Treatment of Fear of Flying: Behavioral, Subjective, and Cardiovascular Effects

Address correspondence and reprint requests to: Dr. Michael Trimmel, Medical University of Vienna, ZPH, Inst. f. Umwelthygiene, Kinderspitalgasse 15, A-1090 Vienna, Austria; [email protected].

Trimmel M, Burger M, Langer G, Trimmel K. Treatment of fear of flying: behavioral, subjective, and cardiovascular effects. Aviat Space Environ Med 2014; 85:550–62.

Background: Nonaviator fear of flying is a common problem usually managed with behaviorally oriented treatment. The unknown time courses of aspects of anxiety and of physiological response were investigated during a 3-d treatment including flights. Methods: Ratings, heart rate (HR), and heart rate variability (HRV; pNN10, pNN50) of 15 Moderate-Anxious and 9 High-Anxious subjects at critical epochs of treatment and on 2 actual flights, on which 9 Controls participated, were compared. Results: All subjects took the flights and displayed a remarkably reduced fear when comparing pre- vs. post-treatment ratings. Repeated ratings showed an increase in relaxation and drops in general somatic and cognitive aspects of anxiety during flights. However, cognitive aspects in High-Anxious did not drop to the level of Controls. Level of anxiety was related to cardiovascular activity almost always during treatment and flights, in particular during takeoffs (average maximum HR of 137 bpm in High-Anxious compared to 118 bpm in Moderate-Anxious and 98 bpm in Controls in the first flight). Moderate-Anxious showed no obvious relationship of cardiovascular response to critical flight epochs, but had the lowest HRV on flights. Discussion: Results indicate that anxiety in Moderate-Anxious is related to flight, but not to single critical epochs of flying, contrary to High-Anxious, for which higher ratings on cognitive aspects of anxiety associated with more physiological load at critical epochs were observed. However, HR and ratings showed a remarkable drop in both treatment groups, in particular in perceived physiological symptoms, indicating that the treatment facilitates coping of fear of flying.


FEAR OF FLYING (FOF) in nonprofessional aviators is a common psychological problem (20), though studies differ on its prevalence. Based on previous research, Van Gerwen and Diekstra (19) asserted the prevalence of FOF in industrialized countries at 10–40%, the 40% including persons being only moderately apprehensive about flying. Others (5) report that 10% of the population completely avoids flying and a 2.6% prevalence in the general population (12), comparable to studies from the Netherlands (9) and the United States (16). There is a need for treatment (21), as FOF may impair mobility and thus private life and business possibilities in many people, and has commercial consequences (7).

Programs to treat FOF are offered worldwide. Programs based on cognitive-behavioral group treatment were found to be efficacious (8,22), but success was only indicated by subjective verbal reports and not by physiological measurements (4). The present approach also considers the physiological response to fear of real flights to give a broader description of the human response, as suggested earlier (2,13).

Other studies analyzing the physiological response in real flights took three spot measurements (1) during flights or the duration of flights was rather short (27,28). Thus, no data of repeated measurements of heart rate (HR) and heart rate variability (HRV) during real flights related to treated FOF are available. The present study investigates the time course of the physiological and psychological response of a FOF treatment including two real flights in persons with different levels of FOF.

In the FOF treatment program of the Austrian Airlines, Freude am Fliegen (“The Joy of Flying”), a team of psychologists, engineers, and Austrian Airlines pilots and flight attendants leads the participants through 3 d of treatment, covering psychological elements and technical information on flying, taking a simulator flight and a round-trip commercial flight. The efficacy of the treatment was supposed to be mirrored by ratings taken before and after the treatment. Detailed information on the time course of the expected change in FOF was acquired by ratings during the treatment and by the analysis of HR and HRV at critical epochs of the treatment and on flights.


The study was conducted in accordance with the World Medical Association’s Declaration of Helsinki. Data were collected noninvasively during a treatment program the subjects would have undertaken had no experiment existed. Each subject provided written informed consent before participating in the study voluntarily and unpaid.

The 24 subjects (ages 20-60 yr, M = 38.5, 15 women) took part in 3 self-referred courses with a group size of up to 15 persons each. The control group consisted of nine members of Austrian Airlines’ ground personnel (ages 23-52 yr, M = 32.4, four women) and participated only in the “graduation flights” at the third day of treatment.


Subjects taking part in the treatment were grouped into 15 Moderate-Anxious (ages 20-60 yr, M = 43.2, 10 women) and 9 High-Anxious (ages 21-55 yr, M = 36.9, 5 women) based on their scores on the Generalized Flight Anxiety scale of the Flight Anxiety Situations questionnaire (23). The raw scores of the subjects in the treatment groups ranged from 0-19 on a scale ranging from 0 (no generalized anxiety) to 28 (extreme generalized anxiety), with the score of 8.5 being the splitting point for the two anxiety groups.

