(Background to the 2000 hearings)
by John D. Perry, Ph.D., BCIAC, FAACS
The Health Care Financing Administration’s Technology Advisory Committee first planned to discuss Biofeedback Therapy for Urinary Incontinence at their August 6-7, 1996 meeting, but the topic was postponed to May 6-7, 1997. It was agreed then that additional information was needed, and that was recently prepared by the Blue Cross/Blue Shield Association’s Technology Evaluation Center. According to HCFA sources, the BC/BS-TEC report is currently being reviewed by the HCFA-TAC staff but action is not expected for "a couple of months" because "frankly, this isn’t one of our highest priorities".
The complete minutes of the 1997 TAC meeting are printed at the end of this paper. It seems valuable, before TAC considers TEC, to see what they already "know" about biofeedback, and how accurate their information is. Statements from the "analysis" section of the minutes will therefore be examined in detail next. TAC wording is in quotes; commentary is not, and is in color and indented (in media that permit that format).
"Analysis: The TAC's discussion of biofeedback for chronic pain concluded that there is ample literature stating that biofeedback for the treatment of headache and chronic pain may be effective therapy."
We note that even while acknowledging the literature, chronic pain is not covered; see minutes below.
"However, members concluded that behavior therapy is the first approach to treating incontinence. The members agreed that for urinary incontinence, Kegel exercises are the time-honored treatment and have an 85% effective rate."
The belief in an "85% effective rate" is extremely troublesome, since (1) it is not clear what is meant by "effective" and (2) in any case, no such rate is supported by the research literature. Unfortunately the AHCPR Guidelines on Urinary Incontinence do more to confuse the issue than clarify it. At stake here is the very important difference between "symptom reduction rate" and "cured/improved rate". And very often this crucial distinction was lost --- on even the Guideline authors, with disastrous results. An historical digression is necessary.
In the early 1980s, Kathy Burgio, Ph.D. established a great standard for comparing the effectiveness of behavioral interventions in the treatment of incontinence, the "symptom reduction rate". Quite simply, it is the [number of leaks before] minus the [number of leaks after], divided by the [number of leaks before]. In other words, 10 leaks before, minus 3 leaks after treatment, is 7 leaks, divided by 10, gives a "symptom reduction rate" of 70%. 10 minus 10 divided by 10 would be 100%, or "cured". Burgio’s own results, for biofeedback, ranged in the 77-94% "symptom reduction" range.
Unfortunately electrical stimulation researchers were unable to come up with success rates even close to Burgio’s biofeedback success, so rather than report disastrous figures, they invented a new statistic, called the "improvement" rate, or "cured/improved" rate. We know what "cured" means, but what does "improved" mean. Various researchers fudged the definition of "improved", but generally hidden in the text will be the formula, which is often (but not always) "at least a 50% symptom reduction". What is the significance of a 50% symptom reduction rate? That means if you required 4 pads a day before therapy, you now require only 2 pads a day. Is that impressive? Hardly. By most standards, such as the International Continence Society's definition, such a person would still be considered moderately incontinent.
So when a researcher reports a "cured/improved" rate of, say, 75%, it means that at least three quarters of the subjects achieved at least a 50% reduction in accidents. Fortunately a few researchers have given enough information that we have been able to calculate in several studies the actual symptom reduction rate as well, and it turns out that the much more meaningful "symptom reduction" rate is usually about one-half the so-called "cured/improved" rate.
This is an extremely important distinction, and one on which the AHCPR guidelines made a number of serious mistakes, making it appear that PME alone or electrical stimulation or vaginal cones were nearly as effective as biofeedback in reducing incontinence. Sometimes the AHCPR even confuses these outcome measures within a single paragraph. For example, on page 38 (1996), Bo, Hagen, Kvarstein et al (1990) is first described as having shown a "significantly greater reduction in incontinence", but in the very next sentence we learn that they had a "60.1 percent cure/improvement rate..", as if these were linguistically-equivalent ways of saying the same thing. (In fact, we don’t even know what their "symptom reduction rate" was.)
