{"paper_id":"462097e7-a121-41b9-b22a-e7dd8eb18c1f","body_text":"ORIGINAL RESEARCH\nReal-World Evaluation of Direct and Indirect\nEconomic Burden Among Endometriosis Patients\nin the United States\nAhmed M. Soliman . Eric Surrey . Machaon Bonafede . James K. Nelson .\nJane Castelli-Haley\nReceived: September 5, 2017 / Published online: February 15, 2018\n/C211The Author(s) 2018. This article is an open access publication\nABSTRACT\nIntroduction: The prevalence of endometriosis\nand the need for treatment in the USA has led to\nthe need to explore the contemporary cost\nburden associated with the disease. This retro-\nspective cohort study compared direct and\nindirect healthcare costs in patients with\nendometriosis to a control group without\nendometriosis.\nMethods: Women aged 18–49 years with\nendometriosis (date of initial diagnosis = index\ndate) were identified in the Truven Health\nMarketScan/C210Commercial database between\n2010 and 2014 and female control patients\nwithout endometriosis were matched by age\nand index year. The following outcomes were\ncompared: healthcare resource utilization\n(HRU) during the 12-month pre- and post-index\nperiods (including inpatient admissions, phar-\nmacy claims, emergency room visits, physician\noffice visits, and obstetrics/gynecology visits),\nannual direct (medical and pharmacy) and\nindirect (absenteeism, short-term disability, and\nlong-term disability) healthcare costs during the\n12-month post-index period (in 2014 US$).\nMultivariate analyses were conducted to esti-\nmate annual total direct and indirect costs,\ncontrolling for demographics, pre-index clinical\ncharacteristics, and pre-index healthcare costs.\nResults: Overall, 113,506 endometriosis\npatients and 927,599 controls were included.\nEndometriosis patients had significantly higher\nHRU during both the pre- and post-index peri-\nods compared to controls ( p\\0.0001, all cate-\ngories of HRU). Approximately two-thirds of\nendometriosis patients underwent an\nendometriosis-related surgical procedure (in-\ncluding laparotomy, laparoscopy, hysterec-\ntomy, oophorectomy, and other excision/\nablation procedures) in the first 12 months\npost-index. Mean annual total adjusted direct\ncosts per endometriosis patient during the\n12-month post-index period was over three\ntimes higher than that for a non-endometriosis\ncontrol [$16,573 (standard deviation\nEnhanced content To view enhanced content for this\narticle go to http://www.medengine.com/Redeem/\nAF3B4F601BFBE18C.\nElectronic supplementary material The online\nversion of this article ( https://doi.org/10.1007/s12325-\n018-0667-3) contains supplementary material, which is\navailable to authorized users.\nA. M. Soliman ( &) /C1J. Castelli-Haley\nAbbvie, Inc, North Chicago, IL, USA\ne-mail: ahmed.m.soliman@abbvie.com\nE. Surrey\nColorado Center for Reproductive Medicine, Lone\nTree, CO, USA\nM. Bonafede\nTruven Health Analytics, An IBM Company,\nCambridge, MA, USA\nJ. K. Nelson\nTruven Health Analytics, An IBM Company, Ann\nArbor, MI, USA\nAdv Ther (2018) 35:408–423\nhttps://doi.org/10.1007/s12325-018-0667-3\n\n\n(SD) = $21,336) vs. $4733 (SD = $14,833);\np\\0.005]. On average, incremental direct and\nindirect 12-month costs per endometriosis\npatient were $10,002 and $2132 compared to\ntheir matched controls ( p\\0.005).\nConclusions: Endometriosis patients incurred\nsignificantly higher direct and indirect health-\ncare costs than non-endometriosis patients.\nFunding: AbbVie Inc.\nKeywords: Administrative claims database;\nEndometriosis; Health economics; Women’s\nhealth; Work loss\nINTRODUCTION\nEndometriosis is a chronic gynecological disor-\nder characterized by the presence of endome-\ntrial-like tissue (glands and stroma) outside the\nuterus [1]. It is estimated that close to 200 mil-\nlion women worldwide will experience\nendometriosis [ 2]. Over 10 million women in\nthe USA have endometriosis with the highest\nprevalence among women aged between 30 and\n40 years [3, 4].\nCurrent therapeutic options for managing\nendometriosis include pharmacological and\nsurgical treatments [5, 6]. While pharmacologi-\ncal agents (including oral contraceptives, non-\nsteroidal anti-inflammatory drugs,\nprogestogens, gestrinone, and gonadotropin-\nreleasing hormone agonists) are often used as\nfirst-line therapy for reducing pain, surgical\ninterventions (including laparoscopic resection,\nlaparotomy, bilateral salpingo-oophorectomy,\nand hysterectomy) are the mainstay of treat-\nment for patients with advanced endometriosis,\nor who remain refractory to pharmacotherapy,\nor those with medication-related side effects\n[7, 8].\nTreatment of endometriosis through phar-\nmacotherapy and/or surgery imposes a poten-\ntially significant economic burden on the US\nhealthcare system [9] and on society in general\nthrough work loss and reductions in the quality\nof life of the sufferers [ 10–12]. The annual\nhealthcare cost burden in the USA associated\nwith endometriosis was estimated to be $22\nbillion in 2002, of which $17.3 billion was due\nto direct medical costs (outpatient and hospi-\ntalization) and $4.7 billion was due to indirect\ncosts (loss of productivity) [ 13]. Later, in a\nmulticenter, prospective study, the annual\nsocietal burden of endometriosis in the USA was\nestimated to be €49.6 billion (equivalent to\n$69.4 billion; US$ October 2010) [14]. The same\nstudy reported that approximately two-thirds of\nthe costs were due to lost productivity and one-\nthird was due to direct healthcare costs. Given\nthe prevalence of endometriosis, there is a need\nto estimate and update the contemporary cost\nburden of endometriosis in a real-world setting.