May 28, 2019
The Prevalence and Forecast Prevalence of Overactive Bladder in the Medicare Population
Abstract
Objectives:
To determine current and future prevalence of overactive bladder (OAB) among Medicare fee-for-service beneficiaries in the United States.
Methods:
Prevalence of OAB in US adults ⩾ 65 years was determined using the 2013 Medicare Beneficiary Part B Carrier Claims File, Part D Drug Event File, and Medicare Beneficiary Annual Summary File. Prevalence for 2027 was forecasted with US Census population projections. Regional projections were based on applying national OAB rates by age, sex, and race/ethnicity to zip code tabulation area beneficiary populations.
Results:
In the 2013 dataset, the prevalence of OAB was 7.2% (male: 7.7%; female: 6.7%). Across demographic categories, prevalence was the highest among those aged more than 74 years (9.3%), identifying as White (7.4%), and residing in urban areas (7.5%). By 2027, OAB is projected to increase by 48.1%.
Discussion:
OAB affects a substantial proportion of the US fee-for-service beneficiary population, with the prevalent population projected to rise substantially by 2027.
Background
Overactive bladder (OAB) is defined by the International Continence Society (ICS) as a chronic condition characterized by urinary urgency, with or without urge incontinence, usually with urinary frequency and nocturia.1 OAB often has severely negative impacts on a patient’s quality of life, limiting an individual’s physical and social activities, resulting in feelings of social seclusion.2–6 Furthermore, individuals with OAB may feel anxiety, frustration, and embarrassment surrounding their OAB, causing a high degree of psychological stress.
It is well documented that the prevalence of OAB increases with age.7 This causes a public health concern, particularly as there is a growing elderly population in the United States (US), which may put substantial financial strain on the health care system. In a recent systematic literature review of the economic burden of OAB in the US, the average OAB-specific direct health care costs were estimated between $656 and $860 per person (2016 USD), which does not include indirect costs or those associated with treating comorbidities.8 Using Medicare claims data from 2003 to 2004, Cisternas et al9 estimated that the aggregate total expenditures associated with OAB ranged from $1.8 to $3.9 billion (2004 USD) per year. Without accounting for prevalence changes in OAB, when these estimates are inflated to 2018 USD (using the US Consumer Price Index), the total expenditures associated with OAB may be as high as $2.4 to $5.2 billion.10
Previous studies have estimated the prevalence of OAB in the US; however, results have varied widely due to the type of study methodologies and definitions applied. In previous studies, definitions of OAB have included the following: use of coping strategies and feelings of urgency within a specific timeframe (regardless of documented clinical symptoms),6 presence of urinary urgency incontinence within a specific timeframe,11–13 or International Classification of Diseases, Revision 9 (ICD-9) codes,9,14 all of which have resulted in a range of prevalence estimates. In cross-sectional studies for which there was a reliance on self-reporting of symptoms and frequency, such as NOBLE (National Overactive Bladder Evaluation), EpiLUTS (Epidemiology of Lower Urinary Tract Symptoms), and OAB-POLL (OAB on Physical and Occupational Limitations), the estimated prevalence of OAB in the US has ranged from 16.5% to 35.6%.6,11–13 When using claims data, OAB prevalence estimates among Medicare beneficiaries have been lower, from 5.7% to 13.6%.9,14 Differences in reported prevalence across study designs may in part be due to underdiagnosis of OAB that involve patient, physician, and systemic barriers to correctly identifying OAB cases. Feelings of anxiety and embarrassment surrounding their OAB may hinder an individual’s willingness to speak to their doctor about symptoms,15–17 or the treatment of other, or more acute, comorbidities may be prioritized over discussing OAB-related symptoms.16 Finally, prior to the introduction of ICD Revision 10 (ICD-10) codes, there was no OAB specific code, as such, database studies relied on a variety of codes to identify OAB patients.14,18,19 As such, prevalence estimates based on claims data may represent patients with more severe or bothersome OAB.
The objectives of this study were to derive contemporary estimates of the prevalence of OAB in the Medicare fee-for-service (FFS) population, as well as to estimate future burden of OAB in this population based on demographic projections to 2027.
Methods
Data source
The data source used in this study was the 2013 Centers for Medicare & Medicaid Services (CMS) 100% Master Beneficiary Summary File including the Base segment (MBSF). The CMS is a federal agency that administers the national health insurance program predominantly to people who are 65 or older and younger persons with disabilities (Medicare) and assists states in the administration of Medicaid, which provides assistance with medical costs to individuals with limited income and resources.20 The MBSF is compiled annually and includes demographics (eg, date of birth/death, sex, race), enrollment status (eg, number of months enrolled in Part A and B used to calculate FFS status), and status of 62 conditions, for all beneficiaries who were alive and enrolled in Medicare for any part of the year.
