Introduction

Acute work-related stress and individuals’ coping abilities with potential traumatic events (Westgard et al.1993; McFarlane et al. 2009) as well as the development of stress-related disorders (Statista 2016) have been investigated by numerous international studies over the past years.

In particular, the stress load and illness rate of emergency service employees are exceptionally high (Badura et al. 2012). Up to date, however, there has been little research about the acute stress load of emergency physicians of the Helicopter Emergency Medical Service (HEMS). For a better understanding of the stress load, the objective stress load on different working days such as an air-rescue day, a clinic day, or personal stress situations needs to be compared.

With respect to the air-rescue stressors, to our knowledge, only three studies have investigated vital stress parameters during the working days of the HEMS EPs. Benzer et al. (1991) explored the change in physiological parameters during 50 emergency operations in the rescue helicopter, which were compared specifically for six main phases: alert, landing, flight to the hospital, hospital, return flight, and end. It turned out that the highest values of the HR were reached in the phases alarm (M = 124.30 bpm, SD = 21.80) and landing (M = 125.10 bpm, SD = 23.10). Also, a high increase was observed while walking toward the patient (M = 131.40 bpm, SD = 22.70). In the phases flight to the hospital, hospital, and return flight, the HR dropped significantly; however, 15 min after the end, it was still higher than before the emergency operation. Based on the HR, the psychological stressors were present especially in the alarm situation and the interaction with the patient(s). It was noted that the perceived stress of EPs on air-rescue days is not congruential with the physiologically measured stress load (HR), because the physical activity in the phase of alarm is quite low, whereby the HR rises from M = 89 bpm to M = 124 bpm (Benzer et al. 1991). The changes in HR have also been investigated in a larger sample (118 emergency operations) of a more recent study (Carchietti et al. 2011). Hereby, exclusively the flight phase was investigated and stronger changes in HR were present in more complicated and longer flights. Unfortunately, in both studies, only the heart rate and not the activation of the autonomic nervous system was investigated.

Apart from in the studies by Benzer et al. (1991) and Carchietti et al. (2011), it should be noted that in an already published paper about the present study (Petrowski et al. 2018), the HR and HRV of EPs during the whole air-rescue day, clinic day, and control day were considered to compare and quantify the workload of the different days based on objective parameters. A separation into phases like in the study by Benzer et al. (1991) was not performed. The highest values for the HR were found for the air-rescue day (M = 84.89 bpm, SD = 11.52 bpm). The lowest standard deviation of all NN intervals (SDNN) and root mean square of the successive differences (RMSSD) values as a sign for a high activation of the ANS were found for the clinical day with M = 101.44 ms (SD = 29.09 ms) and M = 21.99 ms (SD = 8.17 ms). The low–high frequency ratio (LF/HF) shows the highest sympathetic overweight for the clinic day with M = 8.69% (SD = 3.89%).

In contrast, there exist many studies on stress concerning emergency physicians at hospitals (Adams et al. 1998; Dutheil et al. 2012). For the EPs of an emergency department, Adams et al. (1998) showed that the HR dropped from M = 81.3 bpm before to M = 71.5 bpm after an 8-h night shift (average HR was M = 84.7 bpm). In addition, significant changes in SDNN M = 66 ms (SD = 20 ms) before to M = 78 ms (SD = 16 ms) after the shift and in LF/HF from M = 5.2% (SD = 2.7%) before to M = 3.6% (SD = 2.5%) after the shift were observed. Furthermore, differences between two clinic days (24-h shift and 14-h night shift) and a control day that was comprised of a 10-h shift (clerical work) and resting at home were also investigated (Dutheil et al. 2012). During the 24-h shift, on average, the HR was M = ca. 83 bpm (range M = 82–93 bpm; RMSSD M = ca. 27 ms; LF/HF M = ca. 0.73%), whereby during the 14-h night shift M = ca was. 81 bpm (range M = 78–93 bpm; RMSSD M = ca. 31 ms; LF/HF M = ca. 0.70%). On the control day, on average, the HR was M = ca. 75 bpm (range M = 67–84 bpm; RMSSD M = ca. 33 ms; LF/HF M = ca. 0.55%). The Stress-VAS was higher for the two shifts at the hospital compared to the control day with clerical work. The authors conclude that the HRV was not influenced by the number of treated patients but by the number of life-and-death emergencies. These results indicate that a multidimensional interaction of long working days, exhaustion due to shift-work and sleep deficit results in a higher risk for chronic stress and work-related exhaustion. However, in the literature, the EP shifts in the hospital have not yet been compared to a day off work. Particularly for the EPs working in air-rescue shifts, these comparisons have not yet been published.

