Patient characteristics and changes in anxiety symptoms in patients with panic disorder: Post-hoc analysis of the PARADIES cluster randomised trial.

Anxiety disorders are among the most common mental health problems in primary care. The PARADIES (Patient Activation foR Anxiety DIsordErS) intervention combined elements of cognitive behavioural therapy with case management and has demonstrated efficacy. Our aim was to explore patient characteristics, which may influence the course of anxiety symptoms over a 12 months period. Multiple linear regression was used to quantify associations of baseline characteristics (demographics, clinical parameters, medication use) with changes in anxiety symptoms as measured by the Beck anxiety inventory. Treatment modalities (e.g. adherence to appointment schedules) were considered as confounders. We examined univariate associations between dependent and independent variables before considering all independent variables in a multivariate final model. To find the best model to explain BAI score changes, we performed step-wise selection of independent variables based on Akaike information criteria. We tested for interaction terms between treatment allocation (intervention vs control) and independent variables using the multivariate model. We repeated these analyses in control vs intervention groups separately. From the original trial (N = 419), 236 patients (56.3%) were included. In the multivariate model, receiving the intervention (p<0.001), higher anxiety symptom severity (p<0.001) and longer illness duration at baseline (p = 0.033) were significantly associated with changes in anxiety symptom severity to the better while depression severity at baseline (p<0.001) was significantly associated with changes in anxiety symptoms to the worse. In stratified analyses, the control group showed significant associations between depression symptom severity and illness duration with anxiety symptom changes while baseline severity of anxiety symptoms remained significantly associated with anxiety symptom changes in both groups. A brief primary-care-based exposure training combined with case management is effective in a broad range of patients with panic disorder with/without agoraphobia, including those with longer illness duration and co-existing symptoms of depression at baseline.

Introduction

Anxiety disorders have a high prevalence, with a 12-month rate of about 18% and lifetime rates of about 30% [1-3]. They are among the most common mental health problems seen in primary care settings, where patients with anxiety disorders are predominantly managed [4]. In fact, for most patients with panic disorder, the general practitioner (GP) is the first, and often the only contact [5]. Patients with anxiety disorders experience lower quality of life and increased rates of health care utilisation [6, 7]. Treatment options in primary care include pharmacotherapy (antidepressants and anxiolytics if required) as well as psychotherapeutic approaches, such as cognitive behavioural therapy (CBT) [8-10]. There is also some evidence that case management improves outcomes, which encompasses continuous monitoring and proactive support for patients as a collaborative effort of primary care teams [11-14].

The PARADIES intervention was a practice team–supported exposure training and combined evidence-based elements of CBT with case management [5, 15]. Adult patients diagnosed with PDA (ICD-10: F41.0 or F40.01) were included, while those with suicidality, psychotic or substance-related disorders, severe physical impairments, pregnancy or ongoing anxiety-specific psychotherapy were excluded. Intervention group patients received a therapy companion book providing psychoeducation, instructions on how to perform exposer-exercises, and exposure log sheets. Over a 23-week period, four structured GP visits were scheduled. The first three visits introduced patients to the CBT elements, while the fourth appointment provided relapse-prevention information. From the second visit, patients were encouraged to independently perform anxiety exposure exercises at least twice weekly. Structured telephone monitoring was carried out by a practice nurse in order to enhance treatment adherence and ensure regular monitoring of anxiety symptoms. Where monitoring results were concerning, GPs could arrange additional patient contacts and/or adapt exercise plans. Patients in the control group received treatment as usual. In both groups, GPs could administer any supplementary treatment (including pharmacotherapy or referrals) at their own discretion.

The intervention was evaluated in a cluster randomised controlled trial, the findings of which have previously been published [5, 16]. Briefly, the trial included 73 general practices (419 patients). As described in more detail elsewhere [5], practices first recruited patient participants over a three months period and were subsequently randomized to provide care as usual or deliver the PARADIES intervention to participating patients under their care (allocation ratio 1:1). Overall, 36 practices (230 patients) were randomised to the intervention arm and 37 practices (189 patients) to the control arm. Clinical endpoints were the clinical severity of the anxiety symptoms as assessed by the Beck Anxiety Inventory (BAI) [17]. The intention to treat analysis found that symptoms of anxiety improved to a significantly greater extent in the intervention group (p = 0.008) with an intergroup difference in the reduction of the BAI score (range: 0–63) of 4.0 points [−6.9; −1.2] at twelve months.

The PARADIES trial has therefore established that, on average, patients with anxiety disorder benefit from a combination of CBT and case management. The aim of this study was to examine variability in the course of anxiety symptom severity at 12 months follow up and to explore patient characteristics, which may explain that variability. In particular, we investigated as explanatory factors patient demographics (since age, sex and level of education may influence motivation and/or ability to engage with the intervention) [18], anxiety symptom severity and illness duration at baseline, co-morbid depression, multimorbidity and polypharmacy as well as the type and/or number of psychotropic drugs used to treat these (since all of these factors may influence the prognosis of panic disorders) [19].

Methods

Study design

Following descriptive analysis of baseline characteristics and patient level changes in BAI scores, we first pooled data from all intervention and control group patients in order to examine associations between patient baseline characteristics and patient level changes in the BAI score while controlling for treatment group status (intervention vs control). Differential effects of baseline characteristics on anxiety symptom changes in control and intervention groups were subsequently examined via interaction terms and in stratified analyses. Through stratification by use vs non-use of medication commonly used in the treatment of anxiety disorders (antidepressants and benzodiazepines), we further explored the influence of baseline symptoms of anxiety and depression.

The Ethics Committee of the Friedrich-Schiller University Jena granted approval of the study protocol on 17 August 2012 (no. 3484–06/12). All participating physicians and patients gave their written informed consent to participate in the study.

Study population

For the purposes of this study, we included all participants in the PARADIES trial for whom there was complete information on all baseline characteristics of interest (see below) and BAI score changes between baseline and 12 months follow up.

Dependent and independent variables

As the dependent variable (i.e. the outcome measure of interest), we defined the change in severity of anxiety symptoms measured by the self-administered Beck Anxiety Inventory (BAI) [17] over the 12 months study period. The BAI asks patients to rate how severely affected they had been by 21 typical symptoms of anxiety (total score range 0–63) during the previous week.

In order to identify factors associated with changes in anxiety symptoms, we assessed the following categories of variables:

patient demographics—age, sex, years of education;

clinical parameters at baseline—illness duration, depression scale (Patient-Health-Questionnaire (PHQ-9) [20]), multimorbidity (defined as chronic somatic conditions according to Bussche > 3 [21]), Patient Assessment of Chronic Illness Care (PACIC) [22];

medication at baseline—any antidepressants, any benzodiazepines, psychotropic polypharmacy (defined as two or more of the following medicines taken concomitantly: benzodiazepines, z-drugs, opioids, other hypnotics, antidepressants, antipsychotics), and any polypharmacy (defined as 5 or more medicines taken concomitantly);

Given that general practices and patients had a degree of flexibility in delivering and adhering to the intervention, respectively, it is possible that such factors may modify the effect of the intervention and if so, distort the relationship between baseline characteristics and treatment response. To control for such effects, we therefore considered the following treatment modalities:

completion of the intervention as intended (attendance at 4th appointment: yes vs no)

number of monitoring contacts with the case manager (range: 1–10)

number of additional contacts with the GP (range: 0–2).

