Anda di halaman 1dari 7

ORIGINAL CONTRIBUTION

Behavioral Training With and Without


Biofeedback in the Treatment of
Urge Incontinence in Older Women
A Randomized Controlled Trial
Kathryn L. Burgio, PhD Context Previous research on urge urinary incontinence has demonstrated that mul-
Patricia S. Goode, MD ticomponent behavioral training with biofeedback is safe and effective, yet it has not
been established whether biofeedback is an essential component that heightens thera-
Julie L. Locher, PhD
peutic efficacy.
Mary G. Umlauf, PhD Objective To examine the role of biofeedback in a multicomponent behavioral train-
David L. Roth, PhD ing program for urge incontinence in community-dwelling older women.
Holly E. Richter, PhD, MD Design Prospective, randomized controlled trial conducted from April 1, 1995, to
March 30, 2001.
R. Edward Varner, MD
Setting University-based outpatient continence clinic in the United States.
L. Keith Lloyd, MD
Patients A volunteer sample of 222 ambulatory, nondemented, community-
dwelling women aged 55 to 92 years with urge incontinence or mixed incontinence

U
RGE URINARY INCONTINENCE IS
with urge as the predominant pattern. Patients were stratified by race, type of incon-
a common condition that af- tinence (urge only vs mixed), and severity (frequency of accidents).
fects millions of US individu-
Interventions Patients were randomly assigned to receive 8 weeks (4 visits) of bio-
als, especially older wom-
feedback-assisted behavioral training (n=73), 8 weeks (4 visits) of behavioral training
en.1,2 It is usually treated with drugs that without biofeedback (verbal feedback based on vaginal palpation; n=74), or 8 weeks
inhibit detrusor contractions,3 but ad- of self-administered behavioral treatment using a self-help booklet (control condi-
verse effects are common and behav- tion; n=75).
ioral treatments have also proven effec- Main Outcome Measures Reduction in the number of incontinence episodes as
tive by changing voiding habits or documented in bladder diaries, patients’ perceptions and satisfaction, and changes in
teaching new continence skills.4-11 Bio- quality of life.
feedback-assisted behavioral training
Results Intention-to-treat analysis showed that behavioral training with biofeed-
uses biofeedback to teach patients how back yielded a mean 63.1% reduction (SD, 42.7%) in incontinence, verbal feedback
to control the physiologic responses of a mean 69.4% reduction (SD, 32.7%), and the self-help booklet a mean 58.6% re-
the bladder and pelvic floor muscles that duction (SD, 38.8%). The 3 groups were not significantly different from each other
mediate incontinence.5-9,11 It is effective (P=.23). The groups differed significantly regarding patient satisfaction: 75.0% of the
for treating urge incontinence, produc- biofeedback group, 85.5% of the verbal feedback group, and 55.7% of the self-help
ing improvements ranging from 76% to booklet group reported being completely satisfied with treatment (P=.001). Signifi-
86%, is at least as effective as drug cant improvements were seen across all 3 groups on 3 quality-of-life instruments, with
therapy, and in 1 trial, it was more ef- no significant between-group differences.
fective than immediate-release oxybu- Conclusions Biofeedback to teach pelvic floor muscle control, verbal feedback based
tynin chloride.6-9,11 on vaginal palpation, and a self-help booklet in a first-line behavioral training pro-
Biofeedback-assisted behavioral gram all achieved comparable improvements in urge incontinence in community-
dwelling older women. Patients’ perceptions of treatment were significantly better for
training has multiple components. A
the 2 behavioral training interventions.
primary component is teaching pa-
JAMA. 2002;288:2293-2299 www.jama.com
tients how to identify and exercise pel-
vic floor muscles, and most impor-
Author Affiliations are listed at the end of this Burgio, PhD, Birmingham Veterans Affairs Medical
tant, how to use them to prevent urine article. Center, GRECC/11G, 70019th St S, Birmingham, AL
loss by aborting detrusor contractions Corresponding Author and Reprints: Kathryn L. 35233 (e-mail: kburgio@aging.uab.edu).

©2002 American Medical Association. All rights reserved. (Reprinted) JAMA, November 13, 2002—Vol 288, No. 18 2293

