Panic and Agoraphobia Scale (PAS)

Abstract

The Panic and Agoraphobia Scale (PAS) is a standardized psychometric instrument developed by Borwin Bandelow in 1995 to evaluate the severity and treatment-mediated alteration of symptoms in patients with panic disorder, with or without agoraphobia. Constructed to fulfill the diagnostic criteria of both the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R, DSM-IV, and DSM-5) and the International Classification of Diseases (ICD-10 and ICD-11), the PAS exists in parallel observer-rated and self-report formats. The instrument consists of 14 items organized into five core clinical subscales: (1) Panic Attacks, (2) Agoraphobic Avoidance, (3) Anticipatory Anxiety, (4) Disability, and (5) Worries About Health (functional avoidance/somatic preoccupation). Items are scored on a five-point Likert scale (ranging from 0 to 4), with an integrated agoraphobia situational checklist evaluating specific environmental triggers. Empirical evaluations demonstrate robust psychometric properties, including high internal consistency (Cronbach’s α = .86 for the global score; subscale alphas ranging from .66 to .82), strong test-retest reliability, and high concurrent validity when benchmarked against instruments such as the Hamilton Anxiety Rating Scale (HAM-A) and the Panic Disorder Severity Scale (PDSS). With its high sensitivity to therapeutic change, the PAS has established itself as an indispensable outcome measure in psychopharmacological trials and cognitive-behavioral intervention research worldwide.

Keywords

Panic and Agoraphobia Scale, PAS, Panic Disorder, Agoraphobia, Psychometrics, Borwin Bandelow, Anticipatory Anxiety, Functional Impairment, Clinical Assessment, Cognitive Behavioral Therapy, Psychopharmacology, Anxiety Severity

Authors

The Panic and Agoraphobia Scale was developed by Borwin Bandelow, MD, PhD. Dr. Bandelow is an internationally recognized German psychiatrist, psychopharmacologist, and academic researcher serving as a Senior Professor of Psychiatry and Psychotherapy at the Department of Psychiatry and Psychotherapy, University of Göttingen, Germany.

Throughout his career, Dr. Bandelow has held prominent leadership positions, including serving on the executive committees of the World Federation of Societies of Biological Psychiatry (WFSBP) and the German Society for Anxiety Research (Gesellschaft für Angstforschung). His research focuses on the neurobiology, psychopharmacology, and standardized measurement of anxiety disorders, obsessive-compulsive disorder, and borderline personality disorder. He formulated the international clinical treatment guidelines for panic disorder, generalized anxiety disorder, and social anxiety disorder, authoring foundational textbooks, diagnostic manuals, and over 300 peer-reviewed scientific publications.

Purpose

The primary clinical and psychometric objective of the Panic and Agoraphobia Scale is to deliver a reliable, valid, and multidimensional quantification of the severity of panic disorder and agoraphobia. Prior to the construction of the PAS, clinical researchers and psychiatric practitioners frequently relied on broad anxiety measurement instruments, such as the Hamilton Anxiety Rating Scale (HAM-A) or the State-Trait Anxiety Inventory (STAI). Although these legacy scales capture general autonomic arousal and psychic tension, they fail to disaggregate the distinct clinical phenomena that characterize panic disorder: spontaneous panic paroxysms, situational avoidance behaviors, apprehension regarding future attacks (“fear of fear”), somatic misattributions, and secondary socioeconomic disability.

The PAS was explicitly engineered to address these methodological limitations. Its design serves several key purposes:

• Multi-Domain Clinical Profiling: Rather than collapsing panic pathology into a single global index, the PAS segments pathology into five clinically validated dimensions: frequency and severity of acute panic attacks, agoraphobic avoidance patterns, severity of anticipatory anxiety, functional and social disability, and somatic health worries.
• Treatment Outcome Monitoring: The scale was structured specifically for clinical trials and empirical practice to detect subtle and rapid changes across time. It measures therapeutic response to pharmacotherapy (e.g., selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, benzodiazepines) and cognitive behavioral therapy (CBT) involving interoceptive and in vivo exposure.
• Dual-Rater Concordance: The availability of identical self-report (PAS-SR) and clinician-administered/observer-rated (PAS-OR) forms permits investigators to cross-examine subjective patient experiences with objective clinician evaluations, minimizing diagnostic blind spots and expectancy biases.
• Standardized Response and Remission Thresholds: The PAS establishes empirical cut-offs to categorize disease severity (remission, mild, moderate, severe, extremely severe), facilitating standardized meta-analytic comparisons across clinical cohorts.