The investigation followed a 2/3 × 35/30/6/6/2 quasi-experimental design with the factors Group (High-Anxious/Moderate-Anxious/Controls) and Time on Task (as a repeated factor). Time on Task for physiological recordings were 65 epochs, 35 for the treatment period and 30 epochs taken on the real flights (Tables I and II). Time on Task for the Questionnaire on Aspects of Anxiety were six time points of the treatment period and six epochs of the real flights. Other anxiety questionnaires compared pre- and post-treatment scores.


The ambulatory electrocardiographic (ECG) recording devices (Medilog AR12 Holter ECG recorders from Schiller-Engineering, Graz, Austria, using 3M one-way Ag/AgCl electrodes) had to undergo special security procedures mandated by the airlines, airport security offices, and the pilots of study flights.


Austrian Airlines registered subjects, after which questionnaires to assess flight and general anxiety were completed. Subjects came in earlier each day to attach the ECG recorders, which were detached at the end of each day. The experimenter, a clinical psychologist, administered the questionnaires and noted the time of events.

On the first day subjects met for approximately 3 h at the Austrian Airlines headquarters in Vienna, Austria. Information on psychological factors of fear and anxiety was provided, theoretical models of panic and anxiety were presented, including information about the aerodynamics of flying, and the subjects described their personal fear experiences. Epochs 1 to 9 (Table I) were analyzed.

On the second day subjects met at the technical base station at Vienna International Airport in Schwechat, Austria, visiting a mock-up of an Airbus A320 cabin and boarding the flight simulator. There a flight attendant explained the flight crew’s announcements and the safety system. Then a simulated flight was taken, including instructions of the cabin crew, turbulences, and the landing procedure, with a video system showing a virtual environment during the simulation. Subjects learned techniques to deal with their fears, including relaxation techniques. Lunch was held at the airport restaurant with a panoramic view of the runway, together with the pilot and the coach of the program, discussing concerns about flying. The session on that day lasted about 6 h; epochs 10 to 18 (Table I) were analyzed.

On the third day subjects again met at the technical base station at Vienna International Airport. That day started with an oral presentation by an aircraft technician and a guided tour through the aircraft hangar and a close look at an aircraft being there for maintenance. A technician answered questions and responded to the fears of the subjects, then subjects were introduced to the cockpit and the flight instruments, and finally they could try out different seats in the aircraft cabin. Then psychological interventions were repeated and relaxation was practiced. The treatment program continued with lunch in the panoramic restaurant together with the captain, the first officer, and the cabin manager of the imminent flight. After lunch subjects entered the crew building and were informed about the professional training and work of flight attendants before moving on to the departure area. Subjects checked in, passed the security check with a special security allowance for the ECG devices, and waited for boarding in the lounge (epochs 19 to 35; Table I). Control subjects joined the FOF treatment program and were affixed with the ECG equipment.

All subjects boarded an Airbus A320 for the commercial flight to London or Amsterdam, respectively. After arrival at their respective destinations, subjects spent some time in the duty free shopping area before boarding for the flight back to Vienna. The session on that day lasted about 12 h for the flights to Amsterdam and 14 h for the flight to London. Epochs 36 to 65 (Table II) were analyzed. All subjects filled in the Questionnaire on Aspects of Anxiety three times on each flight (epochs 37, 42, 47, 52, 57, and 62 from Table II). Subjects were asked to rate their anxiety and FOF again on the next day and to return their ratings by mail.


Self-report instruments: The Flight Anxiety Situations Questionnaire (FAS) (23) assesses the degree of anxiety experienced in various flight or flight-related situations on a 5-point Likert type scale, ranging from 0 (no anxiety) to 4 (overwhelming anxiety). The FAS consists of 3 subscales: the 12-item Anticipatory Flight Anxiety Scale (FAS-AF) referring to anxiety experienced when anticipating a flight; the 10-item In-Flight Anxiety Scale referring to anxiety experienced during a flight; and the 7-item Generalized Flight Anxiety Scale (FAS-GF) referring to anxiety experienced when confronted with stimuli in connection with flying and airplanes in general. The present study showed a good to excellent internal consistency of these subscales, the Cronbach’s α ranging from 0.84 to 0.96 [comparable to that of van Gerwen et al. (22) reporting a Cronbach’s α above 0.88].

The Flight Anxiety Modality Questionnaire (FAM) (23) assesses the specific modality of flight-related anxiety symptoms on a 5-point Likert type scale, ranging from 1 (not at all) to 5 (very intensely). The FAM consists of two subscales: the 11-item Somatic Modality Scale (FAM-SM) assessing the physical symptoms during a flight, and the 7-item Cognitive Modality Scale (FAM-CM) relating to distressing cognitions during a flight. The present study showed a very good internal consistency with Cronbach’s α = 0.91 for FAM-SM and 0.86 for FAM-CM, meeting a Cronbach’s α of above 0.89 as reported by van Gerwen et al. (22).

The State-Trait Anxiety Inventory (15) assesses state (STAI-S) and trait (STAI-T) anxiety with 20 items each on a 4-point Likert type scale ranging from 1 (not at all) to 4 (very much so) for the STAI-S and from 1 (almost never) to 4 (almost always) for the STAI-T.

