Behav. Res. Ther. Vol. 29, No. 4, PP. 353-355, 1991 Printed in Great Britain. All rights reserved
CASE HISTORIES
OC05-7967/91 $3.00 + 0.00 Copyright 0 1991 Pergamon Press plc
AND SHORTER
COMMUNICATIONS
The role of safety cues in mediating the response to inhalations agoraphobics RONALD
M.
RAPEE,‘*
LESLIE A. TELFER~
of CO, in
and DAVID H. BARLOW~
‘Department of Psychology, University of Queensland, Queensland 4072, Australia ‘Center for Stress and Anxiety Disorders, Albany, New York, U.S.A. (Received
28 December
and
1990)
Summary-Eighteen subjects meeting DSM-III-R criteria for panic disorder with agoraphobia were randomly allocated to either a safety or non-safety condition and underwent 15 min inhalations of 5.5% CO, in air. In the safety condition, both a senior and junior experimenter were present, the junior experimenter was presented as a professional, and subjects were in personal contact with the senior experimenter throughout the procedure. In the non-safety condition, the senior experimenter left on a bogus emergency, the junior experimenter was introduced as a “student” and was dressed casually, and subjects were alone in the experimental room throughout the procedure. Despite these dramatic manipulations, there was no significant difference between groups on either the number of subjects experiencing a panic attack or on pre-inhalation safety scores. However, a median-split analysis on safety scores indicated that those subjects who perceived themselves to be less safe before the inhalation were more likely to experience a panic attack.
Recent conceptualizations of agoraphobia have focussed on the panic attack as the central feature (American Psychiatric Association, 1987; Barlow, 1988; Goldstein & Chambless, 1978). According to this view, agoraphobic avoidance is seen not as a series of specific phobias, but rather as a coherent attempt to minimise the risk of experiencing or being debilitated by a panic attack (Craske, Rapee & Barlow, 1988; Rachman, 1984; Rapee & Murrell, 1988). Rachman (1984) conceptualized this feature of agoraphobia in terms of “an attempt to achieve and maintain a sense of safety” (p. 59). According to this view, agoraphobic avoidance is at least partly mediated by the client’s attempts to stay in contact with stimuli which represent protection from panic attacks. Indeed, safety cues feature prominently in the clinical picture of agoraphobics with many subjects reporting reduced anticipatory anxiety from a broad variety of idiosyncratic stimuli. Of the many objects which serve as safety signals for agoraphobics, probably the most powerful and consistent is the presence of significant others (Chambless, Caputo, Jasin, Gracely & Williams, 1985; Rachman, 1984). However, while it has been often reported clinically that agoraphobics feel less anxious in the presence of safety cues, and that theoretically, this results in less threat from panic attacks, this suggestion has not been demonstrated in the laboratory. In recent years, one of the more popular methods for precipitating anxiety and panic in the latoratory in panic disorder subjects has been the use of biological challenge procedures such as inhalation of 5.5% carbon dioxide (CO,) in air (e.g. Sanderson, Rapee & Barlow, 1989). A number of studies utilising this methodology have described some anecdotal evidence that safety cues (generally presence of the experimenter) can attenuate the response to these procedures. For example, Bonn, Harrison and Rees (1973) reported that a number of their subjects did not panic in response to infusions of sodium lactate because of the presence of the doctor. Similarly, Rapee (1986) found that many panic disorder subjects undergoing a brief period of voluntary hyperventilation reported that the presence of the experimenter and the hospital setting stopped them from panicking. Despite these interesting anecdotal comments together with the finding that a variety of psychological factors can influence the response to biological challenge procedures (Barlow, 1988; Rapee, 1990; Sanderson et al., 1989), there has not yet been a direct empirical test of the influence of safety cues on the response to inhalations of CO,. This was the purpose of the present study.
