Here is the case for the urge in administrative claims database research about overactive bladder therapies.

With an estimated 16.5% of Americans experiencing symptoms from overactive bladder (OAB) and urinary incontinence (UI), this condition deserves continued research focus.1 In this issue of JMCP, D’Souza et al. present a retrospective claims database analysis of adherence, persistence, and switching behaviors, comparing cohorts of patients using different medications to treat OAB.2 D’Souza et al. use standard approaches to calculate adherence, persistence, and switching rates.2,3 Study patients had at least 1 pharmacy claim for either extended-release (ER) or immediate-release (IR) products for tolterodine or oxybutynin during the period from July 1, 1999, to December 31, 2003. Four patient cohorts (ER vs. IR for each of the 2 drugs) were followed from the index date to either discontinuation (defined as a gap in OAB therapy of at least 45 days), a switch to any other OAB medication, or the end of a 1-year follow-up period. The medication possession ratio was calculated as the sum of the total number of days supplied across all pharmacy claims except the last claim, divided by the total number of days from the first fill date to the last fill date. Persistence was measured as the proportion of patients continuing therapy for 12 months without discontinuation or a switch to another OAB drug. The switch rate was calculated as the proportion of patients who changed from the initial index medication to any other OAB treatment, including another study drug, a different dosage form of the same drug (e.g., oxybutynin ER to oxybutynin IR), or other medications, including trospium chloride, oxybutynin patch, flavoxate, hyoscyamine sulfate, or propantheline bromide. D’Souza et al.’s findings pose a challenge to clinicians to find better strategies to diagnose and manage the symptoms of OAB. As revealed in this study, only 55.5% of patients refilled their first prescription for OAB medication and only 13.2% persisted with the index medication for 1 year. Also telling is the finding that medication switching appears to be common practice among these patients; the switch rate was 13.3% for the overall sample and 24.0% of patients with at least 1 refill. No significant persistence advantages were observed for any of the 4 OAB drug products in the study in multivariate analysis, although the odds of adherence with IR drugs were half that for ER drugs (odds ratio [OR] = 0.504; 95% CI, 0.306-0.704; P < 0.001). In any study, an assessment of whether the data are appropriate for the task at hand is necessary. The finding that only 53.7% of the patients had at least 1 OAB diagnosis recorded during the 18-month eligibility period provides evidence that OAB diagnostic information obtained from large databases is not without shortcoming. These low rates for OAB diagnosis pose serious threats to study validity in that D’Souza et al. may have assessed patients with OAB, as well as patients with interstitial cystitis or other UI problems triggered by urinary tract infections. In addition, subjective measures of well being and patient satisfaction, although not reported in claims databases, may also account for patients’ predilection to discontinue medications or use them intermittently according to lifestyle. For example, when medications for OAB are used only during travel, special events, or other activities to minimize the impact of side effects on quality of life, what appears to be discontinuation or non-adherence based on days supply intervals might actually represent a planned pattern of use. The 13.2% persistence rate for the use of these drugs for OAB found by D’Souza et al. at 1 year is extremely low compared with persistence rates for other drug therapy classes. In a study conducted using a similar methodology from 1998 to 2000 in Quebec, 1-year persistence rates for angiotensin-converting enzyme inhibitors in cardiovascular disease ranged from 64% to 72%.4 Similarly, a cross-national study of persistence for antihypertensive medication use in the elderly revealed that approximately 25% of the patients were without medication for at least 180 days during the first year after the initiation of treatment.5 After 6 years, these percentages increased to 41.1%, 36.3%, and 38.2% for patients in Pennsylvania, British Columbia, and the Netherlands, respectively. Thus, persistence rates may suggest but do not necessarily explain the rationale for discontinuing or changing drug therapy. Other factors, such as behavioral modification, dietary change, and exercise, though not reported in databases, could improve patient symptoms, in part explaining the high rate of non-persistence.6 The statistical techniques employed in the study by D’Souza et al. might have been inadequate to assess intermittent drug use that appears to occur with this particular drug class. In addition to controlling for covariates, such as prescription coverage and comorbidities, intermittent, or cyclic medication use, requires special consideration in data analysis, where censoring (failing to complete treatment) and changes in diagnosis, treatment, and eligibility status complicate analyses.7 Analogous to employment and unemployment fluctuations that were originally modeled as duration data,8 similar cycles with OAB medication use may be captured more accurately with an interval-censoring modeling approach to account for variations in medication-taking behavior. In addition, sensitivity analyses, expanding the study grace period of 45 days to include a range from 30 to 90 days may have detected more subtle nuances in medication use. These questions about persistence and adherence become all the more important as newer agents for the management of OAB Here Is the Case for the Urge in Administrative Claims Database Research About Overactive Bladder Therapies