This study used a prospective staggered baseline case-series design. Six Arabic-speaking patients with a primary diagnosis of PD were randomly and equally allocated to an immediate-start group, which began treatment at the end of recruitment, and a delayed-start group, which began treatment 1 month later. During that month, the delayed-start group completed the same self-report measures as the immediate-start group, with 1 to 2 assessments recorded per patient, providing a period of pretreatment observation. All 6 patients then received the same 16-week guided ICBT program using an Arabic-translated protocol delivered by a single trained therapist. Self-report outcomes were collected pretreatment, posttreatment, and 3 months later (follow-up), as well as weekly during the program; clinician-administered assessments were conducted at pretreatment, posttreatment, and follow-up. Individual semistructured interviews were conducted posttreatment to explore patients’ experiences of the program. Given the small sample size and study design, all quantitative outcomes were analyzed descriptively at both the individual and group levels. Patients were recruited in December 2022. The immediate-start group began treatment in January 2023, and the delayed-start group began treatment in February 2023; patients completed treatment between June and September 2023, and data collection, including the follow-up assessments, was completed in November 2023.

We recruited 6 Arabic-speaking patients with PD using social media (eg, Facebook, Instagram, and WhatsApp) and public advertisements. Study inclusion criteria included (1) a primary diagnosis of PD with or without agoraphobia; (2) age 18 years or older; (3) PD duration of at least 2 months; (4) no active suicide risk; (5) absence of current or a history of psychosis or uncontrolled bipolar I disorder; (6) no current psychotherapy for treating PD or history of a complete course of CBT for PD; (7) a stable medication dosage for 3 months; and (8) willingness and ability to use the internet and to read and write in Arabic. All communications with participants were in Arabic.

Interested participants registered on the study website and completed initial symptom screening questionnaires. Eligible candidates were then screened by telephone, and those who passed were invited for a pretreatment clinical interview via Zoom. During this interview, patients were assessed using the Diagnostic Interview for Anxiety, Mood, and Obsessive-Compulsive and Related Neuropsychiatric Disorders (DIAMOND) [19] and the Panic Disorder Severity Scale–Interviewer Edition (PDSS-IE) [20] to confirm PD as the primary diagnosis and to assess PD symptom severity. All patients completed self-report measures of panic symptoms (Panic Disorder Severity Scale–Self-Report [PDSS-SR]) [21], agoraphobic avoidance (Mobility Inventory [MI]) [22], and AS (Anxiety Sensitivity Index-3 [ASI-3]) [23] pretreatment, posttreatment, and 3 months later, as well as weekly during the program. Posttreatment assessments mirrored the pretreatment assessments and were conducted via Zoom. The 3-month follow-up assessments were conducted by telephone, during which the PDSS-IE was administered; patients self-completed the PDSS-SR, ASI-3, and MI online.

This study was approved by the Social Sciences Ethics Committee of the Hebrew University of Jerusalem (approval number 2022‐07202). All participants gave their informed consent before enrollment by completing an online consent form when registering on the study website and provided additional consent during the pretreatment Zoom interview. All data were deidentified prior to analysis: participants were assigned unique numeric identifiers, and identifying information (eg, names and contact details) was stored separately from the study data on a password-protected server accessible only to the research team. Participants did not receive financial or other compensation for their participation.

Patients were treated via an ICBT version for PD using the same treatment protocol from the study by Strauss et al [13], which was translated into Arabic by 2 researchers from the research team who were native Arabic speakers. One translated the materials, and the second reviewed the translation for accuracy and clarity. Adaptations were made to fit the Arab cultural context. For instance, exposure scenarios involving dating or drinking alcoholic beverages with friends were replaced with situations that are common and comparably anxiety-provoking in the local Arab context, such as visiting family and spending time with friends. The language was also adjusted from formal written Arabic to a clearer, more readable style because formal Arabic can be more difficult to follow in everyday reading. The treatment was administered over 16 weeks, taking individual and clinical needs into consideration. The treatment program consisted of 6 modules, including reading texts, worksheets, videos, and assignments. The modules contained (1) an introduction to the treatment program, (2) psychoeducation, (3) cognitive restructuring, (4) interoceptive exposure, (5) in vivo and interoceptive exposure, and (6) a summary and relapse prevention.