The Body Sensations Questionnaire (BSQ), the Agoraphobic Cognitions Questionnaire (ACQ), and the Mobility Inventory (MI) (6) were presented in the combined German questionnaire Fragebogen zu körperbezogenen Ängsten, Kognitionen und Vermeidung (10). The BSQ is a 17-item inventory referring to physical and physiological body responses, rated on a 5-point Likert type scale, ranging from 1 (not worried) to 5 (extremely). The ACQ is a 14-item inventory referring to catastrophic thoughts about the physical and social consequences of panic attacks, rated on a 5-point Likert type scale, ranging from 1 (I never think this) to 5 (always). The MI refers to agoraphobic avoidance behavior. Both 28-item subscales, the subscale Avoidance Accompanied (MIB) and the subscale Avoidance Alone (MIA) are rated on a 5-point Likert type scale, ranging from 1 (never avoid) to 5 (always avoid).

The Questionnaire on Aspects of Anxiety (QAA) consists of 12 items (see Table III for items and factor loadings of the 3 subscales and the item on actual flight anxiety) and was handed out 12 times during the 3 d of the treatment. Responses were given on a labeled (1…7) 7-point Likert scale with the endpoints labeled “Is not true (0%)” and “Is true (100%)” for nearly all items; only for item 12, the endpoints were labeled “not at all” and “extremely high.” Based on the response on the first 11 items at the epoch “F1_Sitting” (Table I), factor analysis (main components) with varimax rotation revealed a factor “relaxation” (items 6, 10, 11; Cronbach’s α = 0.93) explaining 32% of variance and a factor “anxiety” (items 1, 2, 3, 4, 8, and 9) explaining 44% of variance. To describe anxiety on the physiological as well as on the psychological level, two subscales were computed, namely “Somatic Aspects of Anxiety (SAA)” with the items 1, 2, and 3 (Cronbach’s α = 0.84) and “Cognitive Aspects of Anxiety (CAA)” with the items 4, 8, and 9 (Cronbach’s α = 0.88). Item 12 was treated as a single item index on perceived “Actual Flight Anxiety (AFA)” at the various epochs of the treatment program, including real flights.

Physiological recordings: Continuous ambulatory recordings of ECG were conducted and digitized with 4 kHz to identify R-waves from a modified Lead II configuration. Offline analyses were done using the program Medilog Darwin ( for calculating parameters of time domain (including pNN10) (18) and of frequency domain using an FFT algorithm (detrended linear according to the Task Force of the ESC and the NASPE) (17). HR and HRV parameters were computed only for normal-to-normal beat intervals of 5-min epochs, and only for blocks with at least five successive valid PQRST appearances.

Ratings on FAS, FAM, STAI, BSQ, ACQ, and MI

Two × two ANOVAs with the factors Group (High-Anxious/Moderate-Anxious) × Time (pre/post-treatment) with Time as a repeated factor were computed on the FAS, FAM, STAI, ACQ, BSQ, MIB, and MIA (for means and SD see Table IV). Analysis of FAS-AF and FAS-GF showed significant interactions for Group × Time [F(1,22) = 7.56, P = 0.012 and F(1,22) = 33.60, P < 0.001, respectively] and significant effects for Time [F(1,22) = 126.60, P < 0.001 for the FAS-AF and F(1,22) = 76.36, P < 0.001 for the FAS-GF]. Both subscales showed higher levels of anxiety for High-Anxious than Moderate-Anxious before the treatment, and scores were shown to be reduced with no group differences after the treatment [F(1,22) = 2.04, P = 0.167 for FAS-AF and F(1,22) = 29.36, P < 0.001 for FAS-GF]. The FAS In-Flight Anxiety Scale showed reduced scores after the treatment [F(1,22) = 173.93, P < 0.001] with no interaction of Group × Time [F(1,22) = 0.03, P = 0.871] and no effect for Group [F(1,22) = 0.09, P = 0.764].

Analysis of FAM-SM and FAM-CM showed significant effects for Time [F(1,22) = 29.28, P < 0.001 and F(1,22) = 86.99, P < 0.001, respectively] with reduced scores after treatment and no significant effects for Group [F(1,22) = 0.01, P = 0.905 for FAM-SM and F(1,22) = 1.48, P = 0.237 for FAM-CM] or for Group × Time [F(1,22) = 1.22, P = 0.282 for FAM-SM and F(1,22) = 2.46, P = 0.131 for FAM-CM].

Analysis of STAI showed significant effects for Time, with reduced scores after treatment, on both subscales [F(1,22) = 15.51, P = 0.001 for the STAI-T and F(1,22) = 7.70, P = 0.011 for the STAI-S]; no significant effects for Group [F(1,22) = 0.01, P = 0.904 for the STAI-T and F(1,22) = 0.17, P = 0.684 for the STAI-S] or for Group × Time [F(1,22) = 0.43, P = 0.517 for the STAI-T and F(1,22) = 0.09, P = 0773 for the STAI-S] were observed.

