Towards an effective treatment for fibromyalgia: A randomized, controlled trial comparing exposure-based treatment to traditional cognitive behavior therapy
Fibromyalgia (FM) is a common disorder characterized by generalized pain and fatigue as well as many other somatic and psychiatric symptoms. FM leads to large suffering, high disability, and large societal costs. The most evaluated treatment is traditional cognitive behavior therapy (CBT), which has only shown modest effects in FM.
We have previously developed and evaluated exposure-based CBT for patients with FM. Exposure involves to voluntarily and systematically approach avoided situations and stimuli associated with pain and discomfort. Participants (N=140) were randomized to 10 weeks of internet-delivered exposure therapy (iExp) or waitlist control. Results from the trial showed that iExp rendered large effects on FM symptoms, pain, fatigue, disability and pain-related worry, and that iExp is highly cost-effective compared to wait-list. Thus, the results are promising, but a clear limitation is that iExp has not yet been evaluated against an active control condition.
We are now applying for funds to conduct a large-scale randomized controlled trial comparing iExp to traditional CBT. In addition to studying clinical effects we will also evaluate cost-effectiveness and investigate underlying factors that drive the treatment effect, in order to further enhance the treatment.
Positive results from this project would imply that iExp can soon be implemented within regular healthcare and distributed to patients with FM nationally.
Final report
Project purpose and development
The overall purpose of this project was to create empirical evidence for a potentially highly efficacious treatment for fibromyalgia, a chronic pain disorder, that can be swiftly disseminated on a large scale. The specific aims were to investigate efficacy, cost-effectiveness, moderators and mediators of internet-delivered exposure therapy (Exp-CBT) compared to internet-delivered traditional CBT (T-CBT) in a large-scale RCT.
The project has progressed largely according to plan. However, some changes have been made to ensure good feasibility and improved scientific quality. We changed the name of the project from “Randomized controlled trial of internet-based exposure-based cognitive behavioral therapy compared with internet-based traditional cognitive behavioral therapy for fibromyalgia” to “Randomized controlled trial of internet-based cognitive behavioral therapy for fibromyalgia”. This was done to increase the likelihood of blinding the participants, i.e., that the participants would not be able to form an opinion in advance that one of the treatments was assumed to be more effective than the other. We also added four questions to the post-measurement survey, asking whether the participant knew the names of the different treatments, whether they knew or could guess which one the researchers believed to be most effective, and whether they knew which treatment they had been assigned to. This was done to evaluate whether the blinding of participants had been successful.
Another change in the project is that we decided to increase the total N by 14 participants, to a total of 274 (planned N = 260). The change was made during the winter of 2021-22 when we had a short break in treatments during the Christmas holidays in the first two cohorts of the study. After the break, we continued to have a relatively low completion rate for weekly measurements, which risked reducing the study’s power. This contributed to the decision to increase the number of participants to ensure sufficient statistical power.
Minor adjustments to the implementation included the addition of a questionnaire to measure a potential additional mediator, physical activity measured by the Godin Leisure-Time Exercise Questionnaire, to control for increased physical activity in participants who reported less avoidance (i.e., that any treatment effect for participants who reported less pain-related avoidance over time was not mainly due to them becoming more physically active). We also replaced the PHQ-9 form with the PHQ-2 and the GAD-7 with the GAD-2, as these have equivalent psychometric properties but with fewer questions.
Project implementation
The main data collection took place between the fall of 2021 to spring of 2022 through four consecutive cohorts of participants who were randomized to either internet-based exposure-based CBT or internet-based traditional CBT. The cohorts were in groups of 6 (3+3), 64 (32+32), 186 (93+93), and 18 (9+9) participants, respectively. Of note, cohort three coincided with the spread of a new mutation of the SARS-COV-2 virus in the community and a significant proportion of both research subjects and therapists falling ill with COVID-19 during the 10 weeks of treatment. The final post-measurement was collected in early June 2022 and the final 12-month follow-up was collected in May 2023. We received 100% of the data at pre-treatment, 95% at post-treatment, and 86% at the 12-month follow-up after post-treatment assessment, which, in the context of treatment research, is considered a low degree of data loss.
After completing data collection, the data was analyzed according to plan, i.e., regarding clinical effects, mediators, and moderators, as well as cost-effectiveness. We have also published a fourth scientific publication as a result of one of the four psychology student master’s theses written within the framework of the project.
