Efficacy of cognitive and behavioral therapy (CBT) treatment for individuals living with symptoms associated with environmental factors (SAEF): protocol for a randomized controlled trial

Abstract

Background

Treatment for symptoms associated with environmental factors (SAEF), such as electrohypersensitivity or multiple chemical sensitivity, have been sparsely investigated thus far.

Methods

In this parallel, two-armed open-label randomized controlled trial, we will apply a treatment model for SAEF based on cognitive behavioral therapy. Our goal is to recruit N = 68 people with either multiple chemical sensitivity or electrohypersensitivity, or both, by informing our professional network of healthcare practitioners and patient groups. Prospective participants will be randomized into a treatment or a waitlist-control group. The primary outcome is physical health-related quality of life, while attendance, measures of symptom experience, cognitive and emotional processing, and comorbid psychopathology will be assessed as secondary outcomes. Delivered in-person and/or online through videoconferencing tools, the treatment will consist of 12 individual CBT-inspired sessions, guided by specialized psychologists and spanning eight weeks. Patients will be assessed at three timepoints, each 3 months apart, meaning we will have pre- and post-treatment measurements for both groups.

Discussion

This randomized controlled trial will be the first to thoroughly test a cognitive behavioral therapy intervention for individuals living with SAEF. We intend to meet critical unmet needs for specialized healthcare in this population. Additionally, by adding many secondary outcomes that have not, or not sufficiently, been studied in individuals with SAEF, we will gain a more comprehensive understanding of these patients’ thoughts, feelings, and behaviors.

Trial registration

This trial was first prospectively registered at ClinicalTrials.gov, code NCT05973214, on the 26th July 2023.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-025-03564-5.

Background

Individuals may attribute a variety of non-specific symptoms to environmental factors (SAEF) [1], a condition also known as idiopathic environmental intolerance. There are different types of SAEF, such as towards electromagnetic fields (SAEF-EMF, elsewhere called electrohypersensitivity, [2]), chemicals with or without odors (or SAEF-MCS, elsewhere called multiple chemical sensitivity), noise, infrasound emitted by wind turbines, screens, or moisture and mold in buildings [1]. The exact prevalence is hard to estimate and largely depends on case definition. Karvala and colleagues show that the reported rates range from 0.5% to 52%, but more likely, these disorders affect around 2–15% of the general population [3], SAEF-MCS being the most common. Crucially, SAEF are not explained by toxic effects of blamed environmental factors. For example, electromagnetic fields exposure studies showed no relationship between reported symptoms and actual exposure [4–8]. The incidence of SAEF-EMF did not increase despite the steep increase of EMF exposure in our environment in the last years (3G, 4G, 5G, Bluetooth, etc., [9]). Likewise, no relationship between symptom severity and exposure was found for SAEF-MCS [1, 10–12]. It thus seems that most individuals with SAEF inappropriately attribute symptoms to the environment [1, 10, 13–16].

Psychosocial factors likely play an important role in SAEF. Thus, modern health worries [17], climate change anxiety [18], absorption (i.e., the tendency to get deeply involved in sensory experiences) [19], suggestibility [20], and schizotypal or paranoid personality traits [21, 22] were associated with SAEF in cross-sectional studies. Having SAEF is associated with poorer sleep and insomnia [23], and higher health anxiety [24]. Likewise, compared to healthy controls people with SAEF-MCS show higher trait anxiety [25] and a higher focus on autonomic sensations [24, 25]. Longitudinal assessments showed that somatic symptom distress, attribution, stress, anxiety, and depression predict SAEF burden [26, 27]. Furthermore, a large body of evidence suggests that belief of being exposed to an environmental factor [7, 11, 28–30] and symptom expectations [31] predict symptoms, highlighting that nocebo effects largely contribute to SAEF. These psychosocial factors at play in SAEF are shared across distinct environmental factors. Indeed, a large proportion of individuals attribute symptoms to multiple environmental factors: about 30 to 40% of individuals with SAEF-EMF also have SAEF-MCS [32]. In a nutshell, it thus seems that SAEF feed on a vicious cycle that does not depend on actual toxic effects of environmental factors. Rather, occasional symptoms (of any cause) are inappropriately associated with environmental exposure and then amplified through threat-related overgeneralization [10, 19, 24, 33, 34]. This fear-driven generalization entails that the symptoms easily spread across different contexts.