Psychological Construct

The Panic and Agoraphobia Scale operationalizes panic pathology as a complex, five-dimensional psychopathological construct. Each dimension represents a discrete yet interconnected behavioral, cognitive, or affective feature of the syndrome:

1. Panic Attacks (Items 1–4)

This subscale evaluates the core symptomatic manifestation of panic disorder: discrete periods of intense fear or acute discomfort accompanied by autonomic, respiratory, and cognitive symptoms (e.g., tachycardia, diaphoresis, dyspnea, trembling, derealization, fears of dying or losing control). The subscale measures the frequency of attacks within the preceding week (Item 1), the typical severity or symptomatic intensity of the paroxysms (Item 2), the duration of the attacks (Item 3), and the expectancy profile (Item 4), distinguishing between unexpected (spontaneous) and expected (situationally cued) panic episodes.

2. Agoraphobic Avoidance (Items 5–7)

Agoraphobic avoidance captures behavioral restriction driven by the fear of encountering situations from which escape might be difficult or embarrassing, or in which immediate medical assistance might be unavailable in the event of incapacitating panic-like symptoms. Item 5 assesses the general frequency of avoidance behaviors over the past week. Item 6 functions as a comprehensive situational inventory evaluating 24 specific physical settings (e.g., public transport, airplanes, tunnels, supermarkets, crowds, bridges, elevators, being home alone). Item 7 quantifies the subjective functional importance of the situations avoided, assessing whether the avoidance disrupts crucial daily life activities or merely non-essential events.

3. Anticipatory Anxiety (Items 8–9)

Anticipatory anxiety represents the cognitive and affective phenomenon commonly designated as “the fear of fear.” It encompasses persistent apprehension, physiological hypervigilance, and catastrophic rumination concerning the potential emergence of future panic attacks. The PAS isolates this construct across two distinct facets: the chronological duration/frequency of anticipatory rumination throughout the previous week (Item 8) and the subjective emotional intensity of this fear (Item 9).

4. Disability and Functional Impairment (Items 10–12)

Consistent with contemporary diagnostic frameworks emphasizing functional compromise, this subscale quantifies the degree of secondary impairment inflicted by panic and agoraphobia across three primary life domains: interpersonal and familial functioning (Item 10), social life and leisure activities (Item 11), and occupational or household productivity (Item 12). By assessing domain-specific disruption, the PAS mirrors functional scales like the Sheehan Disability Scale (SDS).

5. Worries About Health / Somatic Preoccupations (Items 13–14)

The final subscale targets the catastrophic misinterpretation of somatic sensations and hypochondriacal health anxieties characteristic of panic disorder. Item 13 evaluates persistent catastrophic beliefs that panic attacks will induce fatal medical crises, such as myocardial infarction, cerebrovascular accidents, or syncope. Item 14 assesses somatic attribution bias—specifically, the patient’s skepticism toward medical reassurance and the enduring conviction that their symptoms stem from an undiagnosed organic medical disease rather than an anxiety disorder.

Theoretical Framework

The structural and conceptual architecture of the PAS is deeply grounded in the cognitive model of panic disorder developed by David M. Clark (1986) and the triple vulnerability model articulated by David H. Barlow (2002).

Clark’s Cognitive Formulation

Clark posited that panic attacks occur when individuals possess an enduring cognitive vulnerability to interpret benign somatic and neurovegetative sensations (e.g., palpitations, lightheadedness, dyspnea) as immediate indicators of impending catastrophic outcomes (e.g., death, insanity, collapse). This catastrophic misinterpretation accelerates autonomic arousal via positive feedback loops:

1. A trigger (internal bodily sensation or external stressful context) generates mild physiological shifts.
2. The individual perceives these shifts and appraises them catastrophically.
3. This appraisal elicits acute sympathetic arousal, exacerbating the bodily sensations.
4. The escalation culminates in a full-blown panic attack.

Items 13 and 14 of the PAS directly measure this cognitive distortion, isolating the catastrophic beliefs and resistance to psychological explanations that perpetuate the panic cycle.

Conditioning and Agoraphobic Development

Complementing cognitive formulations, behavioral conditioning theories clarify the transition from spontaneous panic attacks to pervasive agoraphobic avoidance. Through interoceptive conditioning, internal autonomic sensations become conditioned stimuli for anxiety. Simultaneously, through exteroceptive conditioning, external environments (e.g., crowded spaces, enclosed vehicles) where panic has occurred become associated with danger. According to Mowrer’s two-factor theory, avoidance behaviors are negatively reinforced because retreating from or avoiding feared environments produces rapid reductions in anticipatory anxiety. The PAS directly maps this behavioral cascade through its separate evaluation of panic occurrences, anticipatory apprehension, and situational avoidance.

Validity

The psychometric validity of the Panic and Agoraphobia Scale has been extensively evaluated across clinical and research environments internationally.

