This study evaluated a treatment procedure in which 126 incontinent nursing home patients were checked on an hourly basis, asked if they needed toileting assistance (prompted), and socially reinforced for appropriate toileting. Urodynamic analysis (including cystometrogram), provocative stress test, and behavioral assessment revealed that the nursing home patients were severely debilitated, with 65% demonstrating bladder abnormalities, 87% incapable of independent toileting, and 25% failing to score on the Mini-Mental Status Exam (average score, 8.0). The treatment procedures were evaluated with a multiple baseline design in which subjects were randomly divided into immediate or delayed treatment groups after a baseline observation period. During treatment, the frequency of incontinence per 12 hours changed from a baseline average of 3.85 to a treatment average of 1.91. Three behavioral measures that can be easily collected by nursing staff significantly predicted continence levels during treatment (multiple R, 0.79) and change in incontinence during treatment (multiple R, 0.64). These prognostic criteria offer nursing staff a cost-effective method for selecting the most responsive patients for prompted-voiding treatment.