Introduction

Premenstrual dysphoric disorder (PMDD) is a cyclic mood disorder affecting about 2%–5% of women of reproductive age.1 2 PMDD is characterised by recurring emotional, behavioural, cognitive, and somatic symptoms that occur during the luteal premenstrual phase of the menstrual cycle and remit shortly after the onset of menses.3 Core symptoms of PMDD include affective alterations such as emotional lability, irritability or aggressiveness, anxiety and depressed mood,4 which can co-occur with symptoms like breast tenderness or swelling, bloating, lethargy, appetite changes, sleep disturbances and difficulties concentrating.1 Importantly, PMDD is associated with clinically significant impairment in daily functioning, lower quality of life5 and increased suicidal ideation.6–8

Considering the length of reproductive age, PMDD can negatively impact the quality of life of affected women for a considerable part of their lifetime.5 9 Moreover, PMDD entails substantial societal costs in reduced work productivity, work absenteeism and increased healthcare utilisation.10–12 Given the negative impact of PMDD both at the individual and societal levels, effective and accessible treatment options are vital. Although medical treatments for PMDD are available,13 there is a need to develop additional treatment options, as medications are sometimes not well tolerated.14 15 Considering that PMDD is a mood disorder involving various components treatable by psychotherapeutic methods,16 17 it is crucial to explore psychological treatment options more thoroughly.

Since PMDD symptoms are cyclic and occur during the luteal phase of the menstrual cycle and remit with the onset of the follicular phase, it is reasonable to assume that the condition results from the effects of sex hormones on the central nervous system.18 Support for this theory is that symptoms in individuals with PMDD are absent during non-ovulatory cycles, after removal of the ovaries, or after administration of drugs that suppress ovulation.13 Accumulating evidence indicates that elevated sensitivity to physiological hormonal fluctuations in oestrogen and progesterone play an essential role,19 as well as their impact on neurotransmitter systems, including gamma-aminobutyric acid and serotonin.13 In line with its biological underpinnings, first-line treatments for PMDD are pharmacological interventions.13 These include primarily selective serotonin reuptake inhibitors (SSRIs)14 and hormone-based therapies, such as oral contraceptives19 and gonadotropin-releasing hormone inhibitors.20 Even though pharmacological interventions are often effective, a substantial number of patients with PMDD do not respond to treatment or experience residual symptoms.14 19 21 22 Also, many patients with PMDD choose to refrain from or discontinue pharmacological interventions because of side effects (eg, nausea, insomnia, headache, adverse effects on mood and/or sexual function) or negative attitudes towards medications.14 15 Therefore, non-pharmacological treatments are needed as an alternative or an additive treatment strategy to pharmacological interventions.

Cognitive behavioural therapy (CBT) has been put forward as a promising non-pharmacological treatment option for premenstrual symptomatology. This therapy typically involves implementing behavioural and cognitive strategies to limit the impact of symptoms on daily functioning and quality of life. It also seeks to modify behaviours and thought patterns that maintain or worsen symptoms and sometimes involves lifestyle interventions (eg, physical activity and healthy diet).23 A meta-analysis of randomised controlled trials (RCTs) of face-to-face CBT for PMDD and premenstrual syndrome (PMS; a less severe form of PMDD) reported small to medium effect sizes on measures of mood, behaviour, autonomic responses, physical symptoms and functional impairment.22 Based on the current evidence, the American College of Obstetricians and Gynecologists recommended CBT for the management of affective premenstrual symptoms in its latest guidelines,24 and the Royal College of Obstetricians and Gynaecologists states that CBT should be considered routinely as a treatment option for severe PMS.25

Despite the suggested positive prospects of CBT as a non-pharmacological treatment option, it has been repeatedly noted that many studies evaluating its effects have several methodological limitations (eg, small sample sizes, suboptimal methods for diagnostics and assessments, high attrition rates).22–26 Therefore, well-powered studies using appropriate diagnostic procedures and assessments (eg, prospective daily ratings) are needed to accurately discern the efficacy of CBT for premenstrual syndromes. Moreover, several researchers in the field have suggested that classical CBT strategies may not sufficiently target critical parts of PMDD psychopathology, such as emotional and interpersonal dysregulation.27–31 In line with this notion, it is proposed that CBT for PMDD could benefit from incorporating skills training on emotion regulation and interpersonal effectiveness, similar to those used in dialectic behavioural therapy (DBT).32

The rationale for including emotion regulation skills training in CBT for PMDD is based on accumulating findings showing that individuals with PMDD and PMS (vs non or mild PMS) report more difficulties in emotion regulation.17 33 Also, individuals with PMDD/PMS (vs no-PMS) report using less effective emotion regulation,17 34 35 such as rumination (16 36 37, but see 38), behavioural impulsivity,17 harm avoidance39 40 and non-acceptance of emotional responses.33 Self-reported measures of emotion regulation have been corroborated by experimental studies, showing alterations in cognitive-emotional processing in individuals with PMS versus non-PMS.28 35 41 Moreover, preliminary findings from neuroimaging studies suggest that PMDD is associated with altered emotion-related neural processing, including hyperactivation of emotion-processing brain areas and hypoactivation of cortical regions involved in emotion regulation.18 42