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

and occluding the urethra during con- Clinical Evaluation Form Health Survery (SF-36)18 were
tractions that cannot be inhibited. Al- Potential participants who met initial cri- completed by patients at home and re-
though biofeedback is clearly an effec- teria were scheduled for evaluation in an turned with baseline diaries.
tive technique for teaching pelvic floor outpatient continence clinic. The evalu-
muscle control, it is not established ation consisted of a continence and Inclusion and Exclusion Criteria
whether it is an essential component of medical history, physical examination, To be included, patients had to have at
training for urge incontinence or postvoid catheterization for residual least 2 urge accidents per week on aver-
whether muscle control can be taught urine, urodynamic evaluation, hemo- age documented in the 2-week bladder
adequately by other methods, such as globin A1C in the presence of diabetes, diary, and urge incontinence had to be
verbal feedback during pelvic exami- and urinalysis (urine dipstick on clean- the predominant pattern (the number of
nation or written instructions. One catch specimen with microscopic evalu- urge accidents had to exceed the num-
small study (n=27) found that biofeed- ation, if indicated). In addition, the Mini- ber of stress and other accidents). Also,
back did not enhance outcomes.8 Mental State Examination was used to there had to be urodynamic evidence of
The role of biofeedback in the treat- screen for dementia.14 If patients had a bladder dysfunction (detrusor instabil-
ment of urge incontinence is an impor- urinary tract infection (urine colony ity during filling or provocation or maxi-
tant issue because biofeedback is more count ⬎10 000), fecal impaction, se- mal cystometric capacity of ⱕ400 mL).
expensive and slightly more invasive vere atrophic vaginitis, or a correctable Patients were excluded if they had con-
than other teaching methods in that it metabolic problem, they were offered tinual leakage, postvoid residual urine
involves placement of electrodes. Medi- treatment for the condition and recon- volume greater than 150 mL, severe uter-
care reimburses for biofeedback used sidered for study participation at a later ine prolapse past the vaginal introitus,
to treat urinary incontinence, based on date if the incontinence persisted. decompensated congestive heart fail-
studies of stress incontinence, but very Urodynamic testing was performed ure, or impaired mental status (Mini-
little is known about the role of bio- to document bladder dysfunction (for Mental State Examination score ⬍24).
feedback in the treatment of urge in- inclusion) and to classify the type of
continence.12,13 incontinence for stratification (urge Design
The present study used a random- only vs mixed stress and urge). Two- Prior to randomization, participants were
ized controlled trial to test whether channel supine water cystometry was stratified by race (black or white) because
biofeedback enhances the outcome of performed using a 12F double lumen of possible differences in the pelvic
behavioral training for urge inconti- urodynamic catheter, a rectal balloon, floor,19,20 type of incontinence, and sever-
nence in older women. Specifically, it and room temperature sterile water at ity of incontinence. Baseline bladder dia-
evaluated the relative effects of train- a continuous filling rate of 50 mL/min ries and urodynamic test results were
ing with and without biofeedback up to a maximum of 500 mL. Thresh- used to classify incontinence as urge only
compared with a control condition old volumes were recorded for first de- or mixed stress and urge. To be sure that
consisting of self-administered behav- sire to void, strong desire to void, cys- the groups were similar on pretreat-
ioral treatment. tometric capacity (the highest volume ment severity of incontinence, the blad-
achieved), and detrusor contraction. der diaries were used to stratify partici-
METHODS Strength of the external anal sphincter pants as having mild (⬍5 episodes per
Participants was assessed by using manometry. The week), moderate (5-10 episodes
Participants were older community- catheter was removed and several ma- per week), or severe (⬎10 episodes per
dwelling women with persistent urge in- neuvers were performed to provoke week) incontinence. Patients were ran-
continence. They were recruited through urge or stress incontinence. domized to behavioral treatment with
local advertisements, community out- biofeedback, behavioral treatment with-
reach, and professional referrals and then Bladder Diary and Quality- out biofeedback (verbal feedback based
screened by telephone to determine eli- of-Life Measures on vaginal palpation), or a control con-
gibility. To be eligible, patients were at To measure pretreatment frequency of dition consisting of self-administered
least 55 years old, ambulatory, and had incontinence, patients were provided behavioral training.
described a pattern of predominant urge with 2 weeks of bladder diary book-
incontinence that occurred at least twice lets.15 Patients documented the time of Intervention
per week and persisted for at least 3 every void and incontinent episode, the For all patients, treatment was imple-
months. All participants provided in- volume of each episode of urine loss mented for an 8-week period. Patients
formed consent according to proce- (large or small), and the circum- completed a daily bladder diary
dures approved by the University Insti- stances of each episode. The Hopkins throughout treatment.
tutional Review Board for Human Use. Symptom Checklist (SCL-90-R, for psy- Behavioral Training With Biofeed-
The study was conducted between April chological distress),16 Incontinence Im- back. Treatment consisted of 4 clinic
1, 1995, and March 30, 2001. pact Questionnaire,17 and the Short- visits at 2-week intervals during the
2294 JAMA, November 13, 2002—Vol 288, No. 18 (Reprinted) ©2002 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