Overall, SAEF can greatly impact quality of life [35]: Individuals reported a large impact on school or work attendance, travelling, choice of hobbies, participation in community activities, and access to places of worship, grocery stores, doctors’ offices, or libraries [36–38]. Despite this burden, patients with SAEF face difficulties accessing care [39–41]. These people report high unmet care needs [42] and tend to try many types of (conventional and alternative) treatments although no treatments are validated for SAEF to date. In a study on self-rated efficacy of different MCS treatments, most participants had tried avoidance-based (moving residences, chemical avoidance, removing mercury dental fillings, installing air filters, all avoidance behaviors that are not helpful to improve SAEF and contribute to hinder quality of life − 43), safety-behavior-like strategies (prayer, nutritional supplements), or exercise [39]. As most people undergo more than 20 different options for treatment [39], there is substantial iatrogenesis [43]. In turn, difficulties in social connections and employment can lead to further fear and symptom monitoring, which can lead to higher symptom burden [44].

One might think that people with SAEF are less likely to accept psychotherapy since they attribute their symptoms to a biomedical cause. Only a third of participants have reported trying some sort of psychotherapy, usually giving it a very low efficacy rating (ref 23). However, cognitive behavioral therapy (CBT) has shown to be effective in functional or persistent somatic symptoms, which have similar disorder mechanisms [45–47]. A handful of studies have investigated CBT for SAEF so far. One study used a single-case experimental with a baseline (A) and intervention (B) phase across five subjects and showed preliminary support for CBT on symptom distress, functional impairment, and catastrophizing [48]. Four studies showed preliminary support in favor of CBT treatment for SAEF but with important methodological limitations, including heterogeneity of delivered treatment (4 to 16 CBT sessions) and small samples of participants (17 to 22 subjects) [49]. Preliminary results by Lappalainen et al. [50] suggested that online acceptance and commitment therapy can be effective in reducing somatic, depressive and anxiety symptoms [50], but only individuals with SAEF-MCS were included and their number was not sufficient as to assess the treatment thoroughly. Other trials suffer from small sample sizes and do not find effects, but report high acceptance rates from participants [51–53]. Importantly, although SAEF are considered to derive from similar biopsychosocial mechanisms whatever environmental factor(s) is blamed by affected individuals, trials testing CBT programs have focused only on SAEF associated with one environmental factor so far. Outlines exist for a comprehensive cognitive behavioral therapy specifically tailored for SAEF, that is adapted to SAEF of all kinds and includes cognitive tools alongside with mindfulness and progressive exposure to environmental factors and symptoms [54], but they have not been systematically tested thus far.

In the current study, we seek to systematically test this CBT-inspired treatment program [54] by using a waitlist-control design and recruiting participants with SAEF-MCS, SAEF-EMF, or both. It is expected that:

1. CBT-inspired treatment is effective in improving quality of life, especially physical quality of life.

2. Compared to the control group, at T2 the treatment group will score lower on measures of comorbid disorders (depression, anxiety), SAEF-related symptomatology, somatic symptoms, modern health worries, climate change anxiety, intolerance to uncertainty, autonomic stress, and higher on emotional awareness.

Additionally, we will explore whether treatment efficacy would be moderated or mediated by emotional awareness, therapeutic alliance during treatment, medical and demographic characteristics, cognitive flexibility, or other patient characteristics. Attendance rate for the treatment will be examined and higher rates would indicate higher acceptability. The influence of acceptability and intervention adherence on clinical outcomes will be examined, and possible or adverse events and side effects will be monitored.

Methods

Note that a structured summary of trial design and methods, including items from the World Health Organization Trial Registration Data Set, can be found in Supplementary Material A. information regarding trial oversight and monitoring are provided in the Supplementary Material B.

Sample

Effect sizes of somatic symptom reduction in functional physical symptoms range from d = 0.1 [55] to d = 1.21 [56] with CBT. A moderate effect size of d = 0.70 is thus expected. A power analysis for the group comparison between the experimental versus control group using analysis of covariance, an α error of 0.05 and a power of β = 0.80, showed that a sample size of n = 34 per group is needed. Should any participants drop out before T2, we will recruit extra participants to reach our desired sample size. The estimated sample size is suitable regarding prevalence rates of SAEF and our clinical facilities.

Potential participants are eligible if they are above age 18, present a diagnosis of SAEF assessed by a specialist physician (i.e., non-specific symptoms lasting for at least 6 months, that individuals associate with EMF and/or chemicals exposure, that have significant impact on daily life, and that are not explained by another identified disorder), and have a social security scheme. Only participants who provide informed consent will be admitted to the study. Participants will be asked to provide informed consent at the first medical interview at T1, which is conducted by a physician.