The use of ineffective emotion regulation strategies to deal with affective premenstrual symptoms (eg, lashing out, social withdrawal/avoidance behaviour, catastrophising or overeating) can exacerbate them (eg, depressed mood and anxiety) and lead to secondary adverse effects on relationships, function and quality of life (eg, conflicts and absence from work). The most commonly reported impairment by individuals with severe premenstrual symptoms is a negative impact on relationships.43 44 Findings show that emotion-related symptoms such as irritability, anger and sensitivity to rejection are strongly linked to relational impairment in PMDD.31 45 Learning more effective emotion regulation and interpersonal effectiveness skills (eg, effective communication and conflict management) to deal with these symptoms could potentially counteract the deterioration of symptoms and help reduce the negative impact of PMDD on relationships and other life domains. However, knowledge of the effects of improved emotion regulation and interpersonal skills on the reduction of PMDD psychopathology and relational impairment is limited.

Another challenge is that, even if proven to be effective, the accessibility of CBT for PMDD is low, potentially because face-to-face CBT is resource-demanding. As a result, many patients with PMDD do not access CBT, which can be problematic if they decline or do not tolerate pharmacological interventions. Thus, cost-effective and scalable methods to disseminate CBT for PMDD are needed. In this respect, internet-delivered CBT (ICBT) provides a window of opportunity.46 Findings indicate that therapist-guided ICBT is cost-effective47 and has corresponding effects as face-to-face CBT for various conditions.48 Indeed, recent years have seen the development of ICBT for premenstrual syndromes with promising results.49 50 For instance, one well-powered RCT on ICBT for PMDD reported strong effects on the reduction of symptoms as well as PMDD-related psychological and functional impairment.50 Another RCT on ICBT for PMS conducted on a student population showed a significant decrease in premenstrual symptoms after treatment.49 Yet, further research on ICBT for PMDD is warranted to provide robust evidence of its efficacy. Also, knowledge of whether ICBT for PMDD is value for money is lacking since, to our knowledge, no studies have investigated the cost-effectiveness of CBT or any psychological intervention for PMDD.

Study objectives

The current study aims to investigate the effects of therapist-guided ICBT for PMDD, incorporating skills training in emotion regulation and interpersonal effectiveness. The principal research question is whether the current ICBT intervention can reduce the severity of premenstrual symptoms and their impact on everyday life, as well as reduce PMDD-related psychological and functional impairment during the luteal phase at post-treatment (vs baseline). Long-term treatment effects will be assessed 6 and 12 months post-treatment. The secondary research questions include the intervention’s impact on quality of life and difficulties in emotion regulation. In addition, the project aims to evaluate the treatment’s cost-effectiveness compared with a waitlist control (WL) group.

Methods and analysesDesign

The current study is a parallel two-group RCT with 1:1 individual allocation to 8 weeks of therapist-guided self-help ICBT (TG) or WL. Participants in the WL control group will be offered treatment after completing post-treatment assessments. A study overview is presented in figure 1.

Figure 1

The study flow chart shows an overview of the screening procedure (S), random allocation to treatment or waiting list control groups and time of assessments (T). T1a data consist of the same daily ratings of premenstrual symptoms recorded in the third part of the screening procedure (S3). Both T1a and T2a consist of daily symptom ratings over two consecutive menstrual cycles. T1b and T2b refer to preassessment and postassessment of outcomes not collected daily. Outcomes which are not recorded daily will be collected during the luteal phase (T1b, T2b, FU1, FU2). In the follow-up assessments, daily symptom ratings will be collected during only one menstrual cycle. *The WL will repeat the postassessment after receiving treatment. FU, follow-up; ICBT, internet-delivered cognitive behavioural therapy; PMDD, premenstrual dysphoric disorder; WL, waiting list.

The study has been approved by the Swedish Ethical Authorities (2023-00655-01, 2024-00731-02, 2024-03631-02). Written informed consent will be obtained electronically from all participants before enrolment through the same secure internet platform used for data collection and treatment (online supplemental appendix 1). The study is preregistered at ClinicalTrials.gov (NCT06496139). The study protocol was written according to the Standard Protocol Items: Recommendations for Interventional Trials statement.51

Study setting

The trial will be conducted in Sweden. Uppsala University, the trial sponsor, will be responsible for recruitment, data collection, treatment delivery, analyses and dissemination of results. The intervention will be delivered online via a secure internet platform provided by Linköping University, Sweden (Iterapi),52 which will also be used for data collection and storage. An Independent Data Safety and Monitoring Board will monitor the trial (IDSMB; see ‘Monitoring’ section for more details).