8-week period. At each visit, clinic staff times per day. The initial duration of each ercises, how to use their muscles to pre-
reviewed bladder diaries to ensure that individual contraction was determined vent accidents, and how to respond to
entries were clear and interpretable. In- based on the ability demonstrated by urgency. The complete text is pub-
terventions were implemented by nurse each patient in the original training ses- lished in Staying Dry: A Practical Guide
practitioners. During clinic visits, pa- sion. Across sessions, the duration was to Bladder Control.22 Patients were given
tients in the biofeedback group were increased gradually to a maximum of 10 an instruction booklet and an appoint-
taught skills and strategies for prevent- seconds, with an equal period of relax- ment for a return visit in 8 weeks. They
ing incontinence and provided with oral ation between contractions. Patients were were also given a supply of bladder dia-
and written instructions for daily home advised to practice in various positions ries and stamped envelopes for return-
practice. including lying, sitting, and standing, and ing completed diaries biweekly.
In the first visit, anorectal biofeed- whenever possible to integrate the ex-
back was used to help patients identify ercises into other daily activities. They Posttreatment Assessment
pelvic floor muscles and teach them how were instructed to actively contract pel- Following the last intervention visit, pa-
to contract and relax these muscles se- vic floor muscles during activities that tients completed 2 weeks of posttreat-
lectively while keeping abdominal had resulted in incontinence and to prac- ment bladder diaries and a patient sat-
muscles relaxed.6 A 3-balloon probe was tice interruption or slowing of the uri- isfaction questionnaire, and repeated
inserted into the rectum and used to mea- nary stream during voiding once a day. the 3 quality-of-life measures. When
sure external anal sphincter responses, Behavioral Training Without Bio- they returned for their posttreatment
simultaneously with rectal (abdomi- feedback. This treatment included all visit, these materials were collected and
nal) pressures.21 Tracings were dis- the components of behavioral train- patients were asked to repeat urody-
played on a computer monitor. The sec- ing minus the biofeedback. In lieu of namic testing.
ond visit was devoted to teaching patients biofeedback, verbal feedback based on
how to respond adaptively to the sensa- vaginal palpation was used in the first Data Management and Analysis
tion of urgency (urge suppression strat- treatment session to help patients iden- The sample size was calculated to allow
egies).6,8,22 Instead of rushing to the toi- tify and contract pelvic floor muscles. detection of 15% differences in improve-
let, which increases intra-abdominal If, after several attempts, no contrac- ment between groups with 85% power
pressure and exposes patients to visual tion could be detected vaginally, the ex- and a significance level of .05, assum-
cues that can trigger incontinence, pa- aminer placed a finger just inside the ing a 2-sided hypothesis test and a pooled
tients were encouraged to pause, sit down anal opening and gave verbal feed- within-group SD of 20%. The 3 treat-
if possible, relax the entire body, and con- back of voluntary external anal sphinc- ment groups were first compared using
tract pelvic floor muscles repeatedly to ter contraction. Home practice and all ␹2 analysis and analysis of variance to de-
diminish urgency, inhibit detrusor con- other instructions were the same as for termine whether there were any group
traction, and prevent urine loss. When the biofeedback group. If patients did differences before treatment on key vari-
urgency subsided, they were to proceed not improve by at least 50% by their ables. After treatment, the bladder dia-
to the toilet at a normal pace. Patients third visit, the teaching was repeated. ries were used to calculate change in the
with mixed incontinence were also Self-administered Behavioral Train- frequency of incontinence episodes,
taught stress strategies, which con- ing: Control Condition. The control which was the primary outcome mea-
sisted of contracting pelvic floor muscles group received written instructions for sure. The pretreatment and posttreat-
just before and during any physical ac- an 8-week self-help behavioral pro- ment frequency of incontinence were
tivities such as coughing or sneezing that gram, with the same content as the be- used to calculate a percentage reduc-
had triggered stress incontinence. In the havioral training program described tion for each patient ([pretreatment fre-
third visit, patients who had not achieved above, but completely self-adminis- quency − posttreatment frequency]/
at least 50% improvement underwent tered without benefit of professional [pretreatment frequency]⫻100%).6,11
combined bladder-sphincter biofeed- expertise or equipment. It was a step- Thus, 100% represented total conti-
back to teach them to contract pelvic by-step self-help program written for in- nence, 0% represented no improve-
floor muscles against increasing vol- continent individuals who do not have ment, and a negative percentage indi-
umes of fluid, in the presence of increas- access to a professional with this exper- cated regression. One-way analysis of
ing urgency, and during detrusor con- tise or who simply wish to try such a pro- variance was used to test for differences
traction.6 The fourth visit was used to gram on their own. In language geared among the 3 groups on reduction of in-
review progress, “fine-tune” home prac- to a fifth-grade reading level, it presents continence. The analysis was based on
tice, and encourage persistence. basic information about urge and stress intention-to-treat. When patients did not
Recommendations for pelvic floor incontinence, how to complete bladder complete treatment, the most recent
muscle exercises included 45 exercises diaries, how to locate their pelvic floor bladder diaries were used to calculate
every day divided into manageable ses- muscles (including vaginal palpation), outcome, including baseline diaries when
sions, typically sets of 15 exercises, 3 how to do daily pelvic floor muscle ex- no data were available postbaseline.
©2002 American Medical Association. All rights reserved. (Reprinted) JAMA, November 13, 2002—Vol 288, No. 18 2295

