Driving safety among Israeli military physicians in combat units

Abstract

Objective

The objective of this study was to characterize driving safety and the factors affecting it among physicians in combat units in the Israel Defense Forces (IDF), who have high workloads and substantial sleep deprivation, which could influence driving safety.

Methods

This cross-sectional study included physicians in combat units who had a personal vehicle equipped with an advanced driver assistance system (ADAS). The study outcomes included events such as drowsy driving or falling asleep while driving, as well as motor vehicle accidents (MVAs), obtained from self-reports from digital questionnaires, and objective ADAS driving safety scores. Sleep hours, burnout scores (Maslach Burnout Inventory), combat activity levels, and demographic characteristics were obtained through digital questionnaires, and their effects on the outcomes were analyzed.

Results

Sixty-four military combat unit physicians were included in the study. No differences in drowsy driving, MVAs, or ADAS scores were found between the 2 combat activity level groups. The results showed that 82% of participants reported dozing off while driving, and this was positively correlated with accelerations (β = 0.19; P = .004) and negatively correlated (adjusted R² = 21%) with hours of sleep (β = −0.28; P = .001). Eleven percent reported experiencing MVAs, none of whom required hospitalization. The mean ADAS safety score was 87.17 ± 7.54, and this was positively correlated with the cynicism score (β = 1.45; P = .04; adjusted R² = 4.7%). No association between dozing off/falling asleep while driving and the reported MVAs was found (P = .10 and P = .27, respectively).

Conclusion

Physicians in combat units have a low incidence of MVAs and high ADAS scores. This may be attributed to the high safety climate enforced in military units. However, the high rate of dozing off while driving highlights the importance of addressing driving safety in this population.

Keywords:

• Military physicians
• driving safety
• hours of sleep
• burnout
• advanced driver assistance system (ADAS)

Introduction

Motor vehicle accidents (MVAs) are the leading cause of death among various age groups, but mainly among men between the ages of 15 and 24 (Goldberger et al. 2014). Driving safety among young physicians was primarily studied among residents after night shifts. Doctors and nurses are involved in more MVAs on their way home after a shift compared with other health care workers (HaGani et al. 2022). Young surgical and medical residents have been found to frequently be involved in near misses while driving as well as MVAs following night shifts, with these events correlating positively with low quality of life, burnout, depression symptoms, fatigue, and sleepiness (Freedman-Weiss et al. 2021; West et al. 2012).

A few studies found a correlation between lack of sleep or low sleep quality and low performance in a driving simulator, including impaired braking response and less vehicle control inside the lane. Furthermore, sleep deprivation has been found to have a more significant impact on driving quality than alcohol consumption, leading to more lane departures in real life (Lowrie and Brownlow 2020).

Our study aims to examine the driving safety of military physicians in combat units in the Israel Defense Forces (IDF), who, about 5 years ago, were assigned private vehicles to provide medical care to army companies in different locations. We believe that this unique population of military physicians has a high workload, a high level of burnout, and substantial sleep deprivation, which might influence their driving quality and safety compared with other military physicians.

Our study hypotheses were as follows:

1. Military physicians in combat units will have low levels of driving safety.

2. Based on previous studies on medical residents, we hypothesize that military physicians who have a more substantial workload and work in units with higher activity level (which we would associate with less sleep or higher burnout scores; see instruments section in the Methods section) will have lower levels of driving safety, more instances of drowsy driving, and a higher rate of MVAs.

3. Higher levels of sleep deprivation and burnout scores will correlate positively with outcomes of drowsy driving and falling asleep while driving.

4. Higher levels of sleep deprivation and burnout scores will correlate positively with MVAs.

5. Sleep deprivation and higher burnout scores will correlate negatively with the advanced driver assistance system (ADAS) score.