The three most important results of the project, as well as a discussion of the project’s conclusions
Evaluation of treatment efficacy and treatment moderators
Contrary to the study hypothesis, this large-scale randomized study showed that Exp-CBT was not superior to T-CBT in reducing the severity of fibromyalgia during the 10-week treatment period (Cohen’s d = 0.10, P = .544). Both treatments had moderate to large within-group effects (Cohen’s d = 0.74 – 0.84) that persisted up to 12 months after treatment. Exp-CBT had a slightly greater effect on functional impairment, a difference that was greater at the 12-month follow-up (d = 0.33). T-CBT had a slightly greater effect on symptoms of depression, possibly because it includes behavioral activation, which is the gold standard treatment for depression, although this difference was no longer significant at the 12-month follow-up.
Moderator analyses showed that the relative effect of Exp-CBT compared to T-CBT was not moderated by age at baseline, educational level, number of years with the diagnosis, total severity, pain intensity, pain avoidance behaviors, or pain catastrophizing. In an unplanned moderator analysis initiated by the peer review process, we found that participants fared better in T-CBT, but not Exp-CBT, if they had a higher level of physical activity at baseline.
The study is the second largest evaluation of psychological treatment for fibromyalgia and one of the few that has compared two existing psychological treatments for chronic pain. The study underscores the value of traditional CBT for fibromyalgia and has contributed to a much-needed debate on the importance of comparing new treatments with other active treatments and not just waiting list controls or treatment-as-usual. In summary, the study has provided empirical evidence that both online exposure-based CBT and online traditional CBT are associated with marked reductions in symptoms and functional impairment in fibromyalgia, but that online exposure-based CBT cannot be said to be significantly more effective than online traditional CBT.
Mediators of treatment effect
Results from the mediation analyses indicated that a reduction in avoidance behaviors, pain catastrophizing, and hypervigilance were significant mediators of change in both treatments. An increase in pacing and a reduction in overdoing were significant mediators in T-CBT only. Physical activity was not a mediator. In other words, exposure-based and traditional CBT for fibromyalgia, two seemingly different variants of CBT appear to share common treatment mediators, namely pain-related avoidance behavior, catastrophizing and hypervigilance. These results provide important insights into potentially clinically relevant processes in these two treatments.
Cost-effectiveness
This was the first study to assess the cost-effectiveness and cost-utility for online exposure-based CBT compared to online traditional CBT. Results indicate the treatments are fairly similar regarding total costs and effectiveness, but that online exposure therapy may be cost-effective to online traditional CBT, as each incremental treatment responder was associated with a societal cost reduction of $16,884. There was no significant difference in total accumulated costs, but exposure therapy had significantly lower costs for work cutback and domestic loss, whereas T-CBT had slightly lower therapist costs.
Our conclusion from the study is that internet-based exposure-based CBT may be cost-effective compared to traditional CBT. However, given that there were no significant differences in total costs nor clinical effectiveness, we cannot conclude that Exp-CBT is unambiguously cost-effective compared to T-CBT. Given the relatively small difference in costs and effects between EXP-CBT and T-CBT in the current study, one might consider basing decisions between treatments also on other factors, such as patient preferences and the ease of training therapists. Although not addressed in the current study, against this backdrop it is reasonable to believe that both EXP-CBT and T-CBT might be competitive treatment options compared to face-to-face CBT from a cost-effectiveness perspective.
New research questions
An interesting question is why the observed within-group effects in Exp-CBT were slightly lower in this study (d = 0.77) than in the previous waitlist-controlled RCT (d = 0.84) We are therefore now planning a study aimed at investigating what factors might explain the variation in treatment response for Exp-CBT participants. Further research is also needed to distinguish the mechanisms underlying Exp-CBT and T-CBT, to determine the specific contribution of each treatment's components to the overall treatment effect, and to identify better methods for predicting treatment response. Another venue for future research is to test the treatments in a clinical context and compare the cost-effectiveness to treatment as usual.
