OBJECTIVES: To assess the clinical effectiveness and cost-effectiveness of non-surgical treatments for women with stress urinary incontinence (SUI) through systematic review and economic modelling. DATA SOURCES: The Cochrane Incontinence Group Specialised Register, electronic databases and the websites of relevant professional organisations and manufacturers, and the following databases: CINAHL, EMBASE, BIOSIS, Science Citation Index and Social Science Citation Index, Current Controlled Trials, ClinicalTrials.gov and the UKCRN Portfolio Database. STUDY SELECTION: The study comprised three distinct elements. (1) A survey of 188 women with SUI to identify outcomes of importance to them (activities of daily living; sex, hygiene and lifestyle issues; emotional health; and the availability of services). (2) A systematic review and meta-analysis of non-surgical treatments for SUI to find out which are most effective by comparing results of trials (direct pairwise comparisons) and by modelling results (mixed-treatment comparisons - MTCs). A total of 88 randomised controlled trials (RCTs) and quasi-RCTs reporting data from 9721 women were identified, considering five generic interventions [pelvic floor muscle training (PFMT), electrical stimulation (ES), vaginal cones (VCs), bladder training (BT) and serotonin-noradrenaline reuptake inhibitor (SNRI) medications], in many variations and combinations. Data were available for 37 interventions and 68 treatment comparisons by direct pairwise assessment. Mixed-treatment comparison models compared 14 interventions, using data from 55 trials (6608 women). (3) Economic modelling, using a Markov model, to find out which combinations of treatments (treatment pathways) are most cost-effective for SUI. DATA EXTRACTION: Titles and abstracts identified were assessed by one reviewer and full-text copies of all potentially relevant reports independently assessed by two reviewers. Any disagreements were resolved by consensus or arbitration by a third person. RESULTS: Direct pairwise comparison and MTC analysis showed that the treatments were more effective than no treatment. Delivering PFMT in a more intense fashion, either through extra sessions or with biofeedback (BF), appeared to be the most effective treatment [PFMT extra sessions vs no treatment (NT) odds ratio (OR) 10.7, 95% credible interval (CrI) 5.03 to 26.2; PFMT + BF vs NT OR 12.3, 95% CrI 5.35 to 32.7]. Only when success was measured in terms of improvement was there evidence that basic PFMT was better than no treatment (PFMT basic vs NT OR 4.47, 95% CrI 2.03 to 11.9). Analysis of cost-effectiveness showed that for cure rates, the strategy using lifestyle changes and PFMT with extra sessions followed by tension-free vaginal tape (TVT) (lifestyle advice-PFMT extra sessions-TVT) had a probability of greater than 70% of being considered cost-effective for all threshold values for willingness to pay for a QALY up to 50,000 pounds. For improvement rates, lifestyle advice-PFMT extra sessions-TVT had a probability of greater than 50% of being considered cost-effective when society's willingness to pay for an additional QALY was more than 10,000 pounds. The results were most sensitive to changes in the long-term performance of PFMT and also in the relative effectiveness of basic PFMT and PFMT with extra sessions. LIMITATIONS: Although a large number of studies were identified, few data were available for most comparisons and long-term data were sparse. Challenges for evidence synthesis were the lack of consensus on the most appropriate method for assessing incontinence and intervention protocols that were complex and varied considerably across studies. CONCLUSIONS: More intensive forms of PFMT appear worthwhile, but further research is required to define an optimal form of more intensive therapy that is feasible and efficient for the NHS to provide, along with further definitive evidence from large, well-designed studies.