{"paper_id":"c150b5de-8710-4943-8a91-9a7a0a7f7b16","body_text":"1 \nInvestigating Racial Disparities in Drug Prescriptions for Patients with Endometriosis  \nAparajita KASHYAP1 (BA), Ms. Maryam AZIZ2 (BS), Mr. Tony Y. SUN1 (MA), Ms. Sharon \nLIPSKY-GORMAN1 (MS), Dr. Jessica OPOKU-ANANE3 (MD), Dr. Noémie ELHADAD1,2 \n(PhD) \nAffiliations and Locations \n1. Columbia University Irving Medical Center, Department of Biomedical Informatics \n(New York, NY) \n2. Columbia University, Department of Computer Science (New York, NY) \n3. Columbia University Irving Medical Center, Department of Obstetrics and Gynecology \n(New York, NY) \nDisclosure Statement: The authors report no conflicts of interest.  \nFinancial Support: We acknowledge the National Library of Medicine for their support of this \nwork (T15LM007079 and R01LM013043). The funders had no role in the design and conduct of \nthe study; collection, management, analysis, and interpretation of the data; preparation, review, \nor approval of the manuscript; and decision to submit the manuscript for publication. \nPaper presentation information: A preliminary version of this work was presented at the 15th \nWorld Congress on Endometriosis (May 3-6, 2023), Edinburgh, UK and at the National Library \nof Medicine T15 Training Conference (June 21-23, 2023), Palo Alto, US \nCorresponding Author: Aparajita Kashyap, ak4885@cumc.columbia.edu   \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\n  \n  \n2 \nAbstract  \nBackground: Endometriosis is a chronic disease with a long time to diagnosis and several \nknown comorbidities that requires a range of treatments including of pain management and \nhormone-based medications. Racial disparities specific to endometriosis treatments are \nunknown.  \nObjective: We aim to investigate differences in patterns of drug prescriptions specific to \nendometriosis management in Black and White patients prior to diagnosis and after diagnosis of \nendometriosis and compare these differences to racial disparities established in the general \npopulation.  \nStudy Design: We conduct a retrospective cohort study using observational health data from the \nIBM MarketScan® Multi-state Medicaid dataset. We identify a cohort of endometriosis patients \nconsisting of women between the ages of 15 and 49 with an endometriosis-related surgical \nprocedure and a diagnosis code for endometriosis within 30 days of this procedure. Cohort is \nfurther restricted to patients with at least 3 years of continuous observation prior to diagnosis. \nWe identify a non-endometriosis cohort of women between the ages of 15 and 49 with no \nendometriosis diagnosis and at least 1 year of continuous observation. We compare prevalence of \nprescriptions across selected drug classes for Black vs. White endometriosis patients. We further \nexamine prevalence differences in the non-endometriosis cohort and prevalence differences pre- \nand post-diagnosis in the endometriosis cohort.  \nResults: The endometriosis cohort comprised 16,372 endometriosis patients (23.3% Black, \n66.0% White). Of the 28 drug classes examined, 17 were prescribed significantly less in Black \npatients compared to 21 in non-endometriosis cohort (n=3,663,904), and 4 were prescribed \nsignificantly more in Black patients compared to 6 in the non-endometriosis cohort. Of the 17 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n3 \ndrugs prescribed more often in White patients, 16 have larger disparities pre-diagnosis than post-\ndiagnosis. \nConclusions: Our analysis identified significant differences in medication prescriptions between \nWhite and Black patients with endometriosis, notably in hormonal treatments, pain management, \nand treatments for common endometriosis co-morbidities. Racial disparities in drug prescriptions \nare well established in healthcare, and better understanding these disparities in the specific \ncontext of chronic reproductive conditions and chronic pain is important for increasing equity in \ndrug prescription practices. \n \nKeywords: Health disparities, Reproductive healthcare, observational study, biomedical \ninformatics, prescription patterns, pain medication, chronic disease, temporal analysis \n  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n4 \nIntroduction \nEndometriosis is a chronic, inflammatory disease characterized by growth of \nendometrial-like tissue outside the uterus. Identified symptoms of endometriosis vary widely and \ninclude dysmenorrhea, fatigue, non-menstrual abdominopelvic pain, and heavy menstrual \nbleeding1,2. While the precise prevalence of endometriosis is unknown, it affects an estimated 6-\n10% of women of reproductive age2–4.  Current medical interventions for endometriosis depend \non patient prognosis and include analgesics, combined hormonal contraceptives, progestogens, \ngonadotropin-releasing hormone (GnRH) agonists, GnRH antagonists, aromatase inhibitors, and \nsurgical treatments5.  \nRacial health disparities in the United States are pervasive, and research shows that Black \nand Hispanic patients are less likely to receive a diagnosis of endometriosis compared to White \npatients6, as well as less likely to receive surgical treatment for endometriosis7. Racial bias is \nobserved beyond endometriosis diagnosis, such as in pain management; Black and Hispanic \npatients are less likely to be treated for pain and other symptoms that impact quality of life8–11.  \nIn this work we characterize the drug prescription patterns for endometriosis patients and \npotential differences in prevalence across Black and White patients. Because endometriosis is a \ncondition with delays in diagnosis, patients with suspected endometriosis are often prescribed \nfirst-line treatments prior to their diagnosis. We thus characterize treatment patterns in three \nways: prior to diagnosis, after diagnosis, and overall. Furthermore, to account for multiple co-\nmorbidities of endometriosis, we extend our analysis to all medication classes that are \nsignificantly more prescribed in our endometriosis cohort than in a non-endometriosis \ncomparison cohort. To further contextualize this analysis, we compare treatment patterns of \nendometriosis patients with those without endometriosis and measure differences in treatment \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n5 \npatterns across White and Black patients in a non-endometriosis cohort, namely all women in \nreproductive age without any diagnosis of endometriosis. \n  \nMaterials and Methods \nThe analysis and use of the de-identified dataset presented in this work were carried out under \nResearch Protocol AAAO7805 approved by the Columbia University IRB.  \nDataset \nIn this retrospective cohort study, we focus our analysis to patients with Medicaid, the \nU.S. government program that provides health insurance for people with limited income. The \ndata comes from the IBM MarketScan® Multi-state Medicaid dataset, which draws Medicaid \ndata from several states12. The Medicaid dataset contains de-identified longitudinal records of \npatients and includes inpatient and outpatient services, diagnostic history, and drug prescriptions \nfor over 48 million enrollees12. We leverage the data under the OMOP Common Data Model \n(CDM) format, which follows standardized conventions for drugs, therapies, and other medical \nvocabularies13. The database has been used for a variety of observational health studies due to its \nflexibility and robustness14–16. \nCohort Identification \nWe select patients with endometriosis using a validated cohort definition17. The cohort \ndefinition includes women ages 15 to 49 years who have an endometriosis-related surgical \nprocedure (e.g., laparoscopic surgery) as well as a diagnosis of endometriosis 30 days before or \nafter this procedure. The phenotype definition was validated through manual chart review of \n1,400 endometriosis patients yielding 70% sensitivity, 93% specificity, 85% positive predictive \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n6 \nvalue. Age is calculated at time of diagnosis. Only patients with endometriosis with at least three \nyears of continuous observation prior to diagnosis are considered for the endometriosis cohort. \nThe non-endometriosis cohort comprises of any woman aged 15-49 with no diagnostic \ncode for endometriosis at any point in their medical history and at least one year of continuous \nobservation. Age is calculated using the most recent visit date. In the claims database, it is \nunknown for individual patients whether the gender marker represents sex, gender, or both. \nIdentification of medications \nThe Anatomical Therapeutic Chemical (ATC) classification system separates drugs in a \nstandardized manner; we create drug classes using ATC level 3 classifications because they are \nassociated with specific pharmacological functions without specifying chemical structures18,19. \nWe use the OMOP CDM which explicitly maps medications to their ATC level 3 classification. \nThe ATC classifications create a “bridge” between the clinical guidelines for endometriosis and \nlarge-scale observational health data. \nWe narrow the set of drug classes by identifying drug classes that are prescribed \nsignificantly more often in the endometriosis cohort than the non-endometriosis one. For each \ndrug class, the proportion of patients in each cohort with at least one prescription belonging to \nthat drug class is calculated. The statistical significance of the difference between these relative \nprevalence measurements is calculated using a two-sided Z-test with a 0.01 significance level. \nThis procedure identified 28 drug classes that are significantly more prevalent in the \nendometriosis cohort than the non-endometriosis cohort (see Supplemental Table 1); we then \nmeasure prescription prevalence for each of these drug classes.  \nTo understand the clinical relevance of these 28 drug classes, we checked them against \nthe drugs listed in the 2022 ESHRE guidelines for treating endometriosis5. The only \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n7 \nrecommended treatments that are not included in our analysis are GnRH antagonists and \naromatase inhibitors. GnRH antagonists were not prescribed in the Medicaid dataset, and \naromatase inhibitors were excluded because the corresponding ATC 3rd level drug class \n(“hormone antagonists and related agents”) was not prescribed significantly more in patients \nwith endometriosis than in the non-endometriosis cohort. Several other drug classes more \nprevalently prescribed in the endometriosis population are associated with known comorbidities \nor symptoms of endometriosis including gastrointestinal distress, depression, and anxiety23,24.  \nTemporal Analysis \nEndometriosis is known to have extended delays between symptom onset and \ndiagnosis20,21. We therefore aim to better understand the patterns of drug prescription prior to \ndiagnosis (which occurs via laparoscopic surgery) and whether they differ from prescription \npractices after diagnosis. Prescription practices from the pre-diagnosis period can be thought of \nas prescription practices for undiagnosed endometriosis patients, which is an important area of \nstudy given that Black patients are less likely to be diagnosed with endometriosis compared to \nWhite patients6.  \nWe create two temporal subgroups of drug prescriptions in the endometriosis cohort (see \nFigure 1). Subgroup 1 consists of all prescriptions occurred prior to surgical diagnosis. \nSubgroup 2 consists of all prescriptions post-diagnosis. Note that neither subgroup 1 nor \nsubgroup 2 contain prescriptions given on the day of diagnosis; these prescriptions are only \ncounted in the “overall” prescriptions.  \nStatistical Methods \nFor each drug class of interest, the relative prevalence of prescriptions within that drug \nclass was calculated for Black and White endometriosis patients. A person is counted as part of \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n8 \nthe drug-positive group if they received at least one prescription that fell into the given drug \nclass, and relative prevalence is calculated based on the total number of Black (n = 3,814) and \nWhite (n = 10,805) patients in the endometriosis cohort. The difference in prescription \nprevalence is calculated as 100% ∗ (𝑝!\"#$% − 𝑝&'()* ), where 𝑝!\"#$% \tand 𝑝&'()* relative \nprevalence (see Table 2, Table 3).  \nTo identify significant differences in the Black and White sub-populations, we conduct a \ntwo-sided Z-test from the difference in prevalence and the associated standard error. We apply \nthe Bonferroni correction to account for multiple comparisons; we consider the difference \nsignificant when the adjusted p-value < 0.01. This procedure for statistical analysis was repeated \nfor all prescription classes and both temporal subgroups (pre-diagnosis and post-diagnosis). \nAnalysis was carried out using the scipy package in Python22. \nResults \nCohorts  \nThe non-endometriosis cohort comprises 3,663,904 patients (see Table 1). The \nendometriosis cohort (n=16,372) comprised more White (66%) than the non-endometriosis \ncohort (23%) (of note, 50% of patients in the non-endometriosis cohort have no recorded race). \nThe endometriosis cohort also was slightly older (mean age 32.4 ± 8.0 years) than the non-\nendometriosis cohort (mean age 28.7 ± 10.3 years). The median length of observational period \nwas 6.08 (IQR: 5.08 to 8.84) years for the endometriosis cohort and 4.00 (IQR: 2.00 to 6.25) \nyears for the non-endometriosis cohort. For the endometriosis cohort, the median length of \nobservation prior to diagnosis was 4.20 (IQR: 3.26 to 5.73) years, and the median length of \nobservation post-diagnosis was 1.68 (IQR: 0.82 to 2.89) years.  \nDifferences in overall prevalence of drug prescriptions for Black and White patients  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n9 \nOf the 28 drug classes of interest, 17 (61%) are prescribed significantly more often for \nWhite endometriosis patients compared to Black endometriosis patients (see Figure 2). In the \nnon-endometriosis cohort, 21 (75%) are prescribed significantly more for White patients than for \nBlack patients and 6 (21%) are prescribed significantly more for Black patients.  \nThe largest racial differences in endometriosis patients occur for antidepressants \n(prescription prevalence higher by 20.7% in White patients than Black patients, 95% CI [19.8%, \n21.6%]), anxiolytics (20.4%, 95% CI [19.5%, 21.3%]), antiepileptics (18.7%, 95% CI [17.8%, \n19.6%]), and estrogens (10.2%, 95% CI [9.5%, 11.0%]). Additionally, 4 drug classes are \nprescribed significantly more often in Black patients than White patients, with the largest \ndifferences occurring for iron preparations (7.3%, 95% CI [6.6%, 8.0%]), hormones and related \nagents (7.1%, 95% CI [6.2%, 8.1%]), and antiinfectives/antiseptics (7.0%, 95% CI [6.1%, \n7.9%]). Prescription prevalence differences are more pronounced in the endometriosis cohort \nthan the non-endometriosis cohort for anxiolytics, antiepileptics, estrogens, iron preparations, \nand antiinfectives/antiseptics but are less pronounced in the endometriosis cohort for \nantidepressants and hormones and related agents (see Table 2).  \nOut of the drug classes related to established endometriosis treatment options, three \n(opioids, other analgesics and antipyretics, and hormonal contraceptives for systemic use) are \nsignificantly more prevalent amongst White endometriosis patients; the remaining two \n(hormones and related agents, progestogens) are significantly more prevalent amongst Black \nendometriosis patients (see Table 2). These differences are generally consistent with the non-\nendometriosis cohort, where analgesics are more commonly prescribed for White women and \nhormones and related agents are more commonly prescribed in Black women. One major \ndifference between the endometriosis and non-endometriosis cohorts, however, is in hormonal \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n10 \ncontraceptives for systemic use, which are prescribed significantly more in White women with \nendometriosis (3.3%) and significantly more for Black women without endometriosis (2.0%). \nAdditionally, the difference in prescriptions of estrogens is much higher in the endometriosis \ncohort (10.23%) than in the non-endometriosis cohort (1.83%).  \nTemporal differences in drug prescription differences \nWhen observing drug prescription differences across White and Black endometriosis \npatients across temporal subgroups, there is a pattern of pre-diagnosis differences (temporal \nsubgroup 1) exceeding post-diagnosis differences (temporal subgroup 2), especially for drug \nclasses where White patients are prescribed at a higher rate (see Figure 2 and Table 3). Of the \n17 drug classes that fall into this category, 16 (94%) have larger differences prior to diagnosis \ncompared to post-diagnosis. For drugs that are significantly more prevalently prescribed in Black \npatients, the difference in prevalence decreases from pre- to post-diagnosis for iron preparations \nand antiinfectives/antiseptics; the difference in prevalence is greater post-diagnosis for hormones \nand related agents and progestogens.  \nComment \nPrincipal Findings \nOur analysis of drug prescription patterns amongst White and Black patients with \nendometriosis in the Medicaid population shows racial disparities in disease treatment and pain \nmanagement within endometriosis. Differences in prescription prevalence tend to be larger pre-\ndiagnosis compared to post-diagnosis. We also find that drugs associated with comorbidities of \nendometriosis (including gastrointestinal distress, irritable bowel syndrome, thyroid disorders, \nand lower urinary tract symptoms25–27) are prescribed significantly less prevalently in Black \npatients compared to White patients, often with larger disparities in the endometriosis population \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n11 \nthan the non-endometriosis population; this suggests disparities in holistic endometriosis \nmanagement. \nResults in the Context of What is Known \nSeveral disparities identified in this dataset are well-documented in patient populations \nnot specific to endometriosis (e.g. prescriptions of antidepressants28,29, select anxiolytics30, and \npain medication8–11). Prior studies have also observed racial disparities in endometriosis \ndiagnosis6 and surgical treatments7; therefore, it becomes important to consider how lack of \ntreatment and lack of diagnosis may collectively impact patients’ healthcare experiences. \nAdditionally, when considering that most prescription disparities are larger pre-diagnosis \ncompared to post-diagnosis, we conceptualize the pre-diagnosis period as one where the patient \nmay require treatment but has no formal diagnosis to explain the multi-system, chronic \nsymptoms they may be facing.  \nClinical Implications \nThe most notable differences in prescription prevalence occur for antidepressants and \nanxiolytics, which are generally used to treat depression and anxiety. Depression and anxiety are \nknown comorbidities of endometriosis and are associated with worse endometriosis \nsymptoms23,24. Prior studies have found that Black patients experience less rapport-building \nduring visits31 and are additionally less likely to communicate their symptoms (especially those \nrelated to mental health) to healthcare providers31,32; this suggests a gap in physician-patient \ncommunication which could be due to lack of cultural competency and/or perceived racism33. \nPerceived racism34 includes acts or attitudes that may or may not “objectively” considered racist \nbut do contribute to the (subjective) experience of racial prejudice and is known to contribute to \nhealthcare-related stress35–37. Future qualitative work should investigate how endometriosis \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n12 \npatients (who typically have more frequent interactions with the healthcare system) experience \nracism and the impact this has on their healthcare-related stress.  \nWe observe disparities in the prescription of pain medication (both opioid and non-\nopioid) where Black patients are less likely to receive this treatment than White patients. False \nracial assumptions and stereotypes such as the idea that Black patients asking for pain \nmedications are “drug-seeking”38,39 or that Black patients feel less pain than their White \ncounterparts8 may contribute to these disparities. Perceived racism can also impact how Black \npatients experience pain management: in addition to exacerbating healthcare-related stress, the \nracism experienced by Black patients when seeking chronic pain treatment may also contribute \nto hopelessness (negative expectations about one’s present life and future), which can negatively \nimpact pain management10. Additionally, misconceptions and lack of effective communication \nbetween patients and physicians surrounding medication tolerance, side effects, efficacy, and \naddiction has been found to disproportionately impact racial and ethnic minorities37. Our finding \nthat the disparities in pain medication prescriptions is greater in the pre-diagnosis cohort than the \npost-diagnosis subgroups suggests that the reasons for lack of prescriptions are exacerbated pre-\ndiagnosis when patients have symptoms of, but are not explicitly diagnosed with, endometriosis \ncompared to post-diagnosis, when their endometriosis has been formally diagnosed.  \nFinally, when comparing racial prevalence disparities across the endometriosis and non-\nendometriosis cohorts, there was broad consistency; however, several of the drug classes \nassociated with comorbidities of endometriosis (anxiolytics, urologicals, drugs for functional \ngastrointestinal disorders, thyroid preparations) had disparities that were larger in the \nendometriosis than the non-endometriosis cohort. There may be inequitable treatment of Black \nwomen specific to endometriosis, especially during their journey to diagnosis (when all these \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n13 \ndisparities are larger). Notably, the medication class with the highest difference was Estrogens \n(see Table 2). There are potentially multiple indications to drugs under that medication class that \nmay differ in the endometriosis and non-endometriosis cohorts, such as hormone replacement \ntherapy for post-menopausal women vs. fertility treatments for endometriosis patients.  \nResearch Implications \nWe identify several classes of drugs where there may be racial disparities in prescriptions \nspecifically for endometriosis treatment; however, this work does not dive into the context of \nthese drug prescription patterns, nor does it investigate these disparities in the context of patient \nexperiences and quality of life. Future work should focus on better understanding the context of \ndrug prescription patterns, especially co-occurring conditions associated with specific \nprescriptions or self-reported data about pain and how that relates to analgesic prescriptions for \nWhite and Black patients. Additionally, qualitative work should be conducted to understand the \nmultifactorial reasons for our findings, to assess the impact of these disparities on patient lives, \nand to identify methods for mitigating these disparities. Finally, this analysis should be extended \nto other patient populations to identify disparities in other racial and ethnic minorities, as well as \nuninsured patients (who may face even more barriers to care than the Medicaid patients in this \ndataset).  \nStrengths and Limitations \nOne major strength of this study was the use of the Multi-state Medicaid Dataset. \nMedicaid covers 72.5 million Americans, making it the largest source of healthcare in the United \nStates40. Furthermore, in this study we aimed to compare groups of patients with similar access \nto care (e.g., patients with at least three years of continuous observation prior to endometriosis \ndiagnosis) to ensure that this was not a confounding factor in analysis. Due to the wide variety of \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n14 \nhealthcare settings represented within the Medicaid dataset, we believe our findings will be \ngeneralizable across insured populations within the United States; however, international \ngeneralizability may be limited. This work is also unlikely to accurately represent the \nexperiences of patients without insurance, who experience different (and generally more severe) \nbarriers to care, especially in the case of chronic conditions41. Another general limitation of using \nMedicaid data is the amount of missing data regarding racial and ethnic identities. data was only \nrecorded as White, Black/African American, or missing, meaning that other potential racial \ndisparities could not be measured. Similarly, high missingness in ethnicity data rendered us \nunable to measure possible ethnic disparities.  \nIn this work, we use a validated phenotype17 that requires laparoscopic surgery to define \nour endometriosis cohort. However, recent clinical guidelines recommend that treatment be \nexplored concurrently with further diagnostic exploration if a clinical (vaginal) or imaging test \n(ultrasound, MRI) indicates endometriosis5. Therefore, some prescriptions that occur in the “pre-\ndiagnostic” subgroup are likely meant to treat suspected endometriosis. Another limitation of this \nphenotype is the use of a “female” gender marker as inclusion criteria, which may exclude some \ntransgender men with endometriosis from cohort inclusion.  \nThe use of ATC level 3 drug classes enables us to quickly and accurately pull relevant \ndrugs based on their pharmacological functions while remaining robust to the fact that providers \nin different states may vary in their precise prescription patterns42,43.  However, ATC medication \nclasses are intended to reflect the drug indication, which may not always be consistent with why \nthe drug was prescribed.  \nConclusions \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n15 \nIn this work, we document racial disparities in medication prescription practices for \nBlack endometriosis patients; these disparities exist in several aspects of treatment and are \nreflective of Black patients not receiving holistic endometriosis care. We compare these \ndisparities to a non-endometriosis population and find several drug classes where disparities are \nlarger in the endometriosis population, indicating that these disparities are unique to chronic \ndisease management (and potentially to endometriosis). When coupled with the fact that these \nprescription disparities are larger pre-diagnosis compared to post-diagnosis, it becomes \nimportant for clinicians who treat chronic, complex patients to ensure that they are \ncommunicating with their patients and prescribing medications consistently and equitably.  \nData Sharing Statement  \nThe IBM MarketScan® Multi-state Medicaid dataset is available to license at \nhttps://www.ibm.com/watson-health/about/truven-health-analytics (see ref. 12). Code for cohort \ndefinitions is available through the OHDSI Phenotype Library and at \nhttps://github.com/elhadadlab/endochar. The underlying code for this study is available on \nGithub at https://github.com/elhadadlab/endo_disparities.  \nReferences \n1. Sharma S, Tripathi A, Sharma S, Tripathi A. Endometriosis: The Enigma That It Continues \nto Be. IntechOpen; 2022. doi:10.5772/intechopen.108774 \n2. Acién P, Velasco I. Endometriosis: A Disease That Remains Enigmatic. ISRN Obstet \nGynecol. 2013;2013:242149. doi:10.1155/2013/242149 \n3. Agarwal SK, Chapron C, Giudice LC, et al. Clinical diagnosis of endometriosis: a call to \naction. Am J Obstet Gynecol. 2019;220(4):354.e1-354.e12. doi:10.1016/j.ajog.2018.12.039 \n4. Morassutto C, Monasta L, Ricci G, Barbone F, Ronfani L. Incidence and Estimated \nPrevalence of Endometriosis and Adenomyosis in Northeast Italy: A Data Linkage Study. \nPLOS ONE. 2016;11(4):e0154227. doi:10.1371/journal.pone.0154227 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n16 \n5. Becker CM, Bokor A, Heikinheimo O, et al. ESHRE guideline: endometriosis†. Hum Reprod \nOpen. 2022;2022(2):hoac009. doi:10.1093/hropen/hoac009 \n6. Bougie O, Yap MaI, Sikora L, Flaxman T, Singh S. Influence of race/ethnicity on prevalence \nand presentation of endometriosis: a systematic review and meta-analysis. BJOG Int J Obstet \nGynaecol. 2019;126(9):1104-1115. doi:10.1111/1471-0528.15692 \n7. Orlando MS, Luna Russo MA, Richards EG, et al. Racial and ethnic disparities in surgical \ncare for endometriosis across the United States. Am J Obstet Gynecol. 2022;226(6):824.e1-\n824.e11. doi:10.1016/j.ajog.2022.01.021 \n8. Hoffman KM, Trawalter S, Axt JR, Oliver MN. Racial bias in pain assessment and treatment \nrecommendations, and false beliefs about biological differences between blacks and whites. \nProc Natl Acad Sci U S A. 2016;113(16):4296-4301. doi:10.1073/pnas.1516047113 \n9. Ghoshal M, Shapiro H, Todd K, Schatman ME. Chronic Noncancer Pain Management and \nSystemic Racism: Time to Move Toward Equal Care Standards. J Pain Res. 2020;13:2825-\n2836. doi:10.2147/JPR.S287314 \n10. Ezenwa MO, Fleming MF. Racial Disparities in Pain Management in Primary Care. J Health \nDisparities Res Pract. 2012;5(3):12-26. \n11. Schoenthaler A, Williams N. Looking Beneath the Surface: Racial Bias in the Treatment and \nManagement of Pain. JAMA Netw Open. 2022;5(6):e2216281. \ndoi:10.1001/jamanetworkopen.2022.16281 \n12. Truven Health Analytics. Published June 22, 2022. Accessed February 23, 2023. \nhttps://www.ibm.com/watson-health/about/truven-health-analytics \n13. Blacketer C, Defalco FJ, Ryan PB, Rijnbeek PR. Increasing trust in real-world evidence \nthrough evaluation of observational data quality. J Am Med Inform Assoc. 