Methods

Design and participants

We conducted a cross-sectional study among IDF military physicians serving in combat units. The different units in the IDF are categorized according to activity levels related to the risk involved in serving in the unit, where they are stationed, the type of unit, and the number of days per week the soldiers are away from home (IDF 2022). The details of all military physicians were examined but only military physicians in combat units who had a personal vehicle with an ADAS installed were included in the study. Participants who drove an average of less than 100 km per month or who did not respond to the questionnaire were excluded.

After an explanation of the study and its goals, all participants agreed to complete the questionnaire. Informed consent was obtained from all participants. All procedures were performed in compliance with relevant laws and institutional guidelines and after approval by the appropriate institutional committee. The study was conducted in adherence with the tenets of the Declaration of Helsinki.

Regarding activity level, the different units in the IDF were categorized according to the following: The risk involved in serving in the unit and participating in its operations, the unit’s location, and the number of days per week the soldiers were away from home. In the context of military physicians, battalions and special units are activity level A + units and divisions are activity level A units. Battalions are directly responsible for routine as well as emergency border security. Thus, battalion physicians operate in the field and are responsible for the health of the soldiers, often in distant companies. Division physicians are part of an integrative body. Therefore, they spend less time in the field, have a lower workload, and are able to return home more often for weekends and vacations.

Study tools

Questionnaire completed by study participants

We created a Hebrew questionnaire on Google Forms (Zabarsky and Shmueli 2022). The questionnaire included 35 items. A statement on participant anonymity was included at the start of the questionnaire. We asked for the following information: Age, gender, driving time from home to base, family status (single; in a relationship; or in a relationship, with children), duration of current duty, type of duty, rank, activity level, an estimation of the amount of sleep per 5 days of work per week, estimates of the number of times the participant dozed off or fell asleep while driving, and how many MVAs the participant had been involved in while driving in the last 4 months.

The next section of the questionnaire was the burnout questionnaire. It was based on the Maslach Burnout Inventory (Maslach et al. 1996), which was translated to Hebrew in 1988 by Professor Etzion (see Appendix, online supplement). The tool was validated in its Hebrew form in a previous work by Friedman I (Friedman 1999). The questionnaire’s final burnout score was composed of 3 contributing scores: Emotional exhaustion, cynicism, and personal accomplishment.

Advanced driver assistance system safety data

The personal vehicles of all study participants had an ADAS (Ituran 2023). Data recorded by the ADAS included the number of times the driver exceeded the speed limit (speed alarms), the number of sudden decelerations the driver performed (braking events), the number of excessive accelerations (accelerations), the number of times the driver drove fast or dangerously around another vehicle to bypass it (passing events), the number of aberrant turns the driver made (turns), the quantitative weighted mean ADAS safety score (contributed to by all abovementioned components), and the driver’s monthly mileage driven. We calculated the weighted arithmetic mean mileage per month for each driver.

Study outcomes

The primary outcomes of this study were (1) drowsy driving or falling asleep while driving (events per 4-month period); (2) MVAs derived from self-reports from digital questionnaires (events per 4-month period), and (3) the objective ADAS driving safety score and its components (average per 4-month period). We also analyzed the putative effects of various factors on the primary outcomes. These factors included different activity levels, age, gender, family status, driving distance from home to base, monthly driving distance, hours of sleep, and Maslach Burnout Inventory score and its components.

Statistical analysis

Physicians serving at activity level A+ (39 participants) and those serving at activity level A (25 participants) were compared using a t-test for continuous independent variables and outcomes, including age; duration of current duty; average monthly driving distance; driving time from base to home; weekly hours of sleep; average scores for emotional exhaustion, cynicism, personal accomplishment, and total burnout score; dozing off while driving events; falling asleep while driving events; MVAs; serious MVAs; and ADAS safety score and its components (speed alarms, braking events, accelerations, passing events, and turns). Physicians serving at activity level A + and those at activity level A were compared using the chi-square test for categorical variables, including sex, family status, and rank.