Construct and Factorial Validity

Construct validity was initially established by Bandelow (1995, 1999) across heterogeneous cohorts of psychiatric outpatients and clinical trial participants diagnosed with panic disorder with or without agoraphobia. The five-domain model was validated through structural equation modeling, confirming that the identified subscales converge into a unified panic-agoraphobia syndrome while preserving distinct variance corresponding to behavioral avoidance, cognitive appraisal, autonomic attacks, and psychosocial impairment.

Convergent Validity

The PAS exhibits robust convergent validity when compared against established legacy instruments and standardized psychiatric rating scales:

• Panic Disorder Severity Scale (PDSS): Correlations between the PAS total score and the PDSS total score consistently fall within the range of r = .78 to .85, demonstrating high measurement congruence.
• Hamilton Anxiety Rating Scale (HAM-A): Moderate to high correlations (r = .60 to .72) have been reported, reflecting shared measurement of autonomic anxiety alongside the unique variance captured by the PAS regarding agoraphobic avoidance.
• Clinical Global Impression (CGI): Correlations between the PAS and the CGI Severity Scale typically exceed r = .65, while change scores on the PAS correlate strongly with CGI Improvement ratings (r = .70 to .80).
• Mobility Inventory for Agoraphobia (MI): The PAS Agoraphobic Avoidance subscale correlates exceptionally well with the MI “When Accompanied” and “When Alone” subscales (r = .75 to .84).

Discriminant Validity

The PAS reliably differentiates panic disorder from other psychiatric conditions, including generalized anxiety disorder (GAD), social anxiety disorder (SAD), and major depressive disorder (MDD). Patients with GAD or MDD score significantly lower on the Panic Attacks and Agoraphobic Avoidance subscales, even when presenting with elevated scores on the Disability subscale.

Sensitivity to Change (Predictive and Criterion Validity)

A primary psychometric advantage of the PAS is its high sensitivity to therapeutic change. In large-scale, double-blind, placebo-controlled psychopharmacological investigations (Bandelow et al., 2000), the PAS captured statistically significant divergence between active pharmacological agents (e.g., sertraline, venlafaxine) and placebo as early as week two of treatment. The scale demonstrates large effect sizes (Cohen’s d > 1.20) in post-treatment CBT outcome evaluations.

Reliability

Empirical analyses consistently confirm that both the self-report and observer-rated versions of the Panic and Agoraphobia Scale demonstrate excellent reliability metrics:

Internal Consistency

In the normative validation sample of 425 patients with panic disorder with agoraphobia (Bandelow, 1999), the global PAS exhibited a Cronbach’s alpha of .86, indicating high internal consistency. Subscale alphas were reported as follows:

• Panic Attacks: α = .78
• Agoraphobic Avoidance: α = .82
• Anticipatory Anxiety: α = .75
• Disability: α = .79
• Health Worries: α = .66

Subsequent international validation studies (e.g., German, French, Spanish, Turkish, and Arabic adaptations) have replicated total Cronbach’s alpha values ranging from .84 to .91.

Test-Retest Reliability

In stable clinical cohorts evaluated across a 7- to 14-day test-retest interval (under baseline placebo-lead-in or medication-washout conditions), intra-class correlation coefficients (ICC) and Pearson correlation coefficients ranged from r = .85 to .92 for the total score, confirming high temporal stability when clinical symptoms are unchanged.

Inter-Rater Reliability

The clinician-administered version displays exceptional inter-rater concordance. Independent raters scoring live or videotaped standardized clinical interviews have demonstrated inter-rater reliability coefficients exceeding r = .90. Furthermore, concurrent administrations of the observer-rated and self-report versions demonstrate high rater-patient concordance (r = .80 to .88), verifying that patient self-appraisals align closely with objective psychiatric evaluation.

Factor Analysis

Both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) have confirmed the underlying multidimensional architecture of the PAS.

Exploratory Factor Analysis (EFA)

Initial principal components analyses conducted by Bandelow using varimax and oblimin rotations revealed a clean five-factor solution corresponding to the five a priori clinical subscales. These five factors account for over 62% to 68% of the total variance. Factor loadings for individual items onto their respective primary dimensions consistently exceed .55, with minimal cross-loadings:

• Factor 1 (Disability): Strong primary loadings from Items 10, 11, and 12 (> .75).
• Factor 2 (Panic Attacks): Primary loadings from Items 1, 2, and 3 (> .70).
• Factor 3 (Agoraphobic Avoidance): Dominant loadings from Items 5 and 7 (> .68).
• Factor 4 (Anticipatory Anxiety): Strong loadings from Items 8 and 9 (> .72).
• Factor 5 (Health Worries / Somatic Attribution): Substantial loadings from Items 13 and 14 (> .60).

