Correspondence: Tadashi Funayama, MD, PhD, Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, 1 Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya 467-8601, Japan. Email: firstname.lastname@example.org
In-situation safety behaviors play an important role in the maintenance of anxiety because they prevent patients from experiencing unambiguous disconfirmation of their unrealistic beliefs about feared catastrophes. Strategies for identifying particular safety behaviors, however, have not been sufficiently investigated. The aims of the present study were to (i) develop a comprehensive list of safety behaviors seen in panic disorder and to examine their frequency; and (ii) correlate the safety behaviors with panic attack symptoms, agoraphobic situations and treatment response.
The subjects consisted of 46 consecutive patients who participated in group cognitive behavioral treatment (CBT) for panic disorder. All the patients completed a Safety Behavior List that was developed based on experiences with panic disorder patients.
Carrying medications, distracting attention, carrying a plastic bottle, and drinking water were reported by more than half of the patients. The strongest correlations between panic symptoms and safety behaviors were found between symptoms of derealization and listening to music with headphones, paresthesia and pushing a cart while shopping, and nausea and squatting down. The strongest association between agoraphobic situations and safety behaviors was found between the fear of taking a bus or a train alone and moving around. Staying still predicted response to the CBT program, while concentrating on something predicted lack of response.
An approximate guideline has been developed for identifying safety behaviors among patients with panic disorder and should help clinicians use CBT more effectively for these patients.
THE RELATIONSHIP BETWEEN anxiety and phobic avoidance has received considerable attention over the past 40 years. Avoidance and escape behavior may occur in response to threat and anxiety, and in turn these responses may maintain the crucial threat beliefs. In contrast, learning theory teaches us that exposure to fearful situations leads to a reduction in anxiety. This process is known as ‘habituation’. Exposure has long been known to be a highly effective treatment for phobic disorders.
In some phobic patients, however, especially those with social phobias, the improvements obtainable with exposure alone are relatively modest.[1, 2] In fact, individuals with social phobias are repeatedly exposed to social situations during everyday life, without a marked reduction in anxiety. Patients cannot learn to reduce their anxiety sufficiently in fearful situations.
Salkovskis et al. argued that in-situation safety behaviors, or subtle strategies of avoidance to prevent the feared catastrophe while remaining in the feared situation, play an important role in the maintenance of anxiety despite exposure because they prevent phobic people from experiencing an unambiguous disconfirmation of their unrealistic beliefs about feared catastrophes. When they use safety behaviors, phobic individuals tend to attribute the non-occurrence of feared catastrophes to the implementation of their safety behaviors. For example, an individual with a social phobia who turns their eyes away from others for fear of being noticed and laughed at is likely to think, ‘I managed to avoid being noticed and being considered strange because I turned my eyes away’. Engaging in safety behavior subjectively saves the person from his perceived threat, while simultaneously denying the individual a chance to discover how unlikely their imagined catastrophe is. If patients continue to use safety behaviors to ‘prevent’ danger from happening, the logical link to the perceived threat is strengthened and exposure to feared stimuli ends up distorting cognition even further.
A relatively small number of studies have examined safety behaviors in panic disorder.[6-8] In the case of panic disorder, panic attacks occur as a result of catastrophic misinterpretations of bodily sensations, which often are only normal physiological concomitants of anxiety. Safety behaviors are motivated by these catastrophic misinterpretations. Salkovskis et al. listed 10 typical safety behaviors and asked more than 100 patients with panic disorder how often they used each of these behaviors when they were anxious; after correlating the use of these safety behaviors with the patients' catastrophic cognitions, they identified several theoretically predicted associations. For example, a person who is afraid of fainting is likely to hold on to something, while a person who fears having a heart attack would refrain from exercising.
Safety behaviors are often very subtle and idiosyncratic. For example, while some patients with panic disorder who fear palpitation avoid drinking alcohol, others may drink alcohol to reduce their anxiety in public situations. For the former patients, the avoidance of drinking alcohol is a safety behavior, while for the latter patients the consumption of alcohol is a safety behavior. An individual's cognition of the feared catastrophe is what determines whether a particular action is a safety behavior. Likewise, some patients prefer shopping in crowded places because they feel safer when there are more people available to help them, while others go shopping when few customers are present, such as early in the morning or late in the evening, because they then can avoid being witnessed while having a panic attack. Many safety behaviors are quite covert, such as carrying enough money in one's wallet to ensure that they are well prepared should something happen, making sure that one is carrying a cellular phone, or making sure that someone else is at home while the subject is not at home. All these behaviors are normal in the sense that they can occur on a daily basis in anyone's life, but it is the cognitive process of the patient that makes these actions function as safety behaviors. To cover these various safety behaviors comprehensively in clinical practice, awareness of a much larger number of specific behaviors than the 10 described by Salkovskis et al. is needed. Based on our own experiences with panic disorder patients, we have developed and have been updating a safety behavior list, which we use to help patients recognize their own safety behaviors.
