Click here to jump to "Three Functions of Biofeedback".

June 15, 1999

To: Continence Coalition

Re: Suggestions for Sunday's meeting

CC: Biofeedback manufacturers as listed

=========================================

I regret that I will be unable to attend Sunday's meeting to discuss the draft of the CC's Position Statement #2, due to the short notice I received. However, I do have some suggestions which I am herewith forwarding to all those involved in the hope that my comments might prove useful. My suggestions are provided in line with your questions.

You wrote:

"Understandably, HCFA now requires that all recommendations be supported by, at least, peer-reviewed literature and preferably by randomized, controlled clinical trials."

While I can understand the "peer-reviewed literature" position, I think you will lose the race at the starting gate if you accept the idea that biofeedback can -- or should -- be studied in "randomized, controlled clinical trials", at least as those are defined in traditional pharmaceutical research.

Such "clinical trials" are entirely appropriate when studying the effects of chemical agents, where (1) the effect is small and (2) the participation of the subject is to be minimized. Such trials are necessary to filter out factors such as patient motivation, expectation, therapist bias, and similar factors. In drug research, we are interested in the potency of the drug, and not interested in any subject characteristics. Indeed, the appropriate statistic, Analysis of Variance, factors out such influences, which are called "error variance".

Biofeedback training is very different. It is a goal of biofeedback training to (1) motivate the patient to have (2) high expectations of success. This is often facilitated by (3) therapist enthusiasm and encouragement. The goal of biofeedback training is to use the instrumentation to CHANGE the person -- which is totally unlike the goal in drug research.

We want to CHANGE the person from one who neglects and overlooks their pelvic muscles to one who is intimately conscious of and exerts active control over their own pelvic muscles. It is impossible to "blind" the patient to this condition, since our objective is precisely the opposite; to enlighten them to the beneficial results obtained by paying more attention to these muscles.

In the early days of biofeedback, the difference between biofeedback training and "conditioning" was not well understood. This changed rather dramatically in 1986 with the publication of Shellenberger and Greene's monograph, "The Ghost in the Box" (out of print now, but available as "shareware" at my website: http://www.incontinet.com/ghost.htm.) [I receive no income from the distribution of this book, not even to offset my expenses. It is a public service for which I have already received ample expression of gratitude.] "The Ghost" refers to the idea that there is NO AGENT inside the biofeedback "box" that effects ANY change in the subject. The subject changes his/her own biological behavior, using the information "fed back" from the instruments. If the subject doesn't choose to change, the subject does not change.

[In contrast, it is the aim of drug research to show that there IS an "active agent" or "active ingredient" inside the pill, and that this agent DOES effect a change inside the subject, even if they don't know or believe it. Hence, in drug research, the subject can (and should) be "blind".] The disastrous effects of attempted "blind" research are best illustrated in the largest biofeedback study yet published, which also had one of the most dismal success rates of any study ever undertaken. Unfortunately, people who do not understand the nature of biofeedback (such as the BC/BS TEC committee) cite this as proof of the ineffectiveness of biofeedback in treating incontinence. But in my critical review, I have shown that Burns et al violated most of the rules for conducting good biofeedback therapy.

It is inappropriate to base reimbursement decisions on a research program that failed to meet even the minimum standards for research in its own field, biofeedback. For an extensive discussion of the Burns et al research, see my critical review at http://www.incontinet.com/burns.htm.

It is well known that "placebo effects" account for 20 to 50% improvement in the variable under investigation. When the drug effects are in that range, specific numbers are important for comparison. But when the training effects are significantly greater than even the possible placebo effects, it is sufficient to show the difference.

Although she is by no means the most successful biofeedback therapist, because of her credentials and former position with the NIA, the work of Katherine Burgio is often cited as a gold standard of comparison. Burgio has obtained symptom improvement rates of 80 to 94% among various groups of incontinent patients -- far in excess, by double, of any possible "placebo effects". There are many similar or even better biofeedback reports depicted in my review paper, located at http://www.incontinet.com/comparison.htm.