The treatment was administered by a master’s-level psychology student trained in CBT for PD. The therapist received weekly supervision from a senior clinical psychologist to ensure treatment fidelity and address any clinical challenges. The therapist was a native speaker of Arabic and familiar with the cultural context of the participants. Therapist-patient contact occurred in 2 ways. First, after each module, the therapist reviewed the patient’s worksheets and provided written feedback through the platform’s secure messaging system, after which the next module was activated. Second, the therapist initiated contact when a patient had not logged in or progressed for some time, had not completed a weekly questionnaire, or had raised a question. This was usually by telephone call, followed by a platform message if the call was not answered, and occasionally by a brief WhatsApp (Meta Platforms, Inc) or email message. Patients could also message the therapist through the platform at any time. Completers were defined as patients who finished at least the fifth module of the treatment program.

All measures were either previously validated in Arabic or translated for this study by native Arabic speakers using translation and back-translation procedures to ensure semantic accuracy. Minor linguistic adjustments were made to enhance cultural and contextual appropriateness for Arabic-speaking patients.

The PDSS-SR [21] is a 7-item self-report measure assessing PD symptom severity, including panic attack frequency, distress, fears and avoidance related to agoraphobia and bodily sensations, and impairment in work and social functioning. Each item is rated on a 5-point Likert scale (0‐4), yielding a total score ranging from 0 to 28, with higher scores indicating greater PD severity. A score of 7 or higher is commonly used as a threshold for clinically significant PD symptoms. The PDSS-SR has demonstrated good internal consistency (Cronbach α=0.92) [21].

PDSS-IE [20] is a clinician-administered measure consisting of 7 items assessing PD symptom severity over the past week across the same domains as the PDSS-SR. Items are rated on a 5-point Likert scale (0‐4), yielding total scores ranging from 0 to 28, with higher scores indicating greater PD severity. A score of 7 or higher is commonly used as a threshold for clinically significant PD symptoms. The PDSS-IE has demonstrated good internal consistency (Cronbach α=0.88) [20].

The ASI-3 [23] is an 18-item self-report measure of AS, defined as the fear of anxiety-related sensations. It comprises three 6-item subscales assessing physical, cognitive, and social concerns. Items are rated on a 5-point Likert scale (0=very little to 4=very much), yielding a total score ranging from 0 to 72, with higher scores indicating greater AS. The ASI-3 has demonstrated good internal consistency for both the total scale and its subscales (Cronbach α range=0.73-0.91) [23]. Scores of 23 or higher are commonly used as an indicator of elevated AS.

The MI [22] is a 27-item self-report measure assessing agoraphobic avoidance across a range of situations. Each situation is rated twice on a 5-point Likert scale (1=“never avoid” to 5=“always avoid”): once for avoidance when alone and once for avoidance when accompanied, yielding 2 mean scores. Higher scores indicate greater avoidance. The MI has demonstrated excellent internal consistency (Cronbach α>0.90) [22]. A mean MI score of 1.5 or higher is often considered indicative of clinically significant avoidance.

The DIAMOND [19] is a semistructured clinician-administered interview designed to assess DSM-5 (Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition) psychiatric disorders, with particular emphasis on anxiety, mood, obsessive-compulsive, and related conditions. It evaluates primary and comorbid diagnoses using standardized diagnostic criteria and has demonstrated good reliability. In the current study, the DIAMOND was used during pretreatment assessment to confirm PD as the primary diagnosis and assess comorbidities.

At the end of treatment, all patients were invited to participate in an individual semistructured interview about their experience with the program. Four patients agreed to be interviewed: 3 completers were interviewed shortly after the posttreatment assessment, and 1 noncompleter was interviewed after withdrawing. Of the remaining 2 patients, 1 stopped responding to the therapist following a recurrence of symptoms posttreatment, and 1 asked for their interview to be postponed and then did not attend 2 scheduled appointments. The interviews were conducted in Arabic via Zoom by the treating therapist (LK), a graduate student pursuing a master’s in clinical psychology and trained in CBT for PD, who was known to all patients and served as their therapist throughout the program. The interviews lasted approximately 60 minutes and were not audio-recorded; the therapist documented each interview in written notes taken during and immediately after it. The interview covered overall satisfaction with the program and whether it had met the patient’s expectations; the perceived helpfulness and difficulty of the individual treatment modules; the experience of the therapeutic relationship and the platform’s messaging system; suggestions for improving the program; and, for patients who discontinued the program, what might have enabled them to complete it. Nondirective follow-up prompts were used to clarify and expand on patients’ responses.