Until now, little is known about how the burden of the air-rescue day can be classified in relation into a clinic day and a day off (control day). Therefore, the following study compares the stress load of EPs on an air-rescue day with activity phases of a clinic day and a control day. Based on the study by Benzer et al. (1991), one could postulate that the physiological stress in the phases alarm and landing at the operational site on an air-rescue day is higher than in the activity phases of a clinic day and a control day, respectively.

Methods

Study participants

All of the participants are EPs and were recruited at the base of the helicopter emergency medical service (HEMS) of Dresden (Christoph 38 helicopter) with the support of the German Air Rescue Organization (DRF Luftrettung). Overall, N = 20 EPs (N = 17 male and N = 3 female) with a mean age of M = 44.95 (SD = 4.80) took part and completed the study. The work experience was reported with a mean value of 18.95 years (SD = 4.66). With N = 13 more anesthetists than surgeons (N = 7) took part. For more details, see Table 1.

Table 1 Characteristics of participants
Full size table

The participation in the study was on a voluntary basis. All of the participants reported being in good health.

Factors such as smoking, medication, oral contraceptives, etc. which might influence the HRV were collected in the anamnesis questionnaire.

Design and procedure

This field study with a within-subjects design has the objective to compare the physiological stress load of 20 EPs on different days. From June 2015 to December 2015, an air-rescue day, a clinical day, and a day off (control day) were analyzed for each EP; overall, 3 days were analyzed in this study. For the exact division into phases on the air-rescue day (see Table 2), the dates and status messages from all examined emergency operations were provided by the DRF Luftrettung.

Table 2 Classification of the phases of the emergency operations
Full size table

In a second manuscript (Schöniger et al. 2019) of the same study, the HRV was analyzed over the course of the emergency operations with regard to an accumulation of physiological stressors.

On average, in 2015, there were 3.69 emergency operations per day for the HEMS in Dresden (DRF Luftrettung 2016).

For the psychologic assessment, two questionnaires [Trier Inventory of Chronic Stress (TICS) by Schulz and Schlotz (1999) and the Symptom Checklist-90-Revised (SCL-90-R) by Franke (1995)] were filled out shortly before the first testing day.

The Task Force of The European Society of Cardiology (ESC) and The North American Society of Pacing and Electrophysiology (NASPE) (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology 1996) set up recommendations for the implementation and analysis of the HRV which were followed in the study design.

The subjective stress load for every activity was rated on a visual analog scale (VAS) from 1 (low) to 6 (high) on a daily protocol. The EPs were instructed that the daily protocol should always be carried along during the testing day, and that the rating should be done after the phase as soon as possible.

The local Ethics Committee of the Medical Faculty of the Technical University of Dresden, Germany accepted the study protocol (No #EK348092011).

Heart rate variability (HRV)

As a parameter for the activity of the autonomic nervous system (ANS) and the sympathetic and parasympathetic regulation, the HRV (variables RMSSD, SDNN, and LF/HF) was monitored to quantify the workload during the testing days.

While high RMSSD and SDNN as well as low LF/HF values indicate an organism’s good health, the opposite can be the case due to chronic stress, whereby the adaptability to short-term external and internal stress stimuli decreases (Curic et al. 2008).

The time domain parameters RMSSD and SDNN are directly derived from the RR intervals and behave similarly, while the RMSSD is more suitable for non-standardized conditions (Penttilä et al. 2001). At least, there is a high correlation between these parameters and the HR.

For the frequency domain in this study, the LF/HF ratio was used. While HF is an index of parasympathetic activity, LF is associated with sympathetic activation as well as with parasympathetic activation. The quotient is an expression of the sympatho-vagal balance. In comparison with the time domain parameters, the LF/HF is more vulnerable under non-standardized conditions (Penttilä et al. 2001).