Exclusion criteria are having severe depression, delusion disorder, or a substance use disorder; not having access to the internet (needed for self-help tools); participation in another research study on SAEF, or being under an exclusion period because of such a study; not speaking French; benefiting from French state aid (AME); deprivation of liberty by court or administrative order; being hospitalized without consent; being under protective measures; and pregnancy and/or breastfeeding. All patients will undergo a complete medical interview before inclusion in which these criteria are assessed. The intervention will be performed by licensed CBT psychotherapists.

Recruitment has started in February 2024 and will continue until the sample size is achieved. This will likely be at the end of May 2026, meaning the last included patients will have their last assessments in December 2026.

Patients are recruited from the department of Occupational and Environmental Disease at the Hôtel-Dieu Hospital in Paris. By regularly attending and speaking at medical conferences, including board meetings of the French National Network for Monitoring and Prevention of Occupational and Environmental Diseases (RNV3PE), we hope to encourage practitioners in our network to refer their patients. Participants can express interest via dedicated email address or phone line, and are contacted by the study team for further information and to schedule a first appointment visit. As there are no exclusion criteria regarding place of residence, participants might also come from, for example, other parts of France, Wallonia (Belgium) or Luxembourg. To facilitate this, we reimburse travel expenses up to 500€ for all 3 research visits. No other financial compensation is planned.

Materials

Primary outcome

Our primary outcome is the physical component score of the Short-Form Health Survey (SF-12) at T2. The SF-12 measures health-related quality of life over the last month. The questionnaire consists of 12 items with varying answer formats and is scored on a scale from 0 to 100, with higher scores indicating higher quality of life. Two subconstructs are often calculated: the physical component score and the mental component score. It was originally designed by Ware and colleagues [57] and has since been validated in many settings [58–60], where it has repeatedly shown good construct validity and acceptable reliability (Cronbach’s alpha ≥ 0.7). We used the French version, which was created and validated by Gandek and colleagues [61].

Secondary outcomes

To measure symptoms, we used the symptoms subscale of the Quick Environmental Exposure and Sensitivity Inventory (QEESI), which was derived from the longer Environmental Exposure and Sensitivity Inventory [62]. This subscale consists of 10 items, which are answered on a scale of 0 to 10, with higher scores meaning higher symptom burden. Note that SAEF-EHS and SAEF-MCS have overlapping symptoms [26, 32], so the same questionnaire can be used for their respective somatic symptoms, even though this questionnaire originally was not designed with SAEF-EHS in mind. It has been validated in Japanese [63] and Swedish [64], and it shows high sensitivity and specificity in German [65]. In French, the questionnaire was translated by Asselain and colleagues [66].

The Somatic Symptom Disorder – B Criteria Scale (SSD-12) is a questionnaire designed to assess the excessiveness of thoughts, feelings, and behaviors associated with somatic symptom burden, in line with the DSM-5 diagnostic B criteria for somatic symptom disorder [67]. It was designed by Toussaint and colleagues [68]. Respondents may answer the 12 items on 5-point Likert scales, with higher scores meaning more preoccupation with somatic symptoms. Many validations exist, for example in Chinese [69] and French [70], the latter of which we use. In the validation study, this measure had a high internal consistency of α = 0.90.

The Symptom Interpretation Questionnaire is a self-report instrument that measures how respondents attribute somatic symptoms [71], and it consists of three subscales: somatic, psychological, and environmental/normalizing attribution. The instrument comprises 13 overarching items with 4 subquestions (one subquestion per attribution style plus whether the respondent experienced these symptoms in the last three months). Higher scores indicate stronger attribution. This questionnaire has been validated and used in a variety of settings [72–74]– some showing good psychometric properties [75], some with less positive reviews [76]. Since to our knowledge, no validated French version exists yet, we translated this questionnaire ourselves.

The Levels of Emotional Awareness Scale is a scale that conceptualizes emotional awareness as a socio-cognitive skill and tests this through ten vignettes of situations that might cause emotions in a second-person narrator or other people [77]. Participants are asked to describe how the narrator or the other person in the vignette may feel, and their answers are scored on a scale from 1 to 5 using a (computerized) coding manual. A higher score indicates higher emotional awareness. This measure has been validated in German [78], Dutch [79], and French [80]. The French version showed good reliability (α = 0.71) and validity.

