Data sources

We used data from the National Ambulatory Medical Care Survey (NAMCS). This survey is conducted by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention, and includes information regarding ambulatory care, the predominant method of providing health care services in the United States. A complete description of this federally sponsored annual survey can be found at http://www.cdc.gov/nchs . This study used national deidentified data and was determined to be exempt by the Institutional Review Board of Women and Infants Hospital of Rhode Island, Providence, RI.

The NAMCS is an annual sample of outpatient visits to office-based physicians, including private practices, freestanding clinics, public health clinics, family planning clinics, and faculty practices. Approximately 1500 physicians are sampled each year. The basic sampling unit is the patient visit. Data from the NAMCS can be extrapolated to the >1 billion ambulatory visits to physician offices that occur in the United States annually.

Each patient visit in the NAMCS includes up to 3 diagnoses coded using the International Classification of Diseases, Ninth Revision, Clinical Modification ( ICD-9-CM ); the type of visit (referral, new, or established patient); the type of physician seen (primary care vs specialist); and length of visit. PFD-related visits included ICD-9-CM codes for UI ( ICD-9-CM codes 599.81, 599.82, 625.6, 788.3, 788.31, 788.33, 788.37, 788.38, 788.39), POP ( ICD-9-CM codes 618.0–618.9), fecal incontinence ( ICD-9-CM code 787.6), fistula ( ICD-9-CM code 619), and chronic interstitial cystitis and urethral syndrome ( ICD-9-CM codes 595.1, 597.80–597.81). Based on the type of visit and length of time coded for each visit, an Evaluation and Management (E/M) Current Procedural Terminology code was assigned. Patient sex is a separate variable in the NAMCS, and we restricted our study population to only women.

We used the average national Medicare allowances for physician services based on E/M codes to estimate costs and 95% confidence interval (CI). We converted costs to 2006 dollars using the physicians’ services component of the Consumer Price Index. Because Medicare does not provide full coverage for services, beneficiaries pay deductibles and coinsurance expenses under Part B. To capture these costs, we inflated the estimated costs by 38%, based on recommendations of the Centers for Medicare and Medicaid Services Office of the Actuary. To examine trends, we compared average annual costs in 1996–1997 with 2005–2006.

National estimates of costs and SE were obtained using STATA (Stata Corp, College Station, TX) survey sampling commands and survey weights provided in the NAMCS and Taylor series linearized variance estimation. Total costs are presented in millions of dollars. Estimates based on <30 records or with relative SE >30% were considered unreliable and were not reported, per NCHS standards. All analyses were performed using SAS 9.0 (SAS Institute, Cary, NC) and STATA SE 9.0. P < .05 was considered statistically significant.