Likewise, Burns et al (1990) is described in the 1992 AHCPR Guideline as having a 54 percent "improvement rate" (p. 32), when in fact she actually reported a 54% "symptom reduction rate". And Wells, Rink [sic] and Diokno et al (1991) are reported to have had a "cure/improved rate of 77 percent"(p. 32). These errors makes it look like Wells did much better than Burns, when in fact their symptom reduction rate was probably more like 40%, much worse than Burns!
This same sloppy outcome terminology continued in the 1996 AHCPR Guideline, which states "The effects of PME alone have been well documented in the medical literature" (p. 37). They include these out-patient studies:
(1) Dougherty, Bishop, Mooney et al (1993) achieved "62 percent reduction" rate in 65 subjects.
(2) Ferguson, McKey, Bishop et al (1990) obtained a "56- to 58-percent reduction in leakage" as shown by pad weights, with 20 subjects. In addition:
(3) Bo, Hagen, Kvarstein et al (1990) reported a "60.1 percent cure/improvement"(p. 38) rate in an intensely monitored group, but only a 17 percent rate in a control group without therapist supervision, with a total of 52 subjects. However, to compare Bo with Dougherty and Ferguson, we first need to translate the 60 percent "improvement" into a likely 30% symptom reduction rate. And:
(4) Wells, Brink, Diokno et al (1990), with 54 PME subjects, found that "77 percent of the exercise subjects…reported improvement". Since they apparently included all levels of improvement (and not just the traditional ‘50% reduction or better’) it is impossible calculate the true "symptom reduction rate" for the exercise group. Even a generous estimate of "about half" yields a very mediocre 39 percent symptom reduction rate.
In addition to these out-patient studies, the 1996 Guideline reports two other studies which were quite different from the rest. First:
(1) Flynn, Cell, and Luisi (1994) reported a 82 percent "decrease in UI" in a nursing home among 37 patients with "transient" incontinence (as opposed to "established" incontinence like "stress" and "urge"). While interesting, these cannot be compared to the other studies. And:
(2) Benvenuti, Caputo, Bandinelli, et al (1987) obtained a "95 percent reduction in incontinence episodes" (p. 37). Actually, the 95% figure is a AHCPR miscalculation; the correct figure is 91% (11.3 before, 1.0 after, equals 91%). This is, nevertheless, an extraordinary result for "PME alone", but is it really "PME alone"? Unlike the above studies, this protocol involved 20 half-hour contact hours with a physiotherapist, including 10 daily visits during the first two weeks. At these visits and subsequent "re-education training" sessions at 15, 45 and 90 days, least 6 vaginal palpation examinations were done, in three postures, during which the subjects received extensive VERBAL feedback about the condition and function of their pubococcygeus muscles. On-going verbal feedback, which apparently continued until the patient got it right, was NOT a feature of any of the PME- only studies described above. It is inappropriate to use this study to bolster the average for several very different studies which did, in fact, use "PME alone". A separate category should have been established for "verbal feedback", just like "biofeedback", "electrical stimulation", and "vaginal cones". [It is worth noting that although this study achieved the highest success of any non-biofeedback study, it has never in 11 years been replicated. We suspect that that is because the heavy staff overhead made the protocol too expensive.]
A more accurate synopsis of the "PME alone" literature suggests a symptom reduction rate of no more than 50%, based on the first four studies mentioned above. This is very different from an "85% effective rate".
In addition, the Guideline notes that "patients require repeated guidance over an extended period of time to derive maximum benefit from PME" (p. 38). This appears to be an understatement, as the Wells and Bo studies involved six months or more of PME, and the Benvenuti study apparently involved over three months of training. (In contrast, biofeedback therapy generally is accomplished in less than two months, with significantly higher outcomes.)
Finally, it is difficult to accept the TAC definition of Kegel exercise (defined as plain PME) as the "time-honored treatment", since from 1948 to at least 1968 and probably later "Kegel exercises" were defined by Dr. Kegel himself as always using the Kegel perineometer biofeedback device! It was only in the late 1970s or so that nurses began to have trouble obtaining the Kegel perineometer and began the attempt to teach Kegel’s exercises without biofeedback. [see Perry & Hullett, "The Bastardization of Dr. Kegel’s Exercises", Northeast Gerontological Society, 1988; available on the internet at http://www.incontinet.com/articles/art_urin/bastard.htm, for a more complete discussion].