Quantitative outcomes and interview material were handled separately as described below. Given the small sample size and case-series design, all quantitative data were analyzed descriptively without inferential statistical testing. We computed individual patient scores and group-level descriptive statistics (means and ranges) for the PDSS-SR, ASI-3, and MI at pretreatment, posttreatment, and 3-month follow-up, as well as for weekly assessments during treatment. Within-group pretreatment to posttreatment effect sizes (Cohen d) were computed for the 4 completers across all 3 measures. Visual inspection of weekly trajectories was used to characterize patterns of symptom change across treatment modules. For the 2 patients who dropped out, we examined their pretreatment scores and available treatment data to explore potential factors contributing to noncompletion.

The interview material is reported descriptively as the patient-perspective component of the case series. No formal qualitative analysis was undertaken. The therapist reviewed their notes from each interview and summarized what each patient said about her experience of the program. These summaries are presented separately for each patient, following the order of the interview topics, without grouping or comparing material across patients. No coding scheme, category system, or qualitative analytic framework was developed or applied; no second researcher analyzed the material independently; and no claim of thematic saturation or of representativeness beyond these 4 patients is made. The material recorded in the therapist’s notes as the patient’s own words is presented in quotation marks; other material is the therapist’s summary of what the patient conveyed and is presented without quotation marks. The quotations were translated from Arabic into English by a bilingual member of the research team.

Six Arabic-speaking women aged 21 to 50 (mean 34.7, SD 9.8) years were enrolled in the study. All met the criteria for PD, and some had comorbid conditions such as agoraphobia, depression, or generalized anxiety disorder.

The results are presented in 4 parts: treatment engagement and adherence, group-level treatment outcomes, individual patterns of change during treatment, and patients’ perspectives. Individual scores for all 6 patients are presented in Table 1, weekly PDSS-SR scores in Figure 1, and weekly ASI-3 subscale scores in Multimedia Appendix 1.

Table 1 presents engagement indicators and outcome scores for all 6 patients.

The program was completed by 4 of the 6 (67%; patients 1-4) patients. Completers finished within 17 to 22 weeks, logged in to the treatment website 32 to 42 times, and completed most of the assigned worksheets. The 2 noncompleters showed different levels of engagement. Patient 5 dropped out after 10 weeks, having logged in 23 times and completed 1 module; she cited time management difficulties as the main reason. Patient 6 dropped out after 19 weeks, having logged in 34 times and completed 4 of 6 modules, with a 92% worksheet completion rate. During periods of active engagement, she reported reduced avoidance (eg, volunteering, driving in the city, and going out when accompanied), but her engagement fluctuated before she discontinued. Therapist-initiated telephone calls ranged from 1 to 5 per patient, and patients did not always respond to therapist messages.

The therapist’s records indicate that the engagement of patient 4, a completer, was intermittent throughout treatment in the context of child care and household responsibilities, academic assignments, and a period during which her husband and daughter were hospitalized. For patient 6, the records describe family conflict, an unsafe environment, and responsibility for her child as the main obstacles to engagement, alongside frequent and severe panic symptoms.

Among the 4 completers, scores on all 3 measures decreased from pretreatment to posttreatment. Pretreatment to posttreatment within-group effect sizes, computed for the 4 completers only, were very large: PDSS-SR scores decreased from a mean of 13.75 (SD 3.40) to 1.75 (SD 3.50; Cohen d=3.48), ASI-3 scores from 32.75 (SD 18.08) to 1.75 (SD 2.06; d=2.41), and MI scores from 2.22 (SD 1.09) to 0.96 (SD 0.68; d=1.38). Given the small sample, these effect sizes are presented descriptively. At posttreatment, all 4 completers scored below the ASI-3 threshold (≥23), 3 scored below the PDSS-SR threshold (≥7; patient 2 scored 7), and 3 scored below the MI threshold (mean ≥1.5; patient 4 scored 1.60). Of the 2 noncompleters, patient 6 showed a modest worsening on all 3 measures at the time of discontinuation, and patient 5 showed little change on the PDSS-SR and ASI-3; her lower MI score is difficult to interpret because she completed only 1 module and returned weekly assessments inconsistently.

The 3-month follow-up assessment was completed by 3 of the 4 completers. Patients 1 and 2 had PDSS-SR scores of 0 at follow-up. Patient 3 showed a partial return of symptoms: her PDSS-SR score increased from 0 at posttreatment to 7 at follow-up, and her ASI-3 score increased from 3 to 9. She also described infrequent, partial panic attacks that she was able to manage using skills learned in the program. After the posttreatment assessment, patient 4 notified the therapist that her panic attacks had returned. She did not respond to subsequent contact, so no follow-up scores are available, and this recurrence is reported as a clinical observation.