Analysis of ACQ and BSQ showed only significant effects for Time [F(1,22) = 7.89, P = 0.010 for the ACQ and F(1,22) = 16.12, P = 0.001 for the BSQ], with reduced scores after treatment. No significant effects were observed for either Group [F(1,22) = 1.98, P = 0.174 for the ACQ, F(1,22) = 0.26, P = 0.619 for the BSQ] or for Group × Time [F(1,22) = 0.00, P = 0.966 for the ACQ, F(1,22) = 0.06, P = 0.810 for the BSQ]. Both subscales of the MI showed significant effects for Time [F(1,22) = 13.30, P = 0.001 for the MIB and F(1,22) = 10.87, P = 0.003 for the MIA], with reduced scores after treatment. Neither of the scales showed significant effects for Group [F(1,22) = 0.97, P = 0.336 for the MIB, F(1,22) = 0.28, P = 0.605 for the MIA] nor for Group × Time [F(1,22) = 2.61, P = 0.121 for the MIB, F(1,22) = 2.08, P = 0.164 for the MIA].

Ratings on QAA

QAA was analyzed by ANOVA for the time of treatment (epochs 2, 8, 10, 13, 19, and 33 from Table II) with the groups High-Anxious vs. Moderate-Anxious and ANOVA for the time of the two flights (epochs 37, 42, 47, 52, 57, and 62 from Table II) with three levels of Group. ANOVA for the single item index on perceived AFA during treatment showed an effect for Group with higher scores in High-Anxious than in Moderate-Anxious [F(1,22) = 5.03, P = 0.035] and an effect of Time [F(5,110) = 3.12, P = 0.014, ϵ = 0.902], with higher scores on “D3_Crew+20” than on Day 1. No significant interaction of Group × Time appeared [F(5,110) = 0.42, P = 0.816, ϵ = 0.902]. ANOVA for AFA during flights showed a significant Group × Time interaction [F(10,150) = 4.48, P < 0.001, ϵ = 0.748], a main effect for Group [F(2,30) = 12.66, P < 0.001], and an effect of Time [F(5,150) = 17.50, P < 0.001, ϵ = 0.748]. Mean values showed no anxiety in Controls and a stepwise degradation of anxiety in the treatment groups with a significant degradation from “Sitting” to “Takeoff +20” on both flights. After “Sitting” Moderate-Anxious had lower anxiety than High-Anxious in F1. A significant reduction of anxiety appeared after the second takeoff in both treatment groups (see Fig. 1, upper left, for means and CIs).


Means and 95% CIs of ratings on the Questionnaire on Aspects of Anxiety.


ANOVA for the scale Relaxation during treatment showed no statistically significant effects [F(1,22) = 0.02, P = 0.900 for Group, F(5,110) = 1.38, P = 0.247, ϵ = 0.812 for Time, and F(5,110) = 0.68, P = 0.608, ϵ = 0.812 for Group × Time]. ANOVA for the scale Relaxation during flights showed a significant Group × Time interaction [F(10,150) = 3.68, P = 0.001, ϵ = 0.859] as well as main effects for Group [F(2,30) = 17.61, P < 0.001] and for Time [F(5,150) = 16.01, P < 0.001, ϵ = 0.859]. However, no significant difference between treatment groups appeared. High-Anxious and Moderate-Anxious were less relaxed than Controls throughout the real flights, although relaxation was increasing during the flights with a significant increase after the second takeoff (for means and CIs see Fig. 1, upper right).

ANOVA for CAA during treatment indicated higher scores in High-Anxious than in Moderate-Anxious [F(1,22) = 4.60, P = 0.043], a main effect of Time [F(5,110) = 3.46, P = 0.013, ϵ = 0.761], and no significant Group × Time interaction [F(5,110) = 0.73, P = 0.566, ϵ = 0.761], with low scores on Day 1 and high scores at “D2_Simulator.” ANOVA for CAA during flights showed a significant Group × Time interaction [F(10,150) = 3.41, P = 0.002, ϵ = 0.725] as well as main effects for Group [F(2,30) = 9.30, P < 0.001] and Time [F(5,150) = 15.47, P < 0.001, ϵ = 0.725]. Means and CIs (Fig. 1, lower left) show the reductions in CAA in High-Anxious and Moderate-Anxious during flights and the scores of Controls. In both anxiety groups a significant reduction in CAA appeared at F1 (from “F1_Sitting” to “F1_Takeoff+20”), whereas in F2 High-Anxious showed no significant reduction of CAA, contrary to Moderate-Anxious.

ANOVA for SAA during treatment showed an effect of Time [F(5,110) = 7.07, P < 0.001, ϵ = 0.869] with high scores in “D2_Simulator” and low scores at “D1_Beginning+5” and “D3_Theory,” and no other significant effects [F(1,22) = 1.60, P = 0.220 for Group and F(5,110) = 1.19, P = 0.322, ϵ = 0.869 for Group × Time]. ANOVA for SAA during flights showed a significant Group × Time interaction [F(10,150) = 5.16, P = 0.001, ϵ = 0.471] as well as main effects for Group [F(2,30) = 8.07, P = 0.002] and Time [F(5,150) = 21.31, P < 0.001, ϵ = 0.417]. High-Anxious showed higher scores than Moderate-Anxious at “F1_Sitting” and “F1_Takeoff+20.” Both treatment groups showed a significant reduction in SAA from “F1_Sitting” to “F1_Takeoff+20” and from “F1_Takeoff+20” to “F1_Landing-10” in the High-Anxious. SAA were low in F2 and both anxiety groups met values of the Control group [Moderate-Anxious already at the end of F1 (“F1_Landing-10”)]. Controls did not perceive any SAA during the flights (for means and CIs see Fig. 1, lower right).