METHOD Subjects Twenty-one subjects meeting DSM-III-R criteria for panic disorder with agoraphobia (moderate to severe avoidance) were included in the study. Diagnosis was made with the aid of a structured clinical interview, the Anxiety Disorders Interview Schedule-Revised (ADIS-R, Di Nardo & Barlow, 1988) which has been found to result in excellent reliability for this category (K = 0.90; Di Nardo, Rapee, Moras & Barlow, 1991). Subjects were excluded if they reported any of a variety of physical diseases including cardiovascular problems, respiratory disorders, or recurrent migraines. Subjects were randomly allocated to one of two groups, safety or non-safety, however, three subjects refused to undergo the procedure before receiving instructions appropriate to their group. The safety group contained 3 males and 6 females and had a mean age of 36.6 years (SD = 9.5). The non-safety group contained 3 males and 6 females with a mean age of 36.1 years (SD = 9.0). The two groups also appeared to be very similar on duration of their disorder (safety = 114.4 months, *To whom
all correspondence
should
be addressed. 353
354
CASE HISTORIES AND SHORTER COMMUNICATIONS
non-safeety = 100.3 months), Hamiiton anxiety (O-8 scale) (safety = 6.1, non-safety = 6.4).
score (safety = 19.3, non-safety
= 22.6), and global
clinical
severity
rating
Experimental manipulation Due to ethical considerations, it was not possible to have a condition in which subjects were totally alone during the inhalation. Thus, the initial manipulation involved having both a senior and junior experimenter present in an adjoining room or only the junior experimenter present. However, piloting of self-reported safety levels using this manipulation indicated no real decrement in perceived safety in the junior experimenter alone condition. Thus, it was felt necessary to utilise a multifactor manipulation. Three features were manipulated in the two conditions; presence of the senior experimenter, contact between the senior experimenter and subject, and professionalism of the junior experimenter. To summarise the two conditions; in the safety condition the senior experimenter, who was introduced as “Doctor” and described as having extensive experience with the CO, procedure, was present throughout the inhalation, the door between the subject room and experimenter room remained open allowing the senior experimenter to periodically enter the subject room to check on progress, and the junior experimenter dressed in a professional manner and was introduced as a graduate student. In the non-safety condition the senior experimenter explained that an emergency had come up and was not present throughout the inhalation, the door between rooms remained closed throughout the procedure so that the subject was alone (although in contact with the junior experimenter through an intercom), and the junior experimenter was introduced as a student and dressed in a casual manner (t-shirt, stretch pants, and sneakers). Measures Following instructions, but before the beginning of the procedure, subjects were given a questionnaire which included a list of bogus questions about the procedure. Among these questions were the following two questions rated on nine-point scales (0 = not at all, 8 = extremely): “How SAFE do you think you will feel during the procedure?‘; “How competent do you think the experimenter who will be conducting the procedure is?‘. Throughout the procedure, subjects completed a subjective rating of their current anxiety level (O-8 scale) as well as checking any of a list of 12 self statements (six catastrophic, six non-catastrophic) which applied to them at that moment. These ratings were made every 2.5 min during the inhalation and were prompted by the senior experimenter in the safety condition and by a tape recording in the non-safety condition. Immediately following the CO, inhalation, subjects were asked to complete a modification of the diagnostic symptom questionnaire. This measure has been used successfully in a number of previous CO, studies (e.g. Sanderson el al., 1989) and is essentially a checklist of the DSM-III-R panic attack symptoms experienced during the procedure as well as a few additional questions such as intensity of fear experienced, experienced safety, and similarity to usual panic attacks. Apparatus CO, inhalations were delivered through a continuous positive air pressure throughout the entire procedure. The CO, delivery system has been described al., 1989).
gas mask (C-PAP) which the subject wore in detail in other studies (e.g. Sanderson et
Procedure Following the initial diagnostic interview, suitable subjects were asked to return a week later to undergo the procedure. Subjects were given a consent form which was worded as neutrally as possible (e.g. may or may not experience symptoms, may or may not experience anxiety, etc.). They were then introduced to both experimenters and given a standard description of the procedure which was similar to the information in the consent form. Following this, the experimental manipulation was undertaken. That is, subjects in the safety group were told that the senior experimenter would be present, the door would be open, and so on, while subjects in the non-safety group were told that an emergency had arisen necessitating the senior experimenter’s absence, that they would be alone in the room, and so on. Subjects were then attached to the equipment and given 10 min of room air, followed by 15 min of 5.5% CO, in air. Subjects were told that they could stop the inhalation prior to the 15 min if they felt too anxious. Finally, subjects were given the diagnostic symptom questionnaire at the termination of the inhalation and debriefed.