\nThe purpose of the current retrospective study\nwas to quantify incremental direct and indirect\nhealthcare costs in the USA among newly\ndiagnosed endometriosis patients in the year\nfollowing diagnosis compared to those without\nendometriosis.\nMETHODS\nData Source\nThis study used a similar study design and\nanalytic approach of a previously published\nstudy that estimated the direct and indirect cost\nburden associated with undergoing an\nendometriosis-related surgery [ 15]. This retro-\nspective database analysis used healthcare\nclaims data from the Truven Health MarketS-\ncan\n/C210Commercial Claims and Encounters\n(Commercial) and the MarketScan /C210Health and\nProductivity Management (HPM) databases for\nthe period January 1, 2009 through June 30,\n2015. The commercial database contains pooled\ninpatient and outpatient healthcare experience,\nincluding pharmacy claims, of approximately\n40 million employees and their dependents\ncovered under a variety of health plan types in\nthe USA. Claims-based datasets contain rela-\ntively comprehensive real-world data on\nhealthcare with the ability to track large num-\nbers of patients. An important limitation of\nsuch data is that it is created from a system\ndesigned for reimbursement purposes rather\nthan for research so that, for example, there is\nan absence of information on medical expenses\nnot subject to insurance reimbursement such as\nAdv Ther (2018) 35:408–423 409\n\nfor over-the-counter medications. More specific\ninformation on the limitations of the dataset\nare detailed in the ‘‘Discussion’’.\nThe HPM database includes details on\nworkplace absenteeism, short-term disability,\nand workers’ compensation data for a subset of\nindividuals in the commercial database. The\ndata is provided from employer payroll systems\nand disability case records supplied by data\ncontributors to the Commercial database.\nInformation on specific causes of absences or\ndisability are not directly coded in the database.\nAll database records are de-identified and\nwere accessed with protocols compliant with US\npatient confidentiality requirements, including\nthe Health Insurance Portability and Account-\nability Act (HIPAA) of 1996, and thus were\nexempted from institutional review board\napproval. This article does not contain any\nstudies with human participants or animals\nperformed by any of the authors.\nStudy Population\nEndometriosis Cohort\nFemale patients (aged 18–49 years) with a diag-\nnosis of endometriosis (International Classifi-\ncation of Diseases, Ninth Revision, Clinical\nModification (ICD-9-CM) 617.x) on a non-di-\nagnostic medical claim between January 1, 2010\nand June 30, 2014 during the study time frame\nwere initially selected for analysis. The date of\nfirst diagnosis was defined as the index date.\nAdditionally, patients were considered eligible\nfor inclusion if they had continuous health plan\ncoverage for 12 months before (pre-index per-\niod) and 12 months after (follow-up period) the\nindex date. Patients with a diagnosis of\nendometriosis (ICD-9-CM 617.x) in any posi-\ntion on a non-diagnostic medical claim pre-\nceding the index date, or diagnosis of malignant\nneoplasm of female genitourinary organs (ICD-\n9-CM: 179.x-184.x) or radical hysterectomy\nduring the 12-month pre-index period were\nexcluded. ICD-9-CM codes are used here and\nthroughout for identifying diagnostic informa-\ntion because it was the required coding system\nduring the relevant time period on insurance\nreimbursement claims used in the commercial\ndatabase.\nNon-Endometriosis Control Cohort\nPatients in the non-endometriosis cohort were\nidentified using selection criteria similar to the\nendometriosis cohort. Female patients with no\ndiagnosis of endometriosis during the entire\nstudy period (January 1, 2009 through June 30,\n2015) were randomly selected and directly\nmatched 10:1 to endometriosis patients on the\nbasis of data availability in the index year and\nage at index). Index date for the control cohort\nwas randomly assigned during this time period\nwith a date distribution matching the date dis-\ntribution found among endometriosis patients.\nOutcome Measures\nHealthcare resource utilization and costs (all-\ncause and endometriosis-related) were evalu-\nated for the 12-month pre- and post-index\nperiods. The direct healthcare costs incurred\nwere assessed for healthcare services across var-\nious care settings: inpatient, emergency room\n(ER), outpatient services including obstetri-\ncian–gynecologist (OB/GYN) visits, and outpa-\ntient prescriptions. Endometriosis-related\nutilization and expenditures were defined as\n(a) medical claims with a principal diagnosis of\nendometriosis, (b) pharmacy claims for drugs\nused to treat and manage endometriosis (le-\nuprolide, danazol, depot medroxyprogesterone,\noral medroxyprogesterone acetate, levonorges-\ntrel implants and intrauterine devices, mege-\nstrol acetate, progestin-only oral contraceptive\npills, histrelin, goserelin, nafarelin, triptorelin,\nganirelix, degarelix, cetrorelix, abarelix, and\nother oral contraceptive pills), and (c) en-\ndometriosis-related surgeries (number and pro-\nportion of patients with at least one\nendometriosis-related surgery overall, and\nstratified by laparotomy, laparoscopy, hysterec-\ntomy, oophorectomy, and other excision/abla-\ntion procedures). All expenditures were\ninflation adjusted to 2014 US$ using the Medi-\ncal Care component of the Bureau of Labor\nStatistics Consumer Price Index [ 16].