Development of a core outcome set and outcome definitions for studies on uterus-sparing treatments of adenomyosis (COSAR): an international multistakeholder-modified Delphi consensus study

Adenomyosis is a benign condition affecting the uterus, and it can be associated with a significant symptom burden, mainly painful or heavy menstrual periods ( Li et al. , 2014 ), chronic pelvic pain and reduced quality of life (QoL) ( Li et al. , 2014 ; Choi et al. , 2017 ). The disorder has also been associated with reduced fertility and poor reproductive and obstetric outcomes, including an increased risk of miscarriage, pre-eclampsia, pre-term delivery and postpartum haemorrhage ( Tamura et al. , 2017 ; Younes et al. , 2017 ; Bruun et al. , 2018 ; Hashimoto et al. , 2018 ; Bourdon et al. , 2020 ). Adenomyosis is assumed to be present in about 20% of women attending a general gynaecology clinic ( Naftalin et al. , 2012 ), and is also common in women undergoing infertility treatment ( Puente et al. , 2016 ). In recent years, it became evident that young women also suffer from adenomyosis ( Chapron et al. , 2020 ). Despite the seemingly high prevalence of adenomyosis and its clinical relevance, there is a lack of well-designed clinical trials comparing different options for treating adenomyosis. Furthermore, studies investigating therapeutic interventions for adenomyosis have used many different outcomes and outcome measures, making it challenging to perform a meta-analysis and thus severely curtailing the usefulness of research to inform clinical practice and guidelines ( Tellum et al. , 2021a ). As demonstrated in a previous review, only a few studies on adenomyosis report patient-centred outcomes ( Tellum et al. , 2021a ), which questions their benefit to patients. The selection of appropriate outcomes is crucial when designing clinical trials that evaluate the effects of different interventions. Requiring a standardized set of consensus core outcomes minimizes the risk of bias that results from the investigator ‘cherry-picking’ positive or attractive results for inclusion with the omission of negative or less interesting evidence ( Dwan et al. , 2013 ). The development and use of so-called core outcome sets (COS) are widely supported and encouraged in medicine, and has led to the development of a variety of COS under the umbrella of the Core Outcome Measures in Effectiveness Trials (COMET) and Core Outcomes in Women’s and Newborn Health (CROWN) initiatives. In order to fill this void in the field of adenomyosis, the aim of this consensus study was to develop a COS for adenomyosis research (COSAR) on uterus-sparing interventional studies for symptoms associated with adenomyosis, which was defined as the presence of ectopic endometrial cells and stroma within the myometrium. The scope of the COSAR includes all types of treatment of adenomyosis in premenopausal women and can be applied to all kinds of prospective studies. We asked the women who participated in the focus group workshops and the patient advocacy members how would they like to be referred to while taking part in this consensus development (lay consumer, public research partner or patient), and ‘patient’ was the term they preferred.

The Delphi was piloted between May 10 and 15, 2021, the first round was disseminated and remained open between May 20 and July 30, 2021, and the second round between August 29 and November 04, 2021. The questionnaires and the detailed responses sorted by stakeholder group are provided in Supplementary Data Files S1 and S2 . The Steering Committee consensus meeting was held on December 04, 2021. In the first round, 501 respondents from 48 countries completed the survey, including 327 patients, 7 partners, 17 researchers and 150 health care professionals (HCPs). Figure 2 displays an overview of the countries represented; the participants according to country and stakeholder group are listed in Supplementary Table SI . Amongst HCPs, there were two chiropractors, two physical therapists, six radiologists or ultrasonologists, two midwives and one general practitioner; all other respondents were gynaecologists. In the second round, there were 291 participants from 38 countries resulting in a return rate of 58.1% ( Fig. 1 ). There was no statistically significant difference in distribution between the stakeholder groups between the two rounds ( P  > 0.2) ( Fig. 1 ). Owing to the high attrition rate, we analysed the results for differences between the two groups (those that did and did not return in round two) and found none. World map, colours indicating the number of participants from each country . The final COSAR is displayed in Table I and the response rates for the included items provided in Table II . The detailed response rates per outcome are provided in Supplementary Data Files S1 and S2 . The discussion leading to the final COS is outlined below. Definitions for each outcome were determined by the Steering Committee, and they can be found in Table III , where the lay terms used are also listed. Ten members attended the final Steering Committee consensus meeting. The Committee recognized that symptoms suggested by patients as outcomes could be caused by concomitant disorders, such as endometriosis, and not necessarily by adenomyosis. After discussion, the Steering Committee decided it would neither