Dissemination of the project and its results and collaborations
The project has so far resulted in three scientific publications. Results regarding clinical effects were published in December 2023 in the journal PAIN (JIF 5.9). Results regarding mediators and moderators were published in April 2024 in Behaviour Research and Therapy (JIF 4.2). Results regarding cost-effectiveness were published in May 2025 in The Journal of Pain (JIF 4). The project has also been presented at two scientific conferences: Swedish Pain Society Västerås in October 2021, and The 14th Swedish Congress on Internet Interventions (SweSRII) in Stockholm in June 2025. The collaboration took place because the project is a joint effort between Karolinska Institutet and Uppsala University. Monica Buhrman at Uppsala University has been involved in the research design and questions, writing and revising of the manuscripts, recruitment and supervision of therapists at Uppsala University. She is also the author of the treatment manual in the control condition.
List of publications
Hedman-Lagerlöf, M; Gasslander, N; Ahnlund Hoffmann, A; Bragesjö, M; Etzell, A; Ezra, S; Frostell, E; Hedman-Lagerlöf, E; Ivert, C; Liliequist, B; Ljótsson, Bc; Hoppe, J M; Palmgren, J; Spansk, E; Sundström, F; Särnholm, J; Tzavara, G; Buhrman, M; & Axelsson, E. (2023). Effect of exposure-based vs traditional cognitive behavior therapy for fibromyalgia: a two-site single-blind randomized controlled trial. PAIN, 165(6), 1278–1288. PMID: 38131181, DOI: 10.1097/j.pain.0000000000003128
Hedman-Lagerlöf, M., Buhrman, M., Hedman-Lagerlöf, E., Ljótsson, B., & Axelsson, E. (2024). Shared and Distinct Effect Mediators in Exposure-based and Traditional Cognitive Behavior Therapy for Fibromyalgia: Secondary Analysis of a Randomized Controlled Trial. Behaviour Research and Therapy, 178, 104546. PMID: 38685153, DOI: 10.1016/j.brat.2024.104546
Hedman-Lagerlöf, M., Hedman-Lagerlöf, E., Buhrman, M., & Axelsson, E. (2025). Cost-effectiveness and cost-utility of exposure-based vs. traditional cognitive behavior therapy for fibromyalgia: results from a randomized controlled trial. The Journal of Pain, 105375. PMID: 40107589, DOI: 10.1016/j.jpain.2025.105375
Hoffmann, A. A., Frostell, E., Hedman-Lagerlöf, M., McCracken, L. M., & Axelsson, E. (2024). Psychometric evaluation of the Pain Vigilance and Awareness Questionnaire (PVAQ) in fibromyalgia: Confirmatory factor analysis and the development of a Swedish 8-item version. European Journal of Pain, 28(3), 396-407. PMID: 37759360, DOI: 10.1002/ejp.2188
All publications are published with open access, CC-BY-license, as stipulated by RJ. This has been assured by choosing open access during the publication process and specifying CC-BY-license upon request.
Master’s theses in psychology 30 credits
Amanda Etzell och Georgia Tzavara. (2022). ”Katastrofiering hos personer med fibromyalgi: en svensk validering av Pain Catastrophizing Scale”. Uppsala Universitet, Institutionen för psykologi. https://uu.diva-portal.org/smash/record.jsf?dswid=-7477&pid=diva2%3A1750234&c=1&searchType=UNDERGRADUATE&language=sv&query=&af=%5B%5D&aq=%5B%5B%7B%22freeText%22%3A%22amanda+etzell%22%7D%5D%5D&aq2=%5B%5B%5D%5D&aqe=%5B%5D&noOfRows=50&sortOrder=author_sort_asc&sortOrder2=title_sort_asc&onlyFullText=false&sf=all
Edward Spansk. (2022). ”Smärtkatastrofiering och fysisk aktivitet hos fibromyalgipatienter vid olika former av internetbaserad kognitiv beteendeterapi”. Uppsala Universitet, Institutionen för psykologi. https://uu.diva-portal.org/smash/record.jsf?dswid=-7477&pid=diva2%3A1750208&c=1&searchType=UNDERGRADUATE&language=sv&query=&af=%5B%5D&aq=%5B%5B%7B%22freeText%22%3A%22edward+spansk%22%7D%5D%5D&aq2=%5B%5B%5D%5D&aqe=%5B%5D&noOfRows=50&sortOrder=author_sort_asc&sortOrder2=title_sort_asc&onlyFullText=false&sf=all”
Caroline Ivert och Josefin Palmgren. (2022).”Validity and reliability of the Swedish translation of Fatigue Severity Scale assessment tool in a clinical sample of patients with Fibromyalgia”. Karolinska Institutet, Institutionen för klinisk neurovetenskap.