RESULTS Given the relatively small number of subjects, we chose not to artificially limit the conclusions by the use of overly conservative statistics. Nevertheless, it was also important to limit the type one error rate. One solution was to use a composite score as the main dependent variable. Using this criterion, the main dependent variable was described as the
Table
I. Means
and standard deviations for the safety and non-safety groups on their response to CO,
Safety
Variable Number
of somatic symptoms
Number
of cognitive
symptoms
Feeling of fear Number Subjective
of catastrophic
thoughts
anxiety
4.0 (2.1) 0.6 (0.5) 4.0 (2.3) 0.6 (1.1) 4.5
(2.0) Similarity
to usual attacks
3.8 (2.3)
Non-safety 5.1 (3.0) 0.9 (0.9) 3.9 (2.8) 1.2 (1.7) 4.5 (2.2) 3.1 12.5)
Table 2. Means and standard deviaitons for the high and low safety groups on their response to CO,
Variable Number
of somatic symptoms
Number
of cognitive
symptoms
Feeling of fear Number
of catastrophic
thoughts
Subjective
anxiety
Similarity
to usual attacks
High safety
Low safety
4.4 (3.3) ‘0.4’ (0.7) ‘3.1’ (2.8) 0.8 (1.6) 3.7
5.2 (1.9) 1.0 (0.7) 4.8 (1.9) 1.0 (1.3) 5.3 (1.7) 3.6
(2.1) 3.3
CASE HISTORIES
AND
SHORTER
COMMUNICATIONS
355
presence or absence of a panic attack as defined by the following set of criteria which have been used in previous studies (e.g. Sanderson et al., 1989): (i) the experience of four or more DSM-III-R panic attack symptoms, (ii) at least one of which is a cognitive symptom, together with, (iii) a sensation of panic or fear. According to the above criteria, a panic attack was experienced by four subjects in the safety group and four in the non-safety group, a non-significant difference (Fisher’s exact test = 1.0). Means for the specific variables are shown in Table 1. To examine the success of the manipulation, f-tests were conducted on the two pre-inhalation measures, safety and competence. There was no significant difference between the two groups on how safe they felt prior to the inhalation (safety = 4.8; non-safety = 5.3) (t = -0.6, N.S.). Similarly, there was no significant difference between groups in how competent they saw the experimenter (safety = 7.4; non-safety = 6.9) (t = 1.03, N.S.). Since the manipulation did not appear to influence safety levels as measured by the above scales, it was felt that some light could be shed on the question of the role of safety in CO, inhalaltions by conducting a median-split analysis based on pre-inhalufion safety scores. Nine subjects (2 males and 7 females) scored five or less on the safety scale and nine (4 males and 5 females) scored more than five. There were no significant differences between these groups on any of the demographic variables. Seven subjects in the low safety group and one subject in the high safety group experienced a panic attack according to the above criteria. This was a significant difference (Fisher’s exact test = 0.015). Means for the specific variables are shown in Table 2 and seem to indicate generally higher mean scores for the low safety group.
DISCUSSION Given the dramatic and seemingly obvious nature of the experimental manipulation, it was surprising that there was no significant difference between groups on the pre-inhalation measure of safety. These results however, are important since it is a common belief in biological challenge studies that the simple presence of the experimenter can markedly attenuate the response (Bonn er al., 1973; Rapee, 1986). Obviously, this is not the case under all conditions. It is possible that the presence of the experimenter is not an important factor at all or that other factors may interact with this one. One possible factor that may interact with presence of the experimenter could be the intrinsic safety present in the environment. Subjects in the current study were seeking help from a specialist anxiety clinic with a high profile and good reputation in the community. This may have overridden the specific safety cues in the immediate situation. Indeed, pre-inhalation scores indicated that, on the whole, subjects felt more than moderately safe and felt that the experimenter was very competent. It is interesting that the present results are different to our earlier results manipulating an illusion of control (Sanderson et al., 1989). While the concepts of safety and control are highly similar, they can be theoretically distinguished by suggesting that control relies on a personal alteration of a threatening stimulus while safety relies on a manipulation of the threatening stimulus by an external agent. It appears that the former may be an easier construct to manipulate. When subjects were divided according to their personal a priori level of perceived safety, there was a significant influence of this factor on response to CO,. Thus, it is possible that safety cues are an important factor influencing the response to CO, inhalations, but that these cues are more subtle and idiosyncratic than the ones manipulated in the current study. However, it is certainly possible that the results produced by the median-split analysis could be attributable to a general response bias or to the effects of general state or trait anxiety. Thus, while the present results suggest that perceived safety may influence the response to CO, inhlalations, confirmation of this suggestion will have to await studies which produce a more idiosyncratic manipulation or which provide a more threatening environment.