be practical nor clinically meaningful to restrict this COS to women with adenomyosis alone and exclude those with endometriosis or fibroids, as the conditions frequently occur together. Furthermore, the Steering Committee noted a lack of evidence to determine which symptoms would be caused by adenomyosis alone. The overall view was that outcomes should not be prejudged and that the example set by the patients should be followed when discussing the inclusion of symptoms unless there was a strong reason not to. If future studies show that some symptoms are not associated with adenomyosis, they could be excluded at that time. Overview of the core outcome sets for studies on uterus-sparing treatments of adenomyosis, structured into core domains. Outcome(s) applicable in certain study types only. Rating of the outcomes in the final core outcome set, by stakeholder group. Ratings were given on a Likert scale ranging from 1 to 9. ‘Exclusion’ means 70% or more scoring it as 1–3 and fewer than 15% scoring it as 7–9. Consensus for an outcome being included in the core outcome set (COS) required 70% or more scoring it as 7–9 and fewer than 15% to score it as 1–3. HCP, health care professional; OC, outcome. Definitions and lay terms for the outcomes of the core outcome set. superficial incisional, affecting the skin and subcutaneous tissue. deep incisional, affecting the fascial and muscle layers. organ or space infection, which involves any part of the anatomy other than the incision that is opened or manipulated during the surgical procedure, for example joint or peritoneum. We advise to use the WHO adverse drug reaction terminology for the specific drug reactions. WHO classifies ADR into six classes: Type A reactions (dose-related)—exaggerated but otherwise normal pharmacological response to the effects of the medicines given at a therapeutic dose. The reaction is treated by reducing the dose or withholding the medicine and considering alternative therapy. Type B reactions (non-dose related)—bizarre and unpredictable response with no relation to the dose or pharmacological action of the medicine, that is often allergic in nature. They are uncommon but are often severe and cause high mortality. Type C reactions (dose-related and time-related)—chronic (long term) and related to cumulative dose. The reaction is treated by reducing the dose or withholding the medicine, which may have to be withheld for a long time. Type D reactions (time related)—delayed (i.e. have a lag time) after the use of a drug. They are uncommon but their treatment is often intractable. Type E reactions (withdrawal)—these reactions occur soon after the end of use (i.e. withdrawal) and are uncommon. The reaction is treated by reintroducing the medicine and then withdrawing it slowly. Type F reactions (unexpected failure of efficacy)—these reactions occur when there is a failure of efficacy. Such reactions are common, may be dose-related and are often caused by drug interactions. The reaction is treated by increasing the dose and considering the effects of concomitant therapy. AAGL, American Association of Gynecologic Laparoscopists; ESGE, European Society for Gynaecological Endoscopy; WES, World Endometriosis Society; RCOG, The Royal College of Obstetricians and Gynaecologists; NICE, National Institute for Health and Clinical Excellence; COSAR, Core Outcome Set in Adenomyosis Research; FIGO, International Federation of Gynecology and Obstetrics; WHO, World Health Organization; CDC, Centers for Disease Control. Dysmenorrhea, cyclic pelvic pain, dyspareunia, non-cyclic untriggered pain (including pain during ovulation) and feeling bulky/pelvic pressure symptoms reached the threshold for agreement and were included ( Table II ). Dyschezia reached 73% agreement in total, but the threshold for inclusion was only reached among patients when analysed at the stakeholder level. In the final session, there was complete agreement to include this in the final COS. Two items reached >70% agreement amongst patients, but not overall, and were therefore discussed. Pain radiating to legs was suggested as an additional item in the first round and supported by 78.5% of patients, but only <50% of HCP and researchers. Several Steering Committee members pointed out that their clinical experience showed that this symptom was indeed present in many patients with adenomyosis. A proposed, albeit unproven, mechanism for this was pain radiating through the uterosacral ligaments. Inclusion of the item was strongly advocated by the patient representative. In line with the initial discussion as outlined above, it was agreed to define this item as ‘radiating pain to lower back and/or extremities during menstruation’ and to include it in the COS. However, the group pointed out that this item needed future re-evaluation. A new item suggested during the Delphi, bloating, reached only 64.6% agreement overall, with a high agreement (78.5%) amongst patients but less than 50% of HCPs and researchers. After discussion, it was decided that this symptom is too non-specific and not a priority for inclusion in this version of the COS, an approach supported by the patient representative. Several items that were added to the long list during the patient workshop were left undecided after two rounds of the Delphi. These were dysuria, pain when the bladder is full and