\n410 Adv Ther (2018) 35:408–423\n\nWork losses associated with absenteeism (the\nnumber of days absent from work), short-term\ndisability (STD), and long-term disability (LTD)\nin the 12-month post-index period were asses-\nsed for both endometriosis and non-en-\ndometriosis cohorts in terms of the proportion\nof patients with each category of work loss as\nwell as average number of days lost. To assess\nindirect costs due to work loss, a daily wage was\ncalculated using the age-, gender-, and geo-\ngraphic region-adjusted wage rate from the US\nBureau of Labor Statistics from the year 2010.\nAn estimated daily wage rate ($240 per day) was\nconsidered for calculating costs from lost pro-\nductivity due to STD/LTD. As STD/LTD benefit\ntypically does not replace full wages, 70% of\nwages, which approximates the proportion of\ntotal wages and benefits paid by employers in\nthe HPM database, was used.\nPatient Characteristics\nPatient characteristics including age, geo-\ngraphic region (US Census division), urbanicity,\nhealth plan type, and index year were measured\non the index date. In addition, the Deyo\nCharlson Comorbidity Index (DCI), comorbid\nconditions (based on the presence of ICD-9-CM\ndiagnosis), and medication use (based on\nHealthcare Common Procedure Coding System\n(HCPCS) and National Drug Code (NDC) codes)\nwere measured during the 12-month pre-index\nperiod.\nStatistical Analysis\nPatient characteristics, healthcare utilization,\nand costs for each study cohort were analyzed\ndescriptively. Continuous measures were sum-\nmarized as means and standard deviations, and\ncategorical measures were summarized as\ncounts and percentages. Testing for statistical\ndifferences in the study outcomes between\ncohorts was performed using Chi square test for\ncategorical measures, and ANOVA and t tests for\ncontinuous measures.\nThe ‘‘real-world’’ costs as determined from\nthe Commercial and HPM databases are pre-\nsented first. However, the reported cost\ndifferences for the cohorts may be influenced by\ndifferences in their baseline characteristic, so\nadjusted cost estimates were also computed.\nMultivariate analysis was used to compare both\ndirect and indirect (absence, STD, LTD, and\ntotal) cost outcomes in patients with and\nwithout endometriosis. Generalized linear\nmodels (GLMs) were used to estimate incre-\nmental costs for patients with endometriosis\nversus those without. The models used gener-\nalized estimating equations to account for\nwithin-patient correlation and used a gamma\ndistributed error and log link to account for the\ndistributional characteristics of cost data. Esti-\nmation of the standard error of the marginal\ncost was conducted by using the delta method.\nTwo-part models were used for potentially\ninfrequent outcomes and associated costs such\nas total STD costs [17]. All models controlled for\nthe relevant patient characteristics (age, gender,\ngeographic region, and health plan type, DCI,\nand baseline comorbidities) between the two\ncohorts. The recycled prediction method was\nused to estimate the marginal or incremental\ncosts associated with endometriosis. Ap value of\nless than 0.05 was considered statistically\nsignificant.\nRESULTS\nStudy Sample\nA total of 434,809 women with endometriosis\nbetween January 1, 2010 and June 30, 2014\nwere initially identified, and were directly mat-\nched on age and index year in a 1:10 ratio to\ncontrols ( N = 4,348,090) without endometrio-\nsis. Following application of all eligibility crite-\nria (Fig. 1), a total of 113,506 women with\nendometriosis and 927,599 controls were\nincluded in the analysis.\nPatient Characteristics\nPatient characteristics are presented in Table 1.\nPatients with and without endometriosis were\ndemographically similar. Endometriosis\npatients had a greater comorbid burden than\nAdv Ther (2018) 35:408–423 411\n\ndemographically matched controls, as evi-\ndenced by significantly higher mean [standard\ndeviation (SD)] CCI score for endometriosis\npatients in the pre-index period [0.21 (0.64) vs.\n0.16 (0.57); p\\0.0001]. Endometriosis patients\nalso had significantly higher rates of all the\nassessed comorbidities, with gynecological\ncomorbidities being especially common in the\n12 months prior to an endometriosis diagnosis.\nCompared to controls, the endometriosis\ncohort had a higher proportion of patients with\nprescriptions claims for opioids, nonsteroidal\nanti-inflammatory drugs (NSAIDs), antidepres-\nsants, and hormonal contraceptives (all\np\\0.0001).\nHealthcare Utilization and Expenditures\nPatients with endometriosis had a significantly\nhigher rate of healthcare utilization compared\nto their matched controls during both the pre-\nand post-index periods (Table 2, Fig. 2). During\nthe 12-month post-index period, 29% of\npatients in the endometriosis cohort reported\nan all-cause hospital admission compared to 6%\nin the control cohort ( p\\0.0001). More\nendometriosis patients than controls had an ER\nvisit, physician office visit, and OB/GYN visit in\nboth the pre- and post-index periods (all\np\\0.0001), as well as higher average numbers\nof all-cause hospital admissions (post-index\nonly), ER visits (pre- and post-index), and OB/\nGYN visits (pre- and post-index) compared to\ntheir matched controls.\nEndometriosis patients also showed similarly\nincreased utilization in the 12 months prior to\nthe index date as well as post-index, showing a\ngreater proportion of patients with ER visits,\nphysician office visits, and OB/GYN visits\n(Fig. 2).