The MBSF segment also contains a unique beneficiary identification number, linking data in the Medicare Carrier Files, which provides information on treatments made in ambulatory settings recorded in the B Event Files, and reimbursements for prescription drugs recorded in the Part D Event Files.
Study population
Claims data from the Carrier Files identified beneficiaries in 2013 with a primary or secondary ICD-9 diagnosis code of 596.5, 596.51, 788.3, 788.31, 788.33, 788.41, 788.43, 788.63, or 788.9 (Table 1). The selections of ICD-9 codes were derived from a targeted literature review focused on ICD-9 codes used to define OAB cases in the published literature (Table 1).19,22–27 Beneficiaries with a national drug code (NDC) for the use of an oral OAB pharmacotherapy (antimuscarinics or beta-3 antagonists) were also identified as having OAB.
Data analysis
Data from the MBSF, Carrier, and Part D Event Files were linked using unique identifiers, and persons with and without OAB were identified.
Estimates of OAB prevalence (%) for 2013 were calculated by dividing the number of OAB cases identified in the FFS population by the total 2013 FFS beneficiary population ⩾ 65 years. Fisher exact tests were conducted to assess the relationship between OAB status and demographics (age, sex, race/ethnicity, and region), as characterized by odds ratios and the corresponding 95% confidence intervals (CIs).
Estimates of OAB prevalence (percentage and total) for 2027 were obtained by projecting forward the beneficiary population (for patients ⩾ 65 years) and by assuming a constant prevalence rate within each demographic category, with differences in projected values due to changes in population distribution. Projections for the FFS beneficiary population are based on the US Census national population projections from 2014, and the forecasts were geographically disaggregated to zip-code tabulation (ZIP) areas.28 These projections provide estimates for specific age groups (ie, 65-74, 75-84, and ⩾85 years), sex, and race/ethnicity (ie, non-Hispanic White, non-Hispanic Black, Hispanic, non-Hispanic Asian, and other). Projected prevalence by region (and lower geographic scales) was based on projecting forward 2014 ZIP beneficiary population demographics with respect to age, sex, and race/ethnicity, and applying a Bayesian-smoothed relative risk analysis to estimate OAB prevalence in ZIP beneficiary areas. Bayesian smoothing29 is based on a borrowing strength principle30 and is advisable given often small totals of OAB cases (and patient numbers) for some ZIP areas. A sensitivity analysis was made on the ZIP population area projections by varying the assumed rate of aging (data not shown).
All analyses were conducted with WinBUGS,31 Stata 14, and SAS version 9.1.
Results
Overall prevalence of OAB among FFS beneficiaries in 2013
Overall baseline demographic data from FFS beneficiaries in the CMS database (2013), and among those with OAB, are shown in Table 2. A total of 31 763 561 FFS beneficiaries were eligible for the analysis. Based on a combination of diagnoses and prescriptions codes, 2 278 865 out of 31 763 561 FFS beneficiaries in 2013 were identified as having OAB, for a prevalence of 7.2% among those ⩾65 years (Table 2). When only diagnosis codes were used to identify OAB cases, the prevalence was reduced to 2 139 667 of the total 31 763 561 FFS beneficiaries (6.7%; Figure 1). Using prescription claims data, 487 885 FFS beneficiaries were identified through use of OAB pharmacotherapy (1.5%). A total of 348 687 FFS beneficiaries were identified by both diagnosis codes and prescription claims data (1.1%).
Compared with FFS beneficiaries overall, FFS beneficiaries with OAB were more likely to be older: 58.0% of all beneficiaries with OAB were ⩾75 years, compared with 43.6% of beneficiaries without OAB (odds ratio [OR]: 1.785; CI: 1.780, 1.791) (Table 2). With respect to ethnicity, individuals identifying as White were more likely to have been diagnosed with OAB than individuals identifying as Black (7.4% vs 5.8%; OR: 0.766; CI: 0.762, 0.771), or other minorities (5.9% for Hispanics [OR: 0.775; CI: 0.766, 0.783], 6.7% for Asians [OR: 0.894; CI: 0.885, 0.902], or 5.5% others [OR: 0.727; CI: 0.720, 0.734]).
Geographically, OAB was more prevalent in urban populations than rural (7.5% vs 6.0% [OR: 0.791; CI: 0.788, 0.794]) and regionally: the Northeast and South had a higher prevalence of OAB (8.4% and 7.3% [OR: 0.858; CI: 0.855, 0.861], respectively), compared with the Midwest (6.6% [OR: 0.766; CI: 0.763, 0.769]) or West (6.3% [OR: 0.726; CI: 0.723, 0.729]).