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

Differences between the groups on among the 3 treatment groups on the key of participants in the verbal feedback
patient satisfaction and perceptions parameters, with the exception of blad- group achieved at least 50% and 75% re-
were tested using the ␹2 statistic for cat- der capacity. Therefore, this variable was ductions of incontinence, but differ-
egorical variables or the Kruskal- included as a covariate in the primary ences were small and nonsignificant
Wallis test for ordinal variables. Other analysis of treatment outcome. (FIGURE 2).
outcomes measures, including the We also investigated whether base-
Hopkins Symptom Checklist, the In- Reductions of Incontinence line characteristics were associated with
continence Impact Questionnaire, the Before treatment, the weekly fre- treatment outcomes. Treatment out-
SF-36, and bladder capacity, were ex- quency of incontinence was similar come was not related to diuretic use
amined using 3 (treatment group)⫻2 across the 3 groups, although the ver- (P=.40), previous surgery (P =.87), or
(pretreatment vs posttreatment) re- bal feedback group had slightly more uterine prolapse (P = .69). Further-
peated measures analyses of variance. accidents than the other 2 groups (mean more, no interaction was found by
SPSS version 10.0.5 (SPSS Inc, Chi- [SD], 17.3 [16.3] per week compared therapist. Results did not differ sub-
cago, Ill) was used for all statistical with 15.4 [14.2] and 15.1 [13.5] per stantially when we excluded patients
analyses. week, TABLE 2). After treatment, the lost to follow-up after baseline only.
biofeedback and verbal feedback groups
RESULTS were almost identical (6.1 [10.3] and Bladder Capacity
Of 474 women who were evaluated clini- 6.0 [10.7] accidents per week) and 6.7 A total of 48% of patients completed a
cally, 252 were ineligible or did not par- (11.4) accidents per week were re- posttreatment cystometrogram (30 in
ticipate and 222, aged 55 to 92 years, ported in the self-help group. biofeedback, 35 in verbal feedback, and
were randomized (FIGURE 1). The attri- Behavioral treatment with biofeed- 42 in self-help booklet group). This sub-
tion rate was 15.1% in the biofeedback back resulted in a mean (SD) 63.1% sample who completed pretreatment and
group, 12.2% in the verbal feedback (42.7%) reduction in frequency of acci- posttreatment urodynamics was com-
group, and 9.3% in the self-help book- dents, 69.4% (32.7%) reduction in treat- pared with the remaining patients on out-
let group. Twenty-seven patients did not ment with verbal feedback, and 58.6% come (reduction of incontinence) and
complete treatment. All were included (38.8%) reduction in treatment with the the baseline characteristics. Patients who
in the intention-to-treat analysis. Char- self-help booklet. The analysis of covari- completed a posttreatment cystometro-
acteristics of the participants are pre- ance indicated that the 3 groups were not gram had significantly shorter dura-
sented in TABLE 1. Before treatment, significantly different from each other tions of incontinence (P=.006), were
there were no significant differences (P=.23). Similarly, a larger proportion more likely to have a urethrocele
(P=.03), and had greater reductions of
Figure 1. Patient Flow Diagram
incontinence with treatment (P=.005)
than those who did not complete post-
474 Women Evaluated treatment urodynamics. These 2 groups
did not differ significantly on the remain-
252 Excluded ing variables. Bladder capacity in-
93 Did Not Meet Inclusion Criteria (Type of
Urinary Incontinence, No. of Accidents,
creased by a mean 47.8 mL in the bio-
Urodynamic Results, Diary) feedback group, 63.2 mL in the verbal
83 Failed to Return
24 Declined
feedback group, and 37.0 mL in the self-
37 Medical Contraindications help booklet group. The improvements
15 Impaired Mental Status
across all 3 groups were statistically sig-
222 Randomized
nificant (overall, P=.001), but the in-
creases did not differ among the 3 inter-
73 Assigned to Receive Behavioral 74 Assigned to Receive Behavioral 75 Assigned to Receive ventions (P=.54).
Training With Biofeedback Training With Verbal Feedback Self-help Booklet
Patient Satisfaction and
11 Withdrew 9 Withdrew 7 Withdrew Perceptions of Progress
4 Patient Illness 3 Patient Illness 1 Patient Illness
1 Family Illness 2 Family Illness 6 Treatment Required Too Much Several aspects of the patient’s perspec-
5 Treatment Required Too Much 4 Treatment Required Too Much Effort or Time
Effort or Time Effort or Time tive were assessed by a questionnaire
1 Personal Problem (TABLE 3). After completing treatment,
the biofeedback and verbal feedback
62 Completed 8-Week 65 Completed 8-Week 68 Completed 8-Week
Treatment Treatment Treatment
groups were very similar in their de-
scriptions of progress in therapy and
73 Included in Primary Analysis 74 Included in Primary Analysis 75 Included in Primary Analysis comfort level for continuing treat-
ment. For example, 62.3% and 63.2%
2296 JAMA, November 13, 2002—Vol 288, No. 18 (Reprinted) ©2002 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