\nNearly two-thirds of the patients with\nendometriosis underwent an endometriosis-re-\nlated surgical procedure (including hysterec-\ntomy, oophorectomy, laparoscopy, laparotomy,\nand other excision or ablation) compared to\n1.0% in the controls, among which hysterec-\ntomy was the most common (38%) during the\n12-month post-index period. Less than a third\nof endometriosis patients filled relevant thera-\npeutic endometriosis-related prescriptions dur-\ning the post-index period compared to 21% of\npatients in the control cohort (Table 2).\nOverall, endometriosis patients incurred sig-\nnificantly higher annual all-cause expenditure\ncompared to the controls during both pre- and\npost-index periods (Table 2, Fig. 3). Notably, the\nmajority (62%) of the annual costs for\nFig. 1 Sample selection. The effect of applying eligibility criteria to population sample size is shown\n412 Adv Ther (2018) 35:408–423\n\nTable 1 Demographic characteristics (on index date) and clinical characteristics (during pre-index period)\nPatient characteristic Endometriosis patients\n(N 5 113,506)\nControls\n(N 5 927,599)\np value\nAge, mean (SD) 37.9 (7.4) 38.2 (7.5) \\0.0001\nAge group, N (%) \\0.0001\n18–2 7032 (6%) 58,619 (6%)\n25–29 9290 (8%) 65,369 (7%)\n30–34 18,615 (16%) 144,156 (16%)\n35–39 24,826 (22%) 201,881 (22%)\n40–44 28,939 (25%) 242,468 (26%)\n45–49 24,804 (22%) 215,106 (23%)\nPopulation density, N (%) \\0.0001\nUrban 94,243 (83%) 793,610 (86%)\nRural 17,908 (16%) 121,950 (13%)\nUnknown 1355 (1%) 12,039 (1%)\nGeographic region, N (%) \\0.0001\nNortheast 18,299 (16%) 169,818 (18%)\nNorth central 25,764 (23%) 206,305 (22%)\nSouth 47,074 (41%) 355,934 (38%)\nWest 20,931 (18%) 182,997 (20%)\nUnknown 1438 (1%) 12,545 (1%)\nHealth plan type ( N,% ) \\0.0001\nComprehensive 1131 (1%) 8732 (1%)\nEPO 1638 (1%) 14,779 (2%)\nHMO 16,234 (14%) 141,636 (15%)\nPOS 8508 (7%) 65,224 (7%)\nPPO 67,723 (60%) 538,520 (58%)\nPOS with capitation 630 (1%) 5093 (1%)\nCDHP 7674 (7%) 63,097 (7%)\nHDHP 4116 (4%) 40,811 (4%)\nUnknown 5852 (5%) 49,707 (5%)\nDeyo CCI, mean (SD) 0.21 (0.6) 0.16 (0.6) \\0.0001\nComorbid conditions, N (%)\nUpper respiratory infections 35,478 (31%) 220,514 (24%) \\0.0001\nAbdominal/pelvic pain 29,018 (26%) 65,630 (7%) \\0.0001\nAdv Ther (2018) 35:408–423 413\n\nTable 1 continued\nPatient characteristic Endometriosis patients\n(N 5 113,506)\nControls\n(N 5 927,599)\np value\nHypertension 12,640 (11%) 79,943 (9%) \\0.0001\nAnxiety 10,825 (10%) 56,401 (6%) \\0.0001\nDepression 11,745 (10%) 65,587 (7%) \\0.0001\nMigraine 8511 (7%) 36,980 (4%) \\0.0001\nAsthma 6547 (6%) 37,170 (4%) \\0.0001\nHyperlipidemia 6543 (6%) 43,726 (5%) \\0.0001\nCOPD 3026 (3%) 17,613 (2%) \\0.0001\nDiabetes 3927 (3%) 29,682 (3%) \\0.0001\nIrritable bowel syndrome 2741 (2%) 7998 (1%) \\0.0001\nOsteoarthritis 2650 (2%) 16,886 (2%) \\0.0001\nAcute coronary syndrome 898 (1%) 4775 (1%) \\0.0001\nPelvic peritoneal adhesions 1647 (1%) 1821 ( \\1%) \\0.0001\nAnal or rectal pain 484 ( \\1%) 1600 ( \\1%) \\0.0001\nBladder pain 223 ( \\1%) 586 ( \\1%) \\0.0001\nHeart failure 185 ( \\1%) 1441 ( \\1%) 0.54\nOsteoporosis 179 ( \\1%) 1528 ( \\1%) 0.58\nGynecological comorbidities, N (%)\nAny gynecological comorbidity listed below 74,053 (65%) 109,165 (12%) \\0.0001\nUnspecified symptoms of female genital\norgans\n30,057 (26%) 24,033 (3%) \\0.0001\nExcessive or frequent menstruation 25,414 (22%) 28,607 (3%) \\0.0001\nOvarian cysts 20,400 (18%) 16,864 (2%) \\0.0001\nUterine fibroids 18,618 (16%) 17,571 (2%) \\0.0001\nDysmenorrhea 15,540 (14%) 11,086 (1%) \\0.0001\nVaginitis 8477 (7%) 36,065 (4%) \\0.0001\nDyspareunia 4233 (4%) 3299 (0%) \\0.0001\nMetrorrhagia 4053 (4%) 5478 (1%) \\0.0001\nReproductive claims, N (%)\nPregnancy/delivery 6513 (6%) 65,332 (7%) \\0.0001\nInfertility 6127 (5%) 9077 (1%) \\0.0001\nFertility treatments 5462 (5%) 11,076 (1%) \\0.0001\n414 Adv Ther (2018) 35:408–423\n\nendometriosis patients were incurred within\n3 months post-index. In the 12-month post-in-\ndex period, the endometriosis-related expendi-\ntures accounted for about 39% and 1% of the\ntotal healthcare expenditures in the\nendometriosis and control cohorts, respectively\n(Fig. 4). The ‘‘endometriosis-related’’ expendi-\ntures among controls arose from prescriptions\nfor drugs listed in the ‘‘ Methods’’ section as\n‘‘endometriosis-related’’, but many of which\nalso have non-endometriosis indications.\nHealthcare and pharmacy costs tend to increase\nwith the age of the patient, with older\nendometriosis patients (age 40–49) having\n$3171 more in mean costs compared to younger\npatients (age 18–29), whereas the difference\namong controls was $1471 (Table 3). Here and\nthroughout the ‘‘Results’’ sections, all costs are\npresented in 2014 US$.\nWork Loss and Indirect Costs\nWork loss was higher in the endometriosis\ncohort (Table 2), particularly in terms of days of\nabsence and STD. Rates of LTD claims were low\nin both cohorts (although significantly higher\nin the endometriosis cohort both in terms of\nproportion of patients with a claim and total\nnumber of LTD days). This resulted in higher\nmean indirect cost estimates among\nendometriosis patients vs. controls for all cate-\ngories of work loss: $5383 vs. $4224 per patient\nfor the cost of absenteeism, ( p\\0.0001), $1709\nvs. $402 per patient for the cost of STD\n(p\\0.0001), and $54 vs. $26 per patient for the\ncost of LTD.\nAdjusted Direct and Indirect Healthcare\nExpenditures\nCost estimates have been presented which\ndescribe what is found in the real-world data.