Physiological Measures

Physiological measures related to the treatment period and to the real flights were compared by separate ANOVAs. For the measures during the treatment a 2 × 35 ANOVA with Group (High-Anxious/Moderate-Anxious) × Time (epochs 1-35, Table I) with Time as a repeated factor was computed. For the measures during the flights a 3 × 30 ANOVA with Group (High-Anxious/Moderate-Anxious/Controls) × Time (epochs 36-65, Table II) with Time as a repeated factor was computed.

ANOVA of mean HR during treatment indicated a significant main effect for Time [F(34,748) = 17.22, P < 0.001, ϵ = 0.173] and an interaction for Group × Time [F(34,748) = 2.40, P = 0.032, ϵ = 0.173], but no significant main effect for Group [F(1,22) = 2.07, P = 0.164]; means and CIs are shown in Table V and Fig. 2. Results showed that mean HR was higher for High-Anxious than for Moderate-Anxious during treatment, however, with the exception of “D1_Pilot” at D1, “D2_Simulator,” “D2_Relax,” “D2_Relax+5,” and “D2_Relax+10” at D2; and on D3. On the first day mean HR remained rather constant at the beginning, reaching a peak when the pilot gave aerodynamic explanations (statistically significant to the time before and after in Moderate-Anxious) and dropped slightly at the end of D1 in both treatment groups. On D2 mean HR was rather high when stepping into the mock-up (95.3 bpm for the High-Anxious and 91.3 for the Moderate-Anxious), then dropping to 77.7 bpm and 72.5 bpm, respectively, when practicing relaxation techniques. For these epochs HR of High-Anxious met the HR of Moderate-Anxious. On D3, when listening to theoretical explanations, Moderate-Anxious had a HR not significantly different to High-Anxious and HR was in particular comparable at the time of sitting in the aircraft in the hangar (“D3_Sitting”) and for the time when meeting the crew (“D3_Crew” to “D3_Crew+30”).


Means and 95% CIs of mean HR during treatment (upper part) and flights (lower part).


ANOVA of mean HR during real flights indicated a significant main effect for Time [F(29,870) = 27.33, P < 0.001, ϵ = 0.410] and for Group [F(2,30) = 5.10, P = 0.012], and a significant interaction for Group × Time [F(58,870) = 1.93, P = 0.006, ϵ = 0.410]; means and CIs are shown in Table VI and Fig. 2. Results showed a rather comparable pattern for the three groups, with the exception at “F1_Takeoff,” where only High-Anxious displayed a significant increase in HR compared to “F1_Sitting.”

At “F1_Boarding,” 5-min average mean HR reached a high of 115.9 bpm in High-Anxious and 106.8 bpm in Moderate-Anxious, which is comparable to the 108.1 bpm for Controls. During “F1_Takeoff” the average mean HR of High-Anxious increased (significantly to “F1_Sitting”) to 113.2 bpm, that of Moderate-Anxious displayed 97.7 bpm, and Controls had 81.4 bpm. Average mean HR of High-Anxious dropped during the outgoing flight to 94.5 bpm at “Landing-10” and stayed at that level until landing, Moderate-Anxious showed a decrease to 87.6 bpm when landing. Controls showed no statistically significant change in HR after “F1_Sitting.” A comparable pattern to that of the outbound flight appeared on the inbound flight, with a nonsignificant HR decrease to an average mean of 75.4 bpm for Controls and 86.0 bpm for Moderate-Anxious, whereas High-Anxious showed a significant decrease from “F2_Takeoff” to “F2_Takeoff+25” and “F2_Takeoff+30.” However, neither the increase in mean HR of High-Anxious from “F2_Sitting” to “F2_Takeoff” nor the increase from“F2_Takeoff+30” to “F2_Landing-20” reached statistical significance.

ANOVA of maximum HR during treatment indicated significant main effects for Time [F(34,748) = 10.24, P < 0.001, ϵ = 0.303], for Group [F(1,22) = 4.29, P = 0.050], and for Group × Time [F(34,748) = 2.07, P = 0.027, ϵ = 0.303]. Results (see Fig. 3 for means and CIs) showed that maximum HR was higher for High-Anxious (with an average maximum HR of 127.2 bpm at “D2_Simulator” and 130.2 bpm at “D3_Cockpit”) than for Moderate-Anxious (with an average maximum HR of 115.7 bpm at “D2_Simulator” and 109.6 bpm at “D1_Pilot” and at “D3_Cockpit” and when meeting the crew; however, this time with no differences to High-Anxious).


Means and 95% CIs of maximum HR during treatment (upper part) and flights (lower part).