Study population

The targeted population consists of adult individuals fulfilling the diagnostic criteria for PMDD according to the fifth version of the DSM-5.3 The following eligibility criteria are required for participation.

Inclusion criteria

Age between 18 and 45 years.

PMDD diagnosis according to DSM-5.

Menstrual cycle length between 23 and 34 days, that is, 5–8 cycles in the last 6 months.

Sufficient proficiency in Swedish to comprehend the treatment materials.

Access to computer/tablet/mobile phone with internet connection.

Exclusion criteria

Breastfeeding or pregnancy during the previous 3 months.

Initiation of or change in treatment with antidepressants, benzodiazepines, contraceptives or hormones during the last 3 months.

Current or history of a gynaecological disease (eg, endometriosis, polycystic ovary syndrome) that may confound the results.

Ongoing or previous psychological treatment for premenstrual disorders.

Severe mental disorders that may interfere with the person’s ability to complete the treatment or complicate the interpretation of results, for example, psychosis, bipolar disorder, severe eating disorder or severe depression.

Elevated suicide risk (eg, recurrent active suicidal ideation, current suicide plans or previous suicide attempts).

The inclusion of individuals undergoing stable pharmacological treatment for PMDD (eg, SSRI or hormone-based medications) is justified by the substantial proportion of patients with PMDD experiencing residual symptoms.14 21 Including these individuals will increase the generalisability of results and facilitate recruitment. Notably, for inclusion, potential participants undergoing pharmacological treatment must still fulfil the DSM-5 criteria for PMDD, have been on a stable medication regime for the past 3 months and not plan any treatment changes during the trial (ie, until completion of the post-treatment assessment).

Suicidal ideation is not an exclusion criterion, considering its high prevalence in PMDD.6–8 However, participants judged to have an elevated suicide risk, such as reporting recurrent active suicidal ideation (not only related to the menstrual cycle), current suicide plans or previous suicide attempts, will be excluded and referred to appropriate care. Suicidality will be assessed during the clinical interview (telephone) by a licensed clinical psychologist or psychology students under close supervision.

It is essential to note that PMDD can be experienced by all individuals with a domination of female reproductive sex hormone production and is not linked to gender identity. In this study, we often use the term ‘women’ to describe the sample, which is currently the most widely used term in medical research for describing participants with premenstrual complaints. However, individuals fulfilling the DSM-5 diagnostic criteria for PMDD will be included irrespective of their gender identity.

Recruitment and screening

Participants will be recruited from the general Swedish population through advertisements, for example, social media, magazines, university campus billboards and healthcare facility pamphlets. The screening procedure includes three screening (S) steps. The first step (S1) consists of web-based questionnaires, including screening for PMDD using the Premenstrual Screening Tool53 and other eligibility criteria. The second step (S2) constitutes a clinical diagnostic interview (via telephone) assessing preliminary PMDD diagnosis and psychiatric comorbidity. Trained psychologists will conduct clinical interviews. The third step (S3) entails conducting prospective daily ratings of premenstrual symptoms during two consecutive menstrual cycles, as required to establish PMDD diagnosis according to the DSM-5. Prospective daily symptom ratings will be collected using a digital version of the Daily Report of Severity of Problems (DRSP).54 The DRSP is aligned with DSM-5 diagnostic criteria for PMDD that evaluates the intensity of the 11 DSM-5 premenstrual symptoms categories and interference across different life domains (three items: work/studies/home, social life/hobbies and relationships). Items are rated from 1 (‘not at all’) to 6 (‘extreme’). The DRSP has high internal consistency (Cronbach’s α=0.95) and total score test–retest reliability (ICC=0.80).54

The diagnostic procedure for PMDD based on daily ratings will be assessed using the Carolina Premenstrual Assessment Scoring System (C-PASS),55 defining the luteal phase as the week before the onset of menses (day −7 to −1, where day −1 represents the day before the onset of menses) and the follicular phase as day +4 to +10 (where day+1 is the day of menstrual onset). The C-PASS algorithm evaluates each premenstrual symptom in every cycle individually, based on the following four criteria: (1) A symptom must show a relative luteal to follicular phase decrease in severity, defined as a symptom difference of ≥30% between the mean rating of the symptom in the luteal phase and the mean rating in the follicular phase. The current study will calculate relative change using the follicular phase mean of each symptom within the same cycle as the denominator (ie, method 1: luteal phase mean−follicular phase mean)/follicular phase mean)55; (2) A symptom in the follicular phase should not exceed a mean value of ≤3 (‘mild’); (3) The symptom in the luteal phase must show sufficient absolute severity, defined as rating of ≥4 (‘moderate’) and (4) These clinically significant symptoms must show sufficient premenstrual duration, defined as at least 2 days in the luteal phase in which the symptom is rated ≥4. The above four criteria must be met for a minimum of five symptoms (with at least one being a cardinal symptom) in two menstrual cycles to meet the DSM-5 criteria for PMDD.55

Randomisation

A person not involved in the project will carry out 1:1 equal randomisation for the two groups (TG vs WL) using the package blockrand (https://CRAN.R-project.org/package=blockrand) in R software V.4.2.2 (R Core Team, 2022). Randomisation will render a unique study code for each participant.