pain-associated vomiting. In the final session, there was full agreement that all these outcomes were too non-specific, and therefore were not included in the final COS. However, all agreed that these items need more scientific investigation and could be potentially included in the future. The symptom of urinary frequency achieved 71.9% agreement amongst patients but only 63.5% agreement overall. The Steering Committee determined that this is a specific and easily measurable symptom that might serve as a proxy measure of uterine size or disease severity. There was complete agreement for inclusion. The other items in this category, urge symptoms, residual urine and urinary incontinence, did not reach the threshold for exclusion or inclusion overall. The patient representative pointed out it might be less clear with urinary symptoms what is normal. After discussion, there was a unanimous agreement to exclude all three items. Blood flow volume, and duration of bleeding was supported by all participants through round one and included ( Table II ). Other items in this category reached agreement in some stakeholder groups but did not reach the threshold for inclusion overall ( Table II , Supplementary Data File S2 ). However, the Steering Committee pointed out that there is an evidence-based and internationally established system for describing normal and abnormal uterine bleeding (AUB) that should be considered equal to a COS ( Munro et al. , 2018 ). Consequently, the Steering Committee unanimously supported the inclusion of the elements of AUB System 1 of the International Federation of Gynecology and Obstetrics (FIGO) that describe the frequency and regularity of the menstrual cycle, duration and suspected volume of the menstrual period, and the presence of intermenstrual bleeding. The Steering Committee considered unscheduled bleeding on medication that suppresses gonadal steroids to be a side effect that should be monitored under the harms category. The item ‘coital bleeding’ was supported by patients in the second round (74.6% agreement in this group) but not by the participants overall. This item was discussed at length where some Steering Committee members argued for exclusion because the symptom is too non-specific with an uncertain mechanism. However, given the support of patients, the opinion of the patient representative, and the previous example of dyschezia, the item was put to the vote by the Steering Committee, where the majority supported inclusion in principle. However, because the term ‘coital bleeding’ was considered ambiguous, the Steering Committee unanimously supported the alternative and more encompassing description ‘bleeding associated with sexual activity’, in recognition that this symptom could occur outside penetrative sex. Also, most of the Steering Committee voted to move this newly named item to the category ‘quality of life’ under the concept of ‘sexual function’. The Steering Committee suggested mandatory reporting of the items in this category for all studies evaluating interventions designed to improve reproductive outcomes and recommended them when the study design includes women wishing for future pregnancy. There is an existing COS on infertility ( Duffy et al. , 2018 , 2020 ) ( Table I ) that received consensus support in the first Delphi round ( Table II ). While it was acknowledged that not the entire infertility COS might be relevant for adenomyosis research, selective inclusion was an issue, so all items were included in the adenomyosis COS. The Steering Committee, however, decided to modify two of the definitions in the infertility COS ( Duffy et al. , 2020 ). The first was to use the terms ‘live’ rather than ‘viable’ for early pregnancies; and to describe eutopic pregnancies as ‘normally sited (eutopic)’ rather than ‘intrauterine’, definitions that agree with the ESHRE terminology on ectopic pregnancies ( Kirk et al. , 2020 ). The Delphi participants supported the inclusion of three additional items to this domain, as they were considered relevant in the context of adenomyosis. Overall, while 72.9% supported the inclusion of placentation disorders, this support rose to 90.3% amongst HCPs and consequently was included in the COS. Although the mode of delivery reached the threshold of consensus only amongst HCPs, and only in the second round (72.8%, 58.1% overall agreement), the Steering Committee considered it highly relevant, particularly following uterine sparing procedural interventions. Consequently, the mode of delivery is included in the COS. The same rationale led to the inclusion of postpartum haemorrhage, an item that reached 76.7% and 70.6% agreement amongst HCPs and researchers, respectively, but only 62.7% in the overall participant cohort ( Supplementary Data File S2 ). Several haematological and hormonal laboratory parameters were presented through the long list, of which only ferritin (71.3% total) and haemoglobin (70.1% total) reached the threshold for consensus support. However, since these items represent iron deficiency or anaemia outcome measures rather than independent outcomes, the Steering Committee unanimously agreed to include anaemia as an outcome and not recommend any specific biochemical outcomes. Of note, both CA 125 and oestrogen levels were the only two items in round one to be clearly excluded by both HCPs and