Confirmatory Factor Analysis (CFA)

Subsequent confirmatory factor analyses on large clinical samples demonstrated superior goodness-of-fit indices for the five-factor correlated model compared to unidimensional or two-factor models:

• Root Mean Square Error of Approximation (RMSEA): .048 to .056 (indicating close fit)
• Comparative Fit Index (CFI): .94 to .97 (exceeding standard thresholds for excellent fit)
• Tucker-Lewis Index (TLI): .93 to .96
• Standardized Root Mean Square Residual (SRMR): .042

These structural findings support the multidimensional construct: although these five domains correlate moderately with each other (reflecting a unified underlying panic diathesis), collapsing them into a single factor produces substantial loss of diagnostic information.

Instrument / Measurement Tool

The Panic and Agoraphobia Scale is structured as follows:

• Administration Modality: Available in two parallel versions: Clinician-Administered / Observer-Rated (PAS-OR) and Patient Self-Report (PAS-SR). Both versions utilize identical items, anchors, and scoring metrics.
• Timeframe: Evaluates symptom severity over the preceding one week (“the past week”).
• Administration Time: Approximately 5 to 10 minutes for self-report; 10 to 15 minutes for clinical interview.
• Scale Composition: 14 numbered items divided across five core subscales, supplemented by a 24-item agoraphobia situation checklist (Item 6).
• Response Format: Continuous 5-point Likert-type scale, scored from 0 to 4 (e.g., 0 = None / Not true; 4 = Extremely severe / Definitely true).
• Scoring System:
  • Subscale Scores: Calculated by summing the scores of items within each subscale: Panic Attacks (Items 1–4, theoretical range 0–16, though Item 4 is frequently evaluated qualitatively or scaled 0–4), Agoraphobic Avoidance (Items 5 and 7, range 0–8), Anticipatory Anxiety (Items 8–9, range 0–8), Disability (Items 10–12, range 0–12), and Worries About Health (Items 13–14, range 0–8). Item 6 is a descriptive inventory not scored linearly in the sum.
  • Total Score Range: 0 to 52 (excluding Item 6 checklist count).
  • Severity Interpretation Brackets:
    • 0 – 8: Remission / Normal / Borderline
    • 9 – 18: Mild panic and agoraphobia
    • 19 – 28: Moderate panic and agoraphobia
    • 29 – 39: Severe panic and agoraphobia
    • 40 – 52: Extremely severe panic and agoraphobia

Permissions & Fee and Test Year

The Panic and Agoraphobia Scale was originally developed and published in 1995 by Borwin Bandelow, with the comprehensive commercial clinical manual and multi-language psychometric scales formally published in 1999 by Hogrefe & Huber Publishers (now Hogrefe Publishing Group).

The official test manuals, clinician rating forms, and patient self-report sheets are protected under copyright law. Researchers and clinicians wishing to use the official proprietary forms in clinical trials, institutional healthcare settings, or commercial applications must obtain permissions or purchase testing materials directly from Hogrefe Publishing. Academic researchers conducting non-commercial, investigator-initiated clinical studies may obtain permission directly from the author or through standard academic licensing arrangements.

References

Bandelow, B. (1995). Assessing the efficacy of treatments for panic disorder and agoraphobia. II. The Panic and Agoraphobia Scale. International Clinical Psychopharmacology, 10(2), 73–81. https://doi.org/10.1097/00004850-199506000-00003

Bandelow, B. (1999). Panic and Agoraphobia Scale (PAS): Manual. Hogrefe & Huber Publishers.

Bandelow, B., Broocks, A., Pekrun, G., George, A., Meyer, T., Bartmann, U., Hillmer-Vogel, U., & Rüther, E. (2000). The use of the Panic and Agoraphobia Scale (P & A) in a controlled clinical trial. Pharmacopsychiatry, 33(5), 174–181. https://doi.org/10.1055/s-2000-8356

Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). Guilford Press.

Clark, D. M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24(4), 461–470. https://doi.org/10.1016/0005-7967(86)90011-2

Lam, R. W., Michalak, E. E., & Swinson, R. P. (2004). Assessment scales in depression, mania and anxiety. Taylor & Francis. https://doi.org/10.1201/b14115

Shear, M. K., Brown, T. A., Barlow, D. H., Money, R., Sholomskas, D. E., Woods, S. W., Gorman, J. M., & Papp, L. A. (1997). Multicenter collaborative panic disorder severity scale. American Journal of Psychiatry, 154(11), 1571–1575. https://doi.org/10.1176/ajp.154.11.1571

13. Items of the Scale (Questionnaire)