Statistical analyses

To compare intervention and control groups, the t-test or the Mann-Whitney U test was used for metric variables and the χ2 or Fisher exact test for categorical variables. For all regression models, the dependent variable (change in severity of anxiety symptoms over the 12 months study period) was calculated as the difference in the Beck-Anxiety-Inventory [BAI] [5] between T0 and T2, where we subtracted individual BAI sum scores at T2 from individual BAI sum scores at T0.

As the primary analysis, we examined associations between independent and dependent variables while controlling for allocation status (intervention vs control), using multivariate linear regression. We initially examined univariate associations before considering all independent variables in a multivariate final model. All continuous independent variables were mean-centred. In order to find the best model for explaining BAI score changes, we added all independent variables into multivariate models and performed step-wise selection (both directions: forward and backward) based on AIC criteria (R package MASS; function stepAIC). We also tested interaction terms between treatment allocations and independent variables. In order to examine, whether and to which extent the effect of baseline characteristics on changes in anxiety symptom severity differed between the intervention and control group, we repeated the analyses for the intervention and control groups separately. Since drug treatment may influence baseline symptoms of anxiety and depression, we additionally repeated the analyses for users and non-users of benzodiazepines or antidepressants separately. All effects are reported as significant at p < 0.05. Residual plots (residuals vs. fitted values, Q-Q plot, scale-location plot, and residuals vs leverage values) were used to check the assumptions of the linear regression models. Data were analysed using R version 3.6.3 (https://www.r-project.org/).

Results

Of the 419 patients included in the PARADIES trial, we included in the analysis all 236 (56.3%) patients with complete information on all independent (baseline characteristics) and dependent variables (difference of BAI scores at baseline and 12 months follow up). Of these, 128 patients and 108 patients were members of the intervention and control group, respectively. The proportions of patients with complete information were similar in the intervention group (55.7%) and in the control group (57.1%).

In the overall sample, the mean (SD) age was 45.3 (13.4) years and the majority (75.4%) were women. shows that there were no significant differences between included patients of the intervention and control groups and patients were very similar in terms of demographics and clinical parameters at baseline. With respect to medication, the prevalence of polypharmacy and psychotropic polypharmacy were also similar. In both groups, approximately half of the patients were taking antidepressants at baseline and the median defined daily doses (DDD) of antidepressants did not differ between intervention and control patients. The use of benzodiazepine anxiolytics was rare in both groups, albeit somewhat higher in the intervention vs control group (8.6% vs 3.7%).

With respect to intervention delivery parameters, four intervention group patients had missing information on how many and which of the scheduled appointments had been attended. Almost all intervention group patients (96.1%) attended the first appointment, slightly lower proportions attended the second (94.5%) and third appointment (87.5%) while three quarters (75%) attended the fourth appointment.

Changes in anxiety symptom severity over the study period

shows the distribution of changes in BAI scores in the intervention and control groups, where positive values reflect improvements and negative values reflect worsening of anxiety symptom severity. In the intervention group, the mean BAI (SD) score fell from 28.2 (12.3) to 17.3 (12.5), while in the control group it fell from 28.2 (12.4) to 22.1 (13.3) at 12 months follow up. The mean differences in BAI scores were 9.4 (12.4) in the intervention group and 6.0 (14.5) in the control group. In the intervention group, changes in BAI scores ranged from an improvement by 44 points to a worsening by 22 points, and in the control group from an improvement of 44 points to a worsening of 39 points. The majority of patients in the intervention group (n = 103; 80.5%) and in the control group (n = 70; 74.8%) had changes in anxiety symptom severity to the better while 25 (19.5%) intervention group patients and 38 (35.2%) control group patients had no changes in anxiety symptom severity or changes to the worse.

Distribution of changes in anxiety symptom severity between T0 and T2 among patients allocated to intervention and control groups.

Positive values reflect improvements while negative values reflect worsening of anxiety symptoms.

compares the proportions of intervention and control group patients by BAI severity category at baseline and at the end of the study period. While there were no significant differences between intervention and control groups at baseline, the groups differed significantly at 12 months follow up. Over the study period, the proportions of patients with severe symptoms decreased from 51.9% to 38.9% in the control group and more than halved (from 52.3% to 22.7%) in the intervention group, while the proportions of patients with no or minimal symptoms increased from 3.7% to 16.7% in the control group and from 1.6% to 28.1% in the intervention group.

Examination of factors associated with anxiety symptom changes in the whole study population

Tables show the findings of linear regression analyses, while model diagnostics and iterative changes in R2 values are provided in the supporting .

shows the findings of univariate and multivariate regression analyses for all patients, where allocation status (intervention vs control group) is controlled for. In univariate analysis, three variables were significantly associated with changes in anxiety symptom severity to the better over the study period, namely being allocated to the intervention group (univariate regression coefficient 4.2 (95% CI 0.6 to 7.7); p = 0.020), being female (4.1 (0.1 to 8.2); p = 0.047) and having higher anxiety symptom severity (BAI) at baseline (0.5 (0.4 to 0.7); p<0.001).

In multivariate analysis, receiving the intervention (multivariate regression coefficient 5.7 (95% CI 2.6 to 8.8); p<0.001) and higher anxiety symptom severity at baseline (0.8 (0.6 to 1.0); p<0.001) were still significantly associated with changes in anxiety symptom severity to the better. In contrast to the univariate analysis, longer illness duration at baseline (0.2 (0.02 to 0.4); p = 0.033) was significantly associated with changes in anxiety symptom severity to the better, whereas more severe symptoms of comorbid depression at baseline (-0.9 (-1.4 to -0.5); p<0.001) were significantly associated with changes in anxiety symptoms to the worse after adjustment for confounding variables. Three interaction terms were also included in the final model indicating differential effects of baseline anxiety symptoms, baseline symptoms of depression and illness duration on changes in anxiety symptom severity in the intervention versus control group. The final multivariate model included 12 variables and three interaction terms and explained 31.1% of variability in anxiety symptom changes (R2 = 0.3105).

Examination of factors associated with anxiety symptom changes in intervention vs control patients

shows the results of analyses stratified by allocation status (intervention vs control). In the control group, six variables contributed to the final model (female sex, baseline anxiety symptom severity, baseline symptoms of depression, use of benzodiazepines and use of polypharmacy) and explained 38% of variability in anxiety symptom changes (R2 = 0.38). In the intervention group, only two variables (age and baseline anxiety symptom severity) contributed to the final model and explained 21% of variability in anxiety symptom changes (R2 = 0.21).