A low HRV in both domains is associated with an increased risk of mortality (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology 1996) and with mental disorders like depression (Leistedt et al. 2011).

More relevant for the present study are the short-term effects of stress which can reduce the variability of the HRV significantly (Castaldo et al. 2015) as well as the ability to recover and to return to the initial values.

As monitoring system, a BioHarness™ 3 from Zephyr technology (USA) was used for a continuous HRV recording during the three testing days. The system consists of a chest strap with integrated sensor. The readout process of the raw HRV data was implemented with the software BioHarness™ Log Downloader on completion of the measurement. Subsequently, the software Polar ProTrainer 5™ (Polar, Germany) was used to compute the HRV parameters, whereby the automatic filter was set to moderate (minimum protection zone: 6 m2) to eliminate erroneous values of the R–R interval data.

To avoid overlaps during the quickly changing phases of the emergency operations, the duration of all phases was set to 3 min. Hereby, according to the standards of the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, comparability was preserved.

To consider the influence of physical activity, the phases were separated into activity phases and resting phases. Activity phases of the air-rescue day contain only the 3-min phases of the emergency operations and the average of phases 1–7 as extra value. For every EP, three emergency operations were averaged; overall, 60 emergency operations were analyzed. For the activity phases of the clinic day, two phases with operations or therapeutic interventions were chosen and 3 min was extracted, and for the activity phases of the control day, 3 min of two physical activities, such as biking, swimming, or gardening, was chosen. For the resting phases, on all days two phases with less physical activity, such as reading or watching TV, were used, and 3 min were extracted again. The timestamp for the phases of the emergency operations was provided by the DRF Luftrettung; for all other phases, the daily protocols were used. Except for the phases of the emergency operations, for all activity and resting phases, one phase was selected from the morning and one from the afternoon, to balance circadian effects in HRV. The measuring period on the air-rescue day was M = 13.87 h (SD = 1.82 h) and started at M = 8:12 and ended at M = 22:04. For the clinic day, the period was M = 10.58 h (SD = 3.82 h), from M = 7:51 to M = 18:27. On the control day, M = 10.67 h (SD = 3.55 h) was measured, starting at M = 8:47, to M = 19:28.

Psychological assessments

To assess the anticipatory cognitive appraisal of the workload and the acute and chronic stress load, two instruments were used:

  1. 1.

    The TICS by Schulz and Schlotz (1999) according to the concept by Richter and Hacker (1998) consisted of 57 items with a five-point rating scale. Nine related factors of psychosocial chronic stress were retrospectively evaluated for the previous 3 months with a good reliability. The validity was proved, for the norm sample (Petrowski et al. 2012), see Table 1.

  2. 2.

    For the screening of the general psychiatric symptomatology, the SCL-90-R (Franke 1995) was applied. The questionnaire with 90 items contains a 5-point scale ranging from 0 ‘not at all’ to 4 ‘extremely’. It shows a good reliability and validity. For the norm sample (Franke 1995) see Table 1.

Statistical analysis

The heart rate (HR) and the HRV parameters RMSSD, LF/HF and SDNN were recorded for the activity and resting phases of the air-rescue-day, the clinic day, and the control day to draw conclusions of the activity of the ANS. Using SPSS 21, mean values (M) and standard deviation (SD) were computed for all dependent variables. The normal distribution of the data was checked and confirmed prior to performing the inferential statistical analyses.

First (1), to reflect the effects during the phases of the activity of the three testing days for each phase of the emergency operations and for the average of all phases, ANOVAs with repeated measurements (factor: 3 days) as well as the Bonferroni correction (36 tests) were used. The Mauchly’s test of sphericity was used to control the assumption of sphericity. The ANOVA results were corrected by Greenhouse–Geisser whenever required. Two-sided p values of ≤ 0.0014 were considered significant.

Second (2), to reflect the effects during the resting phases of the three testing days, an ANOVA with repeated measurements (factor: 3 days) as well as the Bonferroni correction (four tests) were used. The Mauchly’s test of sphericity was used to control the assumption of sphericity. The ANOVA results were corrected by Greenhouse–Geisser whenever required. Two-sided p values of ≤ 0.0014 were considered significant.