You do not need to cite my paper, "comparison.htm", which is NOT itself peer-reviewed, to refer to the quoted studies, which ARE peer-reviewed, to make the point. My paper should be used as a reference to identify and summarize the peer-reviewed literature itself.

You further state:

As described before, we understand that HCFA is especially interested in the following:

Our own recommendations were published in 1990 and revised in 1995, under the title "The Perry Protocol", which is still available at http://www.incontinet.com/articles/art_urin/pprotoc.htm.

"1. The type of workup that should be done before biofeedback is instituted"

A. MEDICAL REVIEW: Traditionally it has been assumed that nurses, physical therapists clinical social workers and physiological psychologists (biofeedback specialists), when deal with "medical" diagnoses, would obtain the recommendation of a medical doctor before commencing any such treatment.

Nurses: In the late 1980s a very impressive algorithm was established in research at the University of Pennsylvania, [Baigis-Smith & Smith et al, etc.] in which Nurses, operating under physician's standing orders, used simple screening tests to detect red or white blood cells in the patient's urine. Absent counter-indications, they provided appropriate biofeedback therapy.

All Others: In most states, clinical practice acts require the consultation with and referral by medical doctors prior to therapy. However, this is not always required, and sometimes it is more convenient to schedule medical checks simultaneously with on-going biofeedback therapy.

There are no side effects to biofeedback therapy, and provided that medical examination is not delayed, there is no reason to delay the start of biofeedback training. As a practical matter, many patients will run the full course of biofeedback and be cured of their incontinence before a routine new-patient examination can even be scheduled.

The Perry Protocol recommends a "team approach" to the treatment of incontinence, in which nurses, physical therapists, and biofeedback specialists cooperate in treatment of incontinent patients. Using this model, with a consulting physician and the "PENN" protocol above, will provide the most effective service to the largest number of patients.

B. UROLOGICAL PROCEDURES: Most urologists agree that only the most basic evaluation, usually including H&P and a cystoscopic examination, is necessary in order to recommend a patient for biofeedback training. Actually the cystoscopic examination has no bearing on the referral, but provides the patient with the assurance of the absence of more serious conditions, much like an annual chest xray.

There is no justification for the use of urodynamic procedures, since there is not treatment parameter that is effected by the results. Urodynamic evaluations represent an unnecessary expense to the patient and medical system in the case of ordinary urinary incontinence.

C. HISTORY AND PHYSICAL: There is evidence that a comprehensive History and Physical Examination provides sufficient information to justify biofeedback interventions. The primary object of such H&P is to rule out "transitory" forms of incontinence, as well established in the literature.

D. THE EMG EVALUATION: Probably the least understood step, among medical personnel, is the importance of the initial EMG evaluation in determining the need for biofeedback rehabilitation of the pelvic muscles. The EMG evaluation is often confused with the (raw) EMG evaluation conducted during urodynamic testing.

The EMG evaluation, using an inserted vaginal or rectal sensor, measures (1) the resting muscle tension level, (2) the ability to contract the pelvic muscles under voluntary control, (3) the ability to sustain a contraction long enough to avoid urinary (or fecal) accident, and (4) the ability to relax those muscles when no longer required.

Since first introduced by Farrall Instruments in 1982, the basic EMG evaluation has been adopted by virtually all manufacturers interested in the incontinence market. Some manufacturers have introduced minor variations or "improvements" in the original protocol, but these are based on marketing considerations and have never been subject to scientific comparison or evaluation. The most important datum is the value of a 10-second sustained contraction, compared with the resting level for a similar period.

Note: Some years ago it was proposed that surface patch electrodes could be used in place of inserted sensors; however, no evidence was ever present to show that was an acceptable alternative, in the treatment of incontinence, and the original proposal has subsequently been withdrawn by the manufacturer who first promoted it.

The conventional EMG evaluation provides objective data about both phasic and tonic muscle contractions. There is a general consensus about "norms", but no definitive study has ever been done. In general, the pelvic muscles are reasonably assumed to be stronger in young persons, and to gradually decline with age and inactivity.