For people with electromagnetic hypersensitivity, we will use the Electromagnetic Field Sensitivity Scale (EMFSS). The EMFSS is an 11-item questionnaire that can be used to assess electrohypersensitivity [81]. Most statements can be answered on a 6-point Likert scale, except for one 5-point answer format. Higher values indicate higher levels of electrohypersensitivity. Next to the original Swedish study, where it showed good reliability and validity, the EMFSS has also been used in Finnish studies [82, 83]. We are not aware of a French translation or validation, so we translated this questionnaire ourselves.

For the people with MCS, we used the four other subscales of the QEESI. Thus, there were another forty items across the following four subscales: Chemical Intolerances, Other Intolerances, Masking, and Life Impact. Most questions are answered on a scale of 0 to 10, with higher scores meaning higher symptom burden.

Modern health worries (MHW) are defined as “the degree to which individuals are concerned about features of modernity affecting their health” [84]. This construct was first measured by a 27-item scale with four subscales, namely Toxic interventions, Environmental pollution, Tainted food, and Radiation. Answers are given on a 5-point Likert scale, with higher scores indicating higher preoccupation. The original scale had good psychometric properties [84] and was validated in a Swedish study, too [85], but a 12-item shorter form was later developed that respected this four-factor basis and was recommended as a fast alternative [86]. Interestingly, the original and short version of the scale show considerable differences between countries regarding the subscale profiles – which makes sense when considering historic events, traditions, and local health scares. As no French version exists yet, we translated it for the purpose of this study.

In an attempt to standardize measures of climate anxiety, Clayton and Karazsia [87] designed the Climate Change Anxiety scale (CCA) to capture different factors that make up the overall experience of climate change, namely functional impairment resulting from thinking about climate change, as well as a cognitive-emotional experience. It consists of thirteen items which are responded to on a scale from 1 (= never) to 5 (= almost always). Thus, a higher score corresponds to higher trait climate anxiety. It has been validated by multiple authors [87–89], for example also in a French translation [90].

The Hospital Anxiety and Depression Scale (HADS, [91]) is one of the most widely used self-assessments of anxiety and depression in hospital outpatient clinics, where the presence of physical illness could confound the assessment of psychological distress. It deliberately omits items related to somatic symptoms (e.g., fatigue, sleep disturbance, appetite loss), which are common both in depression/anxiety and in other medical conditions, to minimize diagnostic overlap. Ten questions each assess a person’s levels of depression and anxiety, respectively. The answer format consists of a four-point scale and higher scores indicate higher depression/anxiety levels. This instrument has been validated countless times (see Bjelland and colleagues’ work for an overview, [92]) and has a mean reliability of α = 0.83 as well as mean sensitivity and specificity of 0.80. The French version [93] has also been psychometrically evaluated multiple times [94, 95].

The Body Perception Questionnaire (BPQ) was originally developed by Porges [96] and is designed to measure autonomic reactivity and body awareness. We use the Very Short Form [97], which spans only twelve items answered on a five-point scale. People scoring higher on this measure are more likelyto pay more (dysfunctional) attention to the body. The BPQ has also been discussed as a measure of interception, which is the sensory integration and interpretation of bodily signals [98]. The BPQ has been validated many times and shows good validity and internal consistency [99–101]. We used the French version provided by Porges’s official BPQ website [102].

The Intolerance of Uncertainty Scale – Short Form (IU) is a questionnaire measuring the tendency to consider the probability of a negative event happening as unacceptable or intolerable [103]. This shortened version of the full length Intolerance of Uncertainty Scale [104] comprises twelve items accompanied by a five-point Likert scale, with higher values indicating higher uncertainty intolerance. The full-length version showed high internal consistencies (all α ≥ 0.87) in a replication study [105]. The French adaptation is by Mougiama-Daouda and colleagues [106].

Other assessments

The Attitudes Towards Seeking Mental Health Services Inventory (IASMHS, 109) is a modified version of the Attitudes Towards Seeking Professional Psychological Help Scale [107], which measures how people think and feel regarding utilizing psychological healthcare. This 24-item scale assesses 3 internally consistent factors: psychological openness, help-seeking propensity, and indifference to stigma. Participants answer the questions using a 5-item scale (0 = disagree; 4 = agree). Higher scores on the indifference to stigma scale reflect a positive stance towards mental healthcare services, while higher scores on the other two subscales reflect a more negative attitude. Mackenzie et al. cross-validated their results with multiple experiments [108] and the inventory has further been validated and adapted in Ireland [109], in Portugal [110], and in France [111].

