Comparison of Pelvic Floor Physical Therapy Attendance Based on Referring Provider Specialty.

Abstract

Objective: The objective of this study was to determine whether pelvic floor physical therapy (PFPT) attendance differs based on referring provider specialty and identify factors related to PFPT initiation and completion.

Methods: This was an institutional review board-approved retrospective cohort study examining referrals from female pelvic medicine and reconstructive surgery (FPMRS) and non-FPMRS providers at a single academic medical center to affiliated PFPT clinics over a 12-month period. Demographics, referring specialty and diagnoses, prior treatment, and details regarding PFPT attendance were collected. Characteristics between FPMRS and non-FPMRS referrals were compared and multivariate logistic regression analyses were performed to identify factors associated with PFPT initiation and completion.

Results: A total of 497 referrals were placed for PFPT. Compared with non-FPMRS referrals, FPMRS referrals were for patients who were older (54.7 years vs 35.6 years), and had higher parity; more were postmenopausal (56% vs 18%) and had Medicare insurance (22% vs 10%) (all P < 0.001). Most FPMRS referrals were for patients with urinary incontinence (69% vs 31%), whereas non-FPMRS referrals were for patients with pelvic pain (70% vs 27%) (both P < 0.0001). Pelvic floor physical therapy attendance was similar in both groups when comparing rates of initiation (47% vs 45%) and completion (13% vs 16%). In multivariate analysis, factors associated with initiation were age 65 years or older, additional therapy provided at referring visit, private insurance, Asian race, pregnant or postpartum at time of referral, and more than 1 referring diagnosis (all P < 0.05). No factors were associated with completion.

Conclusions: Less than half of the patients referred to PFPT initiate therapy, and only 15% complete PFPT. The populations referred by FPMRS and non-FPMRS providers are different, but ultimately PFPT utilization is similar.