Alice Ahnlund Hoffmann och Elsa Frostell. (2022). ”Psychometric evaluation of a Swedish translation of the Pain Vigilance and Awareness Questionnaire (PVAQ) In a sample of Swedish participants with fibromyalgia.”. Karolinska Institutet, Institutionen för klinisk neurovetenskap.
Press release Karolinska Institutet website
Information about the project's results has also been disseminated via press release at Karolinska Institutet
https://nyheter.ki.se/tva-typer-av-kbt-lika-effektiva-vid-behandling-av-fibromyalgi
The overall purpose of this project was to create empirical evidence for a potentially highly efficacious treatment for fibromyalgia, a chronic pain disorder, that can be swiftly disseminated on a large scale. The specific aims were to investigate efficacy, cost-effectiveness, moderators and mediators of internet-delivered exposure therapy (Exp-CBT) compared to internet-delivered traditional CBT (T-CBT) in a large-scale RCT.
The project has progressed largely according to plan. However, some changes have been made to ensure good feasibility and improved scientific quality. We changed the name of the project from “Randomized controlled trial of internet-based exposure-based cognitive behavioral therapy compared with internet-based traditional cognitive behavioral therapy for fibromyalgia” to “Randomized controlled trial of internet-based cognitive behavioral therapy for fibromyalgia”. This was done to increase the likelihood of blinding the participants, i.e., that the participants would not be able to form an opinion in advance that one of the treatments was assumed to be more effective than the other. We also added four questions to the post-measurement survey, asking whether the participant knew the names of the different treatments, whether they knew or could guess which one the researchers believed to be most effective, and whether they knew which treatment they had been assigned to. This was done to evaluate whether the blinding of participants had been successful.
Another change in the project is that we decided to increase the total N by 14 participants, to a total of 274 (planned N = 260). The change was made during the winter of 2021-22 when we had a short break in treatments during the Christmas holidays in the first two cohorts of the study. After the break, we continued to have a relatively low completion rate for weekly measurements, which risked reducing the study’s power. This contributed to the decision to increase the number of participants to ensure sufficient statistical power.
Minor adjustments to the implementation included the addition of a questionnaire to measure a potential additional mediator, physical activity measured by the Godin Leisure-Time Exercise Questionnaire, to control for increased physical activity in participants who reported less avoidance (i.e., that any treatment effect for participants who reported less pain-related avoidance over time was not mainly due to them becoming more physically active). We also replaced the PHQ-9 form with the PHQ-2 and the GAD-7 with the GAD-2, as these have equivalent psychometric properties but with fewer questions.
Project implementation
The main data collection took place between the fall of 2021 to spring of 2022 through four consecutive cohorts of participants who were randomized to either internet-based exposure-based CBT or internet-based traditional CBT. The cohorts were in groups of 6 (3+3), 64 (32+32), 186 (93+93), and 18 (9+9) participants, respectively. Of note, cohort three coincided with the spread of a new mutation of the SARS-COV-2 virus in the community and a significant proportion of both research subjects and therapists falling ill with COVID-19 during the 10 weeks of treatment. The final post-measurement was collected in early June 2022 and the final 12-month follow-up was collected in May 2023. We received 100% of the data at pre-treatment, 95% at post-treatment, and 86% at the 12-month follow-up after post-treatment assessment, which, in the context of treatment research, is considered a low degree of data loss.
After completing data collection, the data was analyzed according to plan, i.e., regarding clinical effects, mediators, and moderators, as well as cost-effectiveness. We have also published a fourth scientific publication as a result of one of the four psychology student master’s theses written within the framework of the project.
The three most important results of the project, as well as a discussion of the project’s conclusions
Evaluation of treatment efficacy and treatment moderators
Contrary to the study hypothesis, this large-scale randomized study showed that Exp-CBT was not superior to T-CBT in reducing the severity of fibromyalgia during the 10-week treatment period (Cohen’s d = 0.10, P = .544). Both treatments had moderate to large within-group effects (Cohen’s d = 0.74 – 0.84) that persisted up to 12 months after treatment. Exp-CBT had a slightly greater effect on functional impairment, a difference that was greater at the 12-month follow-up (d = 0.33). T-CBT had a slightly greater effect on symptoms of depression, possibly because it includes behavioral activation, which is the gold standard treatment for depression, although this difference was no longer significant at the 12-month follow-up.