The outcomes of drowsy driving, MVAs, and ADAS score and its components were compared according to gender and family status using t-test and analysis of variance, respectively.

We first tested the effects of potential baseline factors on the outcomes among all participants using univariate regression analysis. The effect of dozing off and falling asleep on MVAs was also tested using univariate regression analysis. Subsequently, additional factors that had a significant effect on any of the outcomes in the univariate analysis were included in a multivariate stepwise linear regression analysis. The study sample satisfied the assumptions of regression analysis including normal distribution and homoscedasticity. Multicollinearity of independent variables was ruled out using a variance inflation factor of <2.5 for all variables. Proportions of categorical data are presented as percentages. Continuous data are presented as the mean ± standard deviation (range). Regression results are presented as adjusted square regression coefficients (adjusted R²). A 2-sided P value <.05 was considered statistically significant. The data were analyzed using SPSS v26 (IBM Corp 2019).

Results

Characteristics of participants at activity levels A + and A

Out of 88 military physicians serving in combat units, 64 completed the digital questionnaire and met the study’s inclusion criteria. The first group included 39 physicians serving in activity level A + units. This group included battalion and special unit physicians, who have the highest workload and who get fewer hours of sleep. The second group included 25 military physicians in activity level A units. This group included division medical officers, clinic commanders, and training base physicians. Thus, both groups were similar in every parameter that was not essential to the study.

Physicians serving at activity level A + were younger by 0.9 years (P = .03) and had a lower rank (P = .006) compared to physicians serving at activity level A. No difference in hours of sleep or burnout scores was found between these groups (Tables 1 and 2).

Table 1. Comparison of independent demographic characteristics of physicians by activity level.

Variable	A + activity level	A activity level	P value
Number of participants	39	25	—
Male (%)	66.6	84	.37
Age (years)	27.9 ± 1.7	28.8 ± 1.3	.03*
Family status (%)			.25
Single	41	32
In a relationship	46	32
In a relationship, with children	13	36
Duration of current duty (months)	12.2 ± 4.6	13.3 ± 7.3	.46
Lieutenant rank (%)	69%	28%	.006**
Driving time from base to home (minutes)	105.0 ± 49.3	138.0 ± 85.0	.9
Monthly driving distance	1,986.1 ± 766.5	2,453.8 ± 1,713.3	.21
Hours of sleep per work week	24.5 ± 5.4	26.6 ± 8.9	.3

Continuous variables are presented as mean ± SD. Proportions of categorical data are presented as percentages.

* P < .05. **P < .01 (Significant values are marked in bold).

Table 2. Comparison of burnout scores of physicians by activity level.

Variable	A + activity level	A activity level	P value
Number of participants	39	25	—
Maslach Burnout Inventory scorea	3.7 ± 1.0	3.8 ± 0.7	.48
Emotional exhaustion score	4.9 ± 1.1	5.0 ± 1.1	.78
Cynicism score	3.5 ± 1.4	3.8 ± 1.0	.28
Personal accomplishment score	2.7 ± 0.9	2.8 ± 0.8	.74

^a The questionnaire’s final burnout score is composed of 3 contributing scores: emotional exhaustion, cynicism, and personal accomplishment.

Continuous variables are presented as mean ± SD.

* P < .05. **P < .01.

Dozing off while driving

Eighty-two percent of participants reported at least 1 dozing off event, with an average of 4.25 ± 4.78 events in 4 months (range = 0–20 events). No difference in the number of events was found for different activity levels, sex, or family status (Tables 3–5). Univariate regression analysis demonstrated a significant correlation between weekly hours of sleep (r = −0.354; P = .002), ADAS score (r = −0.29; P = .01), emotional exhaustion (r = 0.29; P = .01), braking events (r = 0.25; P = .02), accelerations (r = 0.30; P = .008), passing events (r = 0.33; P = .004), turns (r = 0.28; P = .01), and dozing off events. Upon multivariate analysis, only hours of sleep (β = −0.28; P = .001) and accelerations (β = 0.19; P = .004) remained significantly correlated with dozing off events (adjusted R² = 21%).