Forecast prevalence of OAB among Medicare FFS beneficiaries (2027)
By 2027, the FFS beneficiary population is estimated to increase by 48.1%, from 31 763 561 to 47 046 001, of whom approximately 3.4 million were estimated to have OAB (Table 3). Geographically, no major change in the pattern of OAB prevalence was projected. By 2027, the Northeast and South regions are projected to continue to have the highest proportion of FFS beneficiaries with OAB, 8.4% and 7.4% (OR: 0.871; CI: 0.868, 0.874), respectively. There were, however, several notable percent increases in OAB (Table 3). FFS beneficiaries who identify as Black, Hispanic, and Asian were all projected to experience a more than 70.0% increase in OAB, compared with only 46.2% among those who identify as White, given overall changes forecast in demographic distributions. Regionally, although, in both the 2013 estimate and 2027 forecast, the Western region had the lowest prevalence of OAB, in the 2027 forecast it had the greatest percent increase in OAB cases, at 56.2%.
In the sensitivity analysis, the increase in OAB cases remained relatively robust over variant aging scenarios. When assumed increases (from 2014 to 2027) in the ⩾65 years population categories were raised by 5%, there was a 51.2% increase in OAB cases, compared with 48.8%. Furthermore, the number of OAB cases in the 2027 projection was 3.45 million, compared with 3.39 million. Alternatively, if a 5% decrease is assumed in the ⩾65 years population, the projection is lowered to 46.4% and 3.34 million OAB cases.
Discussion
The current prevalence data on OAB in the United States are limited and the published results vary, with estimated prevalence in adults ranging from 5.6% to 35.6%, depending on the OAB definition and the data source. In this study, OAB was defined based on a collection of ICD codes that have been commonly used in previous studies to identity OAB,6,9,11–14 as well as by the use of oral OAB medications. The OAB prevalence estimated in this study is similar to previously published estimates of OAB in Medicare data: 8.8% to 13.6% across sensitivity analyses for the 2003 to 2004 data,9 and 5.7% for the 2006 to 2011 data.14 Differences observed in the estimates may be attributable to different combinations of ICD codes and/or medications used in defining OAB (Table 1).16,19 There is not a specific ICD-9 diagnosis code for OAB and therefore OAB must be defined using a combination of codes describing common OAB symptoms and its specific treatments.19,22–27 In a study conducted by Cisternas et al,9 the ICD-9 codes included in their sensitivity analyses may describe incontinence conditions other than OAB, which may account for the higher prevalence estimate. When compared with the base case reported by Cisternas et al,9 which included codes that were part of the ICS definition of OAB, the estimates are more closely aligned with this study (7.2% vs 8.8%). Conversely, the ICD-9 code list used to identify OAB cases by Ganz et al14 was more restricted than that in this study, which may be a contributing factor to the lower prevalence estimate. By limiting the codes to those most commonly used to identify OAB in the literature,19,22–27 this study provides a reliable estimate of the prevalence of OAB among FFS beneficiaries.
In addition to the variability observed across studies using Medicare data, a notable difference has been observed between these studies relative to cross-sectional surveys based on self-report, with the latter yielding substantially higher prevalence estimates, up to 35.6%.6,11–13 Unlike studies that used Medicare data and were thus limited to an older population (ie, ⩾65 years), several cross-sectional surveys estimating OAB prevalence have included younger participants (NOBLE: 18 to ⩾75 years; EpiLUTS: ⩾40 years; OAB-POLL: 18-70 years).6,11–13 Whereas some methodological and population differences may partially explain the variability, it is likely that self-reported symptoms of OAB, often urinary urgency,12,13 are more common than cases that can be confirmed based on medical claims explicitly linked to OAB, due to a combination of underdiagnosis and undercoding.15–17 Another possible explanation is that individuals identified in claims databases represent a more burdened population compared with those in self-reported surveys. In claims data, individuals have recognized that their symptoms are problematic and seek medical attention, representing a felt need for health care,33 which was not a defining characteristic for identifying OAB cases in self-reported surveys mentioned above. Furthermore, there is the possibility that as additional treatment options become available along with anticipated increases in health-care-seeking behaviors among people with urologic conditions,34 the proportion of those accessing health care for OAB will increase along with the size of the affected population also increasing. Thus, although the results reported here based on the prevalence in Medicare FFS beneficiaries are based on robust analysis of the available data as feasible, they likely represent an underestimate of the true prevalence in this population. Based on this lower-bound prevalence estimate, 2027 projections estimated the total number of FFS beneficiaries with OAB to increase 48.8%, to nearly 3.4 million individuals due to population growth and shifting demographics.