A healthcare utilization questionnaire was inspired by the study of Köteles & Simor [112] and adapted for this study. It consists of 6 questions with an open answer format and assesses the number of medical and paramedical visits, therapists’ visits, physical exams, as well as medication and dietary supplements consumed over the last 3 months. This questionnaire is included in our study for descriptive clinical purposes. We translated this questionnaire into French for the purpose of this study.

The Working Alliance Inventory was designed by Bordin and colleagues [113] to measure three dimensions of working alliance: agreement on therapeutic goals, consensus on tasks, and affective bond between client and therapist. The questionnaire consists of eight items measured on a five-point scale (1 = never to 5 = always), as well two open questions for feedback to the therapist. A higher score indicates a better client-therapist relationship. The questionnaire has been used in many settings and languages, but a recent systematic review points out psychometric problems with the instrument [114]. Nevertheless, we think it can still relay how the client sees the therapy. The French version has been validated [115]. It is filled out 3 weeks into treatment by the main therapist.

The Client Satisfaction Questionnaire assesses how satisfied the patients were with their treatment program [116]. It consists of eight items with a four-point answer scale. High scores indicate high treatment satisfaction. It has been positively evaluated and validated [117].

In addition, a physician will conduct a detailed medical interview with the participant. In this interview, the physician will assess comorbidities (depressive, anxiety and post-traumatic stress disorder, somatic symptom disorder), perpetuating factors (lack of physical activity), and risk markers that were not yet assessed in the questionnaires (alexithymia, absorption, schizotypal traits, catastrophizing tendencies, obsessive-compulsive personality, perfectionism, post-traumatic stress, and all-or-nothing thinking) [118].

Behavioral tasks

The Affect and Symptom Paradigm was designed to study how negative affective stimuli influence somatic symptom perception [119]. Negative and neutral pictures are presented to participants and after each picture series, participants fill in a short questionnaire about their current symptoms and affective state [120]. Negative stimuli evoke more somatic symptoms, and this is more pronounced in people with functional disorders and higher everyday symptom burden [121, 122]. The test-retest reliability is high [123]. For this trial, we translated the questions accompanying the stimulus materials into French.

The Attribution Bias task is a paradigm we designed based on earlier exposure tasks [29, 124, 125]. Here, participants are shown a disorder-related stimulus (Wi-Fi-box or a flask filled with transparent substance for SAEF-EHS or SAEF-MCS, respectively) and asked to place it at a distance where they think they will feel symptoms, but not be overwhelmed by them. Then, the Wi-Fi box is connected to electricity or the flask is opened. Patients are exposed for 10 min and repeatedly asked to assess their general wellbeing and symptoms, after which they have a recovery period. Tasks like this have shown high internal and construct validity [7, 11].

The Belief Updating Task was developed to investigate information processing and belief updating in light of desirable or undesirable information [126]. Participants are prompted with an adverse event and asked to estimate how likely it is to happen to them, and then to a similar other person. They are then presented with the real probability, which is considered good news if it is lower than the participant’s guess, or bad news if it is higher. Participants are then asked to re-estimate. Patients with psychopathology update their beliefs differently to healthy controls, likely because they are more influenced by their prior beliefs and experiences [127]. This task has been used in a variety of settings [127].

Intervention design

The intervention was adapted from an established treatment model for SAEF [54] that has never been systematically tested. Next to the CBT elements, the treatment also includes some elements from Mindfulness-Based Cognitive Therapy (MBCT) and Acceptance and Commitment Therapy (ACT). The intervention consists of twelve individual sessions over eight weeks, each of them guided by CBT psychologists (CFD and FZI) specially trained for managing SAEF and supervised by senior psychologists or psychiatrists (VP, OVdB, CL). For accessibility purposes, the therapy will be delivered partly or fully online and/or in person, as patients wish. We hope this will decrease attrition and increase treatment acceptance as well as face validity, given that some individuals with SAEF avoid internet use while others avoid in-person consultations at the hospital because of chemicals/EMF exposure. This also reduces selection bias, as a more heterogeneous group of people is able to participate. Each session will last about 50 min. Table 1 summarizes the topics and content of each session.

Table 1.