ANOVA of maximum HR during real flights indicated significant main effects for Time [F(29,870) = 12.41, P < 0.001, ϵ = 0.542] and for Group [F(2,30) = 8.31, P = 0.001], and a trend for Group × Time [F(58,870) = 1.42, P = 0.070, ϵ = 0.542 (see Fig. 3 for means and CIs)]. Results showed a comparable pattern of maximum HR in all groups during flights, however, with significantly lower values in High-Anxious during the epochs after “F2_Takeoff” compared to the epochs after “F1_Takeoff.” High values were found for boarding (“F1_Boarding,” “F2_Boarding”) and in High-Anxious at “F1_Takeoff” (138 bpm) and at “F2_Landing-20” and “F2_Landing-10.”

ANOVA of pNN50 during treatment indicated a significant main effect of Time [F(34,748) = 9.94, P < 0.001, ϵ = 0.259], but main effect of Group [F(1,22) = 1.62, P = 0.216] and interaction of Group × Time [F(34,748) = 1.54, P = 0.137, ϵ = 0.259] were nonsignificant (see Fig. 4 for means and CIs). Results showed that pNN50 was higher on D2 and during “D3_Sitting” compared to D1 and the rest of D3.


Means and 95% CIs of pNN50 during treatment (upper part) and flights (lower part).


ANOVA of pNN50 during real flights indicated a significant main effect of Time [F(29,870) = 2.28, P = 0.022, ϵ = 0.280], a trend for Group [F(2,30) = 2.76, P = 0.079], and a trend for Group × Time [F(58,870) = 1.55, P = 0.082, ϵ = 0.280 (see Fig. 4 for means and CIs)]. Results showed a lower pNN50 in Moderate-Anxious compared to Controls at most epochs. Treatment groups showed reduced pNN50 at takeoff and thereafter compared to Controls and High-Anxious on F1, also significantly reduced at “F1_Sitting.”

ANOVA of pNN10 during treatment indicated a significant main effect for Time [F(34,748) = 11.75, P < 0.001, ϵ = 0.292] and an interaction for Group × Time [F(34,748) = 3.05, P = 0.001, ϵ = 0.292], whereas the main effect for Group [F(1,22) = 0.09, P = 0.773] was nonsignificant (see Fig. 5 for means and CIs). Results showed that pNN10 was highest (and comparable in both treatment groups) during relaxation (“D2_Relax,” “D2_Relax+5,” and “D2_Relax+10”). On D3, pNN10 was lower in Moderate-Anxious (in particular from “D3_Cockpit” on) compared to D1 and D2, whereas High-Anxious showed less difference over treatment days. In addition, both treatment groups had a significant reduction in pNN10 from “D3_Sitting” and “D3_Sitting+5” to “D3_Cockpit” and to some extent thereafter.


Means and 95% CIs of pNN10 during treatment (upper part) and flights (lower part).


ANOVA of pNN10 during real flights indicated a significant main effect for Time [F(29,870) = 11.35, P < 0.001, ϵ = 0.518] and a trend for Group × Time [F(58,870) = 1.37, P = 0.095, ϵ = 0.518], whereas Group [F(2,30) = 1.72, P = 0.196] was nonsignificant. Results (Fig. 5) showed a comparable pattern for the three groups on both flights with lower pNN10 in Moderate-Anxious compared to Controls in most epochs in flight. In addition, treatment groups showed reduced values at takeoff on both flights compared to Controls, and in High-Anxious values were also significantly reduced in the preceding epoch (“F1_Takeoff” vs. “F1_Sitting”).


All subjects took the actual flights and thus gave a behavioral indication of the efficacy of the treatment. Regarding the experienced anxiety comparing pre- vs. post-treatment, all ratings in the treatment groups showed high levels before treatment, which dropped to remarkablely low levels afterwards. Ratings of FAS-AF, FAS-GF, and FAM-CM showed higher scores in High-Anxious than in Moderate-Anxious at pre-treatment, with this difference disappearing after treatment. A more differentiated result was obtained by the repeated measures of the QAA. For CAA a remarkable reduction was indicated for both treatment groups; however, ratings of High-Anxious were still above Controls at F2. SAA dropped remarkably in both treatment groups in particular during F1, but Relaxation also increased during F2 about the same amount as in F1 in addition. AFA dropped in both treatment groups during flights and seemed to reflect a mixture of CAA, SAA, and Relaxation, whereas CAA, SAA, and Relaxation by themselves seem to reflect different aspects of anxiety.

Mean HR during flights was higher in High-Anxious than in Controls and at most analyzed epochs also higher than in Moderate-Anxious, who also showed a higher HR than Controls, in particular at “F1_Takeoff” and the following 15 min, “F2_Takeoff,” and the last 20 min before landing on F2. In all three groups HR dropped slightly from the beginning of F1 to the end of F2. Group differences were predominant for “F1_Takeoff” with a comparable pattern at “F2_Takeoff.” However, High-Anxious displayed a significant increase of mean HR in the preceding and the following epoch (“F1_Sitting,” “F1_Takoff+5”) only at “F1_Takeoff.” The mean HR of 113 bpm during “F1_Takeoff” is a rather high value for physical inactivity and may be more impressive considering the average maximum HR of 137 bpm. The maximum HR of High-Anxious was 20-40 bpm above Controls and in Moderate-Anxious 7-20 bpm higher than in Controls, indicating physiological load by anxiety.

