Predicting Disease Probability?

Several organizations have tempered their enthusiasm for routine UDS testing. As early as 1996, the former Agency for Health Care Policy and Research (now called the Agency for Healthcare Research and Quality) proposed that some women did not need UDS testing before surgical treatment. In 2006, the British National Institute for Health and Clinical Excellence (NICE) discouraged UDS testing before any conservative therapy and stated that, in select populations, this testing was not routinely recommended before stress incontinence surgery. In 2005 and reaffirmed in 2011, the American College of Obstetrician Gynecologist practice bulletin on urinary incontinence in women stated: “limited data support the need for cystometric testing in the routine or basic evaluation of urinary incontinence.” Most recently, the Cochrane Collaboration summary on UDS testing for the management of urinary incontinence in children and adults concluded that although testing may change clinical decision making, there is insufficient evidence to show that the outcomes of these decisions render better clinical outcomes. These summary conclusions reflect the diagnostic and therapeutic challenges of managing female lower urinary tract disease.

The precise causes of the various forms of lower urinary tract dysfunction are not clearly understood; correspondingly, the diagnostic criteria of lower urinary tract diseases are many times fuzzy. For example, female stress urinary incontinence could be regarded as one area of success both in our understanding of pathophysiology and in our directed treatments. Yet it remains unclear whether female stress urinary incontinence stems from a lack of anatomic support or of intrinsic urethral function; conceivably, a sling may work by providing a backboard or by correcting neuromuscular function. This sort of fundamental lack of understanding compounds across the translational spectrum from theory to clinical practice. The International Continence Society (ICS) posits on expert opinion that the role of UDS testing is to reproduce the lower urinary tract symptom in controlled conditions so that the cause of the symptom can be determined and objective information rendered to the clinician. Echoing the conclusions of the Cochrane Collaboration, objective information may, indeed, be rendered with UDS testing but it is not clear that this information improves clinical outcomes. Furthermore, given the aforementioned problem of the compounding error that begins with our incomplete understanding of lower urinary tract diseases, the hope that patients’ symptoms could be reproduced with UDS testing is illusory.

Test characteristics are often described in terms of accuracy and precision. The distinctions between these 2 concepts are well defined when discussing a game of darts but less so in respect to most medical contexts. Generally, an accurate test is better than a precise test because, in the former case, there is at least the possibility that what is measured reflects some reference. In defining a test’s net benefit, that reference should point to some useful understanding of reality. A test with poor precision renders uncertainty that the result has approximated the reference; the more uncertain, the less useful the test becomes. In the case of UDS testing, both accuracy and precision are problematic.

The ICS states that there is diagnostic agreement between different observers over urodynamic traces in about 80% of cases. This finding is a misrepresentation of the study cited for that estimate. Logically and experimentally, the reliability and agreement of urodynamic interpretations vary with the diagnosis. As expected, physician evaluators agree more often with themselves than with colleagues but, except for stress incontinence, the reliability and agreement estimates for urodynamic interpretations are moderate at best (at best 38% interobserver agreement for detrusor overactivity with incontinence). The lack of clear quantitative referents leads to the precision problems identified for urodynamic interpretations. The lack of such referents only highlights the diagnostic problems of lower urinary tract diseases and, even if these referents existed, they may have no clinical value.

The accuracy of UDS testing is unknowable. Should estimates of lower urinary tract function rely on the results of UDS testing or the patient’s symptoms? Patient history in predicting urodynamic findings is variable. Patient history has a positive predictive value of identifying urodynamic stress incontinence that ranges between 52% and 100%. Still, detrusor overactivity (DO) is identified in 11% to 16% of women with pure stress incontinence symptoms. Conversely, 22% of women with urgency symptoms are found to have urodynamic stress incontinence. Symptoms may suggest some urodynamic findings but clearly not all and not all the time. Overall, the correlation between the clinical and urodynamic diagnosis in classifying urinary incontinence diagnoses is poor, and to assume the urodynamic observations are the referent standard is unwise based on the current evidence.