Table 3. Comparison of outcome variables of physicians by activity level.

Variable	A + activity level	A activity level	P value
Number of participants	39	25	—
Falling asleep while driving (events/4 months)	0.70 ± 3.20	0.60 ± 2.0	.87
Dozing off while driving (events/4 months)	3.90 ± 4.40	4.80 ± 5.40	.49
MVAs (events/4 months)	0.10 ± 0.30	0.16 ± 0.37	.51
ADAS score (average/4 months)	88.35 ± 6.65	85.12 ± 8.64	.1
Speed alarms (average/4 months)	1.09 ± 4.15	1.99 ± 6.53	.51
Accelerations (average/4 months)	4.82 ± 5.25	9.78 ± 11.55	.053
Braking events (average/4 months)	17.3 ± 11.55	23.91 ± 26.99	.25
Passing events (average/4 months)	2.56 ± 3.87	6.74 ± 13.01	.13
Turns (average/4 months)	9.35 ± 8.82	16.26 ± 18.83	.10

Continuous variables are presented as mean ± SD.

* P < .05. **P < .01.

Table 4. Comparison of outcome variables by gender.

Variable	Male	Female	P value
Number of participants	47	17	—
Falling asleep while driving (events/4 months)	0.83 ± 3.2	0.12 ± 0.33	.36
Dozing off while driving (events/4 months)	4.21 ± 4.68	4.38 ± 5.24	.91
MVAs (events/4 months)	0.10 ± 0.30	0.18 ± 0.39	.43
ADAS score (average/4 months)	86.57 ± 7.32	88.88 ± 4.73	.28
Speed alarms (average/4 months)	1.32 ± 4.73	1.64 ± 6.21	.82
Accelerations (average/4 months)	6.18 ± 7.48	7.82 ± 11.05	.49
Braking events (average/4 months)	20.97 ± 20.64	15.41 ± 12.80	.30
Passing events (average/4 months)	4.96 ± 9.93	1.78 ± 2.11	.19
Turns (average/4 months)	13.04 ± 15.24	8.61 ± 7.24	.12

Continuous variables are presented as mean ± SD.

* P < .05. **P < .01.

Table 5. Comparison of outcome variables by relationship status.

Variable	Single	In a relationship	In a relationship, with children	P value
Number of participants (%)	24 (36)	28 (42)	14 (21)	—
Falling asleep while driving (events/4 months)	1.00 ± 4.06	0.50 ± 1.95	0.36 ± 2.75	.73
Dozing off while driving (events/4 months)	4.46 ± 4.68	4.11 ± 5.02	4.14 ± 3.39	.96
MVAs (events/4 months)	0.125 ± 0.34	0.11 ± 0.31	0.14 ± 0.36	.95
ADAS score (average/4 months)	87.75 ± 8.52	87.21 ± 6.83	86.08 ± 7.58	.81
Speed alarms (average/4 months)	1.64 ± 6.53	1.35 ± 4.88	1.06 ± 2.25	.94
Accelerations (average/4 months)	6.78 ± 10.00	6.42 ± 8.77	6.64 ± 4.73	.99
Braking events (average/4 months)	17.11 ± 22.44	18.48 ± 12.01	25.82 ± 23.63	.37
Passing events (average/4 months)	5.08 ± 13.07	3.71 ± 5.51	3.40 ± 4.41	.80
Turns (average/4 months)	10.66 ± 14.63	10.43 ± 10.68	16.97 ± 17.02	.30

Continuous variables are presented as mean ± SD.

* P < .05. **P < .01.