The EMG evaluation should always be conducted prior to any decision about the course of therapy. However, if the H&P has been completed, EMG results are usually easy to predict.

2. What other therapies should be tried prior to biofeedback therapy

None.

Assuming, of course that at the EMG evaluation indicates pelvic muscle strength and/or control inappropriate for the patient's age and general health. For an active, tennis-playing woman aged 25, a pelvic muscle reading of less than 10 microvolts would predict stress incontinence. For an 85 year old, a reading of less than 3 microvolts would be pathological.

Other than eliminating the possible sources of transient urinary incontinence (in the H&P), there are no other therapies that have any bearing on the use of biofeedback.

"...prior to biofeedback therapy."

The term "biofeedback therapy" covers a multitude of sins. There are three distinct FUNCTIONS that have been subsumed under that term: Documentation, Education, and Practice. In the original Perry Protocol, all three functions were included. (But not everyone today uses all three functions, as will be pointed out.)

A. DOCUMENTATION.

In the early stages of insurance-reimbursed biofeedback, it was essential to document the patient's condition in order to justify the use of biofeedback and to secure reimbursement. In fact, the documentation, in the form of a 51785 billing, was the ONLY aspect of the treatment that was reimbursed. During 1980-90, and even later, Medicare did not pay for biofeedback "practice" sessions [although physical therapist were often reimbursed for muscle reeducation work].

The formal EMG Evaluation, showing "flick", "hold", and "endurance" scores, was printed by the computer program and often submitted directly to insurance companies. The steady improvement in weekly scores was accepted as proof that something worthwhile was happening [in the home practice sessions], and the documented reduction in leakage symptoms eventually proved that the patient was no longer suffering from "stress incontinence", albeit without surgical intervention, and at a great cost savings to the insurer. Since the patient was usually "cured" by the time the insurance appeals were handled, the outcome was clear and the reimbursement justified.

One of the original purpose of the "documentation" phase of biofeedback was to convince the patient that there was objective improvement in her muscles, and therefore, it helped to motivate patients to do otherwise boring and repetitious exercises. The precise measurements provided by clinical biofeedback systems were considered an essential part of the patient motivation system of our original biofeedback protocol. Patients were instructed to take their "printout" home and post it on the refrigerator door!

It should be noted that "documentation" did NOT play any significant role in the research-oriented tradition of the NIH laboratories and the work of Kathy Burgio, PhD. Her subjects were NOT required to pay for treatment, and did not, so no such "documentation" was either required or provided to them.

Much confusion and many problems resulted when some private practitioners, including Burgio disciples J. Tries and B. Woolner, to mention only two of the best known examples, transplanted the NIH "free-service" model into the "fee-for-service" world. And as third-generation trainees began to provide services, insurance companies rightly rebelled at the undocumented claims that were being made for services provided.

B. EDUCATION The second function of "biofeedback therapy" is to teach the patient the correct and most effective methods of making pelvic muscle contractions. ALL biofeedback therapies include this function, and this is what makes biofeedback so much more effective than "mere verbal instruction".

Initially, EMG biofeedback was provided by a simple electrical meter (a voltage meter), which was NOT very effective. Then "light-bars" of LEDs provided much more colorful and dramatic feedback of pelvic muscle activity.

With the advent of personal computers, it became possible to display "movement over time", which proved to be significantly more educational than the previous single- axis displays. The traditional laboratory polygraphs used by Burgio and the NIA research also provide the "movement over time", but in a much less visible and convenient format. [The patient is lying in a left-lateral position and viewing the polygraph from the side.] Many of the most well-know biofeedback studies, including for example those of Burgio, Wyman, Burns, MacDonald, etc., provide ONLY the educational component of biofeedback, and NOT the documentation or practice components of traditional EMG biofeedback.

Typically, the educational phase includes only 3 or 4 sessions, or less than two hours of actual "biofeedback" time. The stated aim is to "teach" the patient how to do correct exercises, in the hopes that they will be able to do so on a daily basis at home. It is obvious from the outcomes of such research that a relatively high level of symptom reduction -- 80-90 percent -- is possible using biofeedback only in an educational manner.