researchers, with 22–35% vote for exclusion ( Supplementary Data File S1 ), and almost reached the exclusion threshold overall in the second round with 14.4%. QoL and health-related QoL (HR QoL) are constructs that comprise several domains and sub-items ( Centers for Disease Control and Prevention, 2000 ). However, during the work on the long list, the Steering Committee considered it counterproductive to let the participants vote on each specific item that was identified in this category through the systematic review ( Tellum et al. , 2021a ), as the length of the list and lack of translation might undermine the validity of the results. Consequently, it was concluded that identification of the disease-specific items relating to HR QoL should be carried out under the scope of a different study and that the concepts of QoL should be presented as one single item. HR QoL was overwhelmingly supported for inclusion in the first Delphi round (96.8% agreement), as was sexual function, which achieved consensus with 91% of participants indicating support. As stated above, the newly termed item ‘bleeding during sexual activity’ was included in this category, as is dyspareunia, which can be included in both categories, sexual function or pain. The Steering Committee recommended the inclusion of all items in this category that the participants in the Delphi process supported. Whereas no item is adenomyosis specific, all are generally considered important for clinical trials ( WHO, 2006 ). Consequently, the following items were included in the COS (with the rate of support): patient adherence to treatment (86.4%), patient satisfaction (91%), discomfort during procedure (86%), recovery time (82.8%) (time to full recovery of normal activities), symptom relief rate (91.2%), any symptom recurrence (90.2%), most bothersome symptom recurrence (88%), length of hospital stay (67.5%) and need for re-intervention or a repeat procedure (82%). There was discussion regarding the lesion size and uterine volume, each of which was supported for inclusion by 80.2% and 72.3%, respectively, of the participants. Members of the Steering Committee offered that uterine volume is only a proxy for disease burden, with limited available evidence demonstrating an association with symptoms or outcomes. Similarly, and while lesion size may reflect treatment effects, the inter-rater reliability of lesional metrics remains challenging. While all agreed that disease burden should be measured according to a unified classification, lesion size and uterine volume can serve as interim outcomes pending the development and general acceptance of a consensus, imaging-based, adenomyosis reporting system. It was further agreed that outcomes that are applicable to specific types of interventions only, such as procedure time, technical parameters (type and amount of energy used) or weight of removed tissue, should not be included in the COS. Still, they should be reported as appropriate according to current practice in the respective field. The Steering Committee agreed to exclude health-economic outcomes from the COS as it could be methodologically challenging for many investigators. However, it was acknowledged that, in many environments, the patient-borne treatment costs are an important component of the therapeutic decision-making process and should be reported, despite the difficulty of international comparisons. Consequently, the Steering Committee strongly recommended reporting the results of cost-utility analyses and the overall and patient costs of treatment. The Steering Committee agreed unanimously that harms, infections, and adverse drug reactions should be monitored and reported systematically according to the nature of the intervention ( Table II ). Apart from unscheduled bleeding on hormonal medication, which was proposed in the menstrual bleeding category, the Steering Committee decided not to specify a list of adverse outcomes that should be measured as this would be lengthy, possibly leading to under-documentation of rare unlisted events. Reporting items are not outcomes per se but contain essential information for the interpretation of study results in trials on adenomyosis. The following reporting items were included in the COS (with rates of overall agreement): presence of endometriosis (93%), presence of leiomyomas (80.4%), presence of chronic pelvic pain (96.2%), desire for future pregnancy (85.4%), previous treatment for adenomyosis (87.8%) and classification of adenomyosis (74.5%). The Steering Committee found that recommending a specific classification for endometriosis is outside the project’s mandate but suggests the description of findings according to the recently published expert consensus ( Tomassetti et al. , 2021 ). For leiomyomas, the Steering Committee recommends reporting according to the well-established FIGO-fibroid classification ( Munro et al. , 2018 ). For adenomyosis, adherence to well-defined terminology is recommended until an internationally accepted and validated reporting system is developed ( Van den Bosch et al. , 2015 ; Harmsen et al. , 2022 ). The scope of this work specifically excluded defining adenomyosis diagnostic criteria (imaging, histopathological or other), as the Steering Committee determined that such criteria should be defined by experts based on valid scientific evidence.

