Reliability of self-reported endometriosis and adenomyosis data in online surveys: results from the ComPaRe-Endometriosis e-cohort.

STUDY QUESTION

How reliable are self-reported endometriosis or adenomyosis data compared with medical reports in online patient surveys?

SUMMARY ANSWER

In our sample, reliability was strong to excellent for diagnosis, age at diagnosis, and stage at diagnosis, while reliability for macro-phenotype of endometriosis ranged from fair to substantial.

WHAT IS KNOWN ALREADY

Many online surveys rely on self-reported endometriosis data. Previous research has shown validation rates ranging from 72% to 95% for self-reported endometriosis diagnosis in population cohorts, but this rate is unknown in the context of online surveys among patient populations or patient e-cohorts.

STUDY DESIGN, SIZE, DURATION: We conducted a specific study within ComPaRe-Endometriosis, a French e-cohort launched in 2018 and following up several thousands of women with endometriosis and/or adenomyosis. Within this study, a total of 215 participants sent their medical reports. We compared participants' self-reported diagnosis and disease characteristics, as collected in self-reported questionnaires at baseline in the ComPaRe-Endometriosis cohort, with data from contemporaneous medical reports.

PARTICIPANTS/MATERIALS, SETTING, METHODS: Participants were asked to send their medical reports mentioning diagnosis and its date, and type(s) and stage of endometriosis. We assessed the reliability between baseline self-reported data and medical report data using descriptive statistics, intraclass correlation coefficient (ICC), Cohen's Kappa, and weighted Kappa.

MAIN RESULTS AND THE ROLE OF CHANCE

The study included 215 participants with a mean age of 33.8 years (SD = 7.0). Over half (52.1%) reported endometriosis alone, 8.4% adenomyosis alone, and 39.5% reported both diseases. Endometriosis and adenomyosis diagnoses were correctly reported among 95.9% and 90.3% of participants, respectively. Self-reported age at diagnosis showed excellent reliability with ICC = 0.96 for endometriosis and ICC = 0.91 for adenomyosis. Regarding self-reported disease stage, substantial-to-almost perfect agreement (κw = 0.86-0.78) and excellent reliability (ICC = 0.91-0.94) were observed. For self-reported macro-phenotype of endometriosis, agreement varied: substantial for endometrioma (K = 0.62-0.74), moderate-to-substantial for deep endometriosis (K = 0.40-0.61), and fair-to-moderate for superficial peritoneal endometriosis (K = 0.35-0.43).

LIMITATIONS, REASONS FOR CAUTION: The most motivated participants may have a particular profile in terms of involvement in their care, which constitutes a limitation to the generalizability of the findings.

WIDER IMPLICATIONS OF THE FINDINGS

Several studies assessed the reliability of self-reported endometriosis data in population cohorts and found validation rates ranging from 72% to 95% with medical reports. This figure could vary greatly depending on several criteria (study population, clarity of the question about diagnosis, tool used to confirm diagnosis…). In this work, we found a striking heterogeneity in the way macro-phenotype of disease was described in medical reports. Since many studies rely on self-reported data, this work underscores the importance of harmonizing the report of diagnosis and endometriosis/adenomyosis characteristics in medical reports in order to ease continuity of care and ensure high-quality research on these conditions.

STUDY FUNDING/COMPETING INTEREST(S): The ComPaRe cohort was supported by the Assistance Publique-Hôpitaux de Paris (APHP) and the Université Paris Cité. ComPaRe-Endometriosis was funded through grants from the Fondation de France (#00132975/WB-2022-44780) and the Agence Nationale de la Recherche (#ANR-22-CE36-0016-01) and received donations from patient societies and initiatives (EndoMind, EndoFrance, EndoAction, Les Joyeux Boulomanes Farlédois, La Belle et L'endo, Les Chroniques Endométriques). S.G. was supported by a PhD scholarship from the French Ministry of Research and a grant from the Crédit Agricole Ile-de-France Mécénat. Z.B. was supported by the French National Association for Research and Technology (ANRT) and Lyv Healthcare during their PhD. The funders had no role in the design, interpretation, or decision to publish the study. The other authors report no conflicts of interest.

TRIAL REGISTRATION NUMBER

N/A.