Moderator analyses showed that the relative effect of Exp-CBT compared to T-CBT was not moderated by age at baseline, educational level, number of years with the diagnosis, total severity, pain intensity, pain avoidance behaviors, or pain catastrophizing. In an unplanned moderator analysis initiated by the peer review process, we found that participants fared better in T-CBT, but not Exp-CBT, if they had a higher level of physical activity at baseline.
The study is the second largest evaluation of psychological treatment for fibromyalgia and one of the few that has compared two existing psychological treatments for chronic pain. The study underscores the value of traditional CBT for fibromyalgia and has contributed to a much-needed debate on the importance of comparing new treatments with other active treatments and not just waiting list controls or treatment-as-usual. In summary, the study has provided empirical evidence that both online exposure-based CBT and online traditional CBT are associated with marked reductions in symptoms and functional impairment in fibromyalgia, but that online exposure-based CBT cannot be said to be significantly more effective than online traditional CBT.
Mediators of treatment effect
Results from the mediation analyses indicated that a reduction in avoidance behaviors, pain catastrophizing, and hypervigilance were significant mediators of change in both treatments. An increase in pacing and a reduction in overdoing were significant mediators in T-CBT only. Physical activity was not a mediator. In other words, exposure-based and traditional CBT for fibromyalgia, two seemingly different variants of CBT appear to share common treatment mediators, namely pain-related avoidance behavior, catastrophizing and hypervigilance. These results provide important insights into potentially clinically relevant processes in these two treatments.
Cost-effectiveness
This was the first study to assess the cost-effectiveness and cost-utility for online exposure-based CBT compared to online traditional CBT. Results indicate the treatments are fairly similar regarding total costs and effectiveness, but that online exposure therapy may be cost-effective to online traditional CBT, as each incremental treatment responder was associated with a societal cost reduction of $16,884. There was no significant difference in total accumulated costs, but exposure therapy had significantly lower costs for work cutback and domestic loss, whereas T-CBT had slightly lower therapist costs.
Our conclusion from the study is that internet-based exposure-based CBT may be cost-effective compared to traditional CBT. However, given that there were no significant differences in total costs nor clinical effectiveness, we cannot conclude that Exp-CBT is unambiguously cost-effective compared to T-CBT. Given the relatively small difference in costs and effects between EXP-CBT and T-CBT in the current study, one might consider basing decisions between treatments also on other factors, such as patient preferences and the ease of training therapists. Although not addressed in the current study, against this backdrop it is reasonable to believe that both EXP-CBT and T-CBT might be competitive treatment options compared to face-to-face CBT from a cost-effectiveness perspective.
New research questions
An interesting question is why the observed within-group effects in Exp-CBT were slightly lower in this study (d = 0.77) than in the previous waitlist-controlled RCT (d = 0.84) We are therefore now planning a study aimed at investigating what factors might explain the variation in treatment response for Exp-CBT participants. Further research is also needed to distinguish the mechanisms underlying Exp-CBT and T-CBT, to determine the specific contribution of each treatment's components to the overall treatment effect, and to identify better methods for predicting treatment response. Another venue for future research is to test the treatments in a clinical context and compare the cost-effectiveness to treatment as usual.
Dissemination of the project and its results and collaborations
The project has so far resulted in three scientific publications. Results regarding clinical effects were published in December 2023 in the journal PAIN (JIF 5.9). Results regarding mediators and moderators were published in April 2024 in Behaviour Research and Therapy (JIF 4.2). Results regarding cost-effectiveness were published in May 2025 in The Journal of Pain (JIF 4). The project has also been presented at two scientific conferences: Swedish Pain Society Västerås in October 2021, and The 14th Swedish Congress on Internet Interventions (SweSRII) in Stockholm in June 2025. The collaboration took place because the project is a joint effort between Karolinska Institutet and Uppsala University. Monica Buhrman at Uppsala University has been involved in the research design and questions, writing and revising of the manuscripts, recruitment and supervision of therapists at Uppsala University. She is also the author of the treatment manual in the control condition.