More severe symptoms of anxiety at baseline were significantly associated with changes in anxiety symptoms to the better in both groups, but the effect was stronger in the control vs intervention group (regression coefficient of 0.8 (0.6 to 1.0)) vs 0.5 (0.3 to 0.7). Only in the intervention group was age significantly associated with changes in anxiety symptom severity to the worse (regression coefficient of– 0.2 (-0.3 to -0.02)). By contrast, only in the control group were more severe symptoms of depression at baseline (regression coefficient of—0.9 (-1.4 to -0.4)) and polypharmacy (regression coefficient of—6.0 (-11.8 to -0.2)) significantly associated with changes in anxiety symptom severity to the worse. Treatment modalities in the intervention group (i.e. attended appointments, telephone contacts and number of additional unplanned contacts) did not alter any of the effect estimates for the baseline characteristics investigated.

Examination of factors associated with anxiety symptom changes in users vs non-users of antidepressants or benzodiazepines

shows the results of analyses stratified by use vs non-use of antidepressants or benzodiazepines. Again, allocation status and anxiety symptom severity at baseline were significantly associated with anxiety symptom changes in both groups and polypharmacy contributed to both models, whereas education time was relevant only among non-users and female sex only among users of antidepressants or benzodiazepines, respectively. While in both groups, more severe baseline symptoms of anxiety were significantly associated with changes in anxiety symptoms to the better, only among users of antidepressants or benzodiazepines were more severe baseline symptoms of depression significantly associated with changes in anxiety symptoms to the worse.

Discussion

In this exploratory analysis of anxiety symptom changes in a cohort of patients with panic disorder enrolled in the randomized controlled PARADIES trial, we found considerable variation in anxiety symptom changes ranging from substantial improvements to relevant worsening of symptoms in both intervention and control groups. Symptoms were unchanged or worsened over the 12 months period in 19.5% and 35.2% of intervention and control group patients, respectively. The final multivariate regression model for all patients included 12 variables and explained 31.3% of this variation. Overall, our analyses confirmed the significant intervention effect reported in the pre-specified primary analysis [5]. In analyses stratified by allocation status, six variables contributed to explain variation in anxiety symptoms changes in the control group but only two variables contributed in the intervention group, which re-emphasises that the PARADIES intervention is a decisive driver of anxiety symptom improvements among those receiving it.

Our study found that for patients in both treatment groups, greater anxiety symptom severity at baseline was associated with changes in anxiety symptom severity to the better at 12 months follow up. We could not find other studies which report the changes in symptom severity as the difference in values from measurement at baseline to measurement at follow-up. Generally, the literature so far suggests that patients with more severe symptom severity at baseline tend to have poorer treatment outcomes [23-25], but findings have been mixed for the association of pre-treatment severity with CBT for panic disorder outcomes [26]. On the other hand, a study by Hadjistavropoulos et al. (2016) showed (consistent with our findings) that greater pretreatment condition severity was associated with larger therapy benefits [27]. Given that the latter intervention had higher intensity than the PARADIES trial (12 vs 4 modules of therapist-assisted Internet-delivered cognitive behavior therapy for depression or generalized anxiety), our findings suggest that even lower intensity interventions may benefit patients with more severe symptoms at baseline. We assume that patients who have more severe symptoms, have more potential to improve but may also be more likely to seek and receive treatment and engage with exercises that were part of the intervention [28]. Additionally, our results may partially be explained by regression toward the mean, a statistical effect in which measured values tend to be closer to the population mean than the baseline values due to statistical variability.

A somewhat surprising finding was that longer illness duration was independently associated with changes in anxiety symptoms to the better in the control group (while no such association was found in the intervention group). Seemingly in contrast to our findings, two naturalistic studies have found less favorable outcomes for patients with longer duration of anxiety symptoms without intervention. Ronalds et al. (1997) found that adult patients in general practice with depressive, anxiety or panic disorder (n = 148; DSM-lll-R criteria), more patients showed improvements in anxiety symptom severity when they had an illness duration of less than six months versus six months or more (52% vs 42%) [29]. Penninx et al. (2011) found that in patients with anxiety and/or depression (n = 1209) longer baseline duration of the index disorder was independently associated with lower likelihood of first remission [24]. Nevertheless, the findings of these studies are not directly comparable to ours, since outcomes were measured differently, and in our study, patients were enrolled in a randomised controlled trial. A possible explanation of our findings is that in the control group, patients with longer illness duration may be more likely than those with shorter illness duration to be initiated on effective treatment, whereas all patients in the intervention group were subjected to the same intervention.

It is known that co-morbid anxiety and depression in the same patient negatively affect the clinical course of anxiety symptoms. In previous studies, patients with more severe manifestations of these illnesses typically respond less robustly to treatment than patients with either disorder alone [24, 30, 31]. However, our finding that in stratified analysis, the negative impact of comorbid depression on anxiety symptom severity was limited to the control group, suggests the intervention also yielded a benefit of CBT in patients with more severe symptoms of depression at baseline. This suggests that the intervention is robust to external confounders over 12 months. Moreover, we have previously shown that our short intervention is cost effective regarding total and direct costs as well as disease-specific health care costs [32].

We found medication at baseline not to be significantly associated with anxiety symptom changes in primary analysis (although benzodiazepine use at baseline and polypharmacy were included in the final models for the control group and the study population as a whole). In the case of benzodiazepines, the lack of a significant association may be attributable to the relatively small proportion of study participants using these drugs at baseline. In the case of antidepressants, our findings suggest that actual symptom control of depression is a more important driver for the course of anxiety symptoms than use of antidepressants (which may not sufficiently control depression symptoms).

When we stratified by use vs non-use of benzodiazepines or antidepressants, more severe symptoms of depression at baseline were associated with changes in anxiety symptoms to the worse among users but not among non-users of antidepressants or benzodiazepines at baseline, respectively. This finding suggests that the presence of more severe symptoms of depression in spite of drug treatment marks a group of patients, who may also have a less favorable prognosis of anxiety symptoms.

In terms of demographic variables, female sex was associated with larger changes in anxiety symptoms to the better only in univariate analysis, whereas older age was associated with anxiety symptom changes to the worse among intervention group patients. Previous studies found women to respond more favorably to collaborative care interventions for anxiety and to report a higher commitment to therapy and a stronger belief in the helpfulness of psychotherapy than men [33]. These factors might predict motivation and effort in treatment and also positive clinical outcomes in CBT [34]. Conversely, older age may have the opposite effect [35]. Although female sex was no longer associated with anxiety symptom changes in multivariate analysis, the change in the point estimate was not large (it fell from 4.1 (0.1 to 8.2) to 3.3 (-0.2 to 6.7) in univariate versus multivariate analysis). Our findings can therefore be interpreted as broadly consistent with previous research.

Strengths and limitations

A major strength of our study is that it is one among very few to examine patient characteristics associated with a more and less favourable course of anxiety symptoms and the effect of intervention. As such it enhances our understanding of who might benefit more and less from the PARADIES intervention, which we consider an essential addition to the primary analysis of intervention effectiveness.