Mean HR of High-Anxious during treatment was rather high (about 90 bpm) from D1 to D3, whereas Moderate-Anxious had a lower HR at D1, with an increase to the level of High-Anxious on D3. Both treatment groups displayed the lowest HR during Relaxation (“D2_Relax”), indicating its effectiveness.

HRV as indicated by pNN50 and pNN10 showed high scores in both treatment groups at D2 and for Relaxation (“D2_Relax,” “D2_Relax+5,” and “D2_Relax+10”) and low scores on D3, in particular in Moderate-Anxious during treatment. This pronounced generally reduced HRV in Moderate-Anxious was also observed during flights and is often interpreted as an expression of low parasympathetic/high sympathetic activity (17). At “F1_Takeoff” pNN10 (and also, but less pronounced, pNN50) of High-Anxious was reduced compared to the previous and subsequent epochs (“F1_Sitting” and “F1_Takeoff+5”), indicating (as also indicated by mean HR and maximum HR) that High-Anxious are affected by takeoff, contrary to Moderate-Anxious. However, Moderate-Anxious display, contrary to High-Anxious, a generally lower HRV compared to Controls, which may indicate additional sources of anxiety not directly related to aspects of flying (as also indicated by the reduced response in the maximum HR in Moderate-Anxious contrary to the High-Anxious at “D3_Cockpit”). Considering the physiological aspects of relaxation, there is convergent evidence as indicated by reduced HR and pronounced HRV in both treatment groups. Therefore, the practice of breathing and other relaxation techniques, possibly monitored with the help of pocket devices (26), could be considered for reducing anxiety (7,29).

As for flights, the group differences in HR remained rather stable, contrary to ratings. One may conclude that HR seems not to be strictly related to anxiety, as also reported for flight phobia stress earlier (11). A somewhat comparable pattern to HR in ratings was observed for CAA with a drop of CAA scores during F1, but group differences still remained. It may be concluded that cognitive aspects of anxiety could be a source of the observed group differences (as also indicated by an increase in mean HR and maximum HR at “F2_Landing” in High-Anxious), whereas somatic aspects (SAA) were effectively treated, which let the subjects take the flights. This can be interpreted as an expression of emotional coping, in particular on D2 as indicated by the drop of HR and high values in HRV parameters. A comparable pattern was observed by Bornas et al. (3), finding a rigid response pattern in the unsatisfactory treatment outcome group compared to the satisfactory treatment outcome group.

This study also clearly demonstrates the physiological load people with high FOF are confronted with as they experience a level of anxiety associated with a mean HR (in a 5-min interval) of 113 bpm and an average maximum HR of 137 bpm; however, both scores are somewhat reduced at F2 (although not statistically significant). As turned out in this study, it might appear that FOF may be caused by different sources of anxiety (i.e., more flight-related ones like takeoff or claustrophobic feelings or even other aspects); other differentiations of anxiety should be investigated with a larger sample. Furthermore, it turned out that in particular pNN10, as one member of the pNNxx family (18), seems to be particularly sensitive to examining differences in physiological response under circumstances with a generally high HR, as suggested earlier (14).

Of course there are some shortcomings in the current study. The neglect of gender is certainly a limitation, but that was due to the small sample size, another weakness of the study. Differences in interoceptive response as well as in sensitivity of anxiety (24,25) should be considered as further research topics. Unfortunately there were no baseline data available from before the flights and from after the flights, as subjects’ cooperation was limited. Furthermore, prospective studies might take into consideration follow-up investigations of subjects taking part successfully in the FOF treatment programs in order to explore the long-term sustainability of the treatment.


We want to thank the founder of the treatment program “Freude am Fliegen,” Dr. Robert Wolfger, for his support and help not only during the study but also in obtaining the necessary permits with the authorities.

Authors and affiliations: Dr. phil. Michael Trimmel, Dr. phil. Margit Burger, and Gabriela Langer, Centre of Public Health, and Dr. med. Karin Trimmel, Department of Neurology, Medical University of Vienna, Vienna, Austria.

