Background A study was undertaken to determine outcome and to identify
predictors of success for biofeedback for constipation. Methods Patie
nts who had at least one biofeedback session were evaluated whether or
not they completed a treatment course. Parameters assessed included u
se of cathartics, number of spontaneous bowel movements per week, pres
ence of rectal pain, number of biofeedback sessions and results of ano
rectal physiology. Results A total of 194 patients (59 male, 135 femal
e) of median age 71 (range 11-96) years, including 30 with concomitant
rectal pain, were treated. The median number of spontaneous bowel mov
ements per week before treatment was 0. Some 35 per cent of patients h
ad complete success (three or more spontaneous bowel movements per wee
k with discontinuation of cathartics), 13 per cent had partial success
(fewer than three spontaneous bowel movements per week with continued
use of cathartics) and 51 per cent had no improvement. Neither patien
t age, sex nor duration of symptoms significantly affected outcome. On
ly 18 per cent of patients who had between two and four sessions had c
omplete success, compared with 44 per cent of those who had five or mo
re (P < 0.001). A total of 63 per cent of patients who completed the t
reatment protocol experienced complete success, compared with 25 per c
ent of those who self-discharged (P < 0.0001). Conclusions This large
study indicates that the success rate of biofeedback for patients with
constipation is less than previously reported. However, the success r
ate improves significantly after five or more sessions and is signific
antly related to the patient's willingness to complete treatment.