Pelvic floor problems are contextualized diseases and decision making by patients and physicians is never exclusively summarized by symptom inventory or testing results. This feature alone should give pause to any practice of routine urodynamic testing. The patients’ therapeutic goal should drive the ascertainment of diagnostic information. The focus on how UDS testing impacts clinical outcomes is the subject of at least 2 forthcoming randomized trials: the Value of Urodynamics prior to Stress Incontinence Surgery (VUSIS) and Value of Urodynamic Evaluation (ValUE) trials. The results of the Stress Incontinence Surgical Treatment Efficacy Trial (SISTEr trial) in comparing the efficacy of the Burch urethropexy and pubovaginal sling did not demonstrate that urodynamic testing predicted the treatment outcome or postoperative voiding dysfunction. Yet women with Valsalva leak point pressures or maximum urethral closure pressures in the lowest quartile were nearly 2-fold likely to experience recurrent urine leakage 1 year after an obturator or retropubic midurethral sling. This sort of prognostic information poses a metaphysical problem. Assuming these findings are reproducible in routine clinical practice, how much is prognosis worth? This sort of question cannot be answered without engaging patients.

In the case of pelvic organ prolapse, UDS testing is hoped to clarify who might have lower urinary tract dysfunction following repair. Pelvic organ prolapse does impact lower urinary tract function, yet it is unclear that UDS testing adds much additional diagnostic or prognostic information. The Colpopexy and Urinary Reduction trial (CARE trial) showed that urine leakage with preoperative prolapse reduction, although variable with technique, predicted a higher risk for postoperative stress incontinence at 3 months; yet this same trial recommended prophylactic incontinence surgery in all women undergoing a sacrocolpopexy. Following this logic, a woman’s preoperative continence could be seen as having no diagnostic value in the choice to perform a concurrent incontinence procedure. Indeed, the mean risk of developing postoperative stress incontinence following the surgical repair of vaginal prolapse is around 15% regardless of the preoperative clinical symptoms if a concurrent incontinence procedure is performed. That mean risk jumps to between 59% and 65% without concurrent incontinence surgery in the settings of either positive occult testing in a woman without clinical symptoms of stress incontinence or positive clinical symptoms. Reduction testing to identify occult incontinence is not an effective method to guide surgeon decision making. The use of urodynamic testing with occult testing does not seem to improve on clinical judgment. The decision to perform an incontinence surgery concurrent with any repair of vaginal prolapse is largely without evidenced guidance and, as such, patient choice should prevail.

Common Treatments for Different Problems?

When 2 or more diseases possess the same treatment, the value of any testing that makes these distinctions is dubious. Conservative treatments such as physical therapy, behavior modification, medications, or pessary, can address a variety of lower urinary tract symptoms. For interested patients, these options could be pursued without a precise diagnosis, and the results of any therapeutic trial may render diagnostic information more useful than that discoverable from UDS testing. There are, however, other settings whereby the information gathered from UDS testing makes distinctions without differences. The distinction between types I and II stress urinary incontinence and type III has been based on the urodynamic measure of Valsalva leak point pressure. For years, this distinction drove differential stress incontinence treatment: a Burch urethropexy for types I and II and a pubovaginal sling for type III. The SISTEr trial did not confirm the utility of this distinction.

Digesu and colleagues conducted a large study of 4500 women with symptoms suggestive of overactive bladder (OAB). A postvoid residual urine of more than 100 mL was identified in 8% of women. A third of the patients had DO on UDS testing, with only 28% having OAB symptoms. The investigators argued that without UDS testing, some women could have been treated with anticholinergic medications and made worse and other asymptomatic women would have been denied treatment. An opposite conclusion seems equally reasonable. Besides documenting the poor correlation between DO and OAB, to initiate anticholinergic medications only on urodynamic evidence of DO delays care and denies many symptomatic women of a medication they may find helpful. Hashim and Abrams’ study of patients with OAB found 30% of women with no OAB symptoms had DO on testing. In women with OAB symptoms and no DO, it is unknowable if the test or the diagnosis failed. Treatment considerations, however, necessarily prioritize the symptom, thus, undermining the value of UDS testing. For urinary urgency and frequency, the distinction between OAB and DO does not seem to render a rational difference in the treatment approach.

Per the most recent Cochrane Collaboration review, clinical decision making may be altered by UDS testing, but clinical outcomes do not seem to be altered. Specifically, there is no difference in the number of women with posttreatment urinary incontinence if they did or did not undergo pretreatment UDS testing. The rate at which probability differences in UDS diagnoses will result in meaningful changes in intervention seems to be low and, correspondingly, factors negatively in rendering a positive net patient benefit to routine UDS testing.