REFERENCES American Psychiatric Association (1987). Diagnostic and Statistical Manual of Mental Disorders (3rd edition-revised). Washington, D.C.: APA. Barlow, D. H. (1988). Anxiety and ifs disorders: The nature and treatment of anxiety andpanic. New York: Guilford Press. Bonn, J. A., Harrison, J. & Rees, L. (1973). Lactate infusion in the treatment of “free-floating” anxiety. Canadian Psychiatric Association Journal, 18, 41-45. Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J. &Williams, C. (1985). The mobility inventory for agoraphobia. Behaviour Research and Therapy, 23, 35-44. Craske, M. G., Rapee, R. M. & Barlow, D. H. (1988). The significance of panic-expectancy for individual patterns of avoidance. Behavior Therapy, 19, 577-592. Di Nardo, P. A. & Barlow, D. H. (1988). Anxiety Disorders Interview Schedule-Revised (ADIS-R). Di Nardo, P. A., Rapee, R. M., Moras, K. & Barlow, D. H. (1991). Kappa co-efficients for DSM-III-R anxiety disorders. To be published. Goldstein, A. J. & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47-59. Rachman, S. (1984). Agoraphobia-A safety-signal perspective. Behaviour Research and Therapy, 22, 59-70. Rapee, R. M. (1986). Differential response to hyperventilation in panic disorder and generalised anxiety disorder. Journal of Abnormal Psychology, 95, 24-28. Rapee, R. M. (1990). Psychological mechanisms underlying the response to biological challenge procedures in panic disorder. In McNaughton, N. & Andrews, G. (Eds), Anxiely. Otago: University of Otago Press. Rapee, R. M. & Murrell, E. (1988). Predictors of agoraphobic avoidance. Journal of Anxiety Disorders, 2, 203-217. Sanderson, W. C., Rapee, R. M. & Barlow, D. H. (1989). The influence of an illusion of control on panic attacks induced via inhalation of 5.5% carbon dioxide-enriched air. Archives of General Psychiatry, 46, 157-162.
CASE HISTORIES
OC05-7967/91 $3.00 + 0.00 Copyright 0 1991 Pergamon Press plc
AND SHORTER
COMMUNICATIONS
The role of safety cues in mediating the response to inhalations agoraphobics RONALD
M.
RAPEE,‘*
LESLIE A. TELFER~
of CO, in
and DAVID H. BARLOW~
‘Department of Psychology, University of Queensland, Queensland 4072, Australia ‘Center for Stress and Anxiety Disorders, Albany, New York, U.S.A. (Received
28 December
and
1990)
Summary-Eighteen subjects meeting DSM-III-R criteria for panic disorder with agoraphobia were randomly allocated to either a safety or non-safety condition and underwent 15 min inhalations of 5.5% CO, in air. In the safety condition, both a senior and junior experimenter were present, the junior experimenter was presented as a professional, and subjects were in personal contact with the senior experimenter throughout the procedure. In the non-safety condition, the senior experimenter left on a bogus emergency, the junior experimenter was introduced as a “student” and was dressed casually, and subjects were alone in the experimental room throughout the procedure. Despite these dramatic manipulations, there was no significant difference between groups on either the number of subjects experiencing a panic attack or on pre-inhalation safety scores. However, a median-split analysis on safety scores indicated that those subjects who perceived themselves to be less safe before the inhalation were more likely to experience a panic attack.