Weekly PDSS-SR scores for all 6 patients are shown in Figure 1. Patterns of change varied among the completers. Two completers improved steadily, with most change occurring during the cognitive restructuring and exposure modules. One completer showed more fluctuation, with increases in symptoms often coinciding with her menstrual cycle or other life stressors. The fourth completer improved throughout treatment. Weekly MI scores declined for all 4 completers, with the largest decreases generally coinciding with the interoceptive and in vivo exposure modules. Among the noncompleters, patient 6’s avoidance remained high throughout treatment.

Posttreatment interviews were conducted with 4 of the 6 patients (patients 1, 2, 3, and 5). What each patient said is reported below, separately for each patient, following the order of the interview topics. Patient numbers correspond to those in Table 1.

Patient 1 (completer) described the program positively and said it had met her expectations. She said that “the first modules helped me the most, especially psychoeducation concerning fear and panic attacks, specifically the fact that symptoms are not dangerous and the metaphors about my ability to control my PD,” and found the final summary module least useful because it added little that was new. She described the therapeutic relationship as important, and the relationship via the platform as sufficient and efficient; she said she received support whenever she needed it while experiencing the contact as more practical and formal than face-to-face contact. She asked for “the opportunity to record voice messages since writing needs more effort, and at times it made it harder to convey my message precisely.”

Patient 2 (completer) also said the program had met her expectations. She reported that she was not anxious about forgetting material because she could return to it at any time, and that moving between modules helped her see how they fit together. She found the in vivo exposures the most helpful component and the interoceptive exposures less so and suggested that the 2 exposure modules would work better combined. She described the therapeutic relationship as important and the platform’s messaging system as adequate and said that nonverbal communication was the 1 element she missed, as facial expressions and body gestures were sometimes needed to convey a message accurately. She reported procrastination and difficulty establishing regular working times, since she works better with fixed deadlines. She proposed adding several Zoom meetings at points during the program.

Patient 3 (completer) likewise described the program positively. She identified the in vivo exposure module as the most helpful, described the psychoeducation and cognitive restructuring modules as worth revisiting during treatment, and found the interoceptive exposure module the most demanding. She described the therapeutic relationship as important and the relationship via the platform as sufficient for treatment, but said she believed a face-to-face meeting would produce a stronger therapeutic relationship because of nonverbal communication and the greater frequency of contact. During treatment, she once asked the therapist to telephone her because she found a written explanation hard to follow. She reported difficulty finding time for the program and said she dealt with this by working early in the morning or after her children were asleep. She said she “would like to have the opportunity to contact other patients so we can encourage each other and get validation for our experiences.”

Patient 5 (a noncompleter, interviewed after she withdrew) said the program had met her expectations. She described the therapeutic relationship as important and the platform’s messaging system as adequate. She said she could not sustain the program alongside her other commitments: she began treatment during Ramadan, when she needed to focus on fasting, prayer, and worship, and at the end of a school semester in her work as a teacher. She also reported feeling burdened by the weekly questionnaires. Asked what might have enabled her to continue, she answered: “maybe more practical exercises from the beginning, more Zoom meetings with the therapist, and the possibility to share experiences with other participants in order to get validation and motivation.”

This case series had 2 aims. Regarding the first aim, 4 of 6 (67%) patients completed the treatment program. All 4 completers showed reductions on the PDSS-SR, ASI-3, and MI from pretreatment to posttreatment; of the 3 who were assessed at the 3-month follow-up, 2 maintained their gains. Individual patterns of change varied across patients, with the largest reductions generally coinciding with the exposure-focused modules. Regarding the second aim, the 4 interviewed patients spoke positively about the program. Because these interviews were not formally analyzed and included only 4 of the 6 patients, they are reported descriptively.

The completion rate of 67% (4/6 patients) in our study is comparable to that reported by Strauss et al [13], suggesting similar engagement levels across different cultural contexts. Completers demonstrated high engagement, logging into the treatment website frequently and showing high worksheet completion rates. However, 2 patients dropped out for different reasons. One left due to difficulties managing their time, while the other struggled with severe PD alongside complex life circumstances and interpersonal issues. Both patients also experienced considerable interpersonal issues, raising the question of whether integrating an additional treatment approach, such as psychodynamic or interpersonal therapy, might have been more appropriate for them.