C. BIOFEEDBACK FOR PRACTICE The original Kegel Perineometer, introduced in 1948, was used primarily for thrice-daily home practice, as documented in Kegel's publications and in the literature which accompanied his device. Kegel believed (based, he said, on his patient's progress) that daily biofeedback practice -- morning, noon and night -- was essential to recovery of urinary control. In one article he showed a hypothetical patient's progress, with a setback in muscle strength resulting from a day of skipped biofeedback practice.

Home practice devices have traditionally been the first products introduced by manufacturers interested in serving the incontinent population, starting with the "Personal Perineometer" from Farrall Instruments in 1981, and including the home trainer from Hollister/ InCare in 1988. Thought Technology introduced an inexpensive home trainer, the U-Control, in 1995, and other similar products have been or are about to be announced.

In order to keep the price of these devices reasonable, manufacturers have provided only the "light-bar"

one-dimensional (EMG) scale that was found on the original Personal Perineometer. Since there is no displayed time dimension, these devices are less useful for educational purposes. On the other hand, Jacque Sussett, using ONLY inexpensive home trainers for daily patient practice, produced even better results than Burgio, who used biofeedback only for education.

(see http://www.incontinet.com/comparison.htm for details.)

In summary, "biofeedback therapy" traditionally includes three components: documentation, education, and practice, although many research studies -- and clinical practices -- utilize only the educational function. Additional research is needed to the relative contribution to patient success of each of these components. It should be noted that many clinicians are content to provide a limited subset of components and then tell their patients "that's all we can do for you -- be glad you are down to X pads per day", when in fact that is not true.

"3. What parameters should be documented during a session"

In the Perry Protocol, which first proposed and described the use of EMG biofeedback for the treatment of incontinence, there are two major areas of documentation: (1) Physiological measures and (2) Symptom measures.

(1) Physiological Measures. Since the use of biofeedback for incontinence is predicated on an objective abnormality in the function and control of the pelvic floor muscles, it is essential that changes in this abnormality be documented at every stage of therapy.

Consistent with the concepts first developed in "The Ghost", it is essential to demonstrate "training effects" prior to any expectation of "outcome effects". If the muscles don't improve, we should not expect the symptoms to improve.

Conversely, if symptoms improve but the objective muscle measurements don't, it is clear that some placebo effect is in operation and the biofeedback therapy is NOT being effective -- and therefore, perhaps should not be reimbursed, either.

The parameters of EMG biofeedback evaluation have been discussed above. The initial RESTING LEVEL is essential in order to determine if the patient is relaxed or tense. This is important for effective training, but high pelvic tension is also a component in Sensory Urge Incontinence as well as several Pelvic Pain Dysfunctions, which may co-exist with urinary incontinence.

It should be noted that manometric devices are of no value in determining the patient's resting level, since they cannot differentiate between an anatomically tiny orifice and a functionally tense one. Although Kegel used a manometric device in 1948, there has been significant scientific progress in the study of "relaxation", especially since Benson's work in the 1960s. There is no justification for utilizing a 51-year old technology just because it was the only thing available in 1948.

(2) Symptom Measures. The use of daily bowel and bladder records for ALL patients, first described in the early 1980s by Burgio et al, is absolutely essential to an effective biofeedback-training program. Like any self-improvement program, daily biofeedback practice requires strong motivation and patient discipline. Since there is usually a decline in urinary leakage symptoms after only a few days practice, this change should be documented for the patient to ensure and enhance motivation [I.e., "See, it is working already! You are still wearing pads, but your leaks are down by 20%! Keep it up!] Obviously a complete B&B record will include incidents of Urgency and frequency of urination, as well as a record of fluid intake.