Recent conceptualizations of agoraphobia have focussed on the panic attack as the central feature (American Psychiatric Association, 1987; Barlow, 1988; Goldstein & Chambless, 1978). According to this view, agoraphobic avoidance is seen not as a series of specific phobias, but rather as a coherent attempt to minimise the risk of experiencing or being debilitated by a panic attack (Craske, Rapee & Barlow, 1988; Rachman, 1984; Rapee & Murrell, 1988). Rachman (1984) conceptualized this feature of agoraphobia in terms of “an attempt to achieve and maintain a sense of safety” (p. 59). According to this view, agoraphobic avoidance is at least partly mediated by the client’s attempts to stay in contact with stimuli which represent protection from panic attacks. Indeed, safety cues feature prominently in the clinical picture of agoraphobics with many subjects reporting reduced anticipatory anxiety from a broad variety of idiosyncratic stimuli. Of the many objects which serve as safety signals for agoraphobics, probably the most powerful and consistent is the presence of significant others (Chambless, Caputo, Jasin, Gracely & Williams, 1985; Rachman, 1984). However, while it has been often reported clinically that agoraphobics feel less anxious in the presence of safety cues, and that theoretically, this results in less threat from panic attacks, this suggestion has not been demonstrated in the laboratory. In recent years, one of the more popular methods for precipitating anxiety and panic in the latoratory in panic disorder subjects has been the use of biological challenge procedures such as inhalation of 5.5% carbon dioxide (CO,) in air (e.g. Sanderson, Rapee & Barlow, 1989). A number of studies utilising this methodology have described some anecdotal evidence that safety cues (generally presence of the experimenter) can attenuate the response to these procedures. For example, Bonn, Harrison and Rees (1973) reported that a number of their subjects did not panic in response to infusions of sodium lactate because of the presence of the doctor. Similarly, Rapee (1986) found that many panic disorder subjects undergoing a brief period of voluntary hyperventilation reported that the presence of the experimenter and the hospital setting stopped them from panicking. Despite these interesting anecdotal comments together with the finding that a variety of psychological factors can influence the response to biological challenge procedures (Barlow, 1988; Rapee, 1990; Sanderson et al., 1989), there has not yet been a direct empirical test of the influence of safety cues on the response to inhalations of CO,. This was the purpose of the present study.
METHOD Subjects Twenty-one subjects meeting DSM-III-R criteria for panic disorder with agoraphobia (moderate to severe avoidance) were included in the study. Diagnosis was made with the aid of a structured clinical interview, the Anxiety Disorders Interview Schedule-Revised (ADIS-R, Di Nardo & Barlow, 1988) which has been found to result in excellent reliability for this category (K = 0.90; Di Nardo, Rapee, Moras & Barlow, 1991). Subjects were excluded if they reported any of a variety of physical diseases including cardiovascular problems, respiratory disorders, or recurrent migraines. Subjects were randomly allocated to one of two groups, safety or non-safety, however, three subjects refused to undergo the procedure before receiving instructions appropriate to their group. The safety group contained 3 males and 6 females and had a mean age of 36.6 years (SD = 9.5). The non-safety group contained 3 males and 6 females with a mean age of 36.1 years (SD = 9.0). The two groups also appeared to be very similar on duration of their disorder (safety = 114.4 months, *To whom
all correspondence
should
be addressed. 353
354
CASE HISTORIES AND SHORTER COMMUNICATIONS
non-safeety = 100.3 months), Hamiiton anxiety (O-8 scale) (safety = 6.1, non-safety = 6.4).
score (safety = 19.3, non-safety
= 22.6), and global
clinical
severity
rating
Experimental manipulation Due to ethical considerations, it was not possible to have a condition in which subjects were totally alone during the inhalation. Thus, the initial manipulation involved having both a senior and junior experimenter present in an adjoining room or only the junior experimenter present. However, piloting of self-reported safety levels using this manipulation indicated no real decrement in perceived safety in the junior experimenter alone condition. Thus, it was felt necessary to utilise a multifactor manipulation. Three features were manipulated in the two conditions; presence of the senior experimenter, contact between the senior experimenter and subject, and professionalism of the junior experimenter. To summarise the two conditions; in the safety condition the senior experimenter, who was introduced as “Doctor” and described as having extensive experience with the CO, procedure, was present throughout the inhalation, the door between the subject room and experimenter room remained open allowing the senior experimenter to periodically enter the subject room to check on progress, and the junior experimenter dressed in a professional manner and was introduced as a graduate student. In the non-safety condition the senior experimenter explained that an emergency had come up and was not present throughout the inhalation, the door between rooms remained closed throughout the procedure so that the subject was alone (although in contact with the junior experimenter through an intercom), and the junior experimenter was introduced as a student and dressed in a casual manner (t-shirt, stretch pants, and sneakers). Measures Following instructions, but before the beginning of the procedure, subjects were given a questionnaire which included a list of bogus questions about the procedure. Among these questions were the following two questions rated on nine-point scales (0 = not at all, 8 = extremely): “How SAFE do you think you will feel during the procedure?‘; “How competent do you think the experimenter who will be conducting the procedure is?‘. Throughout the procedure, subjects completed a subjective rating of their current anxiety level (O-8 scale) as well as checking any of a list of 12 self statements (six catastrophic, six non-catastrophic) which applied to them at that moment. These ratings were made every 2.5 min during the inhalation and were prompted by the senior experimenter in the safety condition and by a tape recording in the non-safety condition. Immediately following the CO, inhalation, subjects were asked to complete a modification of the diagnostic symptom questionnaire. This measure has been used successfully in a number of previous CO, studies (e.g. Sanderson el al., 1989) and is essentially a checklist of the DSM-III-R panic attack symptoms experienced during the procedure as well as a few additional questions such as intensity of fear experienced, experienced safety, and similarity to usual panic attacks. Apparatus CO, inhalations were delivered through a continuous positive air pressure throughout the entire procedure. The CO, delivery system has been described al., 1989).