The protocol for COSAR was prospectively published ( Tellum et al. , 2021b ) and the project was registered with the COMET initiative (registration number 1649). A Steering Committee was formed comprising specialists with different expertise in the sub-field of adenomyosis (infertility, surgery, diagnostics, basic science) and a patient advocate. In this setting, a patient advocate represents a patient organization and, in contrast to a patient, their focus is not on their personal, lived experience with a condition. The Steering Committee identified three main stakeholder groups to inform the construct of the COS: researchers with expertise in the field, healthcare professionals involved in diagnosis and therapy (doctors, nurses, physical therapists), and patients and their relatives with lived experience of adenomyosis. Potential participants from the personal network of the Steering Committee were contacted directly, as were researchers with highly relevant publications in the field who were identified through a literature search. A systematic web search was used to create a comprehensive, global list of national gynaecological associations and adenomyosis and endometriosis patient advocacy organizations. They were contacted by the managing team with a request to distribute the Delphi survey amongst their members. Participants were further recruited through social media, congresses, and courses and through the network of the World Endometriosis Society (WES) that was represented on the Steering Committee. A website was created to provide information for all participants ( www.cosar.org ). A structured literature review, performed to identify previously reported outcomes ( Tellum et al. , 2021a ), resulted in the creation of a preliminary long list. Additional items were added through patient (focus group) workshops and the Steering Committee ( Fig. 1 ). Finally, using the taxonomy developed and recommended by the COMET initiative ( Dodd et al. , 2018 ) the Steering Committee structured the long list into core areas by removing redundant items and merging others according to concepts ( Fig. 1 ). Lay terms were identified for each item on the long list, first in a workshop with patients for whom English was their native language and then the terms were modified in face-to-face meetings with non-native speakers living in other countries. Flow chart illustrating the Delphi process with inclusion and exclusion of potential outcomes for studies on uterus-sparing treatments of adenomyosis . The Delphi technique is a well-established approach to answering a research question through the identification of a consensus view across subject experts, and it is recommended by the COMET initiative for establishing COS ( Williamson et al. , 2017 ; Barrett et al. , 2020). An electronic Delphi survey was developed on a web-based platform (Nettskjema, University Information Technology Center, University of Oslo, Norway) and piloted with 18 individuals, representing all stakeholder groups, before the launch. After discussion within the Steering Committee and following advice from COMET, it was decided not to perform translations of the survey into languages other than English. Part of the rationale for this decision was time and available funding, and part was the absence of consensus regarding which of the many languages to select. By using non-native English speakers to assess and modify the lay terms of the long list, the Steering Committee tried to ensure that it would be understood by most people. Items on the long list were presented alphabetically within each core area ( McColl et al. , 2001 ). The consensus process was performed as a 2-step modified Delphi procedure, comprising two survey rounds and a final consultation meeting of the Steering Committee. For round 1, the Delphi survey was distributed via a website link or QR-code through presentations at conferences, courses, social media, and member or individual emails to stakeholders and stakeholder organizations. For round 2, all participants from the first round received an invitation by individual email and then three email reminders if they had not responded. The modified Delphi process allowed participants to leave comments and suggestions for new items in the first round and provided summarized feedback to those who participated in the second round, allowing them to change their score by considering the opinions of others ( Fish et al. , 2020 ). The items that did not reach consensus through both survey rounds were discussed in a semi-structured face-to-face consultation meeting within the Steering Committee. Decisions to include or exclude were made by discussion and majority vote. Each item was graded from 1 to 9 ( De Meyer et al. , 2019 ), with the additional option “I can’t rate the outcome because I don’t know the outcome”. Written anchors were provided to reduce measurement error ( Beckstead, 2014 ; Remus et al. , 2021 ) (1. Extremely unimportant; 2. Very unimportant; 3. Unimportant; 4. Maybe unimportant; 5. Unsure unimportant or important; 6. Maybe important; 7. Important; 8. Very important; 9. Extremely important). Scores of 1–3 signified an outcome of limited importance, scores of 4–6 signified an outcome as important but not critical and scores of 7–9 signified an outcome as critical, as defined by the Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group ( Guyatt et al. , 2011 ). A consensus that an outcome should not be included in the COS was defined as 70% or more participants scoring it as 1–3 and fewer than 15% scoring it as 7–9. Consensus for an outcome being included in the COS required 70% or more scoring it as 7–9 and fewer than 15% to score it as 1–3. If an outcome was included by one stakeholder group but not the others, the item was discussed in the Steering Committee consensus meeting ( Williamson et al. , 2017 ). If no agreement was to be reached by discussion, the decision was determined by majority opinion. All data and Delphi scores were collected securely via the ‘Nettskjema’ platform. All fields were mandatory to avoid missing data. Participants entered their email addresses to avoid and identify duplicate entries. Data were reported using ranking orders, percentages and frequencies. A P -value <0.05 was considered statistically significant. Statistical analysis was performed using IBM SPSS Statistics, Version 22.0 (IBM Corp, Armonk, NY, USA) and Microsoft ® Excel ® , Version 2111 (Microsoft Corp., Redmond, WA, USA). Institutional review board and Personal Data Officer approval were obtained from the Oslo University Hospital. Owing to the nature of this study, approval from the Regional Committee for medical and health research Ethics system in Norway was waived. The participation in the survey was voluntary, and by participating, the participants gave their consent to be included in the study.

