Cognitive behavioral group therapy for panic disorder in a general clinical setting: a prospective cohort study with 12 to 31-years follow-up
BMC Psychiatry volume 20, Article number: 259 (2020)
The long-term prognosis after cognitive behavioral therapy (CBT) in outpatient groups for panic disorder and agoraphobia is not well known. The purpose of this study was to assess long-term outcomes in terms of psychological health, health-related quality of life (HRQoL), quality of life (QoL) and treatment satisfaction after CBT for panic disorder and agoraphobia.
The sample consisted of 68 patients (61% response rate), who were assessed at pretreatment; at the start and end of treatment; and after 3 months, after 1 year, and over the long term (M = 24 years; SD = 5.3; range: 12 to 31 years). The main outcome was the total score on the Phobic Avoidance Rating Scale (PARS-total). At long-term follow-up, HRQoL was measured with the RAND-12 questionnaire, and QoL was measured with two questions from the “Study on European Union Statistics on Income and Living Conditions”. Patient experiences and treatment satisfaction were assessed by the Generic Short Patient Experiences Questionnaire. A marginal longitudinal model was applied to study the main outcome.
The effect size of the long-term change (mean change/ pooled SD) in the PARS-total score was (− 1.6, p < 0.001) and was stable over time. A PARS-total score reduction of 50% was found in 98% of patients at the long-term follow-up. The patients’ HRQoL and QoL were similar to the expected scores for the general Norwegian population. Of the patients, 95% reported high to very high satisfaction with the CBT, and 93% reported large treatment benefits.
To the best of our knowledge, this study has the longest follow-up after group CBT for panic disorder and agoraphobia, showing a good prognosis in ≥93% of the participating patients.
Panic disorder and agoraphobia are disabling disorders often associated with impairments in a wide range of life areas . The first documented treatment for agoraphobia was exposure therapy, and later, the cognitive elements of treatment were developed. The combined form of treatment including both exposure and cognitive therapy was labeled cognitive behavioral therapy (CBT), which is increasingly viewed as the treatment of choice . Antidepressant medication use is well documented , and there is also empirical evidence for less traditional forms of treatment, such as physical exercise .
Several randomized controlled trials have shown good effects of CBT in the treatment of panic disorder with or without agoraphobia [2, 5]. In the majority of these studies, patients were selected with strict exclusion criteria, and therapists received specialized training. Fewer studies have been conducted in the general clinical setting, where patients have been treated by therapists without specialized therapeutic competence. CBT is most often delivered individually, and most scientific studies have addressed individual CBT. Some have found that good results can also be achieved following group CBT , and this format may be more cost effective and feasible in the clinical setting. Nevertheless, the number of studies, especially long-term studies (> 2 years), on CBT in groups is scarce [5, 7,8,9,10,11,12,13,14]. Long term outcome following CBT varies across studies. The first studies showed modest results, indicating that these disorders tend to follow a chronic course [9,10,11,12]. In two long-term studies with patients following inpatient treatment with emphasis of group therapy, 38 patients (71%) were assessed 20 years after a combination of exposure and psychodynamic therapy , while 31 (67%) were investigated after cognitive and guided mastery therapy . The long-term outcomes were very good in both samples.
We found no long-term follow-up study of patients treated with group CBT in the general clinical setting. Moreover, a Cochrane review recommended that studies in this field include broader measures of quality of life (QoL) in addition to standard disease-specific outcomes . Thus, we wanted to address these research needs by extending a prior study (N = 83) published in 1998 . In that study, group CBT was offered to patients with panic disorder belonging to the catchment area of the outpatient clinic of the Central Hospital of Førde in Western Norway, with a rural population of 35,000 people. In general, the patients improved during treatment and maintained those gains at one-year follow-up.
The principal aim of the current study was to investigate the long-term trajectories of outcomes in panic disorder following outpatient CBT in groups in a general clinical setting by extending the 1998 sample and performing a long-term follow-up. The secondary aims were to investigate these patients’ long-term health-related quality of life (HRQoL) and QoL compared to those of the average general Norwegian population and to address patient experiences and treatment satisfaction.