Intervention

The intervention will be delivered online via a secure internet platform provided by Linköping University, Sweden (Iterapi).52 Participants will be able to access the treatment and complete assessments through their computer, tablet or mobile phone. All participants in the TG will be assigned a personal therapist who will be available for support and guidance throughout the treatment. Therapists will be licensed clinical psychologists or psychology students with CBT training enrolled in their last semester of the psychology programme at Uppsala University. Participants will start the intervention at the beginning of their follicular phase.

The intervention consists of an 8-week therapist-guided self-help, internet-delivered treatment based on the principles of CBT. It comprises eight modules, each corresponding to one new module per week. At the end of each week, participants submit homework assignments and receive feedback from their therapist. Therapists and participants will primarily communicate digitally via a secure message function incorporated into the internet treatment platform. Telephone calls may occur if necessary, such as in cases of an adverse event (AE) or technical problems.

The current intervention is inspired by an ICBT protocol for PMDD previously evaluated in a German population with promising results.50 56 The protocol has been modified to incorporate skills training in emotion regulation and interpersonal effectiveness. Skills training has been derived from DBT,32 including mindfulness, acceptance, emotion regulation strategies, distress tolerance and interpersonal effectiveness. As in the German ICBT treatment protocol,50 modules promoting a healthy lifestyle are included, which participants work with in parallel with other treatment modules. However, the lifestyle modules have been made optional so that participants can choose the lifestyle module/modules they wish to work on. A lifestyle module addressing sleep has also been added. An overview of the treatment is summarised in table 1. The treatment will be further adjusted according to feedback from a Swedish PMDD patient panel.

Table 1

Treatment overview

Monitoring of adherence

Therapists will monitor participants’ adherence to treatment at least once a week by checking assignments completed in each module. Accessing a new module will require completing the previous module. Participants will receive a reminder message if they fail to submit their homework assignments on time or have not logged in to the platform in four consecutive days. The message will prompt the participant to continue working with treatment materials and remind them that the therapist is available to answer questions and provide support. The mean number of completed modules will be reported in the study results.

To promote and monitor therapists’ adherence to treatment, therapists will receive training on delivering the intervention from two experienced licensed clinical psychologists, who will also provide weekly and on-demand supervision sessions. Even though the treatment is highly standardised, the principal investigator (PI) will monitor therapists’ adherence to the protocol in the weekly supervision sessions and periodically monitor the messages sent to participants. Standardised responses will be available for common questions on treatment materials. The number of messages and time therapists spend on each participant (eg, writing messages and reviewing homework assignments), as well as the number of telephone calls, will be assessed and reported.

Control group

Given that the evaluation of ICBT for PMDD is still in an early stage, the study will use a waiting list as the control condition. Although an active control condition is generally preferable to reduce bias, a waitlist constitutes a proper comparison group in the early stages of treatment evaluation.57 However, a follow-up trial including an active control condition is needed to isolate treatment-specific from non-specific effects (eg, the amount of attention the patient receives and treatment expectations). Notably, there will not be a comparison group for the 6-month and 12-month post-treatment follow-ups, as the waiting list control group will receive treatment after the post-assessment due to ethical considerations (ie, a longer waiting time was not considered ethically justified).

Outcome measures

A summary of assessments collected throughout the trial is presented in table 2.

Table 2

Schedule of assessments

Primary outcomes

The trial has two primary outcome measures: (1) change in PMDD symptoms and impact on everyday life (prospective daily ratings) and (2) change in psychological impairment and functional impairment. The inclusion of two primary outcomes is justified by the key role of both measures to evaluate the effects of ICBT on PMDD. Even though previous findings suggest that ICBT can reduce PMDD symptoms,22 50 an essential goal of CBT is to mitigate the impact that symptoms have on daily function and mental health.58 Evaluating treatment effects beyond symptom reduction is therefore vital in psychological treatments. Moreover, the same primary outcomes were used in the only currently existing RCT on ICBT for PMDD,50 thus facilitating the comparison of results between studies.

Prospective daily ratings: symptoms and impact on everyday life

The study’s first primary outcome and endpoint consist of baseline (T1a) to post-treatment (T2a) group differences (TG vs WL) in PMDD symptoms and their impact on everyday life during the luteal phase (day −7 to −1, where −1 is the last day before menses). This outcome is measured with prospective daily DRSP ratings54 during two menstrual cycles at baseline and post-treatment, respectively. Notably, in the current study, the interference scale of the DRSP will be referred to as ‘impact in everyday life’ to differentiate it from the functional impairment scale of PMS Impact Questionnaire (PMS-I)58, described next.