2021;28(10):2251-\n2257. doi:10.1093/jamia/ocab132 \n14. Reps JM, Rijnbeek PR, Ryan PB. Identifying the DEAD: Development and Validation of a \nPatient-Level Model to Predict Death Status in Population-Level Claims Data. Drug Saf. \n2019;42(11):1377-1386. doi:10.1007/s40264-019-00827-0 \n15. Nash D, Katcoff H, Faerber J, et al. Impact of Device Miniaturization on Insertable Cardiac \nMonitor Use in the Pediatric Population: An Analysis of the MarketScan Commercial and \nMedicaid Databases. J Am Heart Assoc. 2022;11(16):e024112. \ndoi:10.1161/JAHA.121.024112 \n16. Adamson DM, Chang S, Hansen LG. Health Research Data for the Real World: The \nMarketScan Databases. \n17. Elhadad N, McKillop MM, Schwartz JM, et al. Selection and Characterization of an \nEndometriosis Cohort Across an Observational Health Database Network. MedArxiv. \nPublished online 2023. \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n17 \n18. WHOCC - Structure and principles. Accessed February 23, 2023. \nhttps://www.whocc.no/atc/structure_and_principles/ \n19. Anatomical Therapeutic Chemical (ATC) Classification. Accessed February 23, 2023. \nhttps://www.who.int/tools/atc-ddd-toolkit/atc-classification \n20. Tewhaiti-Smith J, Semprini A, Bush D, et al. An Aotearoa New Zealand survey of the \nimpact and diagnostic delay for endometriosis and chronic pelvic pain. Sci Rep. \n2022;12(1):4425. doi:10.1038/s41598-022-08464-x \n21. Nnoaham KE, Hummelshoj L, Webster P, et al. Impact of endometriosis on quality of life \nand work productivity: a multicenter study across ten countries. Fertil Steril. \n2011;96(2):366-373.e8. doi:10.1016/j.fertnstert.2011.05.090 \n22. Virtanen P, Gommers R, Oliphant TE, et al. SciPy 1.0: fundamental algorithms for scientific \ncomputing in Python. Nat Methods. 2020;17(3):261-272. doi:10.1038/s41592-019-0686-2 \n23. Mirkin D, Murphy-Barron C, Iwasaki K. Actuarial Analysis of Private Payer Administrative \nClaims Data for Women With Endometriosis. J Manag Care Pharm. 2007;13(3):262-272. \ndoi:10.18553/jmcp.2007.13.3.262 \n24. Missmer SA, Tu FF, Agarwal SK, et al. Impact of Endometriosis on Life-Course Potential: \nA Narrative Review</p>. Int J Gen Med. 2021;14:9-25. doi:10.2147/IJGM.S261139 \n25. Holdsworth-Carson SJ, Ng CHM, Dior UP. Editorial: Comorbidities in Women With \nEndometriosis: Risks and Implications. Front Reprod Health. 2022;4. Accessed February 21, \n2023. https://www.frontiersin.org/articles/10.3389/frph.2022.875277 \n26. Gabriel I, Vitonis AF, Missmer SA, et al. Association between endometriosis and lower \nurinary tract symptoms. Fertil Steril. 2022;117(4):822-830. \ndoi:10.1016/j.fertnstert.2022.01.003 \n27. Peyneau M, Kavian N, Chouzenoux S, et al. Role of thyroid dysimmunity and thyroid \nhormones in endometriosis. Proc Natl Acad Sci. 2019;116(24):11894-11899. \ndoi:10.1073/pnas.1820469116 \n28. Remmert JE, Guzman G, Mavandadi S, Oslin D. Racial Disparities in Prescription of \nAntidepressants Among U.S. Veterans Referred to Behavioral Health Care. Psychiatr Serv. \n2022;73(9):984-990. doi:10.1176/appi.ps.202100237 \n29. González HM, Croghan TW, West BT, et al. Antidepressant Use among Blacks and Whites \nin the United States. Psychiatr Serv Wash DC. 2008;59(10):1131-1138. \ndoi:10.1176/appi.ps.59.10.1131 \n30. Cook B, Creedon T, Wang Y, et al. Examining racial/ethnic differences in patterns of \nbenzodiazepine prescription and misuse. Drug Alcohol Depend. 2018;187:29-34. \ndoi:10.1016/j.drugalcdep.2018.02.011 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n18 \n31. Ghods BK, Roter DL, Ford DE, Larson S, Arbelaez JJ, Cooper LA. Patient–Physician \nCommunication in the Primary Care Visits of African Americans and Whites with \nDepression. J Gen Intern Med. 2008;23(5):600. doi:10.1007/s11606-008-0539-7 \n32. Probst JC, Laditka SB, Moore CG, Harun N, Powell MP. Race and ethnicity differences in \nreporting of depressive symptoms. Adm Policy Ment Health. 2007;34(6):519-529. \ndoi:10.1007/s10488-007-0136-9 \n33. Kim M. Racial/Ethnic Disparities in Depression and Its Theoretical Perspectives. Psychiatr \nQ. 2014;85(1):1-8. doi:10.1007/s11126-013-9265-3 \n34. McConahay JB, Hardee BB, Batts V. Has Racism Declined in America? It Depends on Who \nIs Asking and What Is Asked. J Confl Resolut. 1981;25(4):563-579. \n35. Clark R, Anderson NB, Clark VR, Williams DR. Racism as a stressor for African \nAmericans: A biopsychosocial model. Am Psychol. 19991101;54(10):805. \ndoi:10.1037/0003-066X.54.10.805 \n36. Anderson NB, Bulatao RA, Cohen B, National Research Council (US) Panel on Race E. \nSignificance of Perceived Racism: Toward Understanding Ethnic Group Disparities in \nHealth, the Later Years. In: Critical Perspectives on Racial and Ethnic Differences in Health \nin Late Life. National Academies Press (US); 2004. Accessed May 19, 2023. \nhttps://www.ncbi.nlm.nih.gov/books/NBK25531/ \n37. Mpofu JJ. Perceived Racism and Demographic, Mental Health, and Behavioral \nCharacteristics Among High School Students During the COVID-19 Pandemic — \nAdolescent Behaviors and Experiences Survey, United States, January–June 2021. MMWR \nSuppl. 2022;71. doi:10.15585/mmwr.su7103a4 \n38. Cuevas AG, O’Brien K, Saha S. African American experiences in healthcare: “I always feel \nlike I’m getting skipped over.” Health Psychol Off J Div Health Psychol Am Psychol Assoc. \n2016;35(9):987-995. doi:10.1037/hea0000368 \n39. Cuevas AG, O’Brien K, Saha S. What is the key to culturally competent care: Reducing bias \nor cultural tailoring? Psychol Health. 2017;32(4):493-507. \ndoi:10.1080/08870446.2017.1284221 \n40. Medicaid Eligibility | Medicaid. Accessed February 23, 2023. \nhttps://www.medicaid.gov/medicaid/eligibility/index.html \n41. McWilliams JM. Health Consequences of Uninsurance among Adults in the United States: \nRecent Evidence and Implications. Milbank Q. 2009;87(2):443-494. doi:10.1111/j.1468-\n0009.2009.00564.x \n42. Russo V, Orlando V, Monetti VM, et al. Geographical Variation in Medication Prescriptions: \nA Multiregional Drug-Utilization Study. Front Pharmacol. 