Table 1. Baseline Characteristics of the Study Sample*


No. (%)

Biofeedback Verbal Feedback Self-help Booklet Total


Characteristic (n = 73) (n = 74) (n = 75) (N = 222) P Value
Demographics
Age, mean (SD), y 64.8 (7.1) 65.8 (7.6) 65.8 (8.5) 65.4 (7.7) .66
High school graduate† 60 (87.0) 67 (91.8) 66 (93.0) 193 (90.6) .59
Black 11 (15.1) 13 (17.6) 11 (14.7) 35 (15.8) .87
History
Parity, mean (SD) 2.5 (1.7) 3.0 (2.0) 2.7 (2.0) 2.7 (1.9) .33
Duration of symptoms, mean (SD), y 7.1 (7.8) 6.6 (7.7) 6.6 (8.7) 6.8 (8.1) .91
Using diuretics 8 (11.0) 17 (23.0) 15 (20.0) 40 (18.0) .13
Using estrogen 52 (72.2) 51 (68.9) 44 (59.5) 147 (66.5) .23
Previous treatment with medication 16 (21.9) 18 (24.3) 21 (28.0) 55 (24.8) .69
Previous treatment with surgery 16 (21.9) 13 (17.6) 12 (16.0) 41 (18.5) .63
Activity restricted by UI 46 (63.9) 52 (70.3) 44 (59.5) 142 (64.6) .67
Pelvic examination
Urethrocele 20 (27.4) 14 (18.9) 20 (27.0) 54 (24.4) .40
Cystocele, 2° or 3° 30 (41.1) 26 (35.1) 29 (38.7) 85 (38.3) .76
Rectocele, 2° or 3° 13 (17.8) 11 (14.9) 12 (16.0) 36 (16.2) .89
Atrophic mucosa 3 (4.1) 2 (2.7) 0 5 (2.3) .23
Uterine prolapse 4 (5.5) 0 4 (5.3) 8 (3.6) .12
Bladder capacity, mean (SD), mL 282 (117) 238 (100) 266 (105) 262 (109) .04
Type of UI (on diary and urodynamics)
Urge UI only 50 (68.5) 50 (67.6) 50 (66.7) 150 (67.6)
.97
Mixed stress and urge UI 23 (31.5) 24 (32.4) 25 (33.3) 72 (32.4)
Severity classification, accidents per week
Mild (⬍5) 14 (19.2) 14 (18.9) 16 (21.3) 44 (19.8)
Moderate (5-10) 20 (27.4) 21 (28.4) 21 (28.0) 62 (27.9) ⬎.99
Severe (⬎10) 39 (53.4) 39 (52.7) 38 (50.7) 116 (52.3)
*UI indicates urinary incontinence.
†There were 9 cases missing from education because of incomplete patient forms.

described their condition as “much bet-


Table 2. Results of Behavioral Treatment on Frequency of Incontinent Episodes
ter,” respectively, whereas only 30.8%
Verbal
of patients in the self-help group con- Biofeedback Feedback Self-help Booklet P
sidered themselves “much better” (over- Results (n = 73) (n = 74) (n = 75) Value
all, P=.002). On all 5 measures with sig- No. of accidents per week,
mean (SD)
nificant group differences, the verbal Pretreatment 15.1 (13.5) 17.3 (16.3) 15.4 (14.2) .62
feedback group was found to be signifi- Posttreatment 6.1 (10.3) 6.0 (10.7) 6.7 (11.4) .78
cantly better than the self-help booklet Percentage reduction
group (description of progress, P<.001; Mean (SD) 63.1 (42.7) 69.4 (32.7) 58.6 (38.8) .23
accidents are smaller, P=.006; comfort- Median (interquartile 75.0 (−120.0 to 100.0) 82.8 (0 to 100.0) 70.4 (−29.4 to 100.0)
range)
able with treatment, P = .01; satisfac-
tion with progress, P<.001; and restric- progress and patient satisfaction com- SF-36 (all P<.05). These effects indi-
tion of activities, P=.002), and on 3 of pared with the self-help group. cated significant improvements across all
the 5 measures (description of progress, 3 treatment groups. One significant
satisfaction with progress, and restric- Psychological Distress, Impact of group⫻time interaction effect was found
tion of activities), the biofeedback group Incontinence, and Quality of Life on the vitality subscale, indicating that
was also found to be superior to the self- Repeated measures analyses of treat- vitality scores increased more for the
help group (P<.001, P=.03, and P=.047, ment effects revealed statistically signifi- verbal feedback group (P=.01). Other-
respectively). The verbal feedback group cant main effects for pretreatment vs wise, no differential treatment effects
did not differ from the biofeedback posttreatment on 9 of 10 scales of the were observed.
group on any measure. Thus, from the Hopkins Symptom Checklist (P⬍.05; not
patients’ point of view, both verbal feed- hostility, P=.13), on all 4 subscales of COMMENT
back and biofeedback led to better out- the Incontinence Impact Questionnaire This study demonstrates that all 3 be-
comes on important measures of (all P<.001), and on 5 of 8 scales of the havioral interventions were effective for
©2002 American Medical Association. All rights reserved. (Reprinted) JAMA, November 13, 2002—Vol 288, No. 18 2297

