Exploring pre-diagnosis hospital contacts in women with endometriosis using ICD-10: a Danish case-control study

Study question: How does pre-diagnosis use of hospital care differentiate between women later diagnosed with endometriosis and age-matched controls without a diagnosis?

Summary answer: Women with hospital-diagnosed endometriosis had more frequent hospital contacts in the 10 years leading up to the diagnosis compared to women without a diagnosis of endometriosis, and the contacts were related to registered diagnoses in nearly all of the included ICD-10 chapters for the entire period.

What is known already: Only a few studies have investigated the utilization of health care among women with endometriosis in the time before diagnosis, but current research shows that women with endometriosis have a higher utilization compared to women without diagnosed endometriosis. To our knowledge, no study has investigated the type of contact related to the higher utilization by using the ICD-10 diagnoses registered to the hospital contact.

Study design, size, duration: This study was conducted as a national Danish registry-based case-control study of 129 696 women. Cases were women with a first-time hospital-based diagnosis of endometriosis between 1 January 2000 and 31 December 2017.

Participants/materials, setting, methods: Using density sampling, we identified 21 616 cases. Each case was matched on age at the date of diagnosis (index date) to five women without hospital-diagnosed endometriosis (n = 108 080) at the time of matching. The utilization and registered ICD-10 diagnoses related to the hospital contact were included for the 10 years before the index date.

Main results and the role of chance: The probability of having a high number of hospital contacts (six or more) was more common among women with endometriosis (68.6%) compared to women without endometriosis (55.7%) In general, women without endometriosis were more likely to have fewer than six contacts. The diagnoses registered to the contact among cases were related to a greater variety of ICD-10 chapters when compared to controls with the same number of contacts. For nearly all of the included ICD-10 chapters, women with endometriosis were more likely to have a diagnosis over the entire period compared to controls, with the only exception being in the chapter related to pregnancy.

Limitations, reasons for caution: Our results are only applicable for women with hospital-based diagnosed endometriosis since we were not able to include women diagnosed at the general practitioner or private gynecologists. We were not able to make a causal interpretation, as we do not have information on the onset of symptoms of the included diseases. The association may be overestimated due to detection bias. However, a sensitivity analysis only changed the results slightly, indicating a low risk of this bias.

Wider implications of the findings: This study is in accordance with previous studies on the subject, indicating that the utilization of health care prior to endometriosis is not necessarily restricted to endometriosis-related symptoms and that endometriosis can be associated with many other diseases. Future studies may explore hospital contacts and causes/diagnoses following the endometriosis diagnosis to further shed light on whether our results are due to a pattern of multiple pathologies or rather an expression of misdiagnoses among women with endometriosis before diagnosis.

Study funding/competing interest(s): This study is supported by grants from the project Finding Endometriosis using Machine Learning (FEMaLe/101017562), which has received funding from The European Union's Horizon 2020 research and innovation program and Helsefonden (21-B-0141). A.W.H. received grant funding from NIHR, CSO, Roche Diagnostics, and Wellbeing of Women. A.W.H.'s institution received consultation fees from Theramex, Joii, Gesynta, and Gedeon Richter. A.W.H.'s institution received honoraria for lectures from Theramex and Gedeon Richter. A.W.H. is listed as co-inventor on a patent application (UK Patent App No. 2217921.2, International Patent App No. PCT/GB2023/053076). P.T.K.S.'s institution (University of Edinburgh) received consultation fees from Gesynta Pharma AB and BenevolentAI Bio Ltd. P.T.K.S's institution (University of Edinburgh) declares a patent application (UK Patent Application No. 2310300.5. Androgens in diagnostic strategies for endometriosis). P.T.K.S. is a treasurer of the World Endometriosis Society, Fellowship in the Royal Society of Edinburgh, and a Scientific Advisor of the Royal College of Obstetrics and Gynaecology.

Trial registration number: N/A.