French adaptation and validation of the panic disorder severity scale – self-assessment | BMC Psychiatry



Data were drawn from baseline (T0) and post-treatment (T1) assessment of participants in a pragmatic multicenter clinical trial of transdiagnostic group cognitive-behavioral therapy (CBT) for anxiety disorders conducted at the Quebec, Canada. The study was conducted in three integrated health and social services centers, including Quebec City and adjacent regions (an urban and rural population of 737,000); Estrie (i.e. an urban and rural population of 474,000 inhabitants); and the city of Laval (that is, a predominantly urban population of 435,000). The complete methodology and main results of the trial have been described previously [12]. The inclusion criteria for the trial included the following: (1) being between the ages of 18 and 65, (2) being fluent in spoken and written French; (3) DSM-5 diagnostic criteria met for at least one of the following major anxiety disorders: panic disorder, agoraphobia, generalized anxiety disorder and/or social anxiety disorder with a Clinical Severity Index (CSR) ≥ 4 out of Anxiety and Related Disorders Interview Schedule for the DSM-5 (ADIS-5) [13]. Exclusion criteria for the trial included: (1) active suicidal intentions, psychosis, bipolar disorder, and active substance and addictive disorder in the past 12 months; (2) marked cognitive impairment; (3) consultation with a psychiatrist within the past 12 months.

A three-step process was used for patient recruitment. First, we advertised the study to the general population and community care settings through regional, geolocated newspapers Facebook and Google AdWords advertising and bulletin boards. Self-referred individuals completed a brief online screening survey that included Generalized Anxiety Disorder-7 (GAD-7; cutoff value ≥ 8 for anxiety disorders) [14], and other measures. Second, clinical evaluators conducted a brief telephone screening interview with potential participants. Third, face-to-face clinical assessment interviews were conducted with potential candidates to assess their eligibility based on the ADIS-5 (T0) assessment. All participants taking part in the third assessment stage (T0) of the clinical trial and returning the baseline questionnaires, with a main anxiety disorder, whether or not they were enrolled in the trial, were potentially eligible for the secondary validation study. The sample for Objectives 1 and 2 consisted of all T0 participants with a complete dataset. The sample for Objective 3 consisted of participants enrolled in the clinical trial, meeting DSM-5 criteria for panic disorder and with a complete pre-post data set.

The study was approved by the main ethics committee (Research Ethics Committee of the Integrated University Health and Social Services Center of Estrie – Center hospitalier universitaire de Sherbrooke, #MP-22–2016-570 ) and the boards of directors of other institutions (Research Ethics Committee of the University Institute in Mental Health of Quebec, #2017-166; Scientific and Research Ethics Committee – CISSS de Laval, #2016 -2017/C54). All patients gave their written informed consent.


The experimental condition was a transdiagnostic group cognitive behavioral therapy program that included education and self-monitoring, specific cognitive restructuring, graded exposure and response prevention, and generalized cognitive restructuring. [15]. The intervention was delivered to groups of 8 to 10 patients over a 12-week period by two psychologists or psychotherapists. The usual treatment control (TAU) group continued to receive unrestricted usual care throughout the trial.


The initial in-person assessment interview (T0; pre-random assignment) included the consent form, ADIS-5 administration, and a questionnaire on socio-demographics, health care costs, medications, and job performance by a trained evaluator at the doctoral level. Patients also received a participant’s booklet with the self-assessment questionnaires, and the same booklet was used at T0 and T1. Only the measurements used for the secondary validation study are described here.

Diagnostic interview

The ADIS-5 was used to assess the DSM-5 diagnostic criteria for anxiety disorders and comorbid mental disorders [13]. Significant clinical improvement after treatment for panic disorder was defined as a Clinical Severity Score (CSR)