Although the treatment was originally planned for 16 weeks, completers required 17 to 22 weeks to finish the program. This suggests that the original time frame may have been too short for some patients to comfortably complete all 6 modules while managing other life circumstances. Greater scheduling flexibility appeared to support engagement and completion by allowing patients to progress at a pace that better fit their needs. Future implementations may benefit from setting a slightly longer planned treatment window (eg, 18‐22 wk) or building in greater flexibility from the outset rather than treating extensions as exceptions.

Our hypothesis was supported descriptively, with all 4 completers showing reductions in PD symptoms and AS from pretreatment to posttreatment. The maintenance of these improvements at the 3-month follow-up for 2 of the 3 patients assessed is encouraging, suggesting the potential long-term efficacy of ICBT for PD. The greater symptom improvement observed during exposure-focused modules is broadly consistent with CBT models emphasizing exposure as a key mechanism in panic treatment [24]. The varied patterns of improvement observed among patients suggest that personalized treatment approaches may be valuable. While some patients showed steady improvement, others experienced fluctuations related to external factors such as menstrual cycles or life stressors. This variability suggests that future ICBT programs might benefit from built-in flexibility to address individual differences and life circumstances that may impact treatment progress.

The ASI-3 scores observed in this sample can be compared with those reported by Strauss et al [13] for Hebrew-speaking patients who underwent the same ICBT program. Both samples improved on measures of AS and panic symptoms during treatment, and the 2 were broadly comparable at the start of treatment. When individual cases are considered, no consistent pattern emerges: variability appeared more closely tied to differences between individual patients than to language or cultural background, and, with 4 completers and no statistical comparison, we do not interpret any group-level pattern as a cultural difference. This finding differs from Wagner et al [18], who reported that among patients with PD in Australia, those of Arabic background scored higher on fear of dying (physical concerns), whereas those of Australian background scored higher on social concerns such as sweating. This discrepancy may reflect sample size, since larger studies average across patients and may obscure individual differences, whereas our study examined fewer patients in greater depth.

The current study has several limitations. The small sample size, consisting solely of women, limits the generalizability of the findings. Additionally, the reliance on self-reported weekly questionnaires, which were inconsistently completed, requires careful interpretation of the results. Another limitation is the specific inclusion criteria, particularly the decision to include only patients for whom PD was the primary diagnosis. Many PD cases are comorbid with more severe conditions, and excluding such cases may have reduced the generalizability of the results. A further limitation concerns the interview material. The interviews were conducted by the treating therapist, who also wrote the notes on which the account is based and prepared the summaries reported herein. This dual role may have introduced interviewer effects and social desirability bias, in that patients may have been reluctant to voice dissatisfaction directly to the therapist who had treated them, which may have shaped what was recorded and emphasized. The interviews were not audio-recorded, so quotations are as documented in the therapist’s contemporaneous notes rather than verbatim transcriptions; no formal qualitative analysis was undertaken, and no second researcher examined the material independently. Moreover, interviews were conducted with only 4 of the 6 patients, and the 2 who were not interviewed are those whose engagement was most disrupted; therefore, this account is likely to underrepresent the experience of patients who fared least well. Consequently, the interview material should be viewed as an illustrative patient perspective rather than as a systematic analysis of patient experience.

Future research should address these limitations by replicating this study with a larger, more diverse sample, including both sexes, comorbid conditions, and varying levels of PD severity. This would allow for a more thorough examination of cultural differences in PD and AS following treatment. Additionally, given the inconsistent completion of weekly questionnaires and the difficulty of maintaining regular contact with several patients during treatment, incorporating brief scheduled Zoom meetings could help track patient progress more closely, strengthen therapeutic engagement, and reduce missing data. Finally, some interviewed patients mentioned possible changes to the program, such as voice messages, occasional video meetings, and contact with other patients. Because these comments came from 4 patients and were not formally analyzed, they are noted only as ideas that future studies could examine.

This case series provides preliminary evidence that an Arabic-translated guided ICBT program for PD was feasible for 6 Arabic-speaking women in Israel. The program was completed by 4 of the 6 patients, and all 4 showed reductions in panic symptoms, AS, and agoraphobic avoidance; 2 of the 3 assessed at the 3-month follow-up maintained their gains. Given the small, all-female sample, the absence of a control condition, and incomplete follow-up, these findings do not establish efficacy or generalizability, and larger controlled studies with Arabic-speaking patients are needed.