However, our study also faces limitations. First, this is a post-hoc analysis of data from the completed PARADIEs trial and as such it is an exploratory study, which is vulnerable to multiple testing bias. While our findings confirm the findings of the fully powered primary analysis, the identified associations between independent and dependent variables should be interpreted as hypothesis generating rather than hypothesis testing. The sizes of the intervention group (n = 230) and the control group (n = 189) were different due to a higher loss to follow up in the control group, possibly due to the de-blinding of the intervention-status after randomisation. Nevertheless, variables like age, education, depression scale or BAI were well balanced between both groups and all statistical analyses were adjusted for the relevant variables. Second, only 236 of 419 patients (56.3%) could be included in the regression analysis, because BAI was not documented or independent variables were missing. Third, it was not possible to calculate a multilevel model (which is the appropriate method of analysis for cluster-randomized studies) because there were too many practices with only 1–2 patients and our multi-level models did not converge. However, the intraclass correlation coefficient (0.017) was very low in the analysis of the PARADIES trial suggesting minimal cluster effects [5]. Fourth, the final multivariate model only explained approximately a third of the variation in anxiety symptom changes among participants, which suggests there may be additional sources of variation, which we did not measure. Finally, although the intervention focussed on non-pharmacological treatment options, we cannot exclude that differential use of anti-anxiety medication in intervention and control groups during follow up may have contributed to the observed effects.

Conclusion

Managing patients with panic disorder in primary care is a challenging task. Our study shows that a short intervention combining elements of CBT and case management, and delivered by a joint effort of interdisciplinary staff (GPs and non-medical trained staff) is effective over 12 month in a broad range of patients, particularly among those with more severe symptoms, and including those with longer disease duration and co-existing symptoms of panic disorder and depression. Further research is required to identify effective primary care interventions for a considerable proportion of patients (19.5% in the intervention group of this study) whose symptoms remained stable or worsened over the 12 months study period.

Pearson correlation coefficient (for continuous variables)/t-test for categorical variables.

(PDF)

Click here for additional data file.

Regression analysis, R2 and model residuals.

(PDF)

Click here for additional data file.

22 Mar 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Dr. Burak Yulug

Academic Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf\\

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

4. One of the noted authors is a group or consortium (PARADIES study group). In addition to naming the author group, please list the individual authors and affiliations within this group in the acknowledgments section of your manuscript. Please also indicate clearly a lead author for this group along with a contact email address.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is a secondary regression analysis from an already-completed cluster randomized clinical trial. As such, it is an exploratory study, not powered to detect significant treatment differences, and it should be so-stated in the text. The comment "p<.05 is considered significant" means nothing with no adjustment for many, many hypothesis tests. I suggest a comment that "p-values should be interpreted as a guide, rather than as a determinant of significance". Alternatively, if there are confirmatory hypotheses, these should be stated clearly, power analysis given, and Bonferroni adjustments for multiple tests.

I appreciate that they are concerned about analyzing this data set incorrectly, given they are ignoring the cluster effect, and they do indicate some homogeneity across clusters. Finally, any regression course would require all students to do diagnostics for the model, including residual analyses and graphical inspection. I would expect no less in a scientific paper!

Reviewer #2: Dear Editor,

I am stating the evaluations about the study below.

Kind regards.

1. A large part of the introduction describes the PARADIES study, including the independent variables investigated in the study and their relationship with panic disorder in the introduction section.

2. Out of 236 patients in the study, 128 were in the intervention group and 108 were in the control group. In the method section, it is explained in detail how the patients were assigned to the groups.

3. Comparison of both groups (intervention & control) in terms of sociodemographic characteristics (with t test and Chi-square tests for independent groups) and giving p values in Table 1 in the Results section.

4. Half of the patients are receiving antidepressant treatment at the beginning, comparing the drug doses used by the patients using antidepressants in terms of equivalence (For example, 10 mg. Escitalopram = 20 mg. Citalopram = 20 mg. Fluoxetine = 50 mg. Sertalin = 20 mg. Paroxetine).

5. Of the 128 patients, 44 improved, 22 worsened. Of the 106 patients, 44 improved and 39 worsened. Rates are given on worsening only. Giving rates over recovery. (The rate of 74.8% on page 9 should be 64.8%.)

6. Adding the following table to the results section and making its statistical analysis.

The patients in both the intervention group and the control group were divided into groups according to the scores they got from the Beck Anxiety Inventory at the baseline and at the end of the 12-month follow-up, and comparing the two groups.

0-7 points (No anxiety)

8-15 points (Mild anxiety)

16-25 points (Moderate anxiety)

≥ 26 points (Severe anxiety)

7. Showing the correlation analysis with a table in the results section in terms of the relationship between the independent variables included in the regression analysis and the dependent variable (Pearson correlation coefficients).

8. R2 (R square) changes should be given in multiple regression analysis. It shows how much of the variation in the variance (indicates how much the established model determines the investigated relationship) of the relevant independent variable explains. In this study, the model determined 31.1% of the variance. The share of each independent variable in this percentage should be shown by giving the change in R2 (R square).

9. Table 2 and Table 3 need to be redone. Multiple regression analyzes should be performed separately for the intervention group and the control group, and the discussion section should be rewritten according to the results of this analysis (Univariate regression analysis does not need to be specified).

10. One of the dilemmas of the study is that BDZ (Benzodiazepine) and AD (Antidepressant) treatments used in baseline affect both baseline anxiety severity and depression severity. Therefore, the effect of the intervention applied in the study will be understood more accurately in a sample that does not use drugs. It is recommended to exclude the patients using drugs in both the 128-person intervention group and the 106-person control group, and to analyze the symptoms above for the new intervention and control groups consisting of patients who do not use drugs, and to interpret the similarities and differences between the results found with the 128 and 106-person samples in the discussion section.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Abdullah Burak UYGUR

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

20 May 2022

1) Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.

OUR RESPONSE: The manuscript now meets PLOS ONE's style requirements, including those for file naming.

2) We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. In your revised cover letter, please address the following prompts:

a. If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b. If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

c. We will update your Data Availability statement on your behalf to reflect the information you provide.

OUR RESPONSE: The data that support the findings of this study are available from the PARADIES study consortium but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request. Data requests may be directed at Jochen.Gensichen@med.uni-muenchen.de

3) We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

OUR RESPONSE: We now provide this data (univariate association of treatment modalities with BAI changes) in the new table 4.

4) One of the noted authors is a group or consortium (PARADIES study group). In addition to naming the author group, please list the individual authors and affiliations within this group in the acknowledgments section of your manuscript. Please also indicate clearly a lead author for this group along with a contact email address.

OUR RESPONSE: We provide the names of the PARADIES study group at the end of the manuscript in the acknowledgement

Reviewer #1:

1) This is a secondary regression analysis from an already-completed cluster randomized clinical trial. As such, it is an exploratory study, not powered to detect significant treatment differences, and it should be so-stated in the text. The comment "p<.05 is considered significant" means nothing with no adjustment for many, many hypothesis tests. I suggest a comment that "p-values should be interpreted as a guide, rather than as a determinant of significance". Alternatively, if there are confirmatory hypotheses, these should be stated clearly, power analysis given, and Bonferroni adjustments for multiple tests.