Contents of the therapeutic intervention, adapted from Van den Bergh et al. [54]

	Theme	Content
Session 1	Building Engagement	Collection of anamnesis and exploration of the patient’s personal theory of illness. Compassion-based approach to establish a therapeutic alliance. Introduction to CBT; addressing scepticism and reinforcing collaboration.
Session 2	Explaining the stress response system	Linking up the biopsychosocial explanatory model with the patient’s anamnestic information, explaining the stress response, the nocebo effect with a demonstration by provoked hyperventilation exercise, and how avoidance behaviors can contribute to symptoms.
Session 3	Goal Setting	Motivational interview and review of decision balance to highlight the benefits of symptoms and the costs of avoidance. Identification of avoidance patterns and introduction to exposure therapy concepts.
Session 4	Behavior	Exposure and behavioral planning with compassion-based motivational components. Introduction to applied relaxation & relaxation training.
Session 5	Cognitions	Focus on managing negative thoughts and beliefs. Tackle catastrophizing, jumping to conclusion, general health views. Explore alternative interpretations of symptoms and relations to environmental cues.
Session 6	Mindfulness	Psychoeducation on body-mind connection. Introduction to interoceptive and grounding techniques to connect with bodily sensations.
Session 7	Interoception Exposure	Acceptance of bodily symptoms: internal and situational interoceptive exposure.
Session 8	Emotions	Identify and modify emotion driven action tendencies through Socratic methodology. Explorations to increase emotional awareness and refrain emotional avoidance.
Session 9 & 10	Broader life context	Ameliorating broader context of life: Living with uncertainty, targeting psychosocial and psychobiological processes.
Session 11	Relapse Prevention	Setbacks and maintenance planning. Discussing relapse prevention and working out a realistic health concept. Reinforce autonomy and adaptive responses.
Session 12	Conclusion	Reviewing maintenance plan, feedback, updated narrative.

Every session will start with reviewing the patient’s homework and setting a schedule for the present session. To solidify the new behavioral strategies, patients will receive homework assignments. As is usual for CBT, the therapy sessions will be tailored to a patient’s specific problem (i.e., exposure tasks will be based on a patient’s personal, SAEF-related anxiety hierarchy), but will follow the contents and schedule shown in Table 1. Participants cannot switch groups.

Participants may quit the study at any time without any justification or negative consequences for usual care. If a participant wishes to drop out, the lead physician (VP) will schedule a telephone call with the participant to receive feedback on the intervention (if appropriate) and, if the participant wishes to provide this information, reasons for dropping out. Should this be due to any misunderstandings or other interpersonal factors, the lead physician will try to clear this up in the hopes of convincing the participant to continue.

Strategies to improve adherence to interventions

Each session is grounded in a therapeutic posture based on emotional validation, active listening, and maintaining a secure framework to support a strong therapeutic alliance. The therapist acts as a facilitator inviting the patient to explore their experiences with curiosity and compassion.

Self-observation and self-regulation tools are integrated throughout the program (e.g., symptom diary, motivational balance scale, emotion wheels, avoidance hierarchies, and daily interoceptive training).

The therapist may use CBT techniques such as motivational interviewing to engage the participant if they notice that the participant is becoming disengaged. The therapist will also fill out a questionnaire about the therapeutic working alliance.

Relevant concomitant care permitted or prohibited during the trial

Participants may continue receiving usual medical care (e.g., pharmacological treatment or non-CBT supportive therapy) during the trial.

However, initiation of other structured psychological interventions targeting persistent physical symptoms—particularly those based on cognitive behavioral therapy (CBT), mindfulness-based approaches, or similar psychotherapeutic frameworks—is prohibited during the intervention and follow-up periods, to avoid contamination of the treatment effect.

Provisions for post-trial care

Participants interested in further psychotherapy will be given the contact details of psychotherapists experienced in SAEF. Patients are offered regular medical consultations (every 3 months) as a usual care for the waitlist control group, if they do not want to participate or after the treatment program is over.

Procedure

In this monocentric study, participants will complete the study visits at Hôtel-Dieu Hospital (Paris, France). The therapy sessions will be delivered both face-to-face and online via videoconferencing tools.

This is a randomized controlled trial with two parallel groups: an experimental group and a waitlist control group. Participants will be randomly allocated to a study arm in a 1:1 ratio. The allocation algorithm will be activated by the physician during the T1 medical interview. The participants will be assigned to the study arms through a computer algorithm (randomization by blocks of 12 patients) implemented in the REDCap electronic database [128, 129]. Thus, participants will immediately know to which group they are assigned. Allocation concealment will be ensured, as the algorithm will not release the randomization code until the participant has been recruited into the trial, which takes place after completion of the baseline measurement. There is no blinding in this study.

