The Bristol Female Lower Urinary Tract Symptoms questionnaire: development and psychometric testing

Objective To develop a questionnaire that is sensitive to changes in the symptomatology of the female lower urinary tract, particularly urinary incontinence, providing an instrument that can characterize symptom severity, impact on quality of life and evaluate treatment outcome. Patients, subjects and methods Items covering as wide a range of urinary symptoms as possible were devised after consultation with clinicians and a health scientist, a literature review and discussion with patients. Additional items assessed the degree of ‘bother’ that symptoms were causing. Eighty‐five women with clinical symptoms attending for urodynamic assessment and 20 women with nonewere asked to self‐complete the questionnaire. The instrument's validity was assessed by interviewing patients and measuring levels of missing data, comparing symptom scores between clinical and non‐clinical populations and comparison with frequency/volume charts and data from pad tests. The instrument's reliability was assessed by measuring both internal consistency and stability, using a 2‐week test‐retest analysis. Results The questionnaire was completed by the patients with a mean of only 2% of items missing; most questions were easily understood. Construct validity was good, with the instrument easily differentiating clinical and non‐clinical populations. Criterion validity, as tested against frequency/volume charts and pad‐test data, was acceptable, with Kappa coefficients of 0.29–0.79 for frequency/volume data and Spearman rank correlations of 0.50–0.97 and 0.31–0.67 for frequency/volume and pad‐test data, respectively. The reliability of the instrument was good; a Cronbach's alpha of 0.78 indicated that the symptom questions had high internal consistency, while stability was excellent, with 78% of symptoms and problems answered identically on two occasions, and Spearman rank correlations of 0.86 and 0.90, respectively. Conclusion The instrument has good psychometric validity and reliability. The stability demonstrated at baseline and the ability to differentiate clearly between community and clinical populations suggest that it should be ideal for measuring changes following therapeutic intervention. The addition of life‐impact items and a ‘bother’ factor may provide the opportunity to identify those women who wish treatment for their symptoms; this dimension requires further exploration.