Study objective - The aim was to compare the costs and effects of management of intractable urinary incontinence by urinary catheterisation or incontinence pads. Design - This was a prospective, randomised study comparing catheterisation with pads, supplemented by additional data collected from patients with chronic indwelling catheters. Main outcome measures were costs of equipment, nursing time, patient preference, nursing preference, and clinical and bacteriological assessment of urinary infection. Subjects - 78 intractably incontinent elderly female patients were randomly allocated to management by urinary catheter or pads and toileting. Supplementary data on equipment costs and nursing time were collected from 27 patients, of whom 22 were already catheterised at the time of the randomisation and five were catheterised by the nursing staff after the last date for entry into the randomisation. Main results - Of the 38 patients randomised to catheterisation, 14 refused consent so only 24 were catheterised on day 1 of the study. There was a rapid removal of catheters, especially in the first six weeks of the study and only four of the randomised catheter patients completed the full 26 weeks of the study. However, eight of the pads patients were catheterised between the 7th and 22nd week because of deteriorating general condition and all retained their catheters for the remainder of the study period. Of 35 patients who had experienced catheters and pads, 12 expressed a clear preference for catheters, 12 for pads, and 11 were undecided. Nurses were in favour of the use of pads, mainly because of concerns about urinary infection with catheters. Comparing costs for patients managed with catheters (532 patient weeks) or pads (903 patient weeks), catheter patients required less nursing time (15.4 v 29.0 h per patient per week) but equipment costs were higher (£19.20-24.65 v £8.79-11.35 per patient per week), mainly because of the cost of catheter care (£12.75 per patient per week). Asymptomatic bacteriuria was prevalent in both groups but 73% of catheterised patients received treatment for clinical signs of infection compared with 40% of pads patients. Only 30% of patients who were treated had any generalised symptoms of infection. Conclusions - Use of catheters reduce nursing time but may increase weekly equipment costs depending on the cost of laundry. Despite the high dropout rate among patients randomised to catheters a minority of patients (12/35) expressed a clear preference for catheters and we believe that more patients with intractable incontinence should be given a trial of catheterisation to assess acceptability. Bacteriuria was prevalent in pads or catheter patients but no major episodes of invasive infection were noted in either group.