gas mask (C-PAP) which the subject wore in detail in other studies (e.g. Sanderson et
Procedure Following the initial diagnostic interview, suitable subjects were asked to return a week later to undergo the procedure. Subjects were given a consent form which was worded as neutrally as possible (e.g. may or may not experience symptoms, may or may not experience anxiety, etc.). They were then introduced to both experimenters and given a standard description of the procedure which was similar to the information in the consent form. Following this, the experimental manipulation was undertaken. That is, subjects in the safety group were told that the senior experimenter would be present, the door would be open, and so on, while subjects in the non-safety group were told that an emergency had arisen necessitating the senior experimenter’s absence, that they would be alone in the room, and so on. Subjects were then attached to the equipment and given 10 min of room air, followed by 15 min of 5.5% CO, in air. Subjects were told that they could stop the inhalation prior to the 15 min if they felt too anxious. Finally, subjects were given the diagnostic symptom questionnaire at the termination of the inhalation and debriefed.
RESULTS Given the relatively small number of subjects, we chose not to artificially limit the conclusions by the use of overly conservative statistics. Nevertheless, it was also important to limit the type one error rate. One solution was to use a composite score as the main dependent variable. Using this criterion, the main dependent variable was described as the
Table
I. Means
and standard deviations for the safety and non-safety groups on their response to CO,
Safety
Variable Number
of somatic symptoms
Number
of cognitive
symptoms
Feeling of fear Number Subjective
of catastrophic
thoughts
anxiety
4.0 (2.1) 0.6 (0.5) 4.0 (2.3) 0.6 (1.1) 4.5
(2.0) Similarity
to usual attacks
3.8 (2.3)
Non-safety 5.1 (3.0) 0.9 (0.9) 3.9 (2.8) 1.2 (1.7) 4.5 (2.2) 3.1 12.5)
Table 2. Means and standard deviaitons for the high and low safety groups on their response to CO,
Variable Number
of somatic symptoms
Number
of cognitive
symptoms
Feeling of fear Number
of catastrophic
thoughts
Subjective
anxiety
Similarity
to usual attacks
High safety
Low safety
4.4 (3.3) ‘0.4’ (0.7) ‘3.1’ (2.8) 0.8 (1.6) 3.7
5.2 (1.9) 1.0 (0.7) 4.8 (1.9) 1.0 (1.3) 5.3 (1.7) 3.6
(2.1) 3.3
CASE HISTORIES
AND
SHORTER
COMMUNICATIONS
355
presence or absence of a panic attack as defined by the following set of criteria which have been used in previous studies (e.g. Sanderson et al., 1989): (i) the experience of four or more DSM-III-R panic attack symptoms, (ii) at least one of which is a cognitive symptom, together with, (iii) a sensation of panic or fear. According to the above criteria, a panic attack was experienced by four subjects in the safety group and four in the non-safety group, a non-significant difference (Fisher’s exact test = 1.0). Means for the specific variables are shown in Table 1. To examine the success of the manipulation, f-tests were conducted on the two pre-inhalation measures, safety and competence. There was no significant difference between the two groups on how safe they felt prior to the inhalation (safety = 4.8; non-safety = 5.3) (t = -0.6, N.S.). Similarly, there was no significant difference between groups in how competent they saw the experimenter (safety = 7.4; non-safety = 6.9) (t = 1.03, N.S.). Since the manipulation did not appear to influence safety levels as measured by the above scales, it was felt that some light could be shed on the question of the role of safety in CO, inhalaltions by conducting a median-split analysis based on pre-inhalufion safety scores. Nine subjects (2 males and 7 females) scored five or less on the safety scale and nine (4 males and 5 females) scored more than five. There were no significant differences between these groups on any of the demographic variables. Seven subjects in the low safety group and one subject in the high safety group experienced a panic attack according to the above criteria. This was a significant difference (Fisher’s exact test = 0.015). Means for the specific variables are shown in Table 2 and seem to indicate generally higher mean scores for the low safety group.