Benefit versus Harm of Treatments?

The benefit of treatment in the context of lower urinary tract dysfunction is the restoration of normal bladder function. This restoration would include painless urine storage and efficient and painless voiding. The physical harm of any treatment would threaten one of these basic functions. There are, however, nonphysical harms in treatment, such as inconvenience or financial cost. Harm is any cost that subtracts from the benefit of an action (eg, treatment), and the first law of thermodynamics tells us that no action is without some cost. There is a probability range for the possible costs associated with a given action. A risk is ordinarily understood as a cost with low probability but that can subtract more greatly from the actions benefits. Patients can best balance nonphysical costs because the factors that influence them are often unknown to the clinician. Physical costs of treatment, however, are often well known by clinicians, yet their relative valuation by patients still remains unknown and perhaps unknowable. Taken together, with respect to the matter of any testing, if the net treatment costs exceeds benefits, then the value of testing is undermined.

Physical-therapy treatments for lower urinary tract dysfunctions possess little cost beyond financial cost and inconvenience. Over the long run, these nonphysical costs can be significant enough to lead to discontinuation, hence, the long-term net benefit seems to be low for many women. Medical treatments for urinary incontinence have variable effects on lower urinary tract dysfunction; although the benefits may be meager, the physical costs by way of side effects are common. These side effects include dry mouth, constipation, and dizziness, among others. Although some individual responses may tilt toward net treatment benefit, for many women, medications taken to treat lower urinary tract dysfunction seem to be a wash, with nearly 60% discontinuing the medications at 6 months. The financial costs of anticholinergic medications can be an important barrier to continued use and age impacts the severity of side effects. A small net treatment benefit potentially compromises the net benefit of testing if other factors dominate, such as the harms of testing. Given that the patient valuation of benefits from nonsurgical therapies are difficult to estimate, routine UDS testing before such therapies seems unwise.

Surgical interventions for lower urinary tract dysfunction face the same challenges in restoring lower urinary tract dysfunction. The harms of surgical treatment could be understood in terms of immediate operative and postoperative complications as well as longer-term problems. Operative complications include injury to the lower urinary tract, hemorrhage, bowel injury, wound complications, and infection. Postoperative complications include voiding problems, voiding urgency, altered bladder sensation, and altered sexual function among others. As has been already noted earlier, bladder problems are contextualized diseases. Environmental buffers and barriers alter how a given lower urinary tract problem manifests in the life of an individual. For example, postsurgical alterations in sexual function may not be an adverse effect in a nun. Although long-term results (>4 years) of any incontinence surgery are often reported to be greater than 80%, the SISTEr trial documents an alternative reality. The Kaplan-Meier curves for the success of surgical treatment of urinary incontinence never plateau, with the 2-year outcomes for the 2 surgical approaches around 50%. In this same trial, patient satisfaction from surgery was generally high (73%), but the only baseline predictor of lower satisfaction was greater-urge incontinence symptoms. Given that patients undergoing stress incontinence surgery expect not only the resolution of urine leakage but also urgency and frequency symptoms and given that these symptoms are among those that baseline predict postsurgery dissatisfaction, what role does UDS testing have in clarifying the presurgery net treatment benefit? What is in view here is trying to understand the net benefit of surgery in treating urinary incontinence; there are, indeed, objective and subjective outcomes that document patient benefit but there are also various personalized costs. In neither case are these benefits or costs necessarily clarified by UDS testing.

What about sacral nerve stimulation? Following the recommended basic assessment of any woman presenting with lower urinary tract complaints, a diary of bladder function will have been obtained. That diary documents, at a minimum, the criteria used to judge the efficacy of sacral nerve testing. As has already been stated earlier, there is a disconnection between OAB symptoms and DO, yet for obvious reasons, the OAB symptom is the priority in treatment decisions. Given that office sacral nerve testing is a feasible strategy for testing this therapy option, there is the potential for significant net treatment benefit. For appropriate patients, why add the costs of UDS testing?