In this prospective naturalistic study, assessments were carried out by personal interviews at pretreatment, at the first and last treatment sessions, and at the 3-month, 1-year, and 12 to 31 years follow-up (mean follow-up time was 24-years; SD 5.3). The wide range of long-term follow-up time points is due to the merging of the sample from the 1998 study and new patients. The study was approved by the Regional Committee for Medical Research Ethics in Western Norway (REK) and is registered in the REK archive (reference number: 2016/2693) . Patients were recruited from a regional outpatient clinic serving a catchment area of approximately 35,000 people referred from doctors in primary care or somatic departments. The treatment team received all referrals with a probable diagnosis of panic disorder. Those who met the DSM-III-R criteria for panic disorder with or without agoraphobia and did not meet the criteria for psychotic or posttraumatic stress disorders were offered CBT in a group format.
Of the 183 eligible patients who received treatment in the period 1989–2008, we were not able to trace the addresses of 63 patients (34%), while 8 (5%) were deceased. Thus, we invited 112 patients to participate. Of these, 2 (1%) did not want to participate, 9 (5%) wanted to participate but were unable for different reasons, and 32 (17%) did not respond. The final sample consisted of 68 patients, yielding a response rate of 61%, and they met with the therapists at the hospital. The average follow-up time since treatment was 23.8 years (SD, 5.4), ranging from 12 to 31 years. The participant characteristics at the start of treatment are presented in Table 1.
Treatment was delivered in groups of 6 to 10 participants. The groups met each week for 11 four-hour sessions. The families or significant others were invited to one of the sessions. All group sessions were conducted by the same social worker (TO) assisted by various therapists who were either registered psychiatric nurses or registrars (TB and others). The patients in each group had two lectures on anxiety disorders and their treatment by EWM, who also supervised the therapists weekly. At the pretreatment stage, the patients were given a brief outline of the treatment explaining the rationale for exposure and cognitive restructuring. All patients who met the criteria for major depression during the pretreatment visit were offered antidepressants. Benzodiazepines were tapered and stopped within the first weeks of treatment. Patients were encouraged not to use alcohol while in therapy.
Initial sessions were devoted to psychoeducation with special emphasis on the connections between perceived threat, somatic symptoms of arousal, automatic thoughts, and anxious feelings. The physiological symptoms of sympathetic activation were examined in detail, as were feelings of disaster and catastrophic interpretations of somatic symptoms. The patients were assisted in identifying their own vicious cycles of symptoms, thoughts, and feelings and learned to rate their anxiety on a 0–10 scale. The patients were repeatedly encouraged to use diaries for recording anxiety ratings, noting daily homework assignments, discussing dysfunctional thoughts, and personally monitoring the gains made in therapy .
All participants underwent voluntary hyperventilation early in treatment. Behavioural experiments were used to some extent.
The positive benefits of regular physical activity were emphasized, and the ways in which physical activity influences health in general and anxiety specifically were explained. The patients were encouraged to implement a life-long habit of walking at least 30 min daily or undertaking other physical activities of their choice.
Group sessions consisted of four modules
Review of homework, including anxiety ratings during exposure, and discussion of somatic symptoms, dysfunctional thoughts, and coping strategies.
Planning, performing, and reviewing the present day’s individual in vivo exposure at downtown locations, such as attending public offices, riding the omnibus, shopping, walking the streets, or going to a café.
Review of the progress made during the exposure and recapitulation of cognitive theory.
Assignment of daily homework for the week to come
For the long-term follow-up investigation patients met in groups and were offered sandwiches, tea or coffee. Travel expenses were covered. The Phobic Avoidance Rating Scale (PARS) was scored in individual interviews. The other instruments were self-rating and filled in by the patients while sitting in the group.
The main outcome is based on the Phobic Avoidance Rating Scale (PARS), an observer-rated scale that measures the degree of avoidance. Each item is scored on a 5-point scale where 0 represents no avoidance and 4 indicates total avoidance of the situation in question . We defined a 50% reduction in the PARS-total score as the primary indicator of a substantial response to treatment. Reduced avoidance of this magnitude is readily recognized as clinically significant by both the patient and therapist.
Level of depression was assessed by the Beck Depression Inventory (BDI) , the degree of catastrophic interpretations of somatic symptoms by the Body Sensations Questionnaire (BSQ) , the fear of fear by the Agoraphobic Cognitions Questionnaire (ACQ)  and Agoraphobic Cognitions Scale (ACS) , and the degree of phobic avoidance by the Mobility Inventory for Agoraphobia (MIA) , with subscales for avoidance both alone and accompanied by others.