The occurrence of leaks (or frequency of toileting, or urgency) is sometimes considered a "subjective" measure, and researchers suggest that more objective measures, such as pad weights or pad changes are more "objective" signs of change. But incontinence is a highly personal thing; some patients are deeply disturbed by a small amount of leakage, while others are not. Some patients change pads at the first sign of leakage, while others try to get their money's worth out of each pad. The use of a "pad test" at home is not practical, and the number of leakage incidents is a more significant outcome measure in terms of patient concerns.

"4. How many sessions are reasonable as a standard"

Biofeedback training is a skills acquisition task, and the time it takes varies considerably because some patients learn faster than others. The reasonableness of any number of sessions should be based on the amount of progress that is being made.

It is entirely inappropriate, and inconsiderate of the individual differences in patient functioning, to set any fixed number of sessions as the standard.

[To date, no research has included the appropriate variable "treatment levels" in analysis of biofeedback therapy. In every case, a fixed number of sessions were used, without any scientific justification for such decision.] However, if biofeedback techniques are being properly employed, and the patient is carrying out the required daily at-home practice, there should be (1) a steady improvement in weekly clinical EMG evaluation scores and (2) a steady reduction in urinary accidents.

If there is a lack of improvement in either measure from one week to the next, that is a cause of clinical concern, investigation, and remedy. A consistent lack of improvement in a particular patient may be due to falsified home practice records. A consistent lack of substantial improvement in a therapist's patients may indicate faulty instruments, techniques, or simply poor training of the therapist.

When protocols involving Documentation, Education, and biofeedback Practice are employed, an average of 4 sessions over 5-6 weeks is usually sufficient to get 100% Symptom improvement. With anorectal manometrics, most patients will not tolerate more than three to five sessions of training because of the discomfort and pain. In the Perry Protocol, patients are trained for an additional 30 days to ensure "over-learning" of their muscle control, to avoid a relapse into incontinence during transient stressors, such as a bout with flu.

"5. What situations would support additional sessions"

The most significant factor in prolonging therapy is the number of prior surgeries the patient has had. Four or five such operations can easily double the amount of time that is required. (I.e., from 4 to 8 sessions over 8 to 12 weeks.)

Post-proctatectomy patients are very difficult to predict, since some amount of "muscle bulking" is usually required, and that comes from repetitive exercises, not just learning to do exercises correctly. The amount of bulking needed is related to the amount of material removed during surgery; unfortunately, that is seldom documented. 8 to 16 office visits over 2 to 5 months is not uncommon.

"6. What parameters should be followed to demonstrate outcomes"

As discussed above, the "Symptom reduction ratio", first devised and used by Burgio et al, is the most relevant and representative statistic available. It is also the only one that allows description of both individual patients and groups of patients.

"7. And, of course, what are the indications (medical necessity) for biofeedback therapy of the pelvic floor"

Biofeedback Therapy is indicated in all cases of stress and urge urinary incontinence, urinary retention, and fecal incontinence and functional constipation. Some research shows that a minimum level of patient awareness and ability to engage in the required exercise practice is essential. However, a large measure of such awareness and discipline may be supplied by an appropriate caregiver in many cases.

"We had originally hoped that we might have time to at least begin a similar discussion of electrical stimulation. That is still our desire but considering the starting time of our 6/21 meeting, this will probably need to be rescheduled.

There are substantial and significant between biofeedback and electrical stimulation, which is best understood as an 'active agent' therapy and must be evaluated by methods similar to drugs. For instance, the recent advertisements for electromagnetic chair therapy state that the patient does not even have to pay attention to what is happening.

It would seriously damage the argument for biofeedback to attempt to include a very dissimilar modality, e-stim, in the same discussion. Both therapies involve sensors, wires, and black boxes -- but that is only a superficial similarity. The direction of electrical transmission is 180 degrees opposite, and so is the therapeutic mechanism.

"The SUNA / WOCN Continence Coalition wishes to thank each for your quick responses on this issue and looks forward to a productive meeting in Minneapolis. "

Sorry that I could not be there in person, but I trust this will represent my views, and the views of my company, PerryMeter Systems, for now.

/s/ John D. Perry, PhD