This is a superiority framework: we expect the participants in the experimental group to have clinically better outcomes at T2. Health-related quality of life was chosen as a main comparator. Patients with SAEF have very heterogeneous symptoms [10, 26] and are characterized by avoidance behavior [130], which could explain why SAEF are associated with higher trait anxiety [25] and with having a high impact on lifestyle (e.g. job attendance) [36, 38]. Researchers and clinicians in similar areas have therefore called for using health-related quality of life [131–133], because it best fits to the WHO’s definition of health [134]. Using this classic patient reported outcome follows the current state-of-the-art recommendations for health care research and evaluations [135, 136].

Table 2; Fig. 1 serve as schematic overviews of the study schedule. Participants received the information and were screened for eligibility at least one week before T1. T1, T2, and T3 are 3 months (± 2 weeks) apart, meaning that there are approximately four weeks between the end of the treatment and the next measurement. This was decided because the SF-12 measures quality of life in the last four weeks, so the SF-12 measurement is truly a post-treatment measurement. Note that all three research visits take place at the Hôtel-Dieu Hospital in Paris.

Table 2.

Schedule of enrolment, interventions, and all outcome measures

*Only considered a primary outcome at T2, to compare the groups. At T3, this subscale is considered a secondary outcome

Fig. 1.

Trial Flowchart. Note. SAEF-MCS = Symptoms associated with environment factors, Multiple Chemical Sensitivity; SAEF-EMF = Symptoms associated with environmental factors, sensitivity to electromagnetic fields; EG: experimental group; WLCG: waitlist control group

Data management and analysis

Data will be stored on the REDCap platform, which is a secure web platform for data management [128, 129]. The information obtained during the medical interviews will be entered into REDCap directly by the physician, while the questionnaires are delivered in paper-pencil mode and later entered into REDCap by student assistants. The paper-pencil mode was chosen because study participants living with SAEF-EMF might show some difficulties filling up questionnaires on an electronic device. Range checks will be performed by the data analyst. The physical copies of the informed consents and paper-pencil questionnaires, as well as copies of the digital data, will be sealed and archived in the Hôtel-Dieu hospital for 10 years after the end of the trial. The data will be stored pseudonymously to protect the confidentiality of the participants. Any lists containing identifying information will be deleted upon study completion.

For all outcomes, we will store the data at item level and calculate every participant’s individual scores as suggested in their respective manuals, including the subscale scores where applicable. Data distributions will be examined for normality using the Shapiro–Wilk test. In the presence of significant deviations from normality, non-parametric tests (e.g., Wilcoxon signed-rank tests or Mann–Whitney U tests) will be reported alongside parametric analyses.

We plan to analyze the SF-12 physical component score data by using a repeated measures ANOVA with timepoint (T1 and T2) as a within-subjects factor and group (intervention vs. waitlist control) as a between-subjects factor. We will conduct similar rmANOVAs and multilevel models with all secondary outcomes. We also plan to calculate a reliable change Index to determine the proportion of participants showing clinically meaningful improvement. We will be using an intention-to-treat analysis.

Once T3 is completed, we plan to extend this analysis by conducting a multilevel mixed effects model analysis and including group, treatment phase (before, after, follow-up), and timepoint with random intercepts for participants as predictors, as multilevel models are statistically superior to rmANOVAs, but can only be done with three or more waves of data [137].

We will compare the participants with SAEF-MCS and the participants with SAEF-EMF by doing two-sample t-tests, calculating odd’s ratios, and testing our hypotheses in each group separately. All parameters and outcomes will be calculated the same way as for the full sample. No other subgroup analyses are planned.

No interim analyses are planned, but may be requested for the auditing process by the study sponsors. These interim analyses will not have any impact on recruitment, later analysis strategies, or terminating the trial.

We will not be applying multiple testing correction for the reasons put forward by Simmons [138]. We intend to use R version >4.4 for analyses [139]. We will not impute missing data. We plan to use an intention to treat procedure (ITT) for clinical significance, meaning that some analyses might have a higher N than others (also known as pairwise deletion or available case analysis).

Discussion

With this study, we hope to show that a CBT-inspired intervention is an acceptable and effective treatment for individuals living with SAEF. In our procedure, we attempt to have a participatory approach – we have been actively letting our professional network, especially in France, know about our trial and asked for referrals, and we have been in contact with patient advocacy groups. We will also disseminate our results to these groups. This way, we hope to address the underlying problem in SAEF treatment: a lack of knowledge regarding treatment opportunities [140–142].The waitlist design does not control for expectancy or attention effects. However, it was chosen for feasibility reasons and to ensure all participants would eventually receive the intervention, given the current absence of evidence-based care for SAEF.