OUR RESPONSE: We agree with the reviewer. We have now added in a number of places that we conducted an exploratory analysis and now state in the discussion under limitations:

“This is a post-hoc analysis of data from the completed PARADIEs trial and as such it is an exploratory study, which is vulnerable to multiple testing bias. While our findings confirm the findings of the fully powered primary analysis, the identified associations between independent and dependent variables should be interpreted as hypothesis generating rather than hypothesis testing.”

2) I appreciate that they are concerned about analyzing this data set incorrectly, given they are ignoring the cluster effect, and they do indicate some homogeneity across clusters.

OUR RESPONSE: We agree, an analysis with a multi-level model would be the best method. However, due to the lower number of cases in our analysis (n=236) compared to the main study (n=419), only 1-2 patients remain in many centres. Therefore, an analysis with a multi-level model is no longer possible because the models do not converge. Due to the low intraclass correlation coefficient (0.017) in the main study and the associated low cluster effect, we decided to perform the analysis without multi-level models. We now further explain under limitations that our multi level models did not converge.

3) Finally, any regression course would require all students to do diagnostics for the model, including residual analyses and graphical inspection. I would expect no less in a scientific paper!

OUR RESPONSE: We performed diagnostics for all models including the residual plots. We have added the following sentence in the methods section "Residual plots (residuals vs. fitted values, Q-Q plot, scale-location plot, and residuals vs leverage values) were used to check the assumptions of the linear regression models". In addition, we added the residual plots for the multivariate models in the supporting file 2.

Reviewer #2:

1) A large part of the introduction describes the PARADIES study, including the independent variables investigated in the study and their relationship with panic disorder in the introduction section.

OUR RESPONSE: We have now added to the introduction the following paragraph to explain the rationale for the independent variables investigated: In particular, we investigated as influencing factors patient demographics (since age, sex and level of education may influence motivation and/or ability to engage with the intervention (19)), symptoms of other psychiatric co-morbidities and multimorbidity as well as the type and/or number of psychotropic and other drugs used to treat these (since all of these factors may influence the prognosis of panic disorders (20).

2) Out of 236 patients in the study, 128 were in the intervention group and 108 were in the control group. In the method section, it is explained in detail how the patients were assigned to the groups.

OUR RESPONSE: We have now included the following details describing the randomisation process in lines 78 ff: As described in more detail elsewhere (17), practices first recruited patient participants over a three months time period and were subsequently randomised to provide care as usual or deliver the PARADIES intervention to participating patients under their care (allocation ratio 1:1).

3) Comparison of both groups (intervention & control) in terms of sociodemographic characteristics (with t test and Chi-square tests for independent groups) and giving p values in Table 1 in the Results section.

OUR RESPONSE: We now provide the findings of t test and Chi-square tests and provide p values in Table 1 in the results section.

4) Half of the patients are receiving antidepressant treatment at the beginning, comparing the drug doses used by the patients using antidepressants in terms of equivalence (For example, 10 mg. Escitalopram = 20 mg. Citalopram = 20 mg. Fluoxetine = 50 mg. Sertalin = 20 mg. Paroxetine).

OUR RESPONSE: For antidepressant users at baseline, we have calculated the defined daily dose of antidepressants taken and compared them between groups. The data is presented in table 1. We have also included this variable in linear regression analyses and provide the findings in tables 3 to 5.

5) Of the 128 patients, 44 improved, 22 worsened. Of the 106 patients, 44 improved and 39 worsened. Rates are given on worsening only. Giving rates over recovery. (The rate of 74.8% on page 9 should be 64.8%.)

OUR RESPONSE: We think there is a confusion between patients and changes in BAI points, which we hope to clarify by the following rewording: In the intervention group, changes in BAI scores ranged from an improvement by 44 points to a worsening by 22 points, and in the control group from an improvement of 44 points to a worsening of 39 points. We additionally provide the numbers (%) of patients who improved (had changes to the better) and those who had no changes or had changes to the worse.

6) Adding the following table to the results section and making its statistical analysis.

The patients in both the intervention group and the control group were divided into groups according to the scores they got from the Beck Anxiety Inventory at the baseline and at the end of the 12-month follow-up, and comparing the two groups.

0-7 points (No anxiety)

8-15 points (Mild anxiety)

16-25 points (Moderate anxiety)

≥ 26 points (Severe anxiety)

OUR RESPONSE: We agree that the manuscript benefits from this additional analysis. We have added the findings in a new table 2 and summarised them in text as follows: Table 2 compares the proportions of intervention and control group patients by BAI severity category at baseline and at the end of the study period. While there were no significant differences between intervention and control groups at baseline, the groups differed significantly at 12 months follow up. Over the study period, the proportions of patients with severe symptoms decreased from 51.9% to 38.9% in the control group and more than halved (from 52.3% to 22.7%) in the intervention group, while the proportions of patients with no or minimal symptoms increased from 3.7% to 16.7% in the control group and from 1.6% to 28.1% in the intervention group.

7) Showing the correlation analysis with a table in the results section in terms of the relationship between the independent variables included in the regression analysis and the dependent variable (Pearson correlation coefficients).

OUR RESPONSE: We provide the findings of the correlation analysis in a supporting file 1. We suggest that the findings do not substantially add to the univariate relationships between independent and dependent variables reported in the manuscript and are therefore dispensable in the main manuscript.

8) R2 (R square) changes should be given in multiple regression analysis. It shows how much of the variation in the variance (indicates how much the established model determines the investigated relationship) of the relevant independent variable explains. In this study, the model determined 31.1% of the variance. The share of each independent variable in this percentage should be shown by giving the change in R2 (R square).

OUR RESPONSE: In order to find the multivariate model that best explains the change in BAI, we used a stepwise procedure based on the Akaike Information Criteria (AIC). The AIC is similar to the adjusted R squared measure and penalizes when more variables are included in the model. The result of this procedure is a simple model with the optimal number of variables. If present, multicollinearity is also removed by this procedure.

We provide R2 changes in the supporting file 2.

9) Table 2 and Table 3 need to be redone. Multiple regression analyzes should be performed separately for the intervention group and the control group, and the discussion section should be rewritten according to the results of this analysis (Univariate regression analysis does not need to be specified).

OUR RESPONSE: We have carefully considered this comment and would like to propose that we keep the analysis for the study population as a whole (where we adjust for allocation status), but additionally provide the findings of analysis stratified by intervention and control group. This approach maximises power for examining explanatory variables and avoids missing differential effects in intervention and control groups. We have applied this approach in the revised manuscript and made amendments to methods, results and discussion accordingly.

10) One of the dilemmas of the study is that BDZ (Benzodiazepine) and AD (Antidepressant) treatments used in baseline affect both baseline anxiety severity and depression severity. Therefore, the effect of the intervention applied in the study will be understood more accurately in a sample that does not use drugs. It is recommended to exclude the patients using drugs in both the 128-person intervention group and the 106-person control group, and to analyze the symptoms above for the new intervention and control groups consisting of patients who do not use drugs, and to interpret the similarities and differences between the results found with the 128 and 106-person samples in the discussion section.