What about pelvic organ prolapse surgery? As has already been mentioned, the role of UDS testing in the setting of pending pelvic organ prolapse surgery repair is unclear. Barrier testing may have some predictive role in elucidating which patients will leak following vaginal prolapse repair but this does not require UDS testing. Predicting voiding problems following vaginal prolapse surgery is hampered by the lack of poor definitional referents and the multiple overlapping factors that contribute to voiding problems in this population (eg, medications, age, infection, and so forth). Taken together, there are benefits to surgical and nonsurgical treatments of vaginal prolapse but there are also costs and UDS testing does little to clarify either.

The Harm of Testing?

There is little patient risk in undergoing multichannel urodynamic testing. One randomized double-blind study of antibiotic prophylaxis following UDS showed a 23% infection rate among controls. This rate was not significantly less with antibiotic prophylaxis (18%); however, both rates seem higher than common experience. A systematic review on this topic, although citing poor study methodology among the 8 trials reviewed, reported that 13 individuals would need antibiotic prophylaxis to prevent 1 significant bacteriuria of unknown clinical significance. Beyond infection risk, there are patient discomforts associated with the test, yet one study found 81% of women would repeat UDS testing. This Asian survey study of men and women undergoing UDS testing showed that among pain, physical burden, and embarrassment, all were rated low. Embarrassment rated highest, although a male clinician performing the testing among female patients may have biased this outcome. These subjective ratings may also vary between cultures. Nevertheless, the risk of infection and the reported subjective pain and embarrassment of UDS testing seems to be low.

Although the immediate individual harms of UDS testing seem to be low, there are other contexts to understand the harm of testing. Testing could alter clinical decision making away from the best therapy whereby best could be defined not only in terms of clinical outcomes but the costs to achieve a given clinical outcome. Testing also carries costs to the health care system. Within the British National Health System, the financial resources devoted to health care are fixed, meaning that any money spent on one health expense takes it from others. Within such a system, the net societal benefit determines how health resources are used. Understandably, reducing any testing without a clear benefit is imperative in these systems. The NICE position on UDS testing reflects this thinking. Although such a system has not arrived on US shores, it is clear that the era of limitless health spending is drawing to an end. The interest in capitated health care financing will put powerful pressures on any testing that does not render a clear benefit in clinical decision making.

The Physician Benefit

Never do surgery on an unstudied bladder. So goes the refrain heard among physicians who insist on performing urodynamic testing in all women seen for a pelvic floor disorder. When asked about the diagnostic yield of such testing, the refrain is repeated, which suggests that obeisance to authorities as much as anything else compels testing. There is also the sense that the performance of UDS is self-actualizing to a clinician’s subspecialty medical focus that may render a referral benefit. It should also not go unnoticed that there are powerful financial reasons to do urodynamic testing. Medicare reimbursement for multichannel urodynamic testing can exceed $1000 and private insurers would be expected to reimburse much more than that amount. For comparison, Medicare professional reimbursement for an office cystoscopy is less than $200. It would be profoundly naïve to think that financial incentives do not influence how physicians think about urodynamic testing.

There are efficiency considerations for performing UDS testing. For example, Nager and Jasmine note in surgical planning to correct vaginal prolapse, clinical or urodynamic testing can be done to detect occult urinary incontinence. Clinical testing would seem easier for most patients and physicians, yet it is conceivable that UDS testing may offer the most efficient way to identify the relevant information necessary to make a clinical decision acknowledging that in many cases that decision will be no better than had it been made without the results of UDS testing.

It is useful to distinguish between puzzles and mysteries. Mysteries are not clarified with more information, whereas puzzles are only solved with it. For example, Malcolm Gladwell thinks the Enron debacle was a mystery; the trouble was not because Enron’s management did not reveal enough information but rather it was that analysts failed to make sense of the data that were supplied. Judgment is needed to navigate a mystery because additional information may only make things less clear. Female bladder dysfunction is a mystery. What medical progress has meant for the problem of female bladder dysfunction (and for many other settings in medicine) is the transformation of a puzzle into a mystery. Although in the past UDS testing may have rendered information that quickly resolved a puzzle, today, amidst all that we know (and do not know) about bladder dysfunction and its management, judgment is what is most needed by patients and clinicians. Knowledge of patients’ therapeutic goals, a careful history and physical examination, simple office testing, and awareness of what the individual surgeon can and cannot deliver renders sufficient data to base good judgment in many cases of female lower urinary tract dysfunction.