The research area of underlying mechanisms SAEFs is still sparse. While some attempts have been made to gain a deeper understanding beyond symptoms (see for example the Västerbotten Environmental Health Study, [143]), we hope that our secondary outcome variables will broaden our knowledge of the sociodemographic, cognitive, and emotional factors that predispose, precipitate, perpetuate, or protect from SAEFs. For example, modern health worries and SAEFs are often studied together [13, 22, 144–146], while to our knowledge, there are no studies about uncertainty intolerance in SAEF – although our clinical experience lets us believe that there might be a significant connection. Another example is emotional awareness: we were only able to find one study about its role in SAEF, which had inconclusive results [147], while alexithymia is a known risk factor for persistent somatic symptoms [148]. Future studies may investigate whether improvements in quality of life are mediated by changes in emotional awareness, intolerance of uncertainty, or symptom expectations. Long-term follow-up would also help determine whether the benefits are sustained.

Recruitment may be hindered by low perceived psychological causation among patients with SAEF, making CBT less appealing to individuals living with SAEF, as well as by logistical barriers. To address this, we offered both face-to-face and remote treatment options and included financial support for travel costs. These elements may not only reduce attrition but also improve acceptability in a population traditionally underserved in psychosocial research. Given the cultural variability in SAEF attributions and healthcare access, results may not generalize to all populations. However, the transdiagnostic and modular structure of the CBT program could facilitate adaptation across contexts. While therapists are trained and supervised, future studies may benefit from structured fidelity assessments (e.g., audio recordings reviewed against treatment checklists) to ensure adherence and enable analysis of therapist effects. The hybrid delivery model also provides an opportunity to assess whether certain SAEF subgroups (e.g., those with EMF-related fears) show differential engagement with online vs. in-person sessions—an issue that may have implications for scaling up interventions. In summary, we think this study could improve our current understanding of SAEF and their treatment.

Abbreviations

ACT

Acceptance and Commitment Therapy

BPQ-12

Body Perception Questionnaire

CBT

Cognitive Behavioral Therapy

CCA

Climate Change Anxiety Scale

EG

Experimental Group

EMFSS

Electromagnetic Field Sensitivity Scale

HADS

Hospital Anxiety and Depression Scale

ITT

Intention to treat

IU

Intolerance of uncertainty scale

LEAS

Levels of Emotional Awareness Scale

MBCT

Mindfulness-based cognitive therapy

MHWS

Modern Health Worries Scale

QEESI

Quick Environment Exposure Sensitivity Index

rmANOVA

Repeated measures analysis of variance

SAEF

Symptoms associated with environmental factors

SF-12

12-item Short-Form Health Survey

SIQ

Symptom Interpretation Questionnaire

SSD-12

Somatic Symptom Disorder – B Criteria Scale

WLCG

Waitlist Control Group

Authors’ contributions

VP is the lead physician and investigator; he conceived the study, conceptualized the study, led the protocol development and project administration, will perform large parts of the investigation, and supervises the team. TP will be the lead data analyst; she helped with the methodological decisions, wrote large parts of this manuscript, and will oversee data curation, formal analysis, and visualization. IZ is a psychologist, she designed first versions of the intervention and wrote preliminary drafts of the manuscript. MRT is a physician and co-investigator. CFD and FZI are the psychotherapists - they contributed to the intervention design and investigation, and support the project administration. MW, DL, LBC, and OVdB helped conceptualize the study and provided valuable feedback throughout the development and administration of the study. CL is the principal investigator; he conceived the study, acquired funding, validated the feasibility and utility of the protocol, and provided guidance to the junior team members.

Funding

This study was supported by the Agence nationale de sécurité sanitaire de l’alimentation, de l’environnement et du travail (PNREST ANSES-21-RF-17), and the Fondation pour la recherche médicale (ENV202109013786).

Data availability

Investigators and members of the research team will access data after collection is over. To promote transparency and reproducibility, we intend to make the full trial protocol, de-identified participant-level dataset, and statistical analysis code publicly available. Study findings will be published in open-access journals, and data along with analysis scripts will be deposited on OSF (Open Science Framework) or a similar open-access repository.

Declarations

Ethics approval and consent to participate

This research is part of clinical trial C22-19 BELIEF sponsored by Inserm. It was granted approval by local Ethics Committee or “Comité de Protection des Personnes” Ile de France II on 2022 December 16, and registered in a public trials registry (CT 05973214). All study participants give their informed, consent written to participation, in line with French regulations.

Consent for publication

We are willing to provide a model consent form upon request.

Competing interests

The authors declare no competing interests.