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

helping patients identify the pelvic floor almost identical rates of incontinence af- teers who participated in this clinical trial
muscles and use them to prevent epi- ter treatment. The verbal feedback train- may not be representative of the gen-
sodes of urge incontinence. The use of ing did not consist merely of a cursory eral clinical population.
biofeedback did not enhance efficacy pelvic floor muscle contraction during Like any intervention, each of these
more than what was achieved using care- a pelvic examination, but involved a teaching methods may not be the best
ful training with verbal feedback or a de- more comprehensive session in which approach for every patient. Some may
tailed self-help program. In fact, the bio- exercises were carefully and thor- learn more readily by interacting closely
feedback and verbal feedback groups had oughly taught, with time devoted to with the therapist; others may be more
guiding patients through a series of ex- comfortable with the instrumented bio-
ercises. Furthermore, the training was feedback and the more intricate varia-
Figure 2. Reduction in Incontinence at 8 done in the context of an 8-week pro- tions the biofeedback provides. While
Weeks by Intervention Group gram in which patients were taught other many patients may prefer verbal feed-
continence skills and encouraged to per- back over being instrumented, clearly
Biofeedback (n = 73)
sist in their efforts. there are those who cannot identify the
Verbal Feedback (n = 74)
Self-help Booklet (n = 75) In this study, the self-administered be- pelvic floor muscles because of ex-
90
havioral treatment program was also treme weakness, who lack the proprio-
80 very effective. The practitioner should ceptive feedback that allows them to
70 note that this self-help program in- control pelvic floor muscles, and who
60 cluded keeping continuous bladder dia- may do better with biofeedback. Many
Patients, %

50 ries that were mailed in biweekly, and clinicians have observed patients who
40
patients were called if the diaries were cannot identify or adequately control
30
20
not received. Furthermore, when pa- pelvic floor muscles without biofeed-
10 tients were given the self-help booklet, back but subsequently are able to gain
0 they were also given an appointment to control through biofeedback. After
100% >75% >50%
Reduction Reduction Reduction return after 8 weeks, which could have completion of this trial, patients were of-
helped sustain their motivation. It is pos- fered the opportunity to crossover to
Error bars indicate 95% confidence intervals. sible that the highly motivated volun- these treatments. Five patients who com-
pleted treatment with verbal feedback
elected to crossover into the treatment
Table 3. Patient Perceptions of Treatment* with biofeedback. These patients showed
No. (%) a mean 54.2% reduction of inconti-
Verbal Self-help nence after the first treatment and 73.4%
Biofeedback Feedback Booklet P mean reduction after treatment with bio-
Patient Perceptions (n = 53) (n = 57) (n = 65) Value
feedback. Patients who received bio-
Patient description of progress
Much better 33 (62.3) 36 (63.2) 20 (30.8) feedback first were not offered the op-
Better 18 (34.0) 20 (35.1) 36 (55.4) tion of a second intervention in this trial.
⬍.001
About the same 2 (3.8) 1 (1.8) 8 (12.3) Although the biofeedback and ver-
Worse 0 0 1 (1.5) bal feedback interventions were not sig-
Having fewer accidents 51 (96.2) 57 (100.0) 58 (92.1) .09 nificantly more effective than the self-
Accidents are smaller 42 (79.2) 49 (89.1) 42 (67.7) .02 help condition for reducing accidents as
Able to wear less protection 33 (71.7) 40 (83.3) 34 (70.8) .29 documented in bladder diaries, they did
Comfortable enough with treatment 49 (98.0) 54 (100.0) 54 (88.5) .009 result in better outcomes in the pa-
to continue indefinitely tients’ perceptions of and satisfaction
Patient satisfaction with progress with progress. Patient satisfaction with
Completely 39 (75.0) 47 (85.5) 34 (55.7)
Somewhat 12 (23.1) 8 (14.5) 24 (39.3) .001
their progress is likely to be highly re-
Not at all 1 (1.9) 0 3 (4.9)
lated to accident reduction; therefore, it
How much does incontinence
could be that they were less satisfied be-
restrict your activities? cause they did not reduce accidents to
Not at all 36 (69.2) 43 (78.2) 31 (50.8) a critical threshold. However, other as-
.007
Some or all of the time 16 (30.8) 12 (21.8) 30 (49.2) pects of patient satisfaction might have
How disturbed do you feel been affected by their having had less
about the incontinence?
Not at all 26 (49.1) 32 (59.3) 23 (39.0) contact with clinical staff. Personal in-
Somewhat 26 (49.1) 22 (40.7) 32 (54.2) .18 teractions with care providers may con-
Extremely 1 (1.9) 0 4 (6.8) tribute to patient satisfaction through en-
*Not all items were completed on the patient questionnaire; therefore, percentages may not all total 100. couragement and support, which are
2298 JAMA, November 13, 2002—Vol 288, No. 18 (Reprinted) ©2002 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015