"The studies on biofeedback to treat incontinence generally are not randomized or controlled and one of them shows that patients who used a combination of electrical stimulation and biofeedback had a 65% improvement rate. However, these patients also practiced Kegel exercises, making it impossible to distinguish between the effects of biofeedback and Kegel exercises."
It is simply false that most biofeedback studies are not "randomized" – they are. (Clinical series, of course, are not, since they are usually based on actual clinical patients.) In addition, many studies do "control" by comparing PME to PME+biofeedback. And contrary to the last sentence above, it is easy to distinguish between the effects of the PME and PME+ groups by comparing them, which is what Burns, Burgio, and others have done.
The "electrical stimulation and biofeedback study" which obtained 65% improvement rate is difficult to identify. It is probably a reference to Susset et al, J. Urol. 1995, where six weeks of alternating stim and biofeedback produced a "64% overall improvement rate". [Compare that to Susset’s 1990 study, where biofeedback alone produced an 87% symptom reduction rate. Electrical stimulation alone has shown only 30 to 44 percent symptom reduction rates [Perry, "Graphic Comparison of Symptom Reeducation Rates for Behavioral Treatment Options for Stress Incontinence", available on the internet at http://www.incontinet.com/comparison.htm].
"TAC members also noted that biofeedback is most beneficial for the treatment of incontinence in patients who were unable, unwilling, or poorly educated on how to perform Kegel exercises."
Which patients are not included in that group? According to Bump, Hurt, Fantl, and Wyman (Am J Obstet Gynecol, 1991; 165:322-9) "What seems obvious from this study is that simple verbal or written instruction does not represent adequate preparation for a patient who is about to pursue a Kegel exercise program"(p. 326-7). They found such instruction, while "traditional", actually holds potential harm for one quarter of patients.
It is widely agreed in clinical practice that the biggest advantage of biofeedback with home trainers is that the patient is unable to practice self-deception and "forget" to do assigned exercises. Many of the newer devices include computerized memory, which is downloaded to the therapist’s computer at the weekly office visit. Instead of asking "how did you do?", the therapist now says "Let’s see how you did!".
"Although the AHCPR guidelines endorse biofeedback as effective for incontinence, the guidelines do not specify whether it is superior to Kegel exercising."
True, the Guideline authors’ have gone out of their way to conceal the differences between PME, PME+biofeedback, and PME+verbal feedback. However – even limiting our study to the very select group of papers they chose to include -- biofeedback shows "a 75.5 to 82 percent reduction in UI across six studies involving 166 subjects" (p. 39). Moreover, when we look at the five references (and there are only five, not six, studies explicitly named on page 39), we find that the primary sources actually contain references to a range of 75.5 to 95 percent symptom reduction rates; another very significant Guideline error. [For an extended discussion of this and other Guideline errors, see "Critical Review of the Guideline section on Biofeedback", on the internet at http://www.incontinet.com/ahcprbfb.htm.]
In short, based on the studies which the AHCPR itself choose to include, biofeedback accomplishes almost twice as much symptom reduction as PME alone, generally in one-third the time or less.
More vividly, that is the difference between still using many adult products every week even after therapy, versus hardly ever having to use them again! [And that’s the difference between mere "statistical significance" and genuine "social significance".]
"Members believed that incontinence home trainers for urinary incontinence may prove to be useful in the office, but are not appropriate for home use. Home trainers currently are not covered."
Curious it is that the TAC members believe home trainers may be useful in the office but not at home, since there is absolutely NO published evidence to support that position; indeed, no published studies have even investigated that possibility. In fact, as pointed out in Perry, "Graphic Comparison…" (above), the five best biofeedback studies, better by a substantial margin, ALL used home trainers -- at home.
Ironically, what hasn’t been demonstrated is the outcome superiority of clinical office systems. [Their value, however, lies in demonstrating to third party payers the necessity of the biofeedback practice sessions and home trainers. The data from clinical evaluations also plays a major role in boosting the motivation and reward level of the patient. See "The EMG Evaluation", available on the internet at http://www.incontinet.com/emgeval.htm.]