Results

  1. 1.

    Considering the subjective stress load, all values are on a low and for the air-rescue day on a moderate level (Table 3).

    Table 3 Subjective stress load measured by a visual analog scale (VAS), 1 = low, 6 = high
    Full size table
  2. 2.

    Comparing the activity phases of the air-rescue day (emergency operation phase 1–7) with the activity phases of the clinic and the control day, the ANOVA with repeated measures showed significant differences for HR, SDNN and LF/HF (see Table 4).

    Table 4 Comparison of the emergency operation phases from air-rescue day with activity phases from clinical and control day
    Full size table

A significantly higher HR can be found on the air-rescue day for the phase of landing at the operation site (phase 3) with M = 107.30 bpm (SD = 22.66 bpm) than for activity phases of the clinic day (M = 88.28 bpm, SD = 11.81 bpm) and the control day (M = 83.28 bpm, SD = 14.83 bpm).

The SDNN is significantly higher for the phases 1, 2 and the average of all phases than on the clinic and the control day. The highest value was reached in phase 2 with M = 77.52 ms (SD = 40.52 ms), the lowest on the clinic day with M = 3842 ms (SD = 15.16 ms).

The LF/HF was significantly higher for the activities on the clinic day with M = 1281.84% (SD = 587.33%) than the phases 1–5 and the average of all phases of the emergency operations. The lowest LF/HF was found for the activity phases of the control day with M = 693.74% (SD = 404.73%).

  1. 3.

    The comparison of the HRV parameters for the resting phases with the resting phases of clinic day and the control day showed no significant differences (Table 5).

    Table 5 Comparison of the resting phases from air-rescue day with resting phases from clinical and control day
    Full size table

Discussion

The aim of the present study was to investigate the stress load of EPs on an air-rescue day in relation to a clinic day and a control day.

Comparing the values over the entire shift of the study by Petrowski et al. (2018) which uses the same database as the present study, with the values of the studies by Benzer et al. (1991) and Adams et al. (1998), the HR was lower but in line with the study by Dutheil et al. (2012). The SDNN values were higher than the values on the clinic day in the study by Adams et al. (1998). For the RMSSD, the values of all measured days were lower than the values in the study by Dutheil et al. (2012), while the LF/HF was higher on all measured days than in the study by Dutheil et al. (2012).

Based on the literature, one would postulate for the present study that the physiological stress in the phases of alarm and landing at the operational site on an air-rescue day is higher than in the activity phases of a clinic day and control day. The present results showed significant differences in HR and HRV between the different phases of the emergency operations of the air-rescue days and the activity phases of the other days. In the present study, significantly higher HR values were found in the phase of landing at the operation site (phase 3) of the emergency operation on the air-rescue day compared with the activity phases of the clinic day and the control day. This is in line with the results by Benzer et al. (1991), where the highest HR values were also reached early during the emergency operation in the phases of the alarm and landing phases. In addition to the physical activity at the beginning of the emergency operation (walking to the helicopter), the EPs are exposed to psychological stressors such as time pressure, a lack of information about the emergency, and in the phase of landing at the operation site (phase 3) about complications on the flight and landing. An additional stressor might be the anticipatory anxiety dependent on the limited information that is available.

Regarding the present HRV parameters of the activity phases, the air-rescue day and the control day show significantly higher SDNN and RMSSD values and significantly lower LF/HF values than the clinic day. For the EPs of HEMS, to our knowledge, there are still no other studies measuring HRV parameters of air-rescue days available in the literature. Comparing the present HRV values of the air-rescue day with the clinic days in the literature, the SDNN in the air-rescue day activity phases 3–7 as well as the RMSSD for all phases are lower, while the LF/HF for all phases is much higher than LF/HF values in the literature (Adams et al. 1998; Dutheil et al. 2012).

For the resting phases in the present study, there are no significant differences in HR or HRV parameters between the days (air-rescue day, clinic day, and control day). Compared to the literature, the present SDNN and RMSSD values are lower, while the HR is in line and the present LF/HF on the air-rescue day is much higher than the values of the clinic day of the study by Adams et al. (1998) and the clinic day as well as the control day of the study by Dutheil et al. (2012).