Psychological and functional impairment

The second primary outcome and endpoint consist of baseline (T1b) to post-treatment (T2b) group differences (TG vs WL) in PMDD-related psychological and functional impairment during the luteal phase (day −7 to −1), as measured with the PMS-I.58 The PMS-I consists of 18 items rated on a 4-point scale ranging from 1 (not true at all) to 4 (absolutely true) and encompasses two subscales: (1) psychological impairment and (2) functional impairment. The psychological impairment subscale focuses on psychological distress caused by premenstrual symptoms (eg, negative thoughts, difficulties coping with stress and intense emotions). The functional impairment subscale assesses the degree to which symptoms interfere with various daily activities, such as work productivity, social and sexual life, and leisure and physical activities. Thus, functional impairment is assessed more comprehensively in the PMS-I than in the DRSP, which only consists of three general questions. The PMS-I has been shown to have good internal consistency (both scales: Cronbach’s α=0.87−0.90).50 58

Secondary outcomesQuality of life

PMDD can substantially affect quality of life.5 Hence, a secondary outcome constitutes ICBT-related changes in quality of life as measured with the Brunnsviken Brief Quality of Life scale (BBQ).59 The BBQ includes 12 items covering 6 life domains (recreation, philosophy of life, self-regard, creativity, learning and friendship), rated in terms of satisfaction and importance on a scale from 0 (do not agree at all) to 4 (agree completely). The BBQ has good psychometric properties (Cronbach’s α=0.71, test–retest reliability ICC=0.82) and is sensitive to differences between clinical and non-clinical groups.59 The effect of ICBT on quality of life will be assessed by evaluating baseline (T1b) to post-treatment (T2b) group differences (TG vs WL) in BBQ scores.

Emotion regulation

Given the link between PMDD and emotion dysregulation, along with the inclusion of emotion regulation skills training in the ICBT protocol, treatment effects on difficulties in emotion regulation will be evaluated using the Difficulties in Emotion Regulation Scale (DERS-16).60 The DERS-16 measures five dimensions of emotion regulation difficulties: (1) non-acceptance of negative emotions; (2) inability to engage in goal-directed behaviours when distressed; (3) difficulties controlling impulsive behaviours when distressed; (4) limited access to emotion regulation strategies perceived as effective and (5) lack of emotional clarity and is scored on a 5-point scale ranging from 1 (almost never) to 5 (almost always). Validation of the DERS-16 shows excellent internal consistency (Cronbach’s α=0.92–0.95) and good test–retest reliability (ρI=0.85–0.88).60 The effect of ICBT on difficulties in emotion regulation will be assessed by evaluating baseline (T1b) to post-treatment (T2b) group differences (TG vs WL) in DERS-16 total scores.

Health-related quality of life

Participants’ health-related quality of life will be assessed using The Assessing Quality of Life 8 Dimensions (AQOL-8D), which measures quality of life in adults for use in economic evaluation, based on 8 dimensions, including independent living, happiness, mental health, coping, relationships, self-worth, pain and senses) and 35 items with different response levels, each representing increasing levels of severity.61 62

Resource use

Participants’ resource use will be collected using an adapted version of the the Trimbos and Institute of Medical Technology Assessment Cost Questionnaire for Psychiatry (TIC-P) questionnaire.63 This questionnaire collects information on the use of societal resources including healthcare services, medication, social support and assistance, education, as well as absenteeism and presenteeism related to both paid and unpaid work.

Other measures

Sociodemographic data, including age, sex, gender identity, education, occupation, income and family status, will be collected at T1b. Data on reproductive correlates will also be collected, including parity, previous PMDD treatments and age of onset of PMDD. The Credibility/Expectancy Questionnaire (CEQ)64 will be used to evaluate how treatment credibility and expectations impact treatment effects and dropout. The CEQ will be completed at treatment weeks 2 and 4. Treatment satisfaction and AEs will be measured post-treatment using the Client Satisfaction Questionnaire-8.65 The occurrence of AEs will be assessed post-treatment using the Negative Effects Questionnaire (NEQ).66 Participants will also be asked to report the occurrence of AEs related to treatment during treatment week 4.

Follow-up assessments

Follow-up assessments will be conducted 6 and 12 months post-treatment. To limit participant burden, follow-up assessments only include daily ratings (DRSP) over one menstrual cycle. The PMS-I and other outcomes will be collected during the luteal phase.

Blinding

Although complete blinding is not possible in trials of psychological interventions, the following precautions will be taken. Randomisation will be performed externally by a person independent from the study. After group allocation, participants in the TG will be randomly assigned to one of the therapists involved in the trial, also by a person independent from the study. A statistician blinded to group allocation will conduct all statistical analyses.