DISCUSSION Given the dramatic and seemingly obvious nature of the experimental manipulation, it was surprising that there was no significant difference between groups on the pre-inhalation measure of safety. These results however, are important since it is a common belief in biological challenge studies that the simple presence of the experimenter can markedly attenuate the response (Bonn er al., 1973; Rapee, 1986). Obviously, this is not the case under all conditions. It is possible that the presence of the experimenter is not an important factor at all or that other factors may interact with this one. One possible factor that may interact with presence of the experimenter could be the intrinsic safety present in the environment. Subjects in the current study were seeking help from a specialist anxiety clinic with a high profile and good reputation in the community. This may have overridden the specific safety cues in the immediate situation. Indeed, pre-inhalation scores indicated that, on the whole, subjects felt more than moderately safe and felt that the experimenter was very competent. It is interesting that the present results are different to our earlier results manipulating an illusion of control (Sanderson et al., 1989). While the concepts of safety and control are highly similar, they can be theoretically distinguished by suggesting that control relies on a personal alteration of a threatening stimulus while safety relies on a manipulation of the threatening stimulus by an external agent. It appears that the former may be an easier construct to manipulate. When subjects were divided according to their personal a priori level of perceived safety, there was a significant influence of this factor on response to CO,. Thus, it is possible that safety cues are an important factor influencing the response to CO, inhalations, but that these cues are more subtle and idiosyncratic than the ones manipulated in the current study. However, it is certainly possible that the results produced by the median-split analysis could be attributable to a general response bias or to the effects of general state or trait anxiety. Thus, while the present results suggest that perceived safety may influence the response to CO, inhlalations, confirmation of this suggestion will have to await studies which produce a more idiosyncratic manipulation or which provide a more threatening environment.
REFERENCES American Psychiatric Association (1987). Diagnostic and Statistical Manual of Mental Disorders (3rd edition-revised). Washington, D.C.: APA. Barlow, D. H. (1988). Anxiety and ifs disorders: The nature and treatment of anxiety andpanic. New York: Guilford Press. Bonn, J. A., Harrison, J. & Rees, L. (1973). Lactate infusion in the treatment of “free-floating” anxiety. Canadian Psychiatric Association Journal, 18, 41-45. Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J. &Williams, C. (1985). The mobility inventory for agoraphobia. Behaviour Research and Therapy, 23, 35-44. Craske, M. G., Rapee, R. M. & Barlow, D. H. (1988). The significance of panic-expectancy for individual patterns of avoidance. Behavior Therapy, 19, 577-592. Di Nardo, P. A. & Barlow, D. H. (1988). Anxiety Disorders Interview Schedule-Revised (ADIS-R). Di Nardo, P. A., Rapee, R. M., Moras, K. & Barlow, D. H. (1991). Kappa co-efficients for DSM-III-R anxiety disorders. To be published. Goldstein, A. J. & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47-59. Rachman, S. (1984). Agoraphobia-A safety-signal perspective. Behaviour Research and Therapy, 22, 59-70. Rapee, R. M. (1986). Differential response to hyperventilation in panic disorder and generalised anxiety disorder. Journal of Abnormal Psychology, 95, 24-28. Rapee, R. M. (1990). Psychological mechanisms underlying the response to biological challenge procedures in panic disorder. In McNaughton, N. & Andrews, G. (Eds), Anxiely. Otago: University of Otago Press. Rapee, R. M. & Murrell, E. (1988). Predictors of agoraphobic avoidance. Journal of Anxiety Disorders, 2, 203-217. Sanderson, W. C., Rapee, R. M. & Barlow, D. H. (1989). The influence of an illusion of control on panic attacks induced via inhalation of 5.5% carbon dioxide-enriched air. Archives of General Psychiatry, 46, 157-162.