\nAdditionally, multivariate analysis, detailed in\nthe ‘‘Methods’’ section, provided cost estimates\nadjusted for baseline patient characteristics\nwhich may have differed between the cohorts.\nOverall, patients with endometriosis showed a\nsignificant incremental burden when compared\nto those without endometriosis. The mean\nannual incremental direct healthcare cost of\nendometriosis was estimated at $10,002, while\nthe mean incremental indirect costs related to\nabsence hours and STD compared to non-en-\ndometriosis controls were $903 and $1228,\nrespectively, (all p\\0.005; Table 4). Please see\nTables S1–S5 in the supplementary material for\nadditional details on the multivariate results.\nDISCUSSION\nThis retrospective study estimated the incre-\nmental direct and indirect costs associated with\nendometriosis among commercially insured\nTable 1 continued\nPatient characteristic Endometriosis patients\n(N 5 113,506)\nControls\n(N 5 927,599)\np value\nMedications, N (%)\nOpioids 101,557 (89%) 573,082 (62%) \\0.0001\nNSAIDS 84,498 (74%) 468,703 (51%) \\0.0001\nAntidepressants 54,055 (48%) 309,289 (33%) \\0.0001\nEstrogen/progestin oral contraceptives 51,060 (45%) 319,839 (34%) \\0.0001\nCCI Charlson comorbidity index, CDHP consumer driven health plan, COPD chronic obstructive pulmonary disease, EPO\nexclusive provider organization, HDHP high deductible health plan, HMO health maintenance organization, NSAIDS\nnonsteroidal anti-inflammatory drugs, POS point of service, PPO preferred provider organization, SD standard deviation\nAdv Ther (2018) 35:408–423 415\n\nTable 2 Utilization and costs during 12-month post-index period\nOutcome Endometriosis patients\n(N 5 113,506)\nControls\n(N 5 927,599)\np value\nHealthcare utilization (all-cause)\nPatients with inpatient admission, N (%) 33,235 (29%) 60,052 (6%) \\0.0001\nAdmissions per patient, mean (SD) 0.34 (0.62) 0.08 (0.33) \\0.0001\nPatients with emergency room visit, N (%) 35,944 (32%) 168,471 (18%) \\0.0001\nER visits per patient, mean (SD) 0.63 (1.62) 0.28 (0.87) \\0.0001\nPatients with physician office visit, N (%) 109,835 (97%) 804,959 (87%) \\0.0001\nPhysician office visits per patient, mean (SD) 8.0 (7.2) 4.6 (5.2) \\0.0001\nPatients with OB/GYN specialist visit, N (%) 71,846 (63%) 341,295 (37%) \\0.0001\nOB/GYN office visits per patient, mean (SD) 1.8 (2.7) 0.62 (1.2) \\0.0001\nPatients with outpatient prescription claim,\nN (%)\n108,828 (96%) 766,720 (83%) \\0.0001\nPrescriptions per patient, mean (SD) 20.2 (20.7) 12.0 (16.0) \\0.0001\nHealthcare utilization (endometriosis-related)\nPatients with inpatient admission, N (%) 18,495 (16%) 0.0 \\0.0001\nAdmissions per patient, mean (SD) 0.17 (0.38) 0.0 \\0.0001\nPatients with emergency room visit, N (%) 4599 (4%) 0.0 \\0.0001\nER visits per patient, mean (SD) 0.05 (0.25) 0.0 \\0.0001\nPatients with physician office visit, N (%) 51,133 (45%) 0.0 \\0.0001\nPhysician office visits per patient, mean (SD) 0.79 (1.35) 0.0 \\0.0001\nPatients with OB/GYN specialist visit, N (%) 34,966 (31%) 0.0 \\0.0001\nOB/GYN office visits per patient, mean (SD) 0.49 (0.99) 0.0 \\0.0001\nPatients with outpatient prescription claim,\nN (%)\n34,008 (30%) 199,287 (21%) \\0.0001\nPrescriptions per patient, mean (SD) 1.5 (3.2) 1.3 (3.2) \\0.0001\nPatients with endometriosis-related surgery,\nN (%)\n75,935 (66.9%) 10,735 (1.2%) \\0.0001\nLaparotomy 3758 (3.3%) 797 (0.1%) \\0.0001\nLaparoscopy 37,238 (32.8%) 4766 (0.5%) \\0.0001\nHysterectomy 43,030 (37.9%) 5575 (0.6%) \\0.0001\nHysterectomy w/same-day oophorectomy 2206 (1.9%) 195 (0.0%) \\0.0001\nOophorectomy 4050 (3.6%) 484 (0.1%) \\0.0001\nOther excision/ablation 571 (0.5%) 86 (0.0%) \\0.0001\n416 Adv Ther (2018) 35:408–423\n\npatients in the USA and showed that the direct\nhealthcare costs for endometriosis patients\nmeasured both 12 months prior to and after\nendometriosis diagnosis were significantly\nhigher than those in matched controls. Addi-\ntionally, costs due to work loss through absen-\nteeism, STD, and LTD were also higher in\npatients with endometriosis. To our knowledge,\nthis is the first major study that compared both\ndirect and indirect healthcare costs between\npatients with and without endometriosis.\nAnalogous to the findings of previous studies\n[9, 13, 14] this study showed that endometriosis\npatients incurred significantly higher health-\ncare costs compared to controls. The incre-\nmental costs per patient in the first year post-\nTable 2 continued\nOutcome Endometriosis patients\n(N 5 113,506)\nControls\n(N 5 927,599)\np value\nHealthcare costs\nAll-cause, mean (SD) $16,573 (21,336) $4733 (14,833) \\0.0001\nIn terms of per patient per month (SD) $1381 (1778) $394 (1236) \\0.0001\nAll-cause pharmacy costs, mean (SD) $1784 (4696) $1070 (3910) \\0.0001\nEndometriosis-related, mean (SD) $6498 (9046) $89 (497) \\0.0001\nEndometriosis-related pharmacy costs, mean\n(SD)\n$212 (800) $89 (497) \\0.0001\nWork loss\nPatients with absence data, N 1365 10,051\nPatients with absence claim, N (%) 955 (70.0%) 6782 (67.5%) 0.065\nAbsence days, mean (SD) 31.1 (20.9) 24.3 (18.7) \\0.0001\nPatients with short-term disability (STD)\ndata, N\n9701 78,266\nPatients with STD claim, N (%) 3464 (35.7%) 5756 (7.4%) \\0.0001\nSTD days, mean (SD) 14.3 (25.1) 3.3 (16.6) \\0.0001\nPatients with long-term disability (LTD) data,\nN\n8833 68,271\nPatients with LTD claim, N (%) 42 (0.5) 149 (0.2) \\0.0001\nLTD days, mean (SD) 0.4 (8.8) 0.2 (6.2) 0.0036\nIndirect costs due to work loss\nAbsence, mean (SD) $5383 ($3807) $4224 ($3352) \\0.0001\nShort-term disability, mean (SD) $1709 ($3057) $402 ($2064) \\0.0001\nLong-term disability, mean (SD) $54 ($1107) $26 ($770) 0.0035\nAll costs are estimated as per 2014 US$\nOB/GYN obstetrician or gynecologist ,S D standard deviation\nAdv Ther (2018) 35:408–423 417\n\ndiagnosis (over $10,000 with multivariable\nadjustment) are substantial compared to esti-\nmates for other common gynecological ail-\nments such as uterine fibroids ($6873 in 2014\nUS$ [18]) or menorrhagia ($2878 in 2014 US$\n[19]). Notably, over 60% of the total annual\nhealthcare costs for endometriosis patients were\naccrued within 3 months of the index date,\nthereby adding to the existing body of literature\nwhich has reported highest healthcare resource\nutilization and costs in the first year after\nendometriosis diagnosis [20–22].