OUR RESPONSE: We have further stratified the analysis by use of antidepressants or benzodiazepines and provide the findings in the manuscript and made amendments to methods, results and discussion accordingly. We have opted for this approach of adjusting for allocation status rather than a four way stratification by both intervention and control groups because this would have left few individuals in each stratum, which would have limited power even further.

Submitted filename: PONE_D_22_02155_Respons to reviewers.docx

Click here for additional data file.

7 Jul 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Burak Yulug

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: N/A

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: Dear Authors,

The article has some structural problems and this situation continues despite the revision. The independent variables researched in the introduction part are not included enough. Some parts that should be included in the method are included in the introduction.

It is a serious problem that the study did not include only panic disorder patients. In addition, the fact that both groups were treated with medication is a situation that may seriously affect the results of the study, and the daily dose of antidepressant parameter added to the table by the researchers does not relieve these reservations.

In addition, the allocation status, which is at the center of the research, is a binary variable. Although sometimes binary variables are treated as independent variables, classically, dependent and independent variables should be numerical in regression analysis. In this case, it is confusing about the variable in the center of the research.

Based on the above-mentioned reasons, it was deemed appropriate to reject the article.

Best regards.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

29 Jul 2022

PONE-D-22-02155R1

Patient characteristics and changes in anxiety symptoms in patients with panic disorder: Post-hoc analysis of the PARADIES cluster randomised trial

PLOS ONE

1) The article has some structural problems and this situation continues despite the revision. The independent variables researched in the introduction part are not included enough. Some parts that should be included in the method are included in the introduction.

Our response: This study is in an exploratory post hoc analyses of data from the PARDIES trial, and the findings of the primary analysis have been published previously. In our opinion, describing the PARADIES intervention and findings of the trial is essential background information for the reader to understand what the objectives of this paper are. We believe the aim of the analysis is clearly stated, namely “to examine variability in the course of anxiety symptom severity at 12 months follow up and to explore patient characteristics, which may explain that variability.”

With regards to describing the independent variables and its rationale in the introduction, we have added additional information in our previous revision upon the reviewer’s request and have made further specifications in the revised manuscript as follows: “In particular, we investigated as explanatory factors patient demographics (since age, sex and level of education may influence motivation and/or ability to engage with the intervention)[19], anxiety symptom severity and illness duration at baseline, co-morbid depression, multimorbidity and polypharmacy as well as the type and/or number of psychotropic drugs used to treat these (since all of these factors may influence the prognosis of panic disorders) [20].”

We suggest that providing this level of detail in the introduction should be sufficient for the reader to appreciate the rationale for our selection of independent variables. We provide further details on these variables and how they were defined in the methods section.

2a) It is a serious problem that the study did not include only panic disorder patients.

Our response: We would like to clarify that all included patients had panic disorder. We agree that in patients with comorbid depression and panic disorder, it is more difficult to isolate effects on either condition.

However, anxiety and depression share clinical symptoms and causes due to genetic pleiotropy and share psychological, social, and neurobiological risk mechanisms. As a result, comorbidity of anxiety and depression is the rule rather than an exception. For instance, of the individuals with a primary depression diagnosis in the Netherlands Study of Depression and Anxiety (NESDA), 67% had a current and 75% had a lifetime comorbid anxiety disorder diagnosis. Similarly, of those with a primary anxiety disorder diagnosis, 63% had a current and 81% a lifetime depressive disorder diagnosis. Thus, limiting the trial population to those with anxiety disorder but without symptoms of depression would have substantially compromised the applicability/external validity of the findings.

The PARADIES trial has collected data on both depression and anxiety symptoms and all multivariate analyses in this paper took baseline depression symptom severity into account. Our findings suggest that depression symptom severity may mark a group of patients with a less favourable prognosis (as the analysis of control group patients shows) who may nevertheless be at least partially responsive to the PARADIES intervention (as our analysis of intervention group patient shows).

• Choi, K. W., Kim, Y. K., & Jeon, H. J. (2020). Comorbid anxiety and depression: clinical and conceptual consideration and transdiagnostic treatment. Anxiety Disorders, 219-235.

• Demyttenaere, K., & Heirman, E. (2020). The blurred line between anxiety and depression: hesitations on comorbidity, thresholds and hierarchy. International Review of Psychiatry, 32(5-6), 455-465.

• Groen, R.N., Ryan, O., Wigman, J.T., Riese, H., Penninx, B.W., Giltay, E.J., ... & Hartman, C.A. (2020). Comorbidity between depression and anxiety: assessing the role of bridge mental states in dynamic psychological networks. BMC medicine, 18(1), 1-17.

• Hirschfeld R.M.A. The comorbidity of major depression and anxiety disorders: recognition and management in primary care. Prim Care Companion J Clin Psychiatry. 2001;3(6):244–54.

• Lamers F., van Oppen P., Comijs H.C., Smit J.H., Spinhoven P., Van Balkom A.J.L.M., et al. Comorbidity patterns of anxiety and depressive disorders in a large cohort study: the Netherlands Study of Depression and Anxiety (NESD A). J Clin Psychiatry. 2011;72(3):341–8.

• Ter Meulen, W. G., Draisma, S., van Hemert, A. M., Schoevers, R. A., Kupka, R. W., Beekman, A. T., & Penninx, B. W. (2021). Depressive and anxiety disorders in concert–A synthesis of findings on comorbidity in the NESDA study. Journal of affective disorders, 284, 85-97

• Penninx, B., Pine, D., Holmes, Reif, A. Anxiety disorders, Lancet. 2021 March 06; 397(10277): 914–927.

2b) In addition, the fact that both groups were treated with medication is a situation that may seriously affect the results of the study, and the daily dose of antidepressant parameter added to the table by the researchers does not relieve these reservations.

Our response: We agree that including patients with and without anti-anxiety medication introduces heterogeneity in the study population. However, we adjusted for medication use (including dose as requested by the reviewer in his/her last review) at baseline in all analyses. In order to minimise any residual confounding, the reviewer recommended in his/her last review “to exclude the patients using drugs in both the 128-person intervention group and the 106-person control group, and to analyse the symptoms above for the new intervention and control groups consisting of patients who do not use drugs, and to interpret the similarities and differences between the results found with the 128 and 106-person samples in the discussion section.” In our previous revision, we therefore further stratified the analysis by use of antidepressants or benzodiazepines and provide and discuss the findings in the manuscript.

In the current version of the manuscript, we have added the following sentence to the strengths and limitations section in order to address further concerns regarding the medication: “Finally, although the intervention focussed on non-pharmacological treatment options, we cannot exclude that differential use of anti-anxiety medication in intervention and control groups during follow up may have contributed to the observed effects.”