BEHAVIORAL TRAINING IN TREATMENT OF URGE INCONTINENCE

often critical to sustain a patient’s mo- tractions. It may not be necessary to noninstitutionalized patient. J Urol. 1986;136:1022-
1025.
tivation in a behavioral program. achieve such a high degree of control or 2. Hunskaar S, Arnold EP, Burgio KL, et al. Epidemiol-
Because all 3 treatment approaches ap- strength or to sustain a contraction, but ogy and natural history of urinary incontinence. Int Uro-
gynecol J Pelvic Floor Dysfunct. 2000;11:301-319.
peartobeclinicallyusefulandacceptable, only to activate the reflex pathway. The 3. Wein AJ. Pharmacologic options for the overac-
a practical strategy would be to initiate urge suppression strategy, consisting of tive bladder. Urology. 1998;51:43-47.
4. Fantl JA, Wyman JF, McClish DK, et al. Efficacy of
training with an instruction booklet or pelvic floor muscle contraction ad- bladder training in older women with urinary incon-
verbal feedback and reserve biofeedback equate to inhibit the detrusor, may be the tinence. JAMA. 1991;265:609-613.
for those who have difficulty learning pel- most essential component of this therapy. 5. Cardozo LD, Abrams PD, Stanton SL, et al. Idio-
pathic bladder instability treated by biofeedback. Br J
vicfloormusclecontrolinthiswayorwho The finding that biofeedback did not Urol. 1978;50:521-523.
do not progress adequately in their at- enhance the effectiveness of behavioral 6. Burgio KL, Whitehead WE, Engel BT. Urinary in-
continence in the elderly: bladder-sphincter biofeed-
tempts to reduce incontinence in their training for reducing urge incontinence back and toileting skills training. Ann Intern Med. 1985;
daily lives. This approach is consistent indicates that behavioral training has ex- 103:507-515.
7. Burgio KL, Stutzman RE, Engel BT. Behavioral train-
with the current reimbursement policies cellent potential for becoming more ing for post-prostatectomy urinary incontinence. J Urol.
issued by the Centers for Medicare and widely disseminated and may be imple- 1989;141:303-306.
Medicaid Services, which state that bio- mented using existing coding using time 8. Burton JR, Pearce KL, Burgio KL, et al. Behavioral train-
ing for urinary incontinence in elderly ambulatory pa-
feedback is reimbursable after patients spent in patient education and counsel- tients. J Am Geriatr Soc. 1988;36:693-698.
have failed a course of pelvic floor muscle ing as the key factor for determining the 9. Baigis-Smith J, Smith DAJ, Rose M, Newman DK.
Managing urinary incontinence in community-residing
training. However, patients whose efforts specific level of service provided. Be- elderly patients. Gerontologist. 1989;29:229-233.
are not producing results may lose mo- cause verbal feedback and the self-help 10. Elser DM, Wyman JF, McClish DK, et al. The effect
of bladder training, pelvic floor muscle training, or com-
tivation and these patients are likely to program can be implemented without bination training on urodynamic parameters in women
reject alternative behavioral training.13 the equipment and expertise needed to with urinary incontinence. Neurourol Urodyn. 1999;
18:427-436.
This study was specific to the treat- perform biofeedback, they are both ap- 11. Burgio KL, Locher JL, Goode PS, et al. Behavior
ment of urge incontinence and results propriate and practical for use in most vs drug treatment for urge urinary incontinence in older
should not be generalized to stress in- any outpatient clinical practice. women. JAMA. 1998;280:1995-2000.
12. Shepherd AM, Montgomery E, Anderson RS.
continence. Previous studies are incon- Treatment of genuine stress incontinence with a new
Author Affiliations: Department of Veterans Affairs
sistent in determining the role of Medical Center, Birmingham/Atlanta Geriatric Re-
perineometer. Physiotherapy. 1983;69:113.
13. Burgio KL, Robinson JC, Engel BT. The role of bio-
biofeedback in treatment of stress in- search, Education, and Clinical Center, Birmingham, feedback in Kegel exercise training for stress urinary in-
Ala (Drs Burgio and Goode); and School of Medicine
continence.12,13,23,24 Although some re- (Drs Burgio, Goode, Locher, Richter, Varner, and Lloyd),
continence. Am J Obstet Gynecol. 1986;154:58-64.
14. Folstein MF, Folstein SE, McHugh PR. Mini-
search provides evidence that biofeed- Center for Aging (Drs Burgio, Goode, Locher, Mental State: a practical method for grading the cog-