\nIndirect costs associated with endometriosis\nare substantial as well, with average incremental\nindirect costs estimated at US $2132 per patient\n(adjusted estimate) in the 12 months following\nthe index diagnosis. A point to notice here is\nthat claims data can only capture indirect losses\ndue to the absenteeism component; data on\npresenteeism (reduced productivity due to ill-\nness while on the job) was not available. Esti-\nmates from a study of European Union\ncountries suggest that indirect costs due to\npresenteeism can substantially exceed the costs\nof absenteeism, by a factor of 2: €6298 due to\npresenteeism vs. €3280 due to absenteeism [14].\nA study of 10 countries across multiple conti-\nnents estimated that endometriosis resulted in\nthe equivalent work loss of 6.4 h/week due to\npresenteeism and 4.4 h/week due to\nFig. 2 All-cause healthcare utilization in the 12-month pre- and post-index periods. All endometriosis vs. control\ndifferences in utilizations in the pre-index periods and in the post-index period are significant ( p\\0.0001)\nFig. 3 All-cause healthcare costs during 12-month pre- and post-index follow-up period. Values in each time period are\ncumulative cost since index\n418 Adv Ther (2018) 35:408–423\n\nabsenteeism [ 23] and provided indirect cost\nestimates for the USA of about $3200/year for\nabsenteeism due to endometriosis symptoms\nand about $14,800 for presenteeism (values\nconverted into 2014 US$). Instead of using\nrecords of absence and disability claims, that\nstudy based indirect cost estimates on responses\nby patients to versions of the Work Productivity\nand Activity Impairment (WPAI:GH) and Short-\nForm–36 version 2 (SF36v2), and estimated the\nwork loss in terms of lost wages without\nadjustments for reduced pay for short- and\nlong-term disability.\nA previous study characterized the health-\ncare costs and utilization of endometriosis\npatients [22]. Although the prior study followed\ncosts and utilization patterns over a longer\nperiod of time (up to 5 years for a small per-\ncentage of patients), the current analysis\nexpands upon this previous work in several\nways: by reporting endometriosis-related costs\nseparately, examining trends within the first\nFig. 4 Endometriosis-related healthcare costs during 12-month follow-up period. Values over post-period are cumulative\ncost since index. Values in each time period following index date are cumulative costs since index date\nTable 3 Healthcare costs by age during 12-month post-index period\nOutcome Endometriosis patients Controls p value\nAge 18–29 N = 16,322 N = 123,988\nOverall healthcare costs, mean (SD) $14,369 (19,963) $3649 (12,797) \\0.0001\nPharmacy costs, mean (SD) $1646 (4004) $736 (3210) \\0.0001\nAge 30–39 N = 43,441 N = 346,037\nOverall healthcare costs, mean (SD) $16,203 (20,765) $4609 (16,166) \\0.0001\nPharmacy costs, mean (SD) $1858 (4742) $931 (3489) \\0.0001\nAge 40–49 N = 53,743 N = 457,574\nOverall healthcare costs, mean (SD) $17,541 (22,122) $5120 (16,435) \\0.0001\nPharmacy costs, mean (SD) $1766 (4850) $1265 (4349) \\0.0001\nAll costs are estimated as per 2014 US$\nSD standard deviation\nAdv Ther (2018) 35:408–423 419\n\nyear post-diagnosis, and by providing multi-\nvariate models to provide cost estimates adjus-\nted for variations in baseline covariates. It also\ncontains a somewhat broader sample of\npatients, including women up to 49 years of age\n(as opposed to 45 years in Fuldeore et al.’s\nanalysis [22]). Importantly, this study provides\nan update in an evolving field of treatment, and\nconfirms changes in approaches to gynecologi-\ncal treatment.\nA recent analysis shows that hysterectomies\nand oophorectomies (as treatments across all\nindications and not just endometriosis) have\nshown modest decreases in prevalence since the\nyear 2000 with a dramatic shift towards outpa-\ntient instead of inpatient procedures [21]. In an\nanalysis spanning the years 2000–2010, Fulde-\nore et al. [ 22] reported that 40% of\nendometriosis patients received an inpatient\nadmission in their first year. In contrast, in the\ncurrent analysis spanning the years 2010–2014\nonly 29% received an inpatient admission.\nThese numbers include admissions for all rea-\nsons, not just endometriosis, but because sur-\ngical treatment is very common within the first\nyear of diagnosis [ 9, 21], this is very likely rela-\nted to a drop in inpatient treatment for\nendometriosis.\nDespite changes in the treatment landscape,\nsurgery remains a mainstay of endometriosis\ntherapy in our analysis, with 38% of the\npatients undergoing a hysterectomy in the first\nyear post diagnosis, and two-thirds undergoing\na surgical procedure used to treat endometriosis.