• Gensichen, J., Hiller, T. S., Breitbart, J., Brettschneider, C., Teismann, T., Schumacher, U., Lukaschek, K., Schelle, M., Schneider, N., Sommer, M., Wensing, M., König, H. H., Margraf, J., & Jena-PARADISE Study Group (2019). Panic Disorder in Primary Care. Deutsches Arzteblatt international, 116(10), 159–166. https://doi.org/10.3238/arztebl.2019.0159

• Gensichen, J., Hiller, T. S., Breitbart, J., Teismann, T., Brettschneider, C., Schumacher, U., Piwtorak, A., König, H. H., Hoyer, H., Schneider, N., Schelle, M., Blank, W., Thiel, P., Wensing, M., & Margraf, J. (2014). Evaluation of a practice team-supported exposure training for patients with panic disorder with or without agoraphobia in primary care - study protocol of a cluster randomised controlled superiority trial. Trials, 15, 112. https://doi.org/10.1186/1745-6215-15-112

3) In addition, the allocation status, which is at the center of the research, is a binary variable. Although sometimes binary variables are treated as independent variables, classically, dependent and independent variables should be numerical in regression analysis. In this case, it is confusing about the variable in the center of the research.

Our response: We agree with the reviewer that the dependent variable in the linear regression models used must be a metric variable. This requirement is met in all models in the manuscript. However, this requirement does not apply to independent variables. Binary variables can also be considered in regression models as independent variables in addition to metric variables. For example, in most regression models the binary variable "sex" is included as an independent variable. Hence, the inclusion of the variable "allocation status" as an independent variable in the linear regression models is correct from our point of view. Furthermore, the aim of the publication is "to examine variability in the course of anxiety symptom severity at 12 months follow up and to explore patient characteristics, which may explain that variability". Therefore, the objective is not only to show a possible correlation of the change in anxiety symptoms with the variable "allocation status", but also a possible correlation with other "patient characteristics”.

Submitted filename: PONE_R2_answers to reviewer.docx

Click here for additional data file.

20 Sep 2022

Patient characteristics and changes in anxiety symptoms in patients with panic disorder: Post-hoc analysis of the PARADIES cluster randomised trial

PONE-D-22-02155R2

Dear Dr. Lucaschek,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Burak Yulug

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

I have carefully checked your revised paper and agree that major improvements required for the publication has been done.

Reviewers' comments:

21 Sep 2022

PONE-D-22-02155R2

Patient characteristics and changes in anxiety symptoms in patients with panic disorder: Post-hoc analysis of the PARADIES cluster randomised trial

Dear Dr. Lukaschek:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Burak Yulug

Academic Editor

PLOS ONE

Table 1

Descriptive statistics of patient characteristics, GP (practice) characteristics and intervention delivery modalities.

	Total (n = 236)	Intervention Group (n = 128)	Control Group (n = 108)	Comparison intervention vs control p-value
Patient demographics
Age [mean (SD)]	45.3 (13.4)	46.0 (13.0)	44.5 (13.9)	0.3898
Female	178 (75.4%)	93 (72.7%)	85 (78.7%)	0.2824
Education time: years [mean (SD)]	11.1 (3.0)	11.0 (3.0)	11.2 (3.0)	0.6287
Clinical parameters at baseline
Illness duration in months [median (Q1-Q3)]	9 (3–19)	8 (3–19)	9 (3–19.5)	0.7123
Depression scale PHQ-9 T0 [mean (SD)]	11.2 (5.6)	11.4 (5.7)	10.8 (5.5)	0.4163
Multimorbidity (≥3 chronic conditions) [n]	123 (52.1%)	65 (50.8%)	58 (53.7%)	0.6543
BAI [mean (SD)]	27.7 (12.2)	27.1 (11.8)	28.5 (12.6)	0.4068
Patient assessment of chronic illness care (PACIC) [mean (SD)]	6.4 (2.6)	6.2 (2.4)	6.6 (2.7)	0.2896
Medication at baseline
Polypharmacy (≥5 medicines)	48 (20.3%)	25 (19.5%)	23 (21.3%)	0.7372
Psychotropic polypharmacy (≥2 psychotropic medicines)	49 (20.8%)	26 (20.3%)	23 (21.3%)	0.8527
Any benzodiazepine	15 (6.4%)	11 (8.6%)	4 (3.7%)	0.1250
Any antidepressant	108 (45.8%)	52 (48.2%)	56 (51.9%)	0.0846
Antidepressant DDD (median, IQR)	1.00 (0.5, 1.5)	1.00 (0.7, 2.0)	1.00 (0.5, 1.0)	0.1218
Intervention delivery parameters
Appointment 1 performed		123 (96.1%)	-
Appointment 2 performed		121 (94.5%)	-
Appointment 3 performed		112 (87.5%)	-
Appointment 4 performed		96 (75.0%)	-
Telephone contacts [mean (SD)]		9.1 (2.1)
Additional contacts
0		104 (81.3%)	-
1		16 (12.5%)	-
2		3 (2.3%)	-

* significant difference between intervention and control group (p<0.05).

BAI: Beck-Anxiety-Inventory; PACIC: Patient Assessment of Chronic Illness Care (categories: 1 (0%) to 11 (100%)); PHQ: Patient Health Questionnaire; DDD: Defined daily dose.

Table 2

Comparison of intervention and control group patients by BAI severity category at baseline and at the end of the study period.

BAI groups (Score range in points)	Baseline			End of study
	Intervention group	Control Group	p-value	Intervention group	Control Group	p-value
No or minimal (0 to 7)	2 (1.6%)	4 (3.7%)	0.2861	36 (28.1%)	18 (16.7%)	0.0099
Mild (8 to 15)	25 (19.5%)	13 (12.0%)		32 (25.0%)	17 (15.7%)
Moderate (16 to 25)	34 (26.6%)	35 (32.4%)		31 (24.2%)	31 (28.7%)
Severe (≥26 points)	67 (52.3%)	56 (51.9%)		29 (22.7%)	42 (38.9%)

Table 3

Findings of univariate and multivariate linear regression models for all study participants (n = 236).

	Univariate regression coefficient (95% CI)	Multivariate regression coefficient (optimal model after the stepwise procedure; 95% CI)#
Allocated to intervention group (vs treatment as usual)	4.2 (0.6 to 7.7) *	5.7 (2.6 to 8.8)***
Demographics
Age [years]	-0.1 (-0.2 to 0.1)	-0.1 (-0.2 to 0.02)
Female sex (vs male)	4.1 (0.1 to 8.2)*	3.3 (-0.2 to 6.7).
Education time [years]	0.08 (-0.5 to 0.7)
Clinical parameters at baseline
Anxiety symptom severity (BAI T0)	0.5 (0.4 to 0.7)***	0.8 (0.6 to 1.0) ***
Illness Duration [months]	0.01 (-0.1 to 0.1)	0.2 (0.02–0.4)*
Depression scale (PHQ-9)	0.2 (-0.1 to 0.5)	-0.9 (-1.4 to -0.5)***
Multimorbidity (vs not multimorbid)	-1.0 (-4.5 to 2.6)
Patient assessment of chronic illness care (PACIC)	-0.3 (-1.0.1 to 0.4)
Medication use at baseline
Benzodiazepine (yes vs no)	3.7 (-3. to 10.9)	10.4 (-1.3 to 22.0).
Antidepressant (yes vs no)	-2.7 (-6.2 to 0.9)
Antidepressant DDD	0.4 (-2.0 to 2.9)
Polypharmacy (≥5 medicines vs <5)	-1.9 (-6.3 to 2.5)	-3.2 (-7.2 to 0.8)
Psychotropic polypharmacy (≥2 psychotropic medicines vs <2)	1.9 (-2.5 to 6.2)
Interactions
Allocated to intervention group		-0.3 (-0.6 to 0.005).
* Anxiety symptom severity (BAI T0)
Allocated to intervention group		0.8 (0.2 to 1.5)*
* Depression scale (PHQ-9)
Allocated to intervention group		-0.2 (-0.5 to 0.001).
* Illness Duration [months]
R2		0.3105