HRQoL, a multidimensional construct encompassing physical, psychological and social dimensions of health , was measured with the RAND-12 questionnaire, which covers 12 dimensions of functioning and well-being . The RAND-12 has two summary scores: the physical component summary (PCS) and mental component summary (MCS). To calculate the summary scores, we applied a formula allowing the PCS and MCS to be freely correlated . Higher PCS and MCS represent better scores. RAND-12 scores from the general population were obtained from the Norwegian Survey on Living Conditions in 2002 (N = 5396) .
QoL, a global construct encompassing an overall assessment of well-being or life satisfaction , was assessed by two questions: (1) Life satisfaction: “Overall, how satisfied are you with your life nowadays (where zero is not at all satisfied and 10 is completely satisfied)?” (2) Life meaning: “Overall, to what extent do you feel that the things you do in your life are worthwhile (where zero is not at all worthwhile and 10 is completely worthwhile)?” QoL scores from the general population were obtained from the Norwegian part of the European Union Statistics on Income and Living Conditions (EU-SILC) Survey in 2017 (N = 6168) .
Patient experiences and treatment satisfaction were assessed by the Generic Short Patient Experiences Questionnaire (GS-PEQ), which consists of ten generic core questions that cover the essential dimensions of users’ experiences with a range of specialist health care services . The ten GS-PEQ item scores are intended to be used as single indicators for each of the ten specific content areas.
HRQoL, QoL and patient experiences were assessed only at long-term follow-up.
Data are presented as frequencies or percentages and as means and standard deviations or 95% confidence intervals. Exact two-sided p-values are reported. Longitudinal marginal models with an unstructured covariance structure were used in analyses of change, with time as a fixed categorical effect. This method handles missing y-data by estimating outcomes based on all available data. The length of time since starting treatment was not correlated with long-term PARS-total scores (r = 0.06, p = 0.642). Thus, we coded all long-term data (ranging from 12 to 31 years) into one category. The trajectories of the PARS-total score during the first year of follow-up for long-term participants versus non-long-term participants were displayed, and the interaction effect of time*group was tested. Age- and gender-adjusted comparisons of generic health status and QoL between the patient group and the general population were conducted with a one-sample t-test. Effect sizes of changes over time were calculated by using change scores from the longitudinal marginal models divided by the crude pooled standard deviations at the start of treatment and at the last time point. Effect sizes for score differences in the patient group compared to the general population were calculated by using difference scores divided by the standard deviations in the patient group. Effect sizes were interpreted according to the following criteria: trivial (< 0.2), small (0.2 to < 0.5), moderate (0.5 to < 0.8 SD), and large (≥0.8) . The statistical analyses were conducted using SPSS, version 25.0 (SPSS Inc., Chicago, USA).
The PARS score was stable from pretreatment to the start of treatment. The PARS-total score improved during treatment (p < 0.001) (Table 2). The effect size was large at the end of treatment and remained stable thereafter. A reduction of 50% in the PARS-total score was found in 82% of the patients by the end of treatment and in 98% of patients by the 24-year follow-up. All secondary outcomes also improved over time (Table 2; p < 0.001 for all). The standard deviations for all scores decreased over time, indicating that the response to treatment was general. There were no significant differences between the patients who were assessed at the 24-year follow-up and the other patients in terms of their PARS-total trajectories over time (p = 0.647) (Fig. 1).
The patients’ HRQoL and QoL 24 years after CBT were similar to the expected scores in the general population after adjustment for age and gender (Table 3). Details on patient experiences and treatment satisfaction at the 24-year follow-up are presented in Table 4. The patients had high ratings on all items, with 95% reporting high to very high satisfaction with the CBT and 93% reporting large benefits.
To the best of our knowledge, this study has the longest follow-up after group CBT for panic disorder and agoraphobia. The effect size for clinical treatment was large at the end of treatment and remained stable thereafter. A 50% reduction in the PARS-total score was found in 98% of patients at the long-term follow-up, and 93% of patients were satisfied with the outcomes. The patients’ HRQoL and QoL were similar to the expected scores for the general Norwegian population.
Our results correspond well with previous shorter-term studies on the treatment of panic disorder and agoraphobia [2, 4]. The few studies reporting long-term outcomes have diverging results, and the explanations for this is not clear. Our results correspond best to the previous studies with longest follow-up [13, 14]. An interesting common feature of these studies is the focus on the group format. Our clinical experience indicate that the group fellowship was important for many patients. Systematic exposure treatment and cognitive therapy seem to yield substantial and lasting therapeutic gains, and these results can also be achieved in the general clinical setting. Thus, the good results following individual CBT seem to be achieved by group CBT as well. This format is suitable for training new therapists, who can learn the method by working as co-therapists with more experienced professionals.
