In a recent study published in Scientific Reports, researchers evaluated the predictive power of numerous sleep-disordered breathing (SDB) parameters, namely overnight pulse oximeter readings and the Epworth Sleepiness Scale (ESS) in predicting truck collisions attributed to microsleep-related behaviors at the wheel (TC-MRBs).
Study: Risk factors for collisions attributed to microsleep-related behaviors while driving in professional truck drivers. Image Credit: Gorodenkoff/Shutterstock.com
Background
Their findings from a cohort of Japanese truck drivers confirmed by dashcam footage highlight that the 4% oxygen desaturation index (ODI), nadir oxygen saturation (SpO2), and, most significantly, night-time truck driving were associated considerably with TC-MRBs.
Notably, the traditional use of just overnight pulse oximeter readings was found to have poor TC-MRB-predictive power, suggesting the need for combining subjective and objective assessments to improve TC-MRB-predictive accuracy and prevent collisions among professional truck drivers.
SDB and its role in TC-MRBs
Sleep-disordered breathing (SDB) is a medical term for several sleep-related breathing disorders, including obstructive sleep apnea (OSA), central sleep apnea, sleep-related hypoventilation, and hypoxemia.
Characteristic by sleep-obstructing factors, including airway obstruction and sleep interruption due to arousal caused by the absence of respiratory effort, SDB is alarmingly prevalent and is estimated to affect between 15% and 50% of the global population.
Given its association with excessive daytime sleepiness (EDS), cardiovascular diseases (CVDs), dementia, cognitive dysfunction, and metabolic disorders, SDB presents a modifiable risk factor in overnight professions such as professional truck driving.
Unfortunately, SDB among professional truck drivers is substantially higher than the global mean, with between 40.1% and 71.8% of all truck drivers estimated to suffer from the condition.
Previous truck-driver-specific research has identified several adverse outcomes of SDB among this cohort, including hypertension, CVD, anxiety, depression, and metabolic discomfort.
Notably, SDB has been implicated in multiple motor vehicle accidents (MVAs) even among the general population and is assumed to be a key predictor of microsleep-related behaviors at the wheel (TC-MRBs) for truck-driver-induced MVAs.
Traditional clinical interventions against SDB, including diagnosis via full polysomnography (PSG) and treatment using continuous positive pressure (CPAP), have proven useful in significantly decreasing MVAs and CVDs in SDB patients.
Unfortunately, full polysomnography is expensive, time-consuming, and requires specialized equipment and human resources for examination, making it unfeasible for large truck companies.
Therefore, the Epworth Sleepiness Scale (ESS), in tandem with overnight pulse oximeter measures, is often used as proxies for PSG in identifying SDB among truck drivers and, in turn, predicting their MVAs risk.
Unfortunately, these assessments have proven limited in their utility given the ‘self-reported’ nature of current ‘falling asleep at the wheel’ occurrences.
Furthermore, outcomes of studies measuring the predictive accuracy and application of pulse oximeter readings in SDB evaluations remain confounding.
Identifying easy-to-estimate predictors of SDB among this highly at-risk population (truck drivers) would allow for improved MVA mitigation plans and employment policies, thereby benefiting all parties involved.
About the study
The present study is a retrospective, nationwide database evaluation incorporating traditional SpO2 measures, subjective truck-driver sleepiness reports, and objective dashcam footage to evaluate the associations between sleepiness/SDB and TC-MRBs among professional truck drivers.
The study sample cohort was derived from a large Japanese transportation company with more than 400 branches across the country and more than 5,450 actively employed truck drivers.
Participants were enrolled in the study if they were at least 18 and had been involved in none or one suspected sleepiness-associated collision.
Drivers who received medical assessments (especially polysomnography) over the preceding year were excluded from the analyses.
Data collection included sociodemographics, anthropometrics (age, sex, body mass index [BMI]), and medical records (specifically, the average time between health checks and a TC-MRB event). Participants’ systolic and diastolic blood pressure and dashcam footage were acquired.
“In this study, the TC-MRBs group consisted of professional truck drivers who reported during interviews that their truck collisions were caused by falling asleep. After the interviews, we reviewed the 1-min dashcam video footage before the TC-MRBs to confirm that all the professional truck drivers had been involved in TC-MRBs. The dashcam video footage was recorded from the inside and outside of the truck to confirm the behavior of both the truck driver and the other vehicle’s driver.”
TC-MRBs group participants (with one MVA on record) were one-to-one matched with controls (no recorded MVAs; non-TC-MRBs group) to allow for comparisons to identify practices for reducing TC-MRBs among professional truck drivers.
Experimental interventions included overnight pulse oximetry and sleep tests conducted from the convenience of the participant’s homes. Pulse oximetry metrics included the 3% oxygen desaturation index (ODI), 4% ODI, nadir SpO2 (lowest recorded oxygen saturation), and mean SpO2.
“The Gaussian distribution was evaluated using the Shapiro–Wilk test, and the t-test or Mann–Whitney U test was used to compare continuous data. Drivers without TC-MRBs were matched with drivers with TC-MRBs using the propensity matching method (nearest neighbor method, TC-MRBs: non-TC-MRBs = 1:1, caliper: 0.20). The included variables for the matching were age and sex. For comparison, we also calculated the effect size using Cohen’s.”
Study findings and conclusions
Of the 5,454 truck drivers initially included in the study, 862 reported multiple collisions, 4,353 reported collisions not attributed to sleepiness, 165 received recent (<1 year prior) health checkups, and 20 did not provide dashcam footage, all of whom were excluded from the study. Therefore, the final sample size was 108 (N = 54 per study cohort).
Contradicting previous hypotheses, 3% ODI, mean SpO2, ESS scorers, and EDS presence were not statistically associated with TC-MRBs. In contrast, nadir SpO2 and 4% ODI were significantly related to TC-MRBs among evaluated truck drivers.
Unfortunately, when assessed in isolation, the predictive power of overnight pulse oximetry readings was poor.
“The receiver operating characteristic (ROC) curve analysis showed low AUC values with 4% ODI and nadir SpO2 being poor at identifying TC-MRBs, whereas night-time driver was strongly associated with TC-MRBs in the professional driver.”
Study findings suggest that a combination of objective and subjective lines of evidence may perform significantly better than the traditional, solely pulse oximetry diagnostic design.
Specifically, a questionnaire that addresses sleepiness in tandem with home cardiorespiratory monitoring tools or overnight pulse oximetry, combined with dashcam footage, maybe the best non-polysomnography to identify at-risk individuals.
