Rising Inflammatory Bowel Disease Prevalence Highlights the Need for Effective, Cost-Effective Therapies

Abstract

The prevalence of inflammatory bowel disease and costs of care are rising in the United States and highlight the need for effective, cost-effective therapies.

Disease prevalence studies, a cornerstone of health epidemiology, provide important data that influence multiple aspects of health care. Disease prevalence affects clinical decision-making, investment in research, the need for specialized clinical training programs, and patients’ perception of their disease. Furthermore, prevalence studies provide information that can be used to estimate the economic and social burden that a specific disease exacts on society and to shape health policy. For these reasons, accurate estimates of inflammatory bowel disease (IBD) prevalence are critical.

In this issue of Inflammatory Bowel Diseases, Ye et al leverage 2 large administrative databases, the Optum’s Clinformatics DataMart and Truven Health MarketScan, to provide an updated prevalence estimate of IBD in the United States (US).¹ These databases encompass close to 60 million patients, both pediatric and adult, from all 50 states and include inpatient and ambulatory encounters over a 10-year period—all unique strengths to this study. They utilize a case-based algorithm that is dependent strictly on billing codes from patient encounters to identify patients with IBD. Although prevalence was calculated individually in each administrative database, results were pooled using meta-analytic methods. Their results show that using 2016 data, 1 in 209 adults and 1 in 1299 children are affected by IBD in the United States, which approximates to 57 thousand children and 1.2 million adults. They also found IBD prevalence rates to be increasing over time. They report that up to one fifth of adult patients with IBD have indeterminate disease. Taken together, their study provides an estimation of IBD burden in the United States using 2 large administrative databases.

Notably, Ye et al’s prevalence estimate for IBD varies from previously published studies^2–5 (Table 1). Another US claims-based prevalence study estimated an almost 2-fold increase in IBD prevalence in 2009 compared with Ye et al’s analysis of the same year.² Additionally, 2 recent US studies—one a survey-based population study and the other an examination of a large commercial database—estimated much higher prevalence, with the nationally representative survey study predicting around 3 million Americans living with IBD.^{4, 5} How could there be such variability in disease prevalence estimates between these studies? This highlights the inherent challenges in obtaining precise determination of disease prevalence. Prevalence studies require patient cohorts that are sufficiently¹ well characterized so all patients with true disease are included without addition of false-positive diagnoses and² that are diverse, both economically and geographically, to represent the population at-large. In this study, an impressively large number of patients—close to 32 million—were studied. However, Ye et al opted to utilize a highly specific criteria for IBD diagnosis, hoping to minimize inclusion of patients identified falsely as having IBD. This strategy has merit, but, as the authors note, also risks missing patients with true IBD, especially those with stable disease who may not utilize the health care system as frequently as required for inclusion in this study. In fact, to the authors’ credit, they provide a prevalence estimation in their supplementary material based on a more sensitive definition of IBD, which led to an almost 2-fold increase in estimated prevalence of adult IBD—from 478.4 per 100,000 persons to 738.1 per 100,000 persons. Furthermore, as the authors note, their algorithm for case identification has not been validated in the United States. Additionally, there was a limited number of elderly and publicly insured patients included in this study. This may have led to an underestimation of IBD prevalence, as a previous US-based study suggested a higher prevalence in the elderly population and those with public insurance and low-income.⁴ Interestingly, a study from a well-defined geographic region of Scotland utilizing a stringent capture-recapture methodology with manual verification of records noted a significantly higher prevalence of IBD (823 per 100,000 in 2018) compared with earlier estimates and highlights the underestimation of IBD prevalence by relying on in-patient coding alone.⁶

Table 1.

Recent Studies Estimating Prevalence of Inflammatory Bowel Disease in United States

Reference	Study Population/ Database	Sample Size, Study Year	Prevalence of IBD/100,000
2Kappelman MD et al	PharMetrics Choice Patient-Centric Database (IMS Health, Watertown, MA)	12 million; 2009	Children (UC+CD):92 Adults (UC+CD):504
3Shivashankar R et al	Olmsted county, MN, Rochester Epidemiology Project	Approx. 150,000; 2011	Children and Adults (UC+CD): Approx. 533
4Dalhammer JM et al	Nationally representative National Health Interview Survey (NHIS)	33,672 adults (response rate 55.2%); 2015	Adults ≥ 18 years (CD+UC): Approx.1200
1Ye Y et al	Optum’s Clinformatics DataMart (OptumInsight, Eden Prairie, MN) and Truven Health MarketScan (IBM Watson, MI)	Approx. 32.31 million; 2016	Children 2–17 years (UC+CD+IBDU):77 Adults (UC+CD+IBDU): 478.4 Alternate case definition Children 2–17 years (UC+CD+IBDU):85.8 Adults (UC+CD+IBDU): 738.1
5Sheriff MZ et al	Explorys (Explorys Inc., IBM Watson, Cleveland, OH	Approx. 35.4 million; 2013– 2018	Children and Adults (UC+CD): Approx. 795.7

Given these caveats, what can we take away from this study? First, this study illustrates the importance of defining disease in determining the outcome of any prevalence study. Modifications in case definition, for example, can drastically alter prevalence estimates. Second, Ye et al provide further data that the prevalence of IBD is increasing with time. A recent study carried out as part of Crohn’s and Colitis Foundation Initiative analyzed the Optum research database from 2007 to 2016 and, in this large cohort of patients, estimated that patients with IBD incurred 3-fold higher direct costs of care compared with non-IBD controls.⁷ Furthermore, these costs are rising and are likely driven, in part, by therapeutics and certain disease features.⁷ Given the increase in prevalence of IBD noted in the current study and the parallel rising costs of IBD care, it challenges us in the field to continue to discover novel personalized therapies that not only induce deep and long-lasting remission but do so in a cost-effective manner.

Supported by: MD is supported by NIH grant K08DK110421. The funding agencies had no role in the study analysis or writing of the manuscript. Its contents are solely the responsibility of the authors.