Falling asleep while driving

Twenty percent of participants reported falling asleep while driving at least once, with an average of 0.63 ± 2.62 events in 4 months (range = 0–20 events). No difference in the number of events was found for different activity levels, sex, or family status (Tables 3–5). Univariate regression analysis demonstrated a significant negative correlation between weekly hours of sleep (r = −0.531; P < .001) and monthly driving distance (r = −0.21; P = .04) and falling asleep while driving. Upon multivariate analysis, only hours of sleep (β = −0.21; P < .0001) remained significantly correlated with falling asleep while driving (adjusted R² = 27%).

Motor vehicle accidents

Eleven percent of participants experienced an MVA once in 4 months, with no reports of more than 1 MVA or serious MVA requiring a hospital visit. No difference in the number of MVAs was found for different activity levels, sex, or family status (Tables 3–5). Univariate regression analysis demonstrated a significant positive correlation between weekly hours of sleep (r = 0.243; P = .02) and ADAS score (r = 0.25; P = .02) and MVAs. No association between dozing off/falling asleep while driving and the reported MVAs was found (P = .10 and P = .27, respectively).

Upon multivariate analysis, only the ADAS score (β = 0.01; P = .004) remained significantly correlated with MVAs (adjusted R² = 5%).

ADAS safety score and its components

The average ADAS safety score over 4 months was 87.17 ± 7.54 (range = 58.70–100 points). No difference in the score was found for different activity levels, sex, or family status (Tables 3–5). Univariate regression analysis demonstrated a significant correlation between personal accomplishment score (r = −0.21; P = .04), cynicism score (r = 0.25; P = .02), driving time from home (r = −0.23; P = .03), and ADAS safety score. Upon multivariate analysis, only the cynicism score (β = 1.45; P = .04) remained significantly correlated with the ADAS score (adjusted R² = 4.7%).

Univariate regression analysis did not demonstrate any significant correlation between speed alarms and other independent variables.

Univariate regression analysis demonstrated a significant correlation between driving time from home (r = 0.37; P = .001), monthly driving distance (r = −0.39; P = .001), cynicism score (r = −0.2; P = .005), and braking events. Upon multivariate analysis, driving time from home (β = 0.09; P = .006) and monthly driving distance (β = 0.005; P = .003) remained significantly correlated with braking events (adjusted R² = 22.7%).

Univariate regression analysis demonstrated a significant correlation between driving time from home (r = 0.25; P = .02) and accelerations.

Univariate regression analysis demonstrated a significant correlation between driving time from home (r = 0.24; P = .03), monthly driving distance (r = 0.34; P = .003), personal accomplishment score (r = 0.21; P = .04), and passing events. Upon multivariate analysis, only monthly driving distance (β = 0.002; P = .006) remained significantly correlated with passing events (adjusted R² = 9.9%).

Univariate regression analysis demonstrated a significant correlation between driving time from home (r = 0.32; P = .004), monthly driving distance (r = 0.24; P = .03), and turns. Upon multivariate analysis, only driving time from home (β = 0.066; P = .009) remained significantly correlated with turns (adjusted R² = 8.9%).

Discussion

This study aimed to evaluate the driving safety of military physicians in combat units. Driving safety was subjectively assessed by questions regarding the rate of dozing off while driving, falling asleep while driving, and MVAs and objectively assessed by adding the ADAS score and its components for each participant. Overall, we found a low number of driving events and a high average ADAS score.

Regarding the driving safety parameters, the only statistically significant difference between physicians in activity levels A + and A was in accelerations per distance. The amount of sleep correlated negatively with dozing off and falling asleep events, signifying reduced safety with less hours of sleep. The amount of sleep correlated positively with MVAs, paradoxically signifying improved safety with less hours of sleep. The ADAS score correlated with a higher cynicism score, indicating that safer driving was observed with higher burnout.