We have developed a core set of outcomes that should help researchers when designing and reporting the results of future studies on the treatment of adenomyosis. The standardization of reporting will facilitate a better synthesis of evidence and assist patients and clinicians when making decisions regarding the optimal treatment of adenomyosis. The use of a standardized set of outcomes should also stimulate good clinical practice in research and ensure that studies report the outcomes of interest and importance to patients.

Individuals living with adenomyosis and their partners, patient advocates, HCPs and researchers have developed the first COS to standardize outcome selection, collection and reporting for future studies investigating uterus-sparing treatment of adenomyosis in premenopausal women, namely the COSAR. The COSAR is applicable to all uterus-sparing therapeutic interventions, including medical, surgical and other interventional approaches, including those that are guided by imaging techniques. It comprises 50 outcomes, of which 19 are applicable only for certain studies while 10 are reporting items. The strength of this process is the adherence to a recommended and prescriptive methodology, a high number of participants, and a truly global representation of patients and HCPs. Patients were included at all stages of the COSAR development through focus groups and representation on the Steering Committee, ensuring their views were strongly represented. The Steering Committee comprised experts in adenomyosis with different foci of research and clinical interest, such as surgery, imaging, infertility or basic research. This ensured a broad perspective when choosing different outcomes. The COS includes definitions for each outcome to avoid ambiguity in interpretation and includes well-documented existing COS, classifications and definitions ( Munro et al. , 2018 ; Vanhie et al. , 2016 ; Duffy et al. , 2020 ) in the COSAR where possible. Such an approach was designed to ensure a high standard of outcomes and facilitate harmonization of outcomes where conditions overlap. The project and the results are not without limitations. Only 22% of participants were from continents other than Europe, and relatively few patients were from low-income countries. Despite intensive efforts undertaken by the Steering Committee members, it was only possible to engage a small number of participants from Asian countries, a circumstance that may be related to the lack of translation of the survey. As adenomyosis is a benign disorder which requires expert-ultrasound or cost-intensive MRI for diagnosis, the awareness of the condition might be low amongst both patients and health care providers in those regions of the world. Also, a lack of translation of the survey might have limited the active participation of people in non-English speaking countries. If correct, this observation could explain the low rate of engagement with our project and, even for those who did participate, it may have affected their perception of the relevance of some outcomes. However, specific symptoms seem to be universally valid, as the international validation of symptom scoring instruments in gynaecology shows ( Nie et al. , 2017 ; Yeung et al. , 2019 ; Schneider et al. , 2000 ). Some of the potential bias may have been addressed by the inclusion of experts from low and middle-income countries on the Steering Committee, voicing their views and opinions. Another project limitation was the high attrition rate (41.9%), which could weaken the conclusions’ strength. However, our analysis did not indicate that there was an attrition bias. Another concern relates to the observation that many women have concomitant disorders, especially endometriosis, a circumstance that could influence patient perceptions of relevant outcomes to be those that are not adenomyosis-specific. Also, several included outcomes were chosen based on expert opinions and patient preferences despite an absence of evidence confirming their relevance to adenomyosis. The lack of disease-specific QoL evaluation tools for adenomyosis analogous to those developed for similar conditions is also reflected in this knowledge gap. These issues accentuate the urgent need for this COS and further research to identify additional outcomes of relevance for adenomyosis to be included in a future revision of the COSAR. The development of the COSAR is an important step that should improve the quality of future adenomyosis-related clinical investigations, including the performance of systematic review and meta-analysis. Before it can be fully implemented, additional work is needed to define measures for each of the listed outcomes. In addition to dissemination and implementation of the COSAR, it will be necessary to monitor its use in a way that informs future appropriate modifications. Some important tools for adenomyosis research are still missing, including a disease-specific HR-QoL questionnaire and the validation of generic HR-QoL instruments. There is also a need for studies designed to determine which symptoms are adenomyosis-specific and how they affect people’s QoL.

Click here for additional data file. Click here for additional data file. Click here for additional data file.