The main strength of this study is its length of follow-up. Moreover, the study was conducted in the general clinical setting, had few exclusion criteria and included participants with multiple comorbidities. Thus, the results are generalizable to the general clinical setting. Furthermore, we had a broad set of validated measures encapsulating mental health, HRQoL, QoL, and patient experiences and treatment satisfaction. The study’s main limitations are its observational design, and we do not know the long-term prognosis of the 39% of patients who did not participate in the long-term follow-up. Loss of patients is also observed in other studies with a very long follow-up period [13, 14]. Patients may have chosen not to participate in the long-term follow-up because they were doing well and perceived their attendance as unnecessary, because they were doing poorly and did not want further contact, or because they had other reasons unrelated to the study outcomes. However, it is reassuring that the PARS-total trajectories during the first year after treatment were similar in long-term -and other participants. Unfortunately, we have no measure of remission, only of responders. It would also had been nice to have a therapist rating which covered panic/agoraphobia in a broader sense than only avoidance, but these other aspects are covered by self- report instruments.
Our study suggests that CBT in groups is feasible and that the therapeutic gains last for a very long period of time. This finding is good news for a common mental health problem. However, further studies are needed to establish a robust body of evidence on this matter.
Availability of data and materials
The dataset generated during this study is not publicly available, as the patient consent and approval from the Regional Committee for Medical and Health Research Ethics prevent us from sharing individual patient-level data in public repositories. However, the data are available upon reasonable request from the corresponding author.
Cognitive Behavioral Therapy
Health-Related Quality of Life
Quality of Life
Phobic Avoidance Rating Scale
Regional Committee for Medical Research Ethics in Western Norway
Beck Depression Inventory
Body Sensations Questionnaire
Agoraphobic Cognitions Questionnaire
Agoraphobic Cognitions Scale
Mobility Inventory for Agoraphobia
Physical Component Summary
Mental Component Summary
European Union Statistics on Income and Living Conditions
Generic Short Patient Experiences Questionnaire
Barrera TL, Norton PJ. Quality of life impairment in generalized anxiety disorder, social phobia, and panic disorder. J Anxiety Disord. 2009;23(8):1086–90.
van Dis EAM, van Veen SC, Hagenaars MA, Batelaan NM, Bockting CLH, van Den Heuvel RM, Cuijpers P, Engelhard IM. Long-term outcomes of cognitive behavioral therapy for anxiety-related disorders: a systematic review and meta-analysis. JAMA psychiatry. 2019. Online ahead of print.
Nathan PE, Gorman JM. A guide to treatments that work. 4th ed; 2015.
Hovland A, Nordhus IH, Sjobo T, Gjestad BA, Birknes B, Martinsen EW, Torsheim T, Pallesen S. Comparing physical exercise in groups to group cognitive behaviour therapy for the treatment of panic disorder in a randomized controlled trial. Behav Cogn Psychother. 2013;41(4):408–32.
Pompoli A, Furukawa TA, Imai H, Tajika A, Efthimiou O, Salanti G. Psychological therapies for panic disorder with or without agoraphobia in adults: a networkmeta-analysis. Cochrane Db Syst Rev. 2016;4(4):CD011004.
Prats E, Dominguez E, Rosado S, Pailhez G, Bulbena A, Fullana MA. Effectiveness of cognitive-behavioral group therapy for panic disorder in a specialized unit. Actas Esp Psiquiatri. 2014;42(4):176–84.
von Brachel R, Hirschfeld G, Berner A, Willutzki U, Teismann T, Cwik J, Velten J, Schulte D, Margraf J. Long-term effectiveness of cognitive behavioral therapy in routine outpatient care: a 5-to 20-year follow-up study. Psychother Psychosom. 2019;88(4):225–35.
Imai H, Tajika A, Chen P, Pompoli A, Furukawa TA. Psychological therapies versus pharmacological interventions for panic disorder with or without agoraphobia in adults. Cochrane Db Syst Rev. 2016;10(10):CD011170.
Brown TA, Barlow DH. Long-term outcome in cognitive-behavioral treatment of panic disorder: clinical predictors and alternative strategies for assessment. J Consult Clin Psychol. 1995;63(5):754–65.
Durham RC, Chambers JA, Power KG, Sharp DM, Macdonald RR, Major KA, Dow MG, Gumley AI. Long-term outcome of cognitive behaviour therapy clinical trials in Central Scotland. Health Technol Assess. 2005;9(42):1–174.