\nThis high rate of hysterectomy in the presence\nof less invasive surgical interventions and\npharmacotherapy might represent a diagnostic\ndelay necessitating a more invasive treatment\napproach; future research should explore this\nfurther.\nMost newly diagnosed endometriosis\npatients have incurred pharmacy claims for\npain management, and an exceedingly high\nproportion of our sample of endometriosis\npatients, nearly 90%, had at least one prescrip-\ntion for opioids. The use of opioids for pain\nmanagement for endometriosis warrants further\ninvestigation.\nA notable strength of our study is that it\nincluded a large cohort of endometriosis and\nmatched non-endometriosis patients to evalu-\nate the direct and indirect healthcare costs.\nAdditionally, this study provides a current cost\ntrend for endometriosis by using data from a\nmore contemporary time frame than previously\navailable.\nThis study has several limitations inherent to\nadministrative claims data. As patients were\nidentified and variables were captured through\nadministrative claims data, endometriosis-re-\nlated symptoms, which are not recorded as a\ndiagnosis on a medical claim, may not have\nbeen captured. The diagnosis of endometriosis\nand clinical characteristics were determined\nusing ICD-9-CM diagnosis or procedure codes,\nwhich are subject to data coding limitations and\ndata entry error, and it is possible that women\nwith endometriosis may have inadvertently\nTable 4 Multivariable analysis of costs per patient during 12-month post-index period\nType of cost Endometriosis\npatients\nControls Incremental cost\n(endometriosis vs.\ncontrols)\nCost ratio (95%\nconfidence interval)\nTotal healthcare costs, mean (SD) $14,649 ($17,675) $4646 ($5606) $10,002 3.15 (3.12–3.18)\nIndirect costs, mean (SD)\na $6819 (1698) $4687 ($1167) $2132 1.46 (1.38–1.54)\nAbsence costs $5157 ($1312) $4254 ($1082) $903 1.21 (1.16–1.27)\nShort-term disability costs $1632 ($1900) $404 ($471) $1228 4.04 (3.81–4.28)\nAll costs are estimated as per 2014 US$\nAll cohort differences are statistically significant p\\0.001 (endometriosis vs. control cohorts)\nSD standard deviation\na Includes absence, short-term disability, and long-term disability data\n420 Adv Ther (2018) 35:408–423\n\nbeen included in the control group including\nthose with symptoms but for whom a diagnosis\nhad not been made yet or those for whom a\ndiagnosis had not been recorded. The use of\nover-the-counter medications and other self-\nmanagement techniques is not usually captured\nin the claims data, so the use of over-the-\ncounter NSAIDs or other pain management\nmedications is not included in our analysis. The\nindirect costs included in this study are limited\nto absenteeism and disability claims by the\nemployee and do not include reduced presen-\nteeism or potential caregiver burden, as well as\nany other humanistic burden associated with\nthis chronic condition. The absence costs\nthemselves may be conservative, as the disrup-\ntion of missing workers may result in effectively\nmore work loss than is reflected in each\npatient’s missing hours [24]. Endometriosis may\nhave important effects, such as on the psycho-\nlogical well-being of patients [ 25–27], that fall\noutside of the scope of the data used in this\nanalysis. Finally, findings from this observa-\ntional study may be prone to bias from non-\nrandom selection into the treatment group and\nwas limited to endometriosis patients covered\nby Commercial health plans. Therefore, the\nresults are not generalizable to those covered\nunder other types of insurance or who lack\ncoverage.\nCONCLUSION\nPatients with endometriosis had significantly\nhigher direct healthcare costs and indirect costs\n(as measured by absenteeism, short-term dis-\nability, and long-term disability) compared to\npatients without endometriosis in the\n12 months follow-up period. Given its preva-\nlence, this suggests a substantial disease burden\nassociated with endometriosis to the individual,\nhealthcare system, and society.\nACKNOWLEDGEMENTS\nFunding. This study and article processing\nfees (including Open Access fee) were funded by\nAbbVie Inc., and conducted by Truven Health\nAnalytics, USA.\nMedical Writing Assistance. Editorial/writ-\ning assistance for this manuscript was provided\nby Santosh Tiwari, an employee of Truven\nHealth Analytics. AbbVie funded editorial/writ-\ning assistance services.\nAuthorship. All named authors meet the\nInternational Committee of Medical Journal\nEditors (ICMJE) criteria for authorship for this\narticle, take responsibility for the integrity of\nthe work as a whole, and have given their\napproval for this version to be published.\nDisclosures. Ahmed M Soliman is an\nemployee of AbbVie Inc. and may own AbbVie\nstocks/stock options. Jane Castelli-Haley is an\nemployee of AbbVie Inc. and may own AbbVie\nstocks/stock options. Eric Surrey is medical\ndirector at the Colorado Center for Reproduc-\ntive Medicine, has served in a consulting role on\nresearch to AbbVie, and is on the speaker bureau\nfor Ferring Laboratories. Machaon Bonafede is\nan employee of Truven Health Analytics, an\nIBM Company. James K Nelson is an employee\nof Truven Health Analytics, an IBM Company.\nCompliance with Ethics Guidelines. This\narticle does not contain any studies with\nhuman participants or animals performed by\nany of the authors. The research utilized de-\nidentified patient data from administrative\nclaims using codes in compliance with the\nHealth Insurance Portability and Accountability\nAct (HIPAA) of 1996. Thus, it was exempt from\nthe institutional review board approval.