  1. Beckham JC, Vrana SR, May JG, Gustafson DJ, Smith GR. Emotional processing and fear measurement synchrony as indicators of treatment outcome in fear of flying. J Behav Ther Exp Psychiatry 1990; 21:153–62.
  2. Birbaumer N. Neuropsychologie der Angst, Fortschritte der Klinischen Psychologie. München: Urban & Schwarzenberg; 1973.
  3. Bornas X, Riera del Amo A, Tortella-Feliu M, Llabrés J. Heart rate variability profiles and exposure therapy treatment outcome in flight phobia. Appl Psychophysiol Biofeedback 2012; 37:53–62.
  4. Busscher B, van Gerwen LJ, Spinhoven P, de Geus EJ. Physiological reactivity to phobic stimuli in people with fear of flying. J Psychosom Res 2010; 69:309–17.
  5. Capafóns JI, Sosa CD, Viña CM. A reattributional training program as a therapeutic strategy for fear of flying. J Behav Ther Exp Psychiatry 1999; 30:259–72.
  6. Chambless DL, Caputo GC, Bright P, Gallagher R. Assessment of fear in agoraphobics: the body sensations questionnaire and the agoraphobic cognitions questionnaire. J Consult Clin Psychol 1984; 52:1090–7.
  7. Conrad A, Isaac L, Roth WT. The psychophysiology of generalized anxiety disorder: 2. Effects of applied relaxation. Psychophysiology 2008; 45:377–88.
  8. Dean RD, Whitaker KM. Fear of flying: impact on the U.S. air travel industry. J Travel Res 1982; 21:7–17.
  9. Depla MF, ten Have ML, van Balkom AJ, de Graaf R. Specific fears and phobias in the general population: results from the Netherlands Mental Health Survey and Incidence Study (NEMESIS). Soc Psychiatry Psychiatr Epidemiol 2008; 43:200–8.
  10. Ehlers A, Margraf J, Chambless D. Fragebogen zu körperbezogenen Ängsten, Kognitionen und Vermeidung (AKV). Weinheim: Beltz; 1993.
  11. Ekeberg O, Kjeldsen SE, Greenwood DT, Enger E. Correlations between psychological and physiological responses to acute flight phobia stress. Scand J Clin Lab Invest 1990; 50:671–7.
  12. Fredrikson M, Annas P, Fischer H, Wik G. Gender and age differences in the prevalence of specific fears and phobias. Behav Res Ther 1996; 34:33–9.
  13. Lang P. The application of psychophysiological methods to the study of psychotherapy and behavior change. In: Bergin A, Garfield S, eds. Handbook of psychotherapy and behavior change: an empirical analysis. New York: Wiley; 1971:75–125.
  14. Mietus JE, Peng CK, Henry I, Goldsmith RL, Goldberger AL. The pNNx files: re-examining a widely used heart rate variability measure. Heart 2002; 88:378–80.
  15. Spielberger C, Gorsuch R, Luchene R. State-Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychology Press; 1970.
  16. Stinson FS, Dawson DA, Patricia Chou S, Smith S, Goldstein RB, et al. The epidemiology of DSM-IV specific phobia in the USA: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Psychol Med 2007; 37:1047–59.
  17. Task Force of the European Society of Cardiology (ESC) and the North American Society of Pacing Electrophysiology(NASPE). Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Eur Heart J 1996; 17:354–81.
  18. Trimmel K. Sensitivity of HRV parameters including pNNxx proven by short-term exposure to 2700 m altitude. Physiol Meas 2011; 32:275–85.
  19. Van Gerwen LJ, Diekstra RF. Fear of flying treatment programs for passengers: an international review. Aviat Space Environ Med 2000; 71:430–7.
  20. Van Gerwen LJ, Diekstra RF, Arondeus JM, Wolfger R. Fear of flying treatment programs for passengers: an international update. Travel Med Infect Dis 2004; 2:27–35.
  21. Van Gerwen LJ, Spinhoven P, Diekstra RFW, Van Dyck R. Multicomponent standardized treatment programs for fear of flying: description and effectiveness. Cognit Behav Pract 2002; 9:138–49.
  22. Van Gerwen LJ, Spinhoven P, Van Dyck R. Behavioral and cognitive group treatment for fear of flying: a randomized controlled trial. J Behav Ther Exp Psychiatry 2006; 37:358–71.
  23. Van Gerwen LJ, Spinhoven P, Van Dyck R, Diekstra RFW. Construction and psychometric characteristics of two self-report questionnaires for the assessment of fear of flying. Psychol Assess 1999; 11:146–58.
  24. Vanden Bogaerde A, De Raedt R. The moderational role of anxiety sensitivity in flight phobia. J Anxiety Disord 2011; 25:422–6.
  25. Vanden Bogaerde A, Derom E, De Raedt R. Increased interoceptive awareness in fear of flying: sensitivity to suffocation signals. Behav Res Ther 2011; 49:427–32.
  26. Wilhelm FH, Grossman P. Emotions beyond the laboratory: theoretical fundaments, study design, and analytic strategies for advanced ambulatory assessment. Biol Psychol 2010; 84:552–69.
  27. Wilhelm FH, Pfaltz MC, Grossman P, Roth WT. Distinguishing emotional from physical activation in ambulatory psychophysiological monitoring. Biomed Sci Instrum 2006; 42:458–63.
  28. Wilhelm FH, Roth WT. Taking the laboratory to the skies: ambulatory assessment of self-report, autonomic, and respiratory responses in flying phobia. Psychophysiology 1998; 35:596–606.
  29. Wollburg E, Roth WT, Kim S. Effects of breathing training on voluntary hypo- and hyperventilation in patients with panic disorder and episodic anxiety. Appl Psychophysiol Biofeedback 2011; 36:81–91.