Statistical analysisAnalysis populations

All randomised participants will be included in the intention-to-treat (ITT) population, regardless of whether they received or completed treatment. The per-protocol population will be a subgroup of the ITT population containing all participants without a major protocol violation. Sensitivity analyses will also be conducted with (1) participants who have completed at least four mandatory modules (modules 1–4) and (2) participants who have completed at least four mandatory modules and at least one of the optional lifestyle modules. Analyses will also be conducted, including only (3) natural cycling participants.

To further explore the potential differential effects of ICBT, exploratory analyses will be conducted for different symptom clusters (eg, affective symptoms)31 45 and symptom trajectories in terms of onset and/or in combination with severity and severity peak.67 Data will be analysed after the post-assessment (T2a and T2b) and after the completion of the follow-up assessments (T3 and T4, respectively). No interim analyses before the completion of post-assessments are planned.

Analysis of primary outcomeProspective daily symptom ratings

A latent Gaussian process model (LGPM)68 will be employed to test baseline (T1a) to post-treatment (T2a) group differences (TG vs WL), in symptom severity and impact on everyday life during the luteal phase (days −1 to −7), as measured with prospective daily DRSP-ratings (two menstrual cycles, both before and after treatment). A Gaussian process is a stochastic process in which observations occur on a continuous domain, for example, time and space. The underlying function is called a random function in which nearby points are more correlated than distant points. In this study, symptom intensity and impact in everyday life are modelled as random functions of time for each participant. Simultaneously, the underlying effect of ICBT is assumed to be a Gaussian process and, hence, itself a continuous (but latent) function of time across subjects. The effects of ICBT on symptom severity and impact in everyday life are expected to be most pronounced in the luteal phase (days −1 to −7), showing lower symptom severity and impact on everyday life in the TG compared with the WL control group (see figure 2). No baseline or post-treatment group differences are expected during the follicular phase (days +4 to +10). There are several advantages to analysing the data using an LGPM instead of, for instance, a linear mixed model (LMM). First, the continuous, flexible function representing the effect of ICBT on premenstrual symptoms provides a fine-grained map of the symptom development over time. Second, this fine-grained map generally leads to increased power.

Figure 2

Simulated post-treatment symptom trajectories (DRSP total score) of participants in the treatment group (TG) in blue (n=35) and waitlist control group (WL) in black (n=35). Thin lines represent individual participant symptom trajectories, while bold lines show the mean symptom trajectory of each group (TG and WL). A latent Gaussian process model will be applied to evaluate statistically significant effects (p=0.05) of group allocation (TG vs WL) on symptom severity (DRSP total score) and effect sizes at each time point (each day). The plot depicts a simulated true effect of treatment during the luteal phase (day −7 to −1) of medium effect size (Cohen’s d=0.5). The TG is expected to show statistically significantly lower mean symptom severity than the WL during the luteal phase (day −7 to −1). In contrast, no group differences in symptom severity are expected during the follicular phase (day +4 to +10). Simulated data are depicted on a 28-day menstrual cycle. DRSP, daily report of severity of problems.

Psychological and functional impairment

Treatment effects on psychological and functional impairment (PMS-I) will be analysed using LMM. Group (TG vs WL) and time (T1b vs T2b) will be treated as fixed effects variables and participants as a random effects variable. The two PMS-I two subscales, (1) psychological impairment and (2) functional impairment, will be analysed in separate LMMs. Treatment effects will be indicated by significant (α=0.05) group×time interactions, in which the TG is expected to report statistically significantly lower psychological and functional impairment (PMS-I) than the WL control group at T2b.

Analyses of secondary outcomes

Treatment effects on quality of life (BBQ) and difficulties in emotion regulation (DERS-16) will be analysed using LMMs following the same procedure as analyses of PMS-I. In addition, exploratory analyses will be conducted to explore the relationship between difficulties in emotion regulation (DERS-16) and primary outcomes (DRSP and PMS-I) by entering DERS change scores (baseline to post-treatment) into the LGPMs (DRSP) and LMMs (PMS-I).

Covariates

Medication intake will be entered as a covariate in all analyses to control for their influence on results.

Treatment response and deterioration criteria

Treatment response will be evaluated by computing the Reliable Change Index69 for primary outcomes (DRSP and PMS-I) and estimating clinically significant improvement/deterioration.

Remission specifications

Remission will be determined based on symptom ratings over two menstrual cycles post-treatment using the C-PASS algorithm with a 30% change criterion (for details on the C-PASS diagnostic procedure, see the section Recruitment and screening).55 Full remission will be defined as not fulfilling the criteria for PMDD or other menstrual-related mood disorders (MRMD) according to the C-PASS algorithm. Partial remission will be defined as fulfilling the C-PASS algorithm for an MRMD but not for PMDD.55 In the 6-month and 12-month post-treatment follow-ups, remission will be based on the C-PASS algorithm applied to one menstrual cycle, as daily ratings will only be collected during one cycle at follow-ups.