2020;11:418. \ndoi:10.3389/fphar.2020.00418 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n19 \n43. Rosenberg A, Fucile C, White RJ, et al. Visualizing nationwide variation in medicare Part D \nprescribing patterns. BMC Med Inform Decis Mak. 2018;18(1):103. doi:10.1186/s12911-\n018-0670-2 \n \nTables \nTable 1: Descriptive statistics of the endometriosis and non-endometriosis cohorts. \nThis table contains information about the age, race and ethnicity of patients in the endometrsiosi \nand non-endometriosis cohorts. Age is calculated as the date of cohort entry (date of diagnosis \nfor endometriosis patients, date of most recent visit for non-endometriosis patients). \n Endometriosis Cohort  \n(n = 16,372) \nNon-endometriosis Cohort  \n(n = 3,663,904) \nAge   \n15-21 9.9% (n = 1,616) 33.7% (n = 1,233,489) \n22-28 22.2% (n = 3,634) 20.6% (n = 755,365) \n29-35 32.5% (n = 5,323) 18.3% (n = 671,403) \n35-42 23.2% (n = 3,804) 13.6% (n = 499,965) \n42-49 12.2% (n = 1,995) 13.7% (n = 503,682) \nRace   \nBlack or African American 23.3% (n = 3,814) 17.3% (n = 1,199,859) \nWhite 66.0% (n = 10,805) 32.7% (n = 1,829,312) \nNo Matching Concept 10.7% (n = 1,753) 50.0% (n = 634,733) \nEthnicity   \nHispanic or Latino 1.6% (n = 264) 3.8% (n = 138,634) \nNo Matching Concept 98.4% (n = 16,108) 96.2% (n = 3,525,270) \n  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n20 \nTable 2: Overall prescription prevalence in the endometriosis and the non-endometriosis cohort.  \nThis table shows the differences in prescription prevalence (mean prevalence difference) for \nWhite and Black patients in the endometriosis and non-endometriosis cohorts. The standard error \nand adjusted p-value associated with the difference in prevalence is also reported. Significant p-\nvalues (p < 0.01) are emphasized in bold.  \nDrug Class \nEndometriosis Cohort Non-endometriosis Cohort \nMean Prevalence \nDifference (%) \nStandard \nError (%) \nAdjusted \np-value \nMean Prevalence \nDifference (%) \nStandard \nError (%) \nAdjusted  \np-value \nAntidepressants 20.7092674 0.8779937 7.33E-122 22.9035692 0.05304433 0 \nAnxiolytics 20.3798907 0.9113823 1.31E-109 15.6527752 0.05042332 0 \nAntiepileptics 18.6863978 0.92753865 4.07E-89 12.9321962 0.04436682 0 \nEstrogens 10.2351817 0.76278335 6.62E-40 1.8295592 0.02425678 0 \nCough suppressants, \nexcl. Combinations \nwith expectorants 10.2306415 0.7721378 6.34E-39 5.98557177 0.03181319 0 \nUrologicals 8.41948621 0.73720344 4.61E-29 2.94880655 0.02595341 0 \nDrugs for functional \ngastrointestinal \ndisorders 8.11512033 0.83853302 5.25E-21 2.98066657 0.02871509 0 \nCorticosteroids for \nsystemic use, plain 7.61934586 0.67395744 1.73E-28 12.0253099 0.05781293 0 \nHypnotics and \nsedatives 6.73798789 0.88230912 3.12E-13 7.11677719 0.04465059 0 \nThyroid preparations 6.72891005 0.44941487 1.55E-49 4.43773974 0.02185524 0 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n21 \nQuinolone \nantibacterials 6.51072657 0.90775091 1.03E-11 6.41267606 0.04888034 0 \nOther analgesics and \nantipyretics 6.35556622 0.7432395 1.71E-16 7.47203815 0.05709184 0 \nDrugs for peptic ulcer \nand gastro-oesophageal \nreflux disease (GORD) 6.22370589 0.91284682 1.29E-10 7.15618371 0.05234589 0 \nAntiemetics and \nantinauseants 4.93577208 0.63104266 7.30E-14 7.03386181 0.05758966 0 \nBelladonna and \nderivatives, plain 4.5541342 0.74166983 1.15E-08 2.25201094 0.02541702 0 \nAntihistamines for \nsystemic use 4.28910075 0.69098553 7.55E-09 3.91808881 0.05868259 0 \nHormonal \ncontraceptives for \nsystemic use 3.26120892 0.88940093 0.00343917 -1.9807205 0.05820126 1.04E-252 \nOther systemic drugs \nfor obstructive airway \ndiseases 1.75543863 0.64534074 0.0913468 1.57064744 0.02842173 0 \nOther diagnostic agents 1.52705867 0.62911922 0.21296566 0.67426432 0.0239525 3.35E-173 \nAntiinflammatory and \nantirheumatic products, \nnon-steroids 1.35203434 0.43286802 0.02502562 1.17594879 0.0579422 1.98E-90 \nOpioids 1.00063649 0.3378741 0.04285102 4.93323827 0.05817166 0 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n22 \nSelective calcium \nchannel blockers with \ndirect cardiac effects -0.1429571 0.36793811 9.76667522 0.14467697 0.01170009 5.63E-34 \nAntifibrinolytics -0.7107767 0.28571871 0.18001115 0.00790219 0.00613992 2.77322788 \nDrugs for constipation -2.4573705 0.94071259 0.12593019 -1.0426026 0.04477045 8.26E-119 \nProgestogens -4.0913806 0.93452218 0.00016765 -7.8743071 0.05176947 0 \nAntiinfectives and \nantiseptics, excl. \nCombinations with \ncorticosteroids -6.9788283 0.87299686 1.83E-14 -6.3718204 0.0524945 0 \nHormones and related \nagents -7.1121568 0.9398425 5.33E-13 -8.9228181 0.04865653 0 \nIron preparations -7.2843687 0.71144019 1.86E-23 -4.4471093 0.03375195 0 \n \n  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n23 \nTable 3: Disparities in Prescription Prevalence Pre-Diagnosis and Post-Diagnosis.  \nThis table contains all 28 drug classes and the difference in prevalence between White patients \nand Black patients. The associated standard error and p-value are also reported. A positive mean \nprevalence difference indicates that White patients were prescribed a drug within that class more \noften. Significant p-values (p < 0.01) are emphasized in bold.  \nDrug Class \nPre-Diagnosis Post-Diagnosis \nMean Prevalence \nDifference (%) \nStandard \nError (%) \nAdjusted  \np-value \nMean Prevalence \nDifference (%) \nStandard \nError (%) \nAdjusted  \np-value \nAntidepressants 24.2136409 0.91910921 8.27E-152 19.1033773 0.93308442 5.21E-92 \nAnxiolytics 21.6790021 0.92689508 7.77E-120 17.5023318 0.91112685 4.31E-81 \nAntiepileptics 19.0576737 0.8894292 1.05E-100 14.6465251 0.89299566 2.61E-59 \nCorticosteroids for \nsystemic use, plain 13.4112531 0.88354112 6.82E-51 7.09665138 0.90348168 5.61E-14 \nOther analgesics and \nantipyretics 10.647111 0.89842351 2.98E-31 5.48276992 0.9261711 4.51E-08 \nHypnotics and sedatives 10.0331305 0.92767679 4.08E-26 4.99233052 0.94689986 1.89E-06 \nAntiemetics and \nantinauseants 9.36654584 0.88938431 8.66E-25 6.54327582 0.87627634 1.15E-12 \nCough suppressants, \nexcl. Combinations with \nexpectorants 8.58926626 0.64989361 9.88E-39 4.5777826 0.60477066 5.25E-13 \nDrugs for functional \ngastrointestinal disorders 7.06331071 0.72801319 4.13E-21 3.11520408 0.68196701 6.89E-05 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n24 \nAntihistamines for \nsystemic use 6.94421682 0.82975723 8.14E-16 2.29503444 0.93040888 0.19091332 \nDrugs for peptic ulcer \nand gastro-oesophageal \nreflux disease (GORD) 6.79679748 0.94385519 8.36E-12 4.34071438 0.92647094 3.92E-05 \nQuinolone antibacterials 6.63919703 0.94358424 2.77E-11 3.88632803 0.91046271 0.00027548 \nUrologicals 5.58218552 0.56071233 3.34E-22 5.18144241 0.62964183 2.64E-15 \nThyroid preparations 5.44076752 0.40182738 1.27E-40 5.21899767 0.40262683 2.80E-37 \nHormonal contraceptives \nfor systemic use 4.08447171 0.93484337 0.00017462 1.75921784 0.92248195 0.79120025 \nOpioids 3.83223567 0.64179058 3.30E-08 2.73050637 0.66811974 0.00061217 \nBelladonna and \nderivatives, plain 3.71041927 0.64993182 1.59E-07 1.78282791 0.53391252 0.01176382 \nEstrogens 3.70365484 0.5754246 1.71E-09 8.54471765 0.65909065 2.73E-37 \nAntiinflammatory and \nantirheumatic products, \nnon-steroids 2.21986375 0.65859413 0.01050029 0.90768334 0.76028642 3.25540217 \nOther systemic drugs for \nobstructive airway \ndiseases 1.3305698 0.54676884 0.20934261 1.38698549 0.52434092 0.11430042 \nOther diagnostic agents 1.24341988 0.51596245 0.22339418 0.37362264 0.45886946 5.81723098 \nSelective calcium \nchannel blockers with \ndirect cardiac effects -0.0672012 0.31416557 11.628708 -0.1512507 0.26786714 8.01238248 \nAntifibrinolytics -0.2432193 0.22567947 3.9362227 -0.5235776 0.19670476 0.10883203 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n25 \nDrugs for constipation -1.1445405 0.87089075 2.6428188 -3.0364077 0.8782654 0.00763926 \nProgestogens -3.9202865 0.93975007 0.00042341 -4.4411163 0.87031541 4.68E-06 \nIron preparations -5.5722235 0.64193868 5.53E-17 -3.3586467 0.4702192 1.28E-11 \nHormones and related \nagents -6.1206816 0.92499823 5.13E-10 -6.4695481 0.8338995 1.21E-13 \nAntiinfectives and \nantiseptics, excl. \ncombinations with \ncorticosteroids -7.8604257 0.9433205 1.11E-15 -6.7822939 0.94208523 8.48E-12 \n \n  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n26 \nFigures\n \nFigure 1: Temporal subgroups for patient prescriptions.  \nThree example longitudinal records are shown, with prescriptions marked as “x” and a \nlaparoscopic diagnosis of endometriosis marked with a vertical line. As shown in the figure, \nprescriptions ordered prior to diagnosis fall into subgroup 1 (pre-diagnosis, orange) and \nprescriptions ordered after diagnosis fall into subgroup 2 (post-diagnosis, blue). Prescriptions \nfrom the day of diagnosis are not counted in either subgroup.  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n27 \n \nFigure 2: Difference in drug class prescription prevalence for White and Black patients.  \nThis graph shows the percent difference in drug class prescription prevalence between White and \nBlack patients. Error bars indicate standard error, and a difference greater than 0 indicates that \nWhite patients are prescribed a drug from that class with higher prevalence than Black patients. \nThe percent difference is shown across all visits, visits prior to diagnosis (subgroup 1), and visits \nafter diagnosis (subgroup 2).  \n  \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n28 \nSupplementary Material \nSupplemental Table 1: Difference in relative prevalence between endometriosis and \ncomparison cohort.  \nThis table contains the difference in prevalence for the 29 ATC level 3 drug classes we analyze \nin our study. Associated p-values are also shared.  \nDrug Class \nEndometriosis \nCohort: Prevalence \n(%) \nComparison \nCohort: Prevalence \n(%) \nDifference in \nPrevalence \n(%) \n \nAdjusted \np-value \nOther analgesics and \nantipyretics 43.00 36.28 6.71 \n0 \nOpioids 42.09 29.84 12.24 0 \nAntiinflammatory and \nantirheumatic products, \nnon-steroids 38.32 34.44 3.88 \n0 \nCough suppressants, \nexcl. combinations \nwith expectorants 30.46 29.50 0.96 \n1.269E-89 \nAntiemetics and \nantinauseants 30.18 22.40 7.77 \n0 \nAntidepressants 28.33 22.58 5.74 0 \nAntihistamines for \nsystemic use 26.62 32.47 -5.84 \n0 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n29 \nHormonal \ncontraceptives for \nsystemic use 26.61 22.59 4.02 \n1.863E-205 \n \nDrugs for peptic ulcer \nand gastro-oesophageal \nreflux disease (GORD) 22.87 22.06 0.81 \n0.0 \nHypnotics and \nsedatives 22.79 18.39 4.39 \n0.0 \nAnxiolytics 22.48 13.28 9.20 0.0 \nQuinolone \nantibacterials 21.50 17.22 4.27 \n0.0 \nCorticosteroids for \nsystemic use, plain 19.23 15.14 4.08 \n0.0 \nAntiepileptics 18.64 16.51 2.12 0.0 \nProgestogens 14.79 11.01 3.77 0.0 \nDrugs for constipation 13.45 8.98 4.47 4.848E-134 \nHormones and related \nagents 13.42 13.98 -0.56 \n8.297E-272 \nDrugs for functional \ngastrointestinal \ndisorders 10.79 9.89 0.90 \n0.0 \nEstrogens 9.99 5.05 4.94 7.640E-177 \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint \n\n  \n  \n30 \nAntiinfectives and \nantiseptics, excl. \ncombinations with \ncorticosteroids 8.46 4.49 3.97 \n0.0 \nThyroid preparations 7.77 5.31 2.45 7.029E-66 \nUrologicals 7.28 5.50 1.77 2.907E-103 \nOther systemic drugs \nfor obstructive airway \ndiseases 6.33 5.43 0.89 \n0.000859 \nBelladonna and \nderivatives, plain 6.15 5.64 0.50 \n4.858E-269 \nIron preparations 5.18 5.45 -0.27 3.175E-05 \nOther diagnostic agents 4.25 4.21 0.041 2.454E-64 \nAntifibrinolytics 3.64 2.013 1.63 7.487E-50 \nSelective calcium \nchannel blockers with \ndirect cardiac effects 3.38 2.72 0.65 \n4.449E-45 \n \n \nAll rights reserved. No reuse allowed without permission. \n(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. \nThe copyright holder for this preprintthis version posted October 3, 2023. ; https://doi.org/10.1101/2023.10.02.23296435doi: medRxiv preprint","source_license":"CC0","license_restricted":false}