.<0.1;

*<0.05;

**<0.01;

***<0.001;

# independent variables multivariate model: group, age, sex, education time, baseline anxiety symptom severity, illness duration, depression scale, multimorbidity, patient assessment of chronic illness care, benzodiazepine, antidepressant, polypharmacy, psychotropic polypharmacy; DDD: Defined daily dose.

Table 4

Findings of regression analyses stratified by intervention and control status.

	Control group		Intervention group
	Univariate regression coefficient (95% CI)	Multivariate regression coefficient (optimal model after stepwise procedure; 95% CI)#	Univariate regression coefficient (95% CI)	Multivariate regression coefficient (optimal model after stepwise procedure; 95% CI)#
Demographics
Age [years]	-0.02 (-0.2 to 0.2)		-0.1 (-0.3 to 0.0)	-0.2 (-0.3 to -0.02)*
Female sex (vs male)	8.1 (1.1 to 15.1)*	4.5 (-1.3 to 10.3)	1.9 (-2.8 to 6.6)
Education time [years]	0.03 (-1.0 to 1.0)		0.2 (-0.5 to 0.9)
Clinical parameters at baseline
Anxiety symptom severity (BAI T0)	0.6 (0.4 to 0.8)***	0.8 (0.6 to 1.0) ***	0.4 (0.3 to 0.6)***	0.5 (0.3 to 0.6) ***
Illness Duration [months]	0.01 (-0.2 to 0.2)	0.2 (-0.03 to 0.4).	0.0 (-0.1 to 0.2)
Depression scale (PHQ-9)	-0.1 (-0.6 to 0.5)	-0.9 (-1.4 to -0.4)***	0.4 (0.1 to 0.8)*
Multimorbidity (vs not multimorbid)	-0.5 (-6.4 to 5.4)		-1.1 (-5.3 to 3.1)
Patient assessment of chronic illness care (PACIC)	-0.4 (-1.5 to 0.7)		-0.1 (-0.9 to 0.8)
Medication use at baseline
Benzodiazepine (yes vs no)	13.1 (-2.3 to 28.5)	9.6 (-2.8 to 21.9)	-1.0 (-8.5 to 6.5)
Antidepressant DDD	-0.1 (-4.8 to 4.6)		0.7 (-1.9 to 3.4)
Antidepressant (yes vs no)	-4.2 (-10.0 to 1.6)		-0.5 (-4.8 to 3.8)
Polypharmacy (≥5 medicines vs <5)	-3.9 (-11.1 to 3.2)	-6.0 (-11.8 to -0.2)*	0.2 (-5.1 to 5.5)
Psychotropic polypharmacy (≥2 psychotropic medicines vs <2)	3.6 (-3.6 to 10.7)		0.5 (-4.7 to 5.8)
Delivery modalities
Appointment 4 performed	Not applicable		4.7 (-0.4 to 9.9)
Telephone contacts	Not applicable		1.0 (-0.04 to 1.9)
Additional contacts	Not applicable
0			Reference
1			-3.2 (-0.5 to 3.1)
2			4.6 (-9.2 to 18.3)
R2		0.3759		0.2117

<0.1;

*<0.05;

**<0.01;

***<0.001;

# independent variables multivariate model: age, sex, education time, baseline anxiety symptom severity, illness duration, depression scale, multimorbidity, patient assessment of chronic illness care, benzodiazepine, antidepressant, polypharmacy, psychotropic polypharmacy; DDD: Defined daily dose.

Table 5

Findings of regression analyses stratified by use vs non-use of antidepressants and/or benzodiazepines.

	Non-users (n = 120)		Users (n = 116)
	of antidepressants or benzodiazepines		of antidepressants and/or benzodiazepines
	Univariate regression coefficient (95% CI)	Multivariate regression coefficient (optimal model after the stepwise procedure; 95% CI)#	Univariate regression coefficient (95% CI)	Multivariate regression coefficient (optimal model after the stepwise procedure; 95% CI)#
Allocated to intervention group (vs treatment as usual)	3.0 (-1.8 to 7.7)	4.8 (0.8 to 8.8)*	5.0 (-0.2 to 10.3)	5.8 (1.4 to 10.1)**
Demographics
Age [years]	-0.1 (-0.2 to 0.1)		-0.1 (-0.3 to 0.1).
Female sex (vs male)	-0.5 (-6.0 to 5.1)		8.4 (2.5 to 14.3)**	5.5 (0.4 to 10.6)*
Education time [years]	0.2 (-0.7 to 1.1)	0.8 (0.0 to 1.5)*	-0.02 (-0.8 to 0.8)
Clinical parameters at baseline
Anxiety symptom severity (BAI T0)	0.6 (0.4 to 0.7)***	0.6 (0.4 to 0.8) ***	0.5 (0.3 to 0.7)***	0.7 (0.5 to 0.9) ***
Illness Duration [months]	0.01 (-0.2 to 0.2)		0.01 (-0.2 to 0.2)
Depression scale (PHQ-9)	0.7 (0.3 to 1.1)**		-0.1 (-0.6 to 0.3)	-0.6 (-1.1 to -0.2)**
Antidepressant DDD			4.2 (0.2 to 8.1)*
Multimorbidity (vs not multimorbid)	-1.0 (-5.7 to 3.7)		-0.8 (-6.2 to 4.5)
Patient assessment of chronic illness care (PACIC)	-0.7 (-1.6 to 0.3)		0.1 (-0.9 to 1.1)
Medication use at baseline
Polypharmacy (≥5 medicines vs <5)	3.5 (-2.9 to 10.0).	4.4 (-1.0 to 9.9).	-5.4 (-11.5 to 0.7).	-7.1 (-12.1 to -2.1)**
Psychotropic polypharmacy (≥2 psychotropic medicines vs <2)	5.1 (-5.7 to 15.8)		3.1 (-2.4 to 8.6)
R2		0.3039		0.3429

.<0.1;

*<0.05;

**<0.01;

***<0.001;

# independent variables multivariate model: group, age, sex, education time, baseline anxiety symptom severity, illness duration, depression scale, multimorbidity, patient assessment of chronic illness care, benzodiazepine, antidepressant, polypharmacy, psychotropic polypharmacy; DDD: Defined daily dose.