Handling of missing data

The LGPM used to analyse the first primary outcome (prospective daily DRSP ratings) inherently treats missing values in that the smooth functions are associated with larger uncertainty in regions with missing values.70 Hence, correct inference can be made in the presence of missing values. The LMM also inherently handles missing values and will not require imputation.71

Health economic evaluation

The health economic evaluation will adhere to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guidelines for conducting and reporting health economic evaluations (Husereau et al, 2022). The evaluation will include two types of analyses: (1) a cost-effectiveness analysis using the clinically significant improved case as outcome and (2) a cost–utility analysis using quality-adjusted life-years (QALYs) as an outcome.72 The QALY is a measure that comprises both quality of life and length of life and has an intrinsic value for money. QALYs will be estimated using individual scores from the (AQoL-8D)62 and weights derived from the general population.61 Costs will include programme costs and costs related to the consumption of other societal resources. Programme costs will be measured using a microcosting approach, including all cost components needed to set up, maintain and deliver the programme. The societal resources used by participants will be collected using an adapted version of the TIC-P questionnaire.63

Costs will be estimated from two perspectives: (1) a healthcare system perspective and (2) a broader societal perspective. The healthcare system perspective includes intervention costs (total therapist time spent on each participant) and additional costs related to the use of other healthcare resources (healthcare utilisation and medication). The societal perspective includes, in addition to the costs included in the previous perspective, costs related to social support and assistance and productivity losses due to absenteeism and presenteeism related to paid and unpaid work.

Costs will be estimated as the product of resource frequencies and their respective unit cost. Unit costs will be sourced from publicly available sources. Multiple imputations will be employed to account for missing data. Total costs will be calculated for each arm and aggregated over the trial period. Total QALYs for the participants over the trial period will be estimated. Differences in total QALYs and total costs between groups will be analysed using regression models.

Incremental cost-effectiveness ratios (ICERs) will be calculated to display the additional costs related to clinically significant improved cases in the cost-effectiveness analysis, and QALYs gained in the cost–utility analysis. The ICER cost per QALY will be set against a set of values of willingness-to-pay (WTP) for a QALY to determine cost-effectiveness. Uncertainty around the cost and effect data will be explored using non-parametric bootstrapping and represented on cost-effectiveness planes. The probability of the programme being cost-effective at different WTP thresholds will be presented on cost-effectiveness acceptability curves.

Sample size

The fact that research on ICBT for premenstrual disorders is limited49 50 complicates the estimation of expected effect sizes. Also, only one study focusing on PMDD provides data on effect sizes.50 This study reported medium to large effect sizes on treatment-related change in symptom intensity (f2=0.13), impact on everyday life (f2=0.12) and PMDD-related psychological (f2=0.20) and functional impairment (f2=0.26) when compared with a waiting-list control group.50 Meta-analysis of RCTs on face-to-face CBT for PMS and PMDD indicate small to medium effect sizes (range: d+=0.24–0.70) on mood, behaviour-autonomic responses, physical symptoms and functional impairment.22 Based on the limited research and heterogeneity in effect sizes among studies, the sample size calculation for the current study was based on an expected medium effect size (Cohen’s d=0.5, α=0.05, power=80%).

The power analysis for the first primary outcome measure (DRSP), based on analysis with LGPM, rendered a sample size of n=25 (Cohen’s d=0.5, α=0.05, power=90%). The power analysis for the other primary outcome (PMS-I), based on LMM, rendered a sample size of n=123 (Cohen’s d=0.5, α=0.05, power=80%). In line with relatively high attrition rates in internet-delivered interventions,73 the sample size was adjusted to account for an expected attrition rate of 25%, giving a total sample size of n=164.

Monitoring

An IDSMB will be advised of all safety aspects and the study’s inclusion rate. The IDSMB will monitor the trial’s progress to ensure adherence to the protocol and recommend whether to continue, modify or stop the trial. To ensure ongoing assessment and oversight, the IDSMB will oversee the trial in a series of five scheduled meetings: (1) before the start of the trial, (2) during recruitment, (3) at mid-treatment, (4) after the post-treatment assessment and (5) at the end of the trial (after 12 months follow-up). The trial’s PI will attend all scheduled meetings. Additionally, the IDSMB has the authority to convene additional meetings if deemed necessary with or without the PI, such as in response to a SAE or other urgent issues. All members of the IDSMB are entirely independent of the trial, with no financial, professional or personal connections to the study. Each member will provide a conflict of interest statement to affirm their impartiality and commitment to unbiased oversight.