List of publications
Hedman-Lagerlöf, M; Gasslander, N; Ahnlund Hoffmann, A; Bragesjö, M; Etzell, A; Ezra, S; Frostell, E; Hedman-Lagerlöf, E; Ivert, C; Liliequist, B; Ljótsson, Bc; Hoppe, J M; Palmgren, J; Spansk, E; Sundström, F; Särnholm, J; Tzavara, G; Buhrman, M; & Axelsson, E. (2023). Effect of exposure-based vs traditional cognitive behavior therapy for fibromyalgia: a two-site single-blind randomized controlled trial. PAIN, 165(6), 1278–1288. PMID: 38131181, DOI: 10.1097/j.pain.0000000000003128
Hedman-Lagerlöf, M., Buhrman, M., Hedman-Lagerlöf, E., Ljótsson, B., & Axelsson, E. (2024). Shared and Distinct Effect Mediators in Exposure-based and Traditional Cognitive Behavior Therapy for Fibromyalgia: Secondary Analysis of a Randomized Controlled Trial. Behaviour Research and Therapy, 178, 104546. PMID: 38685153, DOI: 10.1016/j.brat.2024.104546
Hedman-Lagerlöf, M., Hedman-Lagerlöf, E., Buhrman, M., & Axelsson, E. (2025). Cost-effectiveness and cost-utility of exposure-based vs. traditional cognitive behavior therapy for fibromyalgia: results from a randomized controlled trial. The Journal of Pain, 105375. PMID: 40107589, DOI: 10.1016/j.jpain.2025.105375
Hoffmann, A. A., Frostell, E., Hedman-Lagerlöf, M., McCracken, L. M., & Axelsson, E. (2024). Psychometric evaluation of the Pain Vigilance and Awareness Questionnaire (PVAQ) in fibromyalgia: Confirmatory factor analysis and the development of a Swedish 8-item version. European Journal of Pain, 28(3), 396-407. PMID: 37759360, DOI: 10.1002/ejp.2188
All publications are published with open access, CC-BY-license, as stipulated by RJ. This has been assured by choosing open access during the publication process and specifying CC-BY-license upon request.
Master’s theses in psychology 30 credits
Amanda Etzell och Georgia Tzavara. (2022). ”Katastrofiering hos personer med fibromyalgi: en svensk validering av Pain Catastrophizing Scale”. Uppsala Universitet, Institutionen för psykologi. https://uu.diva-portal.org/smash/record.jsf?dswid=-7477&pid=diva2%3A1750234&c=1&searchType=UNDERGRADUATE&language=sv&query=&af=%5B%5D&aq=%5B%5B%7B%22freeText%22%3A%22amanda+etzell%22%7D%5D%5D&aq2=%5B%5B%5D%5D&aqe=%5B%5D&noOfRows=50&sortOrder=author_sort_asc&sortOrder2=title_sort_asc&onlyFullText=false&sf=all
Edward Spansk. (2022). ”Smärtkatastrofiering och fysisk aktivitet hos fibromyalgipatienter vid olika former av internetbaserad kognitiv beteendeterapi”. Uppsala Universitet, Institutionen för psykologi. https://uu.diva-portal.org/smash/record.jsf?dswid=-7477&pid=diva2%3A1750208&c=1&searchType=UNDERGRADUATE&language=sv&query=&af=%5B%5D&aq=%5B%5B%7B%22freeText%22%3A%22edward+spansk%22%7D%5D%5D&aq2=%5B%5B%5D%5D&aqe=%5B%5D&noOfRows=50&sortOrder=author_sort_asc&sortOrder2=title_sort_asc&onlyFullText=false&sf=all”
Caroline Ivert och Josefin Palmgren. (2022).”Validity and reliability of the Swedish translation of Fatigue Severity Scale assessment tool in a clinical sample of patients with Fibromyalgia”. Karolinska Institutet, Institutionen för klinisk neurovetenskap.
Alice Ahnlund Hoffmann och Elsa Frostell. (2022). ”Psychometric evaluation of a Swedish translation of the Pain Vigilance and Awareness Questionnaire (PVAQ) In a sample of Swedish participants with fibromyalgia.”. Karolinska Institutet, Institutionen för klinisk neurovetenskap.
Press release Karolinska Institutet website
Information about the project's results has also been disseminated via press release at Karolinska Institutet
https://nyheter.ki.se/tva-typer-av-kbt-lika-effektiva-vid-behandling-av-fibromyalgi