\nData Availability. The datasets generated\nduring and/or analyzed during the current\nstudy are not publicly available because of their\nproprietary nature but are available from the\ncorresponding author on reasonable request.\nOpen Access. This article is distributed\nunder the terms of the Creative Commons\nAttribution-NonCommercial 4.0 International\nLicense ( http://creativecommons.org/licenses/\nby-nc/4.0/), which permits any\nAdv Ther (2018) 35:408–423 421\n\nnoncommercial use, distribution, and repro-\nduction in any medium, provided you give\nappropriate credit to the original author(s) and\nthe source, provide a link to the Creative\nCommons license, and indicate if changes were\nmade.\nREFERENCES\n1. Burney RO, Giudice LC. Pathogenesis and patho-\nphysiology of endometriosis. Fertil Steril.\n2012;98:511–9.\n2. Adamson GD, Kennedy S, Hummelshoj L. Creating\nsolutions in endometriosis: global collaboration\nthrough the World Endometriosis Research Foun-\ndation. J Endometr Pelvic Pain Disord, 2010;2(1):\n3–6. http://www.nichd.nih.gov/health/topics/\nendometri/Pages/default.aspx. Accessed 12 July\n2017.\n3. Luciano AA, LaMonica R, Luciano DE. Strategies\nand steps for the surgical management of\nendometriosis. OBG Manag. 2011;23:34–50.\n4. Juneau biosciences, LLC. End to endo. Genetic\nendometriosis research study. http://www.\nendtoendo.com/Endometriosis_Overview_End_to_\nEndometriosis.html. Accessed 12 July 2017.\n5. Armstrong C. ACOG updates guideline on diagnosis\nand treatment of endometriosis. Am Fam Phys.\n2011;83:84–5.\n6. Streuli I, de Ziegler D, Santulli P, et al. An update on\nthe pharmacological management of endometrio-\nsis. Expert Opin Pharmacother. 2013;14:291–305.\n7. Olive DL, Schwartz LB. Endometriosis. N Engl J\nMed. 1993;328:1759–69.\n8. Bedaiwy MA, Barker NM. Evidence based surgical\nmanagement of endometriosis. Middle East Fertil\nSoc J. 2012;17:57–60.\n9. Soliman AM, Yang H, Du EX, Kelley C, Winkel C.\nThe direct and indirect costs associated with\nendometriosis: a systematic literature review. Hum\nReprod. 2016;31:712–22.\n10. Gao X, YC Y, Outley J, Simon J, Botteman M,\nSpalding J. Health-related quality of life burden of\nwomen with endometriosis: a literature review.\nCurr Med Res Opin. 2006;22(9):1787–97.\n11. Soliman AM, Coyne KS, Zaiser E, Castelli-Haley J,\nSnabes MC, Surrey ES. The effect of endometriosis\nsymptoms on absenteeism and presenteeism in the\nworkplace and at home. J Manag Care Spec Pharm.\n2017;23(7):745–54.\n12. Soliman AM, Coyne KS, Gries KS, Castelli-Haley J,\nFuldeore MJ. The burden of endometriosis symp-\ntoms on health-related quality of life in women in\nthe United States: a cross-sectional study. J Psycho-\nsom Obstet Gynaecol. 2017;38:238–48 .\n13. Simoens S, Hummelshoj L, D’Hooghe T.\nEndometriosis cost estimates and methodological\nperspective. Human Reprod Update.\n2007;13:394–404.\n14. Simoens S, Dunselman G, Dirksen C, et al. The\nburden of endometriosis: costs and quality of life of\nwomen with endometriosis and treated in referral\ncenters. Hum Reprod. 2012;27:1292–9.\n15. Soliman AM, Taylor H, Bonafede M, Nelson JK,\nCastelli-Haley J. Incremental direct and indirect\ncost burden attributed to endometriosis surgeries in\nthe United States. Fertil Steril.\n2017;107(5):1181–90.\n16. United States Department of Labor: bureau of labor\nstatistics. Consumer price index (CPI). https://\nwww.bls.gov/cpi/. Accessed 12 July 2017.\n17. Mullahy J. Much ado about two: reconsidering\nretransformation and the two-part model in health\neconometrics. J Health Econ. 1998;17:247–81.\n18. Fuldeore M, Yang H, Soliman AM, Winkel C.\nHealthcare utilization and costs among women\ndiagnosed with uterine fibroids: a longitudinal\nevaluation for 5 years pre- and post-diagnosis. Curr\nMed Res Opin. 2015;31(9):1719–31.\n19. Jensen JT, Lefebvre P, Laliberte´ F, et al. Cost burden\nand treatment patterns associated with manage-\nment of heavy menstrual bleeding. J Womens\nHealth (Larchmt). 2012;21(5):539–47.\n20. Mirkin D, Murphy-Barron C, Iwasaki K. Actuarial\nanalysis of private payer administrative claims data\nfor women with endometriosis. J Manag Care\nPharm. 2007;13:262–72.\n21. Fuldeore M, Chwalisz K, Marx S, et al. Surgical\nprocedures and their cost estimates among women\nwith newly diagnosed endometriosis: a US database\nstudy. J Med Econ. 2011;14:115–23.\n22. Fuldeore M, Yang H, Du EX, Soliman AM, Wu EQ,\nWinkel C. Healthcare utilization and costs in\nwomen diagnosed with endometriosis before and\nafter diagnosis: a longitudinal analysis of claims\ndatabases. Fertil Steril. 2015;103:163–71.\n422 Adv Ther (2018) 35:408–423\n\n23. Nnoaham KE, Hummelshoj L, Webster P, et al.\nImpact of endometriosis on quality of life and work\nproductivity: a multicenter study across ten coun-\ntries. Fertil Steril. 2011;96:366–73.\n24. Nicholson S, Pauly MV, Polsky D, et al. Measuring\nthe effects of work loss on productivity with team\nproduction. Health Econ. 2006;15:111–23.\n25. Facchin F, Barbara G, Saita E, et al. Impact of\nendometriosis on quality of life and mental health:\npelvic pain makes the difference. J Psychosom\nObstet Gynaecol. 2015;36(4):135–41.\n26. Lagana` AS, La Rosa VL, Rapisarda AMC, et al. Anx-\niety and depression in patients with endometriosis:\nimpact and management challenges. Int J Womens\nHealth. 2017;9:323–30.\n27. Vitale SG, La Rosa VL, Rapisarda AMC, Lagana ` AS.\nImpact of endometriosis on quality of life and\npsychological well-being. J Psychosom Obstet\nGynaecol. 2017;38(4):317–9.\nAdv Ther (2018) 35:408–423 423","source_license":"CC0","license_restricted":false}