Using this projection and the estimated costs reported by Powell et al,8 in 2027, the estimated OAB-specific costs may range from 2.2 to 2.9 billion (without considering for inflation), and the all-cause total direct costs may range from 27.7 to 52.8 billion among FFS beneficiaries ⩾65 years alone.
Numerous prior studies of OAB prevalence have explored the relationship between OAB prevalence and race/ethnicity to some degree, although that work has been largely limited to underpowered secondary analysis and to studies with sample sizes too limited to allow robust conclusions to be drawn in this area.11 Across studies, OAB is most common among those identifying as White (2.7%-33.1%), usually by a substantial margin over those identifying as African American (0.5%-4.9%) or Hispanic (0.6%-3.7%).6,9,12–14 Those identifying as Asian have generally been the group indicating the lowest OAB prevalence (0.8%-1.0%).12,13 In this study, the prevalence of OAB in the 2013 dataset was again found to be the highest among FFS beneficiaries who identified as White; however, the differences across racial/ethnic groups were not as marked as previously reported; in particular, the prevalence among Asians was reported to be 6.7%, notably higher than that in prior studies. In the 2027 projections, there was a substantial percent increase projected in the number of OAB cases among FFS beneficiaries identifying as Asian (74.4%), Hispanic (78.6%), Black (71.3%), and other (95.8%), representing more than 470 000 new cases of OAB among FFS beneficiaries, highlighting what may be an important demographic shift in OAB diagnosis in the future.
An understanding of OAB prevalence by region has been identified as a gap in the current literature. To date, only Cisternas et al9 have examined OAB prevalence from this perspective. Their findings indicated that OAB was the most prevalent in the South.9 Results from this study indicate that the prevalence of OAB among FFS beneficiaries was the highest in the Northeast (8.4%) and the South (7.3%), and that living in these regions showed a higher odds ratio of OAB. In addition, although, in both the 2013 dataset and the 2027 projection, the West region was estimated to have the fewest number of cases of OAB among FFS beneficiaries, this region is projected to experience the greatest percent increase in OAB cases (56.2%), from 372 339 cases in 2013 to 581 734 cases. These results provide important information on the racial/ethnic and geographical heterogeneity of OAB in the US. Although not examined in this analysis, a possible explanation for geographic differences may be related to a growing elderly population in the Western regions. Due to the paucity of data on geographic difference in OAB prevalence, future research in this area is warranted.
The study includes a number of methodological limitations. Reliance on claims data to estimate prevalence is challenged by data coding limitations. As it is recognized that OAB is an underdiagnosed condition,15,17 the true number of OAB cases may be higher. To help overcome this, the NDC data were also used to identify additional probable cases of OAB based on oral OAB medication prescriptions.35 This methodology has been previously applied and published.35 This analysis was limited to a single year; as such, beneficiaries diagnosed with OAB prior to 2013, but not seeking treatment during 2013 would not have been captured. In addition, this analysis was limited to Medicare FFS beneficiaries and may not be generalizable to the prevalence of OAB in other populations, including those who are younger than 65 years, or who are covered by Medicaid or a Medicare Advantage plan. Although the census data used to make future projections are based on initial population estimates, they rely on assumptions regarding births, deaths, and net migration, and as such are subject to increasing uncertainty over time. However, some sensitivity analyses were considered, and the results were robust.
A notable strength of this study includes the large and nationally representative sample achieved through the use of the 2013 CMS 100% MBSF. This permitted construction of a detailed picture of OAB prevalence in the United States delineated by age, sex, ethnicity/race, and geography, enhancing the national generalizability of these results in the Medicare FFS population.
The findings from this retrospective study suggest that OAB is and will continue to impact millions of Medicare beneficiaries in the United States. With a better understanding of the predicted prevalence changes in OAB, health care decision makers and clinicians can better prepare for what will continue to be a burden on the health care system.
Acknowledgments
The authors would like to thank Karissa Johnston, Meagan Harwood, and Kath Quayle for drafting, reviewing, and editing this manuscript. Kath Quayle provided consultation services to, and Karissa Johnston and Meagan Harwood are employees of, Broadstreet Health Economics & Outcomes Research, with received payment from Astellas in the conduct of this study. The data collection, analysis, and manuscript preparation were conducted by the National Minority Quality Forum.
Author’s Note
Katherine Gooch was formally with Astellas Pharma Global Development, Inc. at time of study conduct.
Declaration of Conflicting Interests:
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding:
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was initiated and funded by Astellas Pharma Global Development, Inc. The sponsors collaboratively participated in study design, interpretation of data, and decision for article publication.