Our first hypothesis was that military physicians would demonstrate low driving safety. This hypothesis was refuted. Only 11% of all participants reported 1 MVA in 4 months, and no participants reported more than 1. Eighty-two percent reported at least 1 dozing off event. No participants reported a serious MVA requiring examination in a hospital. Our study’s finding of overall high level of driving safety contrasts with the findings of HaGani et al. (2022), which included 60 physicians. Of the 60 physicians in HaGani et al. (2022) study, 24 reported being involved in MVAs, of whom 40% resulted in physical harm, additionally 70% of the physicians in HaGani et al. (2022) study reported dozing off while driving. Possible explanations for the high level of driving safety in our study are intense supervision over driving quality using the ADAS and the investment in safety education by the army, especially in combat/field units.

The ADAS used in the IDF is produced by the Israeli company Ituran based on Global Positioning System technology using an acceleration sensor (Ituran 2023) It supplies the different driving quality parameters and total scores we introduced in this study. Unique to the IDF, the driver’s direct commander receives an SMS message noting every time the driver exceeds the speed limit. Different units use this supervision tool in different ways, from punishment for poor driving up to rewards for good driving.

In addition, safety education is mandatory in all IDF units for every soldier and officer. Twice a year each unit holds a conference, including lectures about driving safety, relevant to the appropriate season and weather conditions.

These measures are in contrast to physicians working in hospitals, which generally do not invest in driving safety education.

Our second hypothesis was that military physicians operating with a higher workload and in higher activity level units would have a lower level of driving safety, a greater tendency to fall asleep while driving, and a higher rate of MVAs. This hypothesis was based on studies performed on medical residents that found that MVAs and drowsy driving correlate with sleep deprivation and higher burnout (West et al. 2012; Lowrie and Brownlow 2020; Freedman-Weiss et al. 2021; McManus and Stavrinos 2021).

This hypothesis was not supported in this study. First, our assumption that physicians in activity level A + units sleep less and drive greater distances was incorrect. They slept less but drove less compared to the physicians in activity level A units, and these differences were statistically insignificant.

We did not find any statistically significant difference between the 2 groups regarding driving outcomes except accelerations per distance. It is possible that such differences exist but were not found because of the small sample of this study. A possible explanation for the finding that there were no differences is that the 2 groups shared similar independent variables and therefore the driving outcomes were similar. Alternatively, although battalion physicians are more sleep deprived, this deprivation is neutralized by the emphasis on safety education in their units.

Our third hypothesis was that sleep deprivation and higher burnout scores would correlate positively with the driving outcomes of dozing off and falling asleep events. Both of these outcomes correlated with sleep deprivation. However, regarding burnout, the only correlation was between dozing off and emotional exhaustion score—1 component of the burnout questionnaire—and only on univariate analysis.

This result adds to previous studies following night shifts residents who demonstrated drowsier driving (Anderson et al. 2018; Green et al. 2020).

HaGani et al. (2022) studied drowsy driving among 90 health care workers. In their study, drowsy driving correlated with higher burnout parameters (higher emotional exhaustion and lower personal accomplishment), especially among physicians. We did not find these correlations (to be statistically significant), except for the one for dozing off and emotional exhaustion.

Our fourth hypothesis was that MVAs would correlate with sleep deprivation and higher burnout scores. In a study by West et al. (2012), among 340 residents, fatigue, sleepiness, and higher burnout parameters correlated significantly with MVAs. In our study, MVAs correlated positively with hours of sleep and ADAS score; upon multivariate analysis, only the ADAS score remained. It is possible that we could not reinforce our hypothesis because of the small sample size and the low number of MVAs in the sample.

Our fifth hypothesis was that sleep deprivation and higher burnout scores would correlate negatively with the ADAS score. This hypothesis is based on previous studies. Lowrie and Brownlow (2020) used a driving simulator to compare 30 participants under conditions of sufficient sleep, sleep deprivation, and alcohol consumption. When sleep deprived, the participants had lower simulator scores compared to their sufficient sleep state and after alcohol consumption. Talusan et al. (2014) compared 24 residents in the Night Floats shifts system on 12-h shifts to 34 residents in the regular night shifts system on 28-h shifts; the second group experienced greater sleep deprivation, and it was found that they had longer braking times.