The aims of the present study were to (i) develop a comprehensive list of safety behaviors seen in panic disorder and to examine their frequency; (ii) correlate the safety behaviors with panic attack symptoms so that clinicians may use clinical symptoms as cues to their patients' latent safety behaviors; (iii) examine associations between safety behaviors and agoraphobic situations that patients avoid to help clinicians better recognize individual patient safety behaviors; and (iv) examine whether particular safety behaviors can predict a better or worse outcome at the end of cognitive behavioral treatment (CBT).
The subjects were 46 consecutive patients who sought CBT for panic disorder at Nagoya City University Hospital Department of Psychiatry between May 2005 and June 2007. All the patients met the following inclusion criteria: (i) principal Axis I diagnosis of panic disorder with or without agoraphobia according to the DSM-IV, as assessed using a structured clinical interview for DSM-IV (SCID); (ii) absence of current psychosis, bipolar disorder and substance-use disorder; (iii) adequate motivation to undergo CBT (the patients had various reasons for preferring CBT including: insufficient effectiveness of medication, refusal of medication because of pregnancy, etc.); (iv) absence of anxiolytic benzodiazepine medication prior to the commencement of the CBT sessions, because these drugs may interact negatively with exposure treatment; and (v) completion of the Safety Behavior List. The exclusion criteria were as follows: (i) presence of urgent suicidal thoughts; and (ii) presence of an Axis II disorder that would interfere with group treatment. The use of antidepressants was permitted throughout the CBT period because these drugs do not interfere with CBT. The regular use of benzodiazepines as a hypnotic was permitted. Patients with other comorbid anxiety disorders and/or major depressive disorders were admitted to the CBT program after the symptoms of these disorders had abated and they were able to participate in the program.
The patients provided written informed consent after receiving a full explanation of the study's purposes and procedures. The study protocol was approved by the Ethics Committee of Nagoya City University Graduate School of Medical Sciences.
The CBT program used in this study was based on the program developed by the Clinical Research Unit for Anxiety and Depression at the University of New South Wales in Sydney, Australia. The treatment consisted of five major components: (i) psycho-education concerning the nature, causes and maintenance of anxiety; (ii) breathing control; (iii) cognitive restructuring; (iv) graded situational exposure without safety behaviors to reduce agoraphobic avoidance; and (v) interoceptive exposure to reduce patient fears of somatic sensations. Patients were assigned homework after each session.
In the fourth or fifth sessions, graded situational exposure was introduced. After explaining how avoidance contributes to the maintenance of anxiety, patients were introduced to graded situational exposure tasks. Safety behaviors were explained within this framework using the following explanation from the CBT manual.
A safety behavior is any behavior that you perform while in a situation that you fear and without which your anxiety would be greater. Thus, although you appear to remain within the feared situation, you are actually subtly avoiding it. Such behaviors function like a charm. Safety behaviors that are often used by patients suffering from panic disorder include the following:
You do not go out without your medication, although you seldom take it.
You make absolutely sure that you have your cellular phone with you when you drive far from home.
You make it a rule to stand near the doorway when taking a train.
The patients were then asked to list the safety behaviors that they currently performed on a regular basis using the Safety Behavior List.
The patients were then requested to engage in their exposure tasks without performing such safety behaviors (in other words, the graded tasks were constructed so as to be able to be done without the safety behaviors in a graded manner). Our program has been demonstrated to be highly effective in reducing the severity of panic disorder and agoraphobic avoidance.
Before enrolment in the CBT program, all the participants completed a semi-structured interview for anxiety and mood disorders according to the SCID. This interview was conducted by one of the clinicians conducting the CBT program. Then, at the start of the program, all the patients completed the Fear Questionnaire–Agoraphobic Subscale (FQ-Ag) and the Panic Disorder Severity Scale (PDSS). During the course of the CBT sessions, the patients identified their safety behaviors by completing the Safety Behavior List. The PDSS was completed again at the end of the CBT program.
Safety Behavior List
The list was developed by our department based on existing reports[7, 14-16] and our own clinical experiences. The list contained 25 items, each of which was answered using ‘yes’ or ‘no’. All the items are listed in Figure 1. The list was presented as a self-reported questionnaire, and the patients were able to add any behavior not on the list that they suspected of being a safety behavior. We started using this list in May 2005.
