In order to determine if a bladder-relaxant drug could enhance the effectiveness of a toileting program for incontinence among functionally impaired nursing home patients, 15 patients with detrusor hyperreflexia were enrolled in a 6-week trial during which they were put on a habit training protocol alone for 2 weeks, habit training (HT) and a placebo for 2 weeks, and then HT and oxybutynin for 2 weeks (2.5 mg three times daily for 1 week, then 5 mg three times daily for 1 week). Bacteriuria, which was present without other symptoms in 11 of the 13 patients who completed the trial, was treated before HT began. Eradication of the bacteriuria resulted in a small, statistically significant increase in cystometric bladder capacity (44 mL; 95% confidence interval 16 to 72 mL), but the detrusor hyperreflexia and frequent episodes of incontinence persisted. HT alone reduced the mean percentage of every-two-hour checks which were wet from 43% to 32% (95% confidence interval of the difference -1% to -21%; P = .04 by a paired t test); addition of placebo and active drug did not result in any further reduction. Cystometric bladder capacity had not changed significantly by the end of the trial (measured while patients were still on active drug), and there was no consistent relationship between changes in bladder capacity and changes in wet checks. Although our sample size is small, the data are consistent with other studies in similar populations. The relatively practical and inexpensive interventions we employed in this study will need to be modified or intensified in order to substantially reduce the frequency of incontinence among functionally disabled nursing home patients.