeding is involved (Australian Institute of Family Studies, 2017). Consequently, maternal sleep is more closely linked to that of their infants than paternal sleep (Horwitz et al., 2023), and mothers of young children may experience greater disruptions in sleep patterns compared with fathers (Richter et al., 2019). Furthermore, when driving with an infant passenger, mothers may be more inclined to respond to infant signals such as crying, as they have lower tolerance for child distress than fathers (Kahn et al., 2018; Sadeh et al., 2016). Analysis of crash data from the USA suggests that young females are at higher risk of crashes when driving with an infant passenger than with no passengers, yet this risk decreases when another adult is present (Maasalo et al., 2017). This finding is interpreted as a protective effect of an adult passenger, who can attend to the infant, enabling the driver to focus on the task of operating the vehicle. Still, these data do not indicate whether female drivers were in fact mothers, nor do they permit comparison between mother and father drivers. Given that fathers of young children with insomnia may also experience sleep disruption, further research is needed to assess driving performance in this population.
Alongside its strengths, including the objective assessment of driving performance within a controlled environment and utilization of two distinct control groups, this study has several limitations. First, its design does not allow for definitive conclusions regarding causality, as sleep was not directly manipulated. Furthermore, sleep was assessed using self-report measures rather than objective (e.g. via actigraphy) outcomes. However, given the substantial differences in reported sleep duration (~1.8 hr) and fragmentation (~1.1 awakening and ~77 min of nocturnal wakefulness), as well as sleepiness and insomnia symptomology between the clinical and control groups, it is highly plausible that inadequate sleep underlies the poorer driving outcomes observed in mothers of infants with insomnia. Second, the small sample size may have limited this study's power to detect differences between groups. Despite trends in the expected direction, the group effects were not significant for the mean number of lane crossings, SDSP and average speed. Nonetheless, the sample was sufficient to yield significant differences in maximal speed and SDLP, which is considered the most relevant and sensitive metric in investigations of drowsy driving (Verster & Roth, 2011). Future studies are therefore warranted to assess the scope of these driving-related risks with larger participant samples. Third, childless women were younger than mothers in the clinical and non-clinical group. Thus, differences in driving performance may have been partially due to age or driving experience differences. However, the significant differences in most metrics between the clinical group and both control groups suggest that the differences were specifically related to the experience of mothering an infant with sleep difficulties.
Fourth, the simulated driving scenario employed in this study encompassed a 25-min journey along a monotonous highway-like route, conducted in the morning hours. While short morning driving episodes have been demonstrated as commonplace among parents (Armstrong et al., 2015; Koppel et al., 2011), the potential for more pronounced driving impairment may emerge when considering longer durations or driving during later hours. Future studies may thus wish to use scenarios spanning varying time frames and road characteristics, driven during later periods of the day, to assess the risks faced by parents under different driving conditions. Finally, participants in our study were asked to refrain from consuming caffeine prior to attending the laboratory, yet we were unable to objectively validate caffeine abstinence via salivary or plasma concentrations, and only validated compliance via self-report. Acute caffeine intake is an effective countermeasure to cognitive and physical impairments associated with sleep loss, including driving performance (Irwin et al., 2020). Thus, future studies could validate adherence to pre-trial caffeine abstinence procedures to ensure precise evaluations of impairment magnitude are obtained.
Notwithstanding these limitations, the results of this study indicate that mothers of infants with insomnia may experience impaired vehicular control, that could potentially lead to increased risk of fatigue-related motor vehicle crashes. Given the serious nature of their consequences, these findings have significant implications for public health and safety. Further research is necessary to validate and quantify the extent of road safety risks linked to infant insomnia. To promote safe driving and reduce the risk of motor vehicle crashes, healthcare providers and policymakers should be aware of these hazards, encourage at-risk mothers to seek treatment for their infant's insomnia, and develop measures to mitigate driving-related risks for this population.
PATIENT CONSENT STATEMENT
All participants provided written informed consent.
AUTHOR CONTRIBUTIONS
Michal Kahn: Conceptualization; investigation; writing – original draft; methodology; visualization; formal analysis; project administration; data curation. Christopher Irwin: Supervision; methodology; writing – review and editing; data curation; software. Meg Pillion: Conceptualization; investigation; writing – review and editing; methodology; project administration. Hannah Whittall: Conceptualization; writing – review and editing; methodology; investigation. Josh Fitton: Writing – review and editing; investigation. Madeline Sprajcer: Writing – review and editing. Michael Gradisar: Conceptualization; supervision; funding acquisition; writing – review and editing; methodology; resources.
FUNDING INFORMATION
This study was supported by funding from Flinders University.
CONFLICT OF INTEREST STATEMENT
The authors declare that they have no conflict of interest.
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