At the resting phases of the control day, the SDNN and RMSSD values are lower, while the HR is higher and the LF/HF at the control day is much higher when compared with the values of the clinic day of the study by Adams et al. (1998) as well as the clinic day and the control day of the study by Dutheil et al. (2012).

Based on these objective parameters, the resting phases seem to be equally regenerative independent of working day or control day. For this reason, the recovery mechanisms of the EPs seem to be healthy.

For a better classification of the present HRV values, the reference to norm values of the HRV was investigated, which was defined in a study by Löllgen (1999) in a 24-h analysis of a healthy population sample. The SDNN should be M = 141 ms (SD = 139 ms), which is higher than the values of the air-rescue day, the clinic day, and the control day. The RMSSD should be M = 27 ms (SD = 12 ms) and is also higher than on the air-rescue day, the clinic day, and the control day. The LF/HF ratio (M = 1.5–2.0%) was much lower than on the air-rescue day, the clinic day, and the control day. These findings indicate a high ANS activation of the EPs.

According to Penttilä et al. (2001), the RMSSD is the best parameter for non-standardized conditions, since it is robust against the effects of breath rates on the HRV. Unfortunately, for this parameter, no significant changes between the days could be shown.

However, it is noticeable that the values are much lower, e.g., for the phases 1–7 of the emergency operations (M = 19.15 ms, SD 13.67 ms) than the norm values by Löllgen (1999) with M = 27 ms, SD = 12 ms) and even lower than the age-corrected values by Umetani et al. (1998) with M = 31 ms, SD 11 ms for the group of 40–49-year olds. On the clinic day, the RMSSD is even lower.

With regard to Castaldo et al. (2015), where a decrease of the HRV as the effect of acute mental stress could be shown leading to cognitive dysfunctions, cardiovascular disorders, depression, and a performance reduction, we would assume that the decrease of the HRV in the present study was caused by the same mechanism. In our view, the duration of these acute mental stress events must be kept as short as possible to avoid chronification.

Comparing the results of the HRV measurement with the subjective stress level, it is noticeable that the stress on the clinic day is rated lower than on the air-rescue day, even if the HRV is decreased more. We consider the height-subjective stress values in the resting phases of the control day most likely as a statistical outlier as, first, it covers a wide-ranging spectrum of activities and, second, no correlation in the HRV values is apparent.

Our results suggest that the HRV can be used as an additional parameter for stress-monitoring, particularly to showing subconscious processes of the ANS which were not covered by the questionnaires.

As a strength of this study, we would highlight the simultaneous stress recording of the EPs over several days (air-rescue day, clinic day, and control day) to quantify and compare the subjective and objective stress load.

As its main limitation, the small sample size of N = 20 EPs must be mentioned. In the natural setting, where this study took place, the comparability of emergency operations and clinic-shifts is limited. The Task Force of NASPE and ESC (1996) demands 5-min measuring periods instead of 3 min for more precise HRV results. This recommendation could not be considered in this study, since overlaps of the rapidly changing phases of the emergency operations had to be avoided.

Furthermore, for a more appropriate analysis for the consideration of the correlation of HR and HRV, multivariable models would have been necessary; however, they were not practicable for this study due to the small sample size.

In future research, a higher number of emergency operations would be necessary to compare long-term effects over the course of entrants to EPs with long-term professional experience. Further research would be necessary to understand long-term developments. Therefore, the stress load at the beginning of the professional life needs to be investigated. Would a young EP experience a similar stress load like the experienced EPs of the present study? If so, does the stress load remain stable at this high level and what are the consequences to the health of the EP?

Conclusion

We found significant differences between the activity phases with respect to SDNN and LF/HF. The SDNN was significantly higher on air-rescue days compared with clinic and control days, respectively, and the LF/HF ratio was significantly higher on clinic days compared to air-rescue and control days, respectively. Hereby, a significant activation of the EPs ANS could be shown, while the subjective stress load was assessed mainly as low. The highest ANS activation was found on the clinic day. In the resting phases, no significant differences were found between the testing days. We interpret this as a good sign for the ability of regeneration between the phases of high stress loads. For the investigation of the long-term effects, further research is needed.