Data management

Data will be collected with electronic case report forms (e-CRF). Pseudonymised data will be stored in an electronic data capturing system (EDC) hosted by Linköping University (Sweden). The EDC has been employed in numerous trials and complies with Swedish and European data security legislation.52 Data will be recorded using study code and without identifiable data in the CRF to ensure anonymity. Only personnel involved in the trial will have access to the EDC. Logging into the EDC requires double authentification. All changes made in the data within the EDC are traceable. The database will be locked for changes on the completion of the trial. Except for diagnostic status, all data will be registered directly from participants online, limiting potential errors and bias due to manual data input by research personnel.

After data collection and publication of results, data may be shared with other researchers on request, provided that the request complies with data safety regulations. Any shared data will be fully deidentified and will not contain any information that could potentially identify individual participants. Additionally, statistical code will be made available.

Adverse events

No SAEs have been reported in previous trials of ICBT for premenstrual syndromes.49 50 However, potential AEs and SAEs will be continuously monitored by the PI and the therapists delivering the treatment. Participants will have regular contact with their therapist (at least once per week) and are instructed to contact the therapist in case of (S)AEs. In addition, participants will be asked to fill out an online questionnaire about (S)AEs at mid-treatment (treatment week 4). Information on the occurrence of (S)AEs will also be collected post-treatment (T2b; NEQ). Dropout questionnaires will be conducted to evaluate the association between discontinuation from the trial and potential (S)AEs. Participants who discontinue participation due to (S)AEs will be referred to adequate treatment options (eg, outpatient clinic).

All therapists will receive training on managing AEs and SAEs. Therapists will receive weekly supervision by experienced clinical psychologists, who can also be reached on demand by the therapists onsite or by telephone. AEs include (1) occurrence of clinically significant symptoms of a severe mental disorder (eg, manic or psychotic symptoms and substance abuse), (2) clinically significant worsening of symptoms and (3) unforeseen hospitalisation related to psychiatric problems. SAEs are defined as (1) death (suicide or other cause of death), (2) suicide attempt/self-harm, (3) harm to others, (4) life-threatening events and (5) events that led to physical disability. All SAEs and AEs are required to be reported to the PI within 24 hours of notice of the event. The IDSMB will be notified of all SAEs within the same time frame (24 hours). The relation of (S)AEs and the treatment will be documented and rated as (1) certain, (2) likely, (3) possible, (4) unlikely, (5) no relationship and (6) unknown.

End of protocol treatment

The trial may be terminated in the event of serious or several AEs related to the treatment or if administrative complications preclude adherence to the study protocol (eg, difficulties in recruitment). The need to end the trial will be discussed and determined with the advice of the IDSMB. The time of termination and reasons for ending the trial will be reported in the clinical registration (ClinicalTrials.gov, NCT06496139).

Individual participants’ involvement in the study will be terminated if they are no longer willing to participate or withdraw their consent. Participants can withdraw from the study at any time without providing a reason. Withdrawing from the treatment will not result in disadvantages or negative consequences for the participant. The primary investigator and IDSMB can also decide to end the participation of individual participants for medical reasons or the occurrence of severe mental health symptoms (eg, psychosis, mania/hypomania and suicidality).

Patient and public involvement

An advisory PMDD-patient panel has been established. Patients will review and give feedback on treatment and the burden and time required to participate in the trial. Feedback will be collected through a semistructured interview. The treatment will be modified to meet the needs of the patient group and enhance treatment acceptance and tolerability. The panel will also be advised about recruitment strategies and how to promote treatment adherence and minimise dropout rates. Moreover, the panel will be regularly informed about the trial’s progress and provide advice during the treatment phase. Finally, the study will be discussed with the panel before publication of results, including interpretation of results, suggestions for treatment improvement, open research questions and treatment implementation.

A subsample of participants enrolled in the trial will also be invited to provide feedback on the treatment and study participation on completion of the treatment period. Feedback will be collected through semistructured interviews. To improve the treatment in terms of adherence and acceptability, participants who choose to terminate their participation in the study prematurely will be asked if they are willing to fill out a ‘dropout’ questionnaire assessing factors leading them to drop out. Participants who drop out of treatment will also be asked if they are willing to complete post-treatment and follow-up assessments.

Ethics and dissemination

The study has been approved by the Swedish Ethical Review Authority (ID: 2020-05777; Amendments: 2024-00731-02, 2024-03631-02), which includes the ethical aspects of the study, safety rules, the participant information sheet and the informed consent form. Informed consent will be obtained online from all participants before enrolment. Amendments to the protocol will be submitted to the Ethical Review Authority for approval and reported in the study’s clinical trial registration at ClinicalTrials.gov (ID: NCT06496139). Significant deviations from the protocol will be communicated when reporting the study results (eg, publication in peer-reviewed journals). The study will be conducted in accordance with the Declaration of Helsinki. The IDSMB and PI will monitor adherence to the protocol.

The study results will be disseminated in open-access peer-reviewed journals and scientific conferences. The peer-reviewed publication of study results will be sent to all trial par