Fear Questionnaire – Agoraphobic Subscale
The FQ-Ag is a self-reported instrument for measuring the severity of agoraphobic avoidance of five typical situations for agoraphobia. Each situation is rated on a 9-point scale ranging from 0 (would not avoid it) to 8 (always avoid it) to show how much he/she would avoid each of the listed situations because of fear or other unpleasant feelings. The test–retest reliability and factor validity of this instrument have been confirmed.
Panic Disorder Severity Scale
The PDSS is an interview-based, seven-item scale for assessing the severity of panic disorder; clinicians rate the severity of the seven features of panic disorder using a scale from 0 (none) to 4 (extreme). The seven areas covered by the scale include the frequency of panic attacks, distress during panic attacks, anticipatory anxiety, agoraphobic fear/avoidance, interoceptive fear/avoidance, work impairment/distress, and impairment of social functioning. Adequate inter-rater reliability and validity have been reported for both the original and Japanese versions.[18, 19]
All statistical analyses were performed using the SPSS program, version 11.5 (SPSS, Chicago, IL, USA). The significance level set in this study was P ≤ 0.05 (two-tailed test).
Frequency of safety behaviors in panic disorder
We first listed all the items in the Safety Behavior List according to the frequency of use among the patients. We also listed the extra behaviors that the patients had volunteered.
Safety behaviors and panic attack symptoms
The associations between panic symptoms and safety behaviors were estimated by calculating the odds ratio (OR) and its 95% confidence interval for a 2 × 2 table between each panic symptom and each safety behavior. OR >1 indicates that patients with that particular panic symptom tend to use that particular safety behavior, while OR <1 indicates that patients who do not show that particular panic symptom tend to use that safety behavior.
Safety behaviors and agoraphobic situations
The associations between agoraphobic situations and safety behaviors were similarly estimated using an OR. The FQ-Ag were dichotomized with a cut-off score ≥4 (scores between 4 ‘frequently avoid it’ and 8 ‘always avoid it’ were regarded as indicating the presence of the fear).
Safety behaviors and response to treatment
The response to treatment was defined as a ≥40% reduction in the PDSS in accordance with the validated interpretation guidelines. The associations between safety behaviors and response to treatment were similarly estimated using OR.
Of the 52 patients who participated in the CBT program between May 2005 and June 2007, the safety behavior list was completed by 46. Of the 46 patients, the average number of safety behaviors per patient was 7.8 ± 5.0. Table 1 lists the baseline subject demographic and clinical characteristics.
Table 1. Baseline demographic and clinical characteristics
Mean ± SD or n (%)
CBT, cognitive behavioral treatment; FQ-Ag, Fear Questionnaire–Agoraphobic Subscale.
37.3 ± 11.1
Age at onset (years)
31.5 ± 10.4
FQ-Ag baseline score
9.7 ± 7.0
Medication at start of CBT session
Figure 1 shows the frequency of the reported safety behaviors. Table 2 lists all the other safety behaviors volunteered by the subjects and the frequencies of these behaviors.
Table 2. Other safety behaviors
Whispering songs softly
n = 2
Carrying towels and handkerchiefs
n = 2
Carrying proof of insurance and patient registration card
n = 1
Exposing oneself to the wind
n = 1
Crossing one's legs
n = 1
n = 1
n = 1
n = 1
Lifting a pillow
n = 1
Taking more clothes
n = 1
Eating nothing before going out
n = 1
Searching for the way back and checking the timetable beforehand
n = 1
Driving by oneself
n = 1
n = 1
Carrying a cooling agent
n = 1
Carrying a light meal
n = 1
Looking for a lavatory
n = 1
Baring one's arm
n = 1
n = 1
Taking a shower
n = 1
Avoiding the corners of a room
n = 1
Sleeping on one's stomach
n = 1
n = 1
Checking her husband's schedule
n = 1
Playing a game on a cellular phone
n = 1
n = 1
Walking with downcast eyes
n = 1
Table 3 lists the statistically significant associations between panic symptoms and safety behaviors. The strongest correlation was found between symptoms of derealization and listening to music with headphones, paresthesia and pushing a cart while shopping, and nausea and squatting down.
Table 3. Safety behaviors significantly associated with panic criteria symptoms
CI, confidence interval; OR, odds ratio.