In contrast to the TAC assertion, the 1996 Guideline states clearly that "a home training device may provide an added benefit to clinical biofeedback visits…" (p. 40)
Finally, it should be noted that home trainers were covered by HCFA for several years in the late 1980s; then coverage mysteriously disappeared.
"There was general agreement that biofeedback for headaches and urinary incontinence are the biggest biofeedback issues and that HCFA should request an assessment of biofeedback's effectiveness for these indications."
It appears that the BC/BS-TEC report was produced in response to this agreement. The TEC report is the subject of another essay. It is not known if a similar TEC report is under construction for headaches.
Complete text of the minutes pertaining to biofeedback is reproduced below this line:
------------------------------------------------------------------------------------------------------------------------------------ TECHNOLOGY ADVISORY COMMITTEE
May 6 and 7, 1997
MINUTES
Note: These minutes highlight the general discussions and conclusions of members of the Technology Advisory Committee (TAC). They do not necessarily reflect unanimous agreement of all members and are not meant to establish national policy or to interfere with the individual carrier's or intermediary's responsibility to use discretion in deciding coverage issues absent national policy. [break]
J. BIOFEEDBACK CLARIFICATION
At their most recent semiannual conference, the Carrier Medical Directors raised the issue of coverage of biofeedback for a variety of applications. In one workgroup, a question surfaced about whether biofeedback could be allowed for use in urinary or fecal incontinence. It was noted that several carriers do allow coverage for these uses. AHCPR booklet (#2, 1996 update) on treatment of urinary incontinence in adults repeatedly lists biofeedback as an acceptable and effective behavioral technique for this condition.
In the chronic pain management workgroup, the CMDs discussed claims made by psychologists that biofeedback can be useful in treating chronic pain, such as headache and backache. The CMDs noted that biofeedback has also been used in the management of hypertension, Raynaud's phenomenon, esophageal spasm, tinnitus associated with Meniere's disease, temporomandibular joint dysfunction, and urinary and fecal incontinence. The effectiveness of biofeedback in some instances is uncertain or unproven.
CIM §35-27 provides that "biofeedback therapy is covered under Medicare only when it is reasonable and necessary for the individual patient for muscle reeducation of specific muscle groups or for training pathological muscle abnormalities of spasticity, incapacitating muscle spasm, or weakness, and more conventional treatments (heat, cold, massage, exercise, support) have not been successful. This therapy is not covered for treatment of ordinary muscle tension states or for psychosomatic conditions. (See HCFA-Pub. 14-3, §§2200ff, 2215, and 4161; HCFA-Pub. 13-3, §§133.3, 3148, and 3148, and 3149; HCFA-Pub. 10, §§242 and 242.5 for special physical therapy requirements. See also §S35-20 and §65-8)."
In CIM §35-21 and possibly § 35-21.1, coverage of biofeedback for the management of chronic pain in pain rehabilitation programs is not allowed.
Analysis: The TAC's discussion of biofeedback for chronic pain concluded that there is ample literature stating that biofeedback for the treatment of headache and chronic pain may be effective therapy. However, members concluded that behavior therapy is the first approach to treating incontinence. The members agreed that for urinary incontinence, Kegel exercises are the time-honored treatment and have an 85% effective rate. The studies on biofeedback to treat incontinence generally are not randomized or controlled and one of them shows that patients who used a combination of electrical stimulation and biofeedback had a 65% improvement rate. However, these patients also practiced Kegel exercises, making it impossible to distinguish between the effects of biofeedback and Kegel exercises. TAC members also noted that biofeedback is most beneficial for the treatment of incontinence in patients who were unable, unwilling, or poorly educated on how to perform Kegel exercises. Although the AHCPR guidelines endorse biofeedback as effective for incontinence, the guidelines do not specify whether it is superior to Kegel exercising.
Members believed that incontinence home trainers for urinary incontinence may prove to be useful in the office, but are not appropriate for home use. Home trainers currently are not covered.
There was general agreement that biofeedback for headaches and urinary incontinence are the biggest biofeedback issues and that HCFA should request an assessment of biofeedback's effectiveness for these indications.
This document, which was last revised on May 04, 2005, is located at: http://www.incontinet.com/whathcfa.htm
© 1998 by Incontinet. All rights reserved.
We Want to Know What You Think!
