Pelvic floor muscle exercises (PFMEs) are an effective treatment for women with stress urinary incontinence (SUI), and biofeedback is often used to maximize its effect. However, conflicting results about the significant advantages of adding biofeedback to PFMEs on managing SUI have been reported, and studies provide little insight about the reasons for inconsistent results. This systematic review aimed to: 1) explore the effect of PFMEs plus biofeedback protocols compared to PFMEs only protocols on pelvic floor muscle strength (PFMS), objective symptom alleviation, subjective symptom alleviation, and quality of life, and 2) describe the intervention protocols to assess the potential factors influencing inconsistent outcomes.

PRISMA guidelines were followed[1]. PubMed, CINAHL, and Embase databases were searched for eligible studies published from January 1, 2001 to September 15, 2017. Search strategies were developed through consultation with a health science librarian. Randomized controlled trials (RCTs) and quasi-experimental trials having at least two groups, i.e., PFMEs plus biofeedback versus PFMEs only, of women diagnosed with SUI were selected. Studies of pregnant women and perinatal women were excluded because their unique treatment needs fell outside the scope of this review. Narrative reviews, commentaries, case reports, and study protocols were excluded from this review.

Of 1757 studies identified by searching strategies, seven studies (all RCTs) met the inclusion criteria. Six were 2-arm RCTs and one was 3-arm RCT with a blank control group.188 women were in the PFMEs groups, 193 in the PFMEs plus biofeedback group, and 16 in a blank control group. Four studies lacked sufficient sample size for statistical inference, and only one study fit 5 of the 6 criteria for Cochrane risk of bias[2](see Figure 1). Women’s age in seven studies ranged from 46.6 to 58.3 years old, and four studies reported symptoms duration of SUI ranged from 5 to 9.7 years.

Large variations in intervention protocols were evident across studies including:1) content of  education, i.e., information about PFM anatomy/function; 2) mode of education, i.e., individual versus group education; 3) healthcare organization-based training, i.e., methods used to verify initiation of muscle contraction, training plans and revisit appointments with physiotherapists or other health care professionals; 4) support for transition from healthcare organization-based training to home-based training; 5) home-based training, i.e., training plans that differ across studies and from those delivered in healthcare organizations, partial or no adherence monitoring and promoting strategies; and 6) the lengths of intervention programs (see Table1). 
To measure outcomes, five studies reported PFMS by using objective tools (i.e., electromyography, perineometry or pressure transducer) or the subjective measure (i.e., Modified Oxford scale[3]). Three studies used the pad-test to evaluate objective symptom alleviation. Subjective symptom measures used in five studies were self-administered questions with no report of reliability and validity. Three studies reported quality of life using, “Incontinence Quality of Life Instrument”, “International Consultation on Incontinence Questionnaire -Short Form”, or “King Health Questionnaire”. 
For PFMS, four studies showed no significant effect from biofeedback, and one study using Modified Oxford scale showed sustained significant biofeedback effect during and at the end of the intervention (both time frames, p<0.05). There was limited evidence (one study with p<0.05 after the intervention) about the additive effect of biofeedback to PFMEs on alleviating objective symptoms using the pad-test. In addition, no evidence was presented on the benefit of adding biofeedback to PFMEs on alleviating subjective symptom and improving quality of life during intervention and/or follow-up period.

Evidence about the additive effect of biofeedback is limited in this review due to lack of statistical power, high risk of bias for the RCTs, variations in intervention protocols, and hidden bias (e.g., lack of adherence monitoring and reliable assessment tools embedded in included studies). Moreover, variations in the intervention protocols and assessment tools limit the generalization of findings across studies. Rigorous RCTs, using similar protocols and rigorous trial methodology, are required to clarify the effect of adding biofeedback to PFMEs in the treatment of women with SUI.

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2010). Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. International journal of surgery, 8(5), 336-341.
Higgins, J.P. and S. Green, Cochrane handbook for systematic reviews of interventions, 2005, version.
Laycock, J. and D. Jerwood, Pelvic floor muscle assessment: the PERFECT scheme. Physiotherapy, 2001. 87(12): p. 631-642.