Durham RC, Higgins C, Chambers JA, Swan JS, Dow MG. Long-term outcome of eight clinical trials of CBT for anxiety disorders: symptom profile of sustained recovery and treatment-resistant groups. J Affect Disord. 2012;136(3):875–81.
Tyrer P, Seivewright H, Johnson T. The Nottingham study of neurotic disorder: predictors of 12-year outcome of dysthymic, panic and generalized anxiety disorder. Psychol Med. 2004;34(8):1385–94.
Hoffart A, Hedley LM, Svanoe K, Langkaas TF, Sexton H. Agoraphobia with and without panic disorder a 20-year follow-up of integrated exposure and psychodynamic therapy. J Nerv Ment Dis. 2016;204(2):100–7.
Hoffart A, Hedley LM, Svanoe K, Sexton H. Cognitive and guided mastery therapies for panic disorder with agoraphobia: 18-year long-term outcome and predictors of long-term change. Clin Psychol Psychother. 2016;23(1):1–13.
Martinsen EW, Olsen T, Tonset E, Nyland KE, Aarre TF. Cognitive-behavioral group therapy for panic disorder in the general clinical setting: a naturalistic study with 1-year follow-up. J Clin Psychiat. 1998;59(8):437–42.
Langtidsforløpet ved panikklidelse [https://helseforskning.etikkom.no/prosjekterirek/prosjektregister/prosjekt?p_document_id=710071&p_parent_id=729946&_ikbLanguageCode=n]. Accessed 2 Jan 2020.
Clark DM. A cognitive approach to panic. Behav Res Ther. 1986;24(4):461–70.
Hoffart A, Friis S, Martinsen EW. The phobic avoidance rating-scale - a psychometric evaluation of an interview-based scale. Psychiat Dev. 1989;7(1):71–81.
Beck AT, Steer RA, Ball R, Ranieri WF. Comparison of Beck depression inventories-IA and -II in psychiatric outpatients. J Pers Assess. 1996;67(3):588–97.
Chambless DL, Caputo GC, Bright P, Gallagher R. Assessment of fear of fear in agoraphobics - the body sensations questionnaire and the agoraphobic cognitions questionnaire. J Consult Clin Psych. 1984;52(6):1090–7.
Hoffart A, Friis S, Martinsen EW. Assessment of fear of fear among agoraphobic patients - the agoraphobic cognitions scale. J Psychopathol Behav. 1992;14(2):175–87.
Chambless DL, Caputo GC, Jasin SE, Gracely EJ, Williams C. The mobility inventory for agoraphobia. Behav Res Ther. 1985;23(1):35–44.
Cramer JA, Spilker B. Quality of life and pharmacoeconomics an introduction. Philadelphia: Lippincott-Raven; 1998.
Farivar SS, Cunningham WE, Hays RD. Correlated physical and mental health summary scores for the SF-36 and SF-12 Health Survey, V.I. Health Qual Life Outcomes. 2007;5:54.
Norwegian Surveys on Living Conditions [https://nsd.no/nsddata/serier/levekaarundersokelser.html]. Accessed 2 Jan 2020.
Sjetne IS, Bjertnaes OA, Olsen RV, Iversen HH, Bukholm G. The generic short patient experiences questionnaire (GS-PEQ): identification of core items from a survey in Norway. BMC Health Serv Res. 2011;11.
Cohen J. Statistical Power analysis for the behavioral sciences., 1 edn. New York: Academic Press; 1988.
We would like to thank Karl Ove Hufthammer, (Department of Research and Development, Haukeland University Hospital, Bergen, Norway) for supporting the statistical analyses. We also thanks Kristen Hagen (OCD-Team, Molde Hospital, Molde, Norway) for providing useful comments on the manuscript.
This work was supported by internal grants from Førde Hospital Trust, Førde, Norway. The funding body paid for the costs of the data collection but had no influence on the design of the study, analysis or interpretation of data.
Ethics approval and consent to participate
The study was approved by the Regional Committee for Medical Research Ethics in Western Norway (reference number: 2016/2693). The participants provided informed written consent to participate in this study.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Bilet, T., Olsen, T., Andersen, J.R. et al. Cognitive behavioral group therapy for panic disorder in a general clinical setting: a prospective cohort study with 12 to 31-years follow-up. BMC Psychiatry 20, 259 (2020). https://doi.org/10.1186/s12888-020-02679-w