Shortness of breath
Chest pain or discomfort
Carrying a plastic bottle
Listening to music using headphones
Fear of losing control
Fear of dying
Focusing attention on something
Pushing a cart while shopping
Chills or hot flushes
Looking for an escape route
Table 4 lists the statistically significant associations between agoraphobic situations and safety behaviors. Here, the strongest associations were found between the fear of taking a bus or a train alone and moving around and the fear of entering a crowded shop and holding onto something.
Table 4. Safety behaviors significantly associated with agoraphobic situations
CI, confidence interval; FQ, Fear Questionnaire; OR, odds ratio.
Taking a bus or a train alone
Sitting close to the door in a train or a bus
Walking through crowded streets alone
Taking medication before going out
Entering a crowded shop
Holding onto something
Traveling far from home alone
Carrying extra money
Table 5 lists all the safety behaviors that predicted a response to treatment at the end of the CBT program. Staying still predicted a response to the CBT program, while concentrating on something predicted lack of response.
Table 5. Safety behaviors and response to treatment
CI, confidence interval; OR, odds ratio.
Concentrating on something
This study is the most comprehensive report to date on the safety behaviors of patients with panic disorder. We examined 25 candidate behaviors among 46 patients with panic disorder.
First, we listed the typical safety behaviors and determined the frequency with which each behavior was reported. For example, carrying medications, distracting attention, carrying a plastic bottle, and drinking water were reported by more than half of the patients. These safety behaviors need to be routinely screened for during CBT for patients with panic disorder. We also listed some of the more idiosyncratic safety behaviors that have not attracted much attention previously. Awareness of such infrequent behaviors should help clinicians to recognize these behaviors as safety behaviors more readily.
Second, we identified some strong associations between safety behaviors and particular panic symptoms as well as particular agoraphobic situations. Previous studies have shown that exposure to feared situations is more effective without the use of safety behaviors,[8, 16] but failed to suggest how to better identify these behaviors. The associations identified in the present study should help clinicians specifically inquire after and identify particular safety behaviors. For example, patients who report derealization during panic attacks should be asked whether they listen to music using headphones during feared situations, and patients who are fearful of taking a bus or a train alone should be asked whether they have the habit of moving around in this situation.
In contrast, the present study failed to identify specific safety behaviors for such typical panic symptoms as palpitation, chest pain, shortness of breath, or choking feelings. Safety behaviors associated with the fear of dying or the fear of losing control were also not identified. We speculated that patients with these symptoms certainly do engage in safety behaviors but that these behaviors tend to be very idiosyncratic, preventing any trends from being identified.
Salkovskis et al. discussed the association between catastrophic cognitions and in-situation safety behaviors. For example, the fear of fainting was associated with seeking support by holding onto both objects and people. The fear of having a heart attack was associated with sitting down, keeping still, and asking nearby people for help. Predicted relationships between panic-related cognitions and safety behaviors were found, and the cognitive theory of panic disorder was shown to suggest better approaches to CBT. Although theoretical relationships are important, concrete strategies for identifying particular safety behaviors are needed for clinical practice. The present study was able to suggest useful tools for identifying safety behaviors based on triggers of panic attacks and features of agoraphobia.
A number of safety behaviors were found to predict a stronger or poorer response to CBT. Although we emphasized the importance of discontinuing all safety behaviors, some of these behaviors were more difficult to stop than others. For example, ‘Staying still’, which was found to predict a better response to CBT, may be easier to discontinue than ‘Concentrating on something’, which represents an internal covert behavior that is hard for patients to stop and difficult for therapists to ascertain.
The present study has some possible limitations. First, the test–retest reliability of the Safety Behavior List has not yet been established, although the present study can be considered to contribute to the validation of the scale. Second, the sample size was relatively small, and the confidence intervals for the observed OR were wide as a result. Third, the sample was limited to patients who wished to receive CBT at one particular department of psychiatry in Japan, and the generalizability of the present findings to other settings and cultures remains to be examined. Last, the present descriptive study could not answer whether the CBT program used in conjunction with the Safety Behavior List was more effective, enabling greater habituation or swifter disconfirmation, than a program that does not emphasize the discontinuation of safety behaviors. A future experimental study contrasting exposure to feared situations while using safety behaviors and exposure without using safety behaviors may benefit from the use of the extensive Safety Behavior List.
The present study has provided an approximate guideline for identifying safety behaviors among patients with panic disorder and should enable clinicians to provide CBT more effectively for these patients. We expect that the present study will make it easier for clinicians to identify safety behaviors and to augment the therapeutic gains.
This study was supported by a Grant-in-Aid from the Ministry of Health, Labor and Welfare, Japan. The authors declare that they have no competing interests.