back results in higher success rates than Umlauf, Roth, Richter, Varner, and Lloyd), School of nitive state of patients for the clinician. J Psychiatr Res.
Nursing (Dr Umlauf ), and School of Public Health (Dr 1975;12:189-198.
training without biofeedback,12,13 other Roth), University of Alabama at Birmingham. 15. Locher JL, Goode PS, Roth DL, et al. Reliability
studies are equivocal.23,24 Thus, con- Author Contributions: Study concept and design: assessment of the bladder diary for urinary inconti-
Burgio, Goode, Umlauf, Varner, Lloyd. nence in older women. J Gerontol A Biol Sci Med Sci.
clusions regarding the treatment of Acquisition of data: Burgio, Goode. 2001;56:M32-M35.
stress incontinence should be re- Analysis and interpretation of data: Burgio, Goode, 16. Derogatis LR. The SCL-90-R Administration, Scor-
Locher, Roth, Richter, Varner, Lloyd.
served for more definitive studies. Drafting of the manuscript: Burgio, Goode, Umlauf,
ing, and Procedures Manual II. Towson, Md: Clinical
Psychometric Research; 1983.
Furthermore, there is reason to be- Roth. 17. Shumaker SA, Wyman JF, Uebersax JS, et al. Health-
Critical revision of the manuscript for important in-
lieve that biofeedback may play differ- tellectual content: Burgio, Goode, Locher, Roth,
related quality of life measures for women with urinary
incontinence. Qual Life Res. 1994;3:291-306.
ent roles in the treatment of urge vs stress Richter, Varner, Lloyd. 18. Ware JE Jr, Sherbourne CD. The MOS 36-item short-
incontinence. It is clear that biofeed- Statistical expertise: Locher, Roth. form health survey (SF-36), I: conceptual framework and
Obtained funding: Burgio, Goode, Locher, Umlauf, item selection. Med Care. 1992;30:473-483.
back makes it possible for patients to gain Varner. 19. Bump RC. Racial comparisons and contrasts in uri-
better control over pelvic floor muscle Administrative, technical, or material support: Burgio, nary incontinence and pelvic organ prolapse. Obstet
Goode, Locher, Richter, Lloyd.
contraction, especially in the ability to Study supervision: Burgio.
Gynecol. 1993;81:421-425.
20. Howard D, Davies PS, DeLancey JO, Small Y. Dif-
maximize force and to sustain contrac- Funding/Support: This work was supported by grant ferences in perineal lacerations in black and white pri-
AG RO108010 from the National Institute on Aging,
tions, which is important for building National Institutes of Health, Bethesda, Md.
miparas. Obstet Gynecol. 2000;96:622-624.
21. Engel BT, Nikoomanesh P, Schuster MM. Oper-
strength. Treating stress incontinence re- Previous Presentation: Presented at the American Uro- ant conditioning of rectosphincteric responses in the
lies on voluntary periurethral contrac- gynecologic Society 22nd Annual Scientific Meeting, treatment of fecal incontinence. N Engl J Med. 1974;
Chicago, Ill, October 26, 2001. 290:646-649.
tions to occlude the urethra. Thus, Acknowledgment: We thank Beverly Badger, MSN, 22. Burgio KL, Pearce KL, Lucco A. Staying Dry: A Prac-
strength, the ability to sustain contrac- CRNP, Aline Sabol, RN, Lisa Farris, MSN, CRNP, Na- tical Guide to Bladder Control. Baltimore, Md: The
talie Baker, MSN, CRNP, and Rebecca L. Bryant, MSN, Johns Hopkins University Press; 1989.
tions, and a higher degree of muscle con- CRNP, for their roles in implementing treatment pro- 23. Burns PA, Pranikoff K, Nochajski TH, et al. A com-
trol would seem to be important for pre- tocols; Justin Christie, BA, and Claudia Hardy, MPA, parison of effectiveness of biofeedback and pelvic
for recruitment; and Kate Clark, BS, and Artisha Moore,
venting stress accidents. In treating urge BA, for data management.
muscle exercise treatment of stress incontinence in older
community-dwelling women. J Gerontol. 1993;48:
incontinence, mechanical occlusion of M167-M174.
the urethra may be a less important func- REFERENCES
24. Berghmans LCM, Frederiks CMA, de Bie RA. Ef-
ficacy of biofeedback when included with pelvic floor
tion of pelvic floor muscle contraction 1. Diokno AC, Brock BM, Brown D. Prevalence of uri- muscle exercise treatment for genuine stress inconti-
than the fact that it inhibits detrusor con- nary incontinence and other urologic symptoms in the nence. Neurourol Urodyn. 1996;15:37-52.

©2002 American Medical Association. All rights reserved. (Reprinted) JAMA, November 13, 2002—Vol 288, No. 18 2299

Downloaded From: http://jama.jamanetwork.com/ by a University of Sussex Library User on 12/07/2015