PDSS-SR Translation

Carry out an intercultural translation of the PDSS-SR [2] in French for Canada, we followed previously published guidelines [16, 17]. Two direct translations of the original PDSS-SR into English were made, one by a professional translator and the other by a research assistant. The reason for having two direct translations is that it allows the two translations to be compared, which helps identify translation errors and varying interpretations of the wording of the items in the source language. [17]. The two translated questionnaires were then reconciled into a single French-Canadian version through consensual discussion. The reconciliation process allowed any differences between the direct translations to be compared and the preferred wording was integrated into the final version. Minor differences were observed in the translation but were easily resolved by an expert committee consisting of the research team and collaborators involved in the translation process. In addition, this process minimizes the risk that the final French-Canadian version will be biased in favor of a translator’s writing style. [17]. Following this, two back-translations were carried out by English speakers to ensure that all questions retain their original meaning when returned in the source language. These back-translations were compared to the original English version of the PDSS-SR, and minor revisions to the French wording were made to ensure conceptual equivalence of all items. The instrument’s instructions, questions, and response options were reviewed to ensure that synonyms and phrases had similar meanings and were culturally equivalent, to resolve remaining grammatical difficulties, and to affirm the adequacy of the level of reading.

Other self-assessment questionnaires

The patient booklet included instruments with good psychometric properties and validated in French. Measures of anxiety symptoms included Beck Anxiety Inventory (BAI) [18, 19]the Agoraphobia Mobility Inventory (MIA) [20, 21]the Inventory of social phobias (TWIRN) [22, 23] and the Penn State Concern Questionnaire (PSWQ) [24, 25]. Depressive symptoms were assessed with the Patient Health Questionnaire (PHQ-9) [26, 27]. The Sheehan Disability Scale (SDS) has been used as a measure of disability in work, social life and family life [28]. The Insomnia Severity Index (ISI) was also administered to examine the severity of insomnia and associated difficulties [29, 30]. Only self-reported questionnaires used for this validation study were listed, but others were used in the main study.


Data analytics

The factor structure of the PDSS-SR was examined with exploratory factor analysis (EFA) and with confirmatory factor analysis (CFA). For the EFA, the Promax rotation was used, factors with eigenvalues ​​> 1 were retained and a loading factor greater than 0.40 was retained to explain the target factor. A first analysis was carried out without imposing a factor and a second by imposing two factors according to previous research. Then CFA with a factor was used to examine the adequacy of the structure obtained with EFA. The diagonal weighted least squares estimation method was used as it was more appropriate for ordinal data [31]. Model fit was assessed with standardized residual mean square root (SRMR), adjusted goodness of fit index (AGFI), and normed fit index (NFI). SRMR less than 0.08, AGFI greater than 0.95 and NFI greater than 0.95 indicate good model fit [32].

The internal consistency of the French-Canadian PDSS-SR was assessed using Cronbach’s α. Convergent and discriminant validities were examined with Spearman correlations as several variables were not normally distributed [33]. Correlations were assessed for the total score and items of the PDSS-SR with ADIS-5 CSR for anxiety and depressive disorders, as well as different self-report questionnaires. Based on the concept of panic disorder and previous studies [5, 6, 8, 10], this examined convergent validity with another measure of panic disorder (ADIS-5) and related agoraphobia (ADIS-5, MIA), anxiety symptoms (BAI), as well as disability (SDS) . For discriminant validity, lower correlations were expected with other questionnaires that measure different constructs (ADIS-5, ISI, PHQ-9, SPIN, PSWQ), although comorbidity should be considered in interpreting the constructs. results. Receiver operating characteristic (ROC) curves for the PDSS-SR were calculated to assess the score’s performance in distinguishing participants with and without panic disorder with the PDSS-SR full scale. We also conducted an exploratory analysis of the 2-item version recommended by Forsell et al. [11]. Youden’s index was used to define the best thresholds.

To examine the sensitivity to change, a linear mixed regression model was carried out with the method of maximum likelihood and the correlation matrix of the variance components on the PDSS-SR taking into account the intervention group, time and interaction (change between groups). Within-group and within-site variability of therapy was controlled with a random effect on participants nested in the therapy groups nested across the three sites. To control for between-group variability, a second random effect on nested treatment groups across the three sites was added. Since the residuals were normally distributed, bootstrap resampling was not necessary. Responsiveness to change was analyzed for two scenarios of improvement at baseline (T0) and post-treatment (T1) periods in participants with panic disorder enrolled in the clinical trial, including comparison of experimental and active control and comparison of participants with clinically significant symptoms. improvement for panic disorder based on ADIS-5 CSR score for panic disorder. CFA analyzes were performed with SAS version 9, other analyzes were performed with IBM SPSS Statistics version 25. GraphPad Prism version 7 was used to create the figures.


About Author

Comments are closed.