In our study, the ADAS score correlated negatively with a reported driving time from base to home and personal accomplishment and positively with cynicism. Upon multivariate analysis, the only correlation that remained was a positive correlation with the cynicism score.

We believe the explanation for this counterintuitive correlation lies in work conversation and mobile phone use while driving. Physicians with greater job burnout, who feel less meaningfulness in their work and thus more cynicism, tend to use their phones to engage in work-related conversations while driving. Another possible explanation is that physicians in units with strict discipline and safety education opportunities feel more burned out from their interactions with that unit (degree of discipline in the unit as a confounder).

Regarding the finding of lack of association between dozing off/falling asleep while driving and reported MVAs, we believe this correlation does exist but was not found because of the small sample size, which resulted in an overall small number of MVAs reported in the study.

Limitations

Uniqueness, strengths, and weaknesses of the study include the following:

1. Our study is unique because of its participants. When examining the literature, we could not find studies examining the driving safety of this population. The military physician population is unique because they have the professional workload of a resident, together with the responsibility of an army officer, commanding soldiers and operating in a strict and unpredictable military workplace (Jewish Virtual Library 2022).

2. All participants were military physicians in field units. This means the 2 groups we compared were similar in many parameters, thereby avoiding many possible confounders. However, this comparison revealed differences that were too small to be statistically significant between the 2 groups.

3. We used both subjective reported data and objective data.

Limitations of this study included the following:

1. Reasons for noncompliance with study questionnaire filling (which result in exclusion from participation in the study) were high workload and time management abilities. Both confound the answers to the questionnaire, and therefore, the study is inherently biased. Another reason for noncompliance was doubt about anonymity, meaning that those who did participate might have underreported driving events.

2. Participants’ medical conditions were not taken into account, which might have confounded the results. For example, early pregnancy and depression can influence tiredness, amount and quality of sleep, concentration, and driving outcomes.

3. A significant limitation of the study is the small sample size, which influenced the ability to take note of statistically significant differences that exist in real life.

4. Another limitation of the study was the short period in which it took place. Both objective driving data and subjective questionnaire answers were limited to a 4-month period. The advantage was minimization of recall bias. The disadvantage was overall low numbers of driving outcomes. Similar to the small sample size, it influenced the ability to take note of statistically significant differences that exist in real life.

Practical implications of this study are as follows:

1. The military physicians in the study had demographic characteristics similar to Israeli medical residents but had better driving safety outcomes. Therefore, we recommend researching the reasons for this difference and suggesting ways to improve driving safety among medical residents.

2. From the results of the study, it is possible to formulate a recommendation regarding a sufficient amount of sleep for the study population. Sleep deprivation has shown a connection to several driving safety indicators among military doctors, including dozing off and falling asleep while driving.

3. Use of the ADAS score to promote driving safety by screening units needing interventions.

The purpose of this study was to characterize driving safety among military physicians in combat units. Overall, the physicians exhibited a high driving safety profile, reflected by a low number of MVAs and high ADAS safety scores. We attribute these safe driving patterns to increased supervision of driving behaviors using the ADAS and the investment of resources in safety education required by the army.

We did not find a statistically significant difference in driving outcomes between activity levels A and A+. Fewer hours of sleep correlated with drowsy driving and falling asleep while driving. Higher acceleration scores correlated with drowsy driving. MVAs correlated positively with ADAS scores. High ADAS scores correlated with the cynicism score.

Ethics approval

All procedures were performed in compliance with relevant laws and institutional guidelines and the appropriate institutional committee has approved them. Informed consent was obtained from all participants. The study was conducted in adherence to the tenets of the Helsinki Declaration.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability

Data reported in this work are available upon request from the corresponding author.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.