Chapter 4Discussion

KQ1: Prevalence and Incidence of OAB

Methodologic issues. Data related to prevalence and incidence of OAB and urge urinary incontinence is notably coherent given the immense technical challenges to comparability. Response rates vary widely across studies with no clear patterns relating type of questionnaire (phone, mailed, administered) or population to completeness of response. Most research teams documented exhausting conventional options for improving response rates and many presented thoughtful analyses of non-response. While estimates, especially of self-reported symptom complexes, are very sensitive to the operational definitions used, these measures were able to be compared when the authors provided information about details like wording of survey items and the required frequency of an event to meet criterion definitions. With some exceptions, such information provided sufficient context to provide a global and United States picture of OAB and urge incontinence.

Publications that appeared after the ICS consensus definitions had more similar results, even when they used operational definitions that differed from ICS. In general, greater detail was provided in the publications after the ICS standardization to support comparisons across studies. We found no evidence that studies supported by pharmaceutical companies returned higher estimates for similar measures, with the exception of a single United States publication reporting on a cluster of consumer surveys done in malls. The estimates from those surveys were high enough to be considered outliers and the operational definitions used to derive the estimates were not provided in order to determine how their results compared to others.³⁶ There was no readily apparent relationship between method of administration (e.g., mailed survey, administered, or other methods) and findings which is reassuring.

Summary of prevalence and incidence findings. Overactive bladder and urge urinary incontinence occur in women of all ages around the world. The type of healthcare system appears not to clearly relate to the proportion of women affected at any point in time, suggesting that universal access to care such as in the United Kingdom and Canada versus more socioeconomically correlated access in the United States, is not the driver of whether or not women currently report symptoms. Several factors work in synergy to determine whether a woman with OAB symptoms is counted among prevalent OAB cases: degree of bother of symptoms, which varies widely; individual decision to seek care, which is shown to be a low proportion of those affected (13 percent of nurses in the Nurses Health Study); availability of care; and the degree to which an individual woman is distressed by the symptoms or interprets them as a normal part of aging.

Conservatively estimated more than 10 to 15 percent of adult, community-dwelling women in the United States meet criteria for OAB and as many as 5 to 10 percent have urinary incontinence associated with urgency. OAB and continence status are not static; both OAB and urge incontinence can resolve, though the data suggest most women are affected for multiple years at minimum if they develop symptoms.^{64, 77} How often resolution is related to treatment, individual behavioral change, or the natural history of the condition is completely undocumented. Evidence to describe the relationship of UUI or OAB with age is varied. The preponderance of the literature suggests that OAB risk is more directly associated with increasing age than urge incontinence. For both conditions the relationship may be complex, such that rates of increase in the proportion of women who are symptomatic are relatively consistent until a threshold age. The most probable threshold ages at which rates plateau are well past the average age of menopause and reported to fall between 60 and 75 years of age.

Because of the population-based design of these reports, often details like recent urinary tract infections or childbirth were not able to be taken into account, though future researchers could incorporate self-reported status in inclusion and exclusion for the studies as some of the strongest publications in this literature did. Likewise, future research on incidence and regression of symptoms would do well to attempt to account for intercurrent diagnosis and whether or not an individual has received treatment.

The epidemiology of OAB and urge incontinence lead us to conclude, as have others, that the conditions are common, occur across the lifespan, and that providers of all types – primary care, specialists, and advanced practice nurses and health educators – will be called on to advise patients and provide care.¹⁹⁶ Opportunities for detection and treatment in clinical settings should be frequent and attention should be addressed to the degree to which patients find the symptoms distressing including impact of self-image, quality of life, and sexual function.

KQ2: Outcomes of Treatments for OAB

Pharmacologic treatments.

Methodologic consideration for pharmacologic treatment studies. The content of this literature is predominantly of fair to poor quality and strength of the evidence is at best moderate, and improved only by the consistency with which medications for OAB are shown to have modest advantage over placebo. The duration of followup during the masked portion of trials is short, rarely longer than 12 weeks. Given the longevity of treatment likely to be used by patients this is concerning and limits the generalizability to clinical practice. No investigators were able to determine whether or not there is a time period after which individuals can discontinue treatment and still retain some or all of the benefits of treatment reduction. Statistical and methodologic concerns are addressed in detail in the Future Research Section of this chapter. After review and analysis of 110 studies, of which four were good quality, 75 fair and 31 poor, with 68 RCTs, the strength of the evidence for managing OAB with pharmacologic treatment is weak to moderate for short term outcomes and weak for long term outcomes and harms.

Findings from pharmacologic treatment studies. All pharmacologic treatments were effective at improving one or more OAB symptoms when compared to placebo. Reductions ranged from 0.9 to 4.6 in urge urinary incontinence episodes per day across all drug treatments and from 0.7 to 4.2 in voids per day. Study by study, extended release formulations achieved better effects than immediate release, although statistical significance varied. No one drug was definitively superior to others by preponderance of evidence, including comparison of newer selective agents to older antimuscarinics. As estimated by meta-analysis extended release forms (taken once a day) reduce UUI by 1.78 (95 percent CI: 1.61, 1.94) episodes per day, and voids by 2.24 (95 percent CI: 2.03, 2.46) per day. Immediate release forms (taken twice or more a day) reduce UUI by 1.46 (95 percent CI: 1.28, 1.64), and voids by 2.17 (95 percent CI: 1.81, 2.54). Of note, placebo reduces UUI by 1.08 (95 percent CI: 0.86, 1.30), and voids by 1.48 (95 percent CI: 1.19, 1.71). Even in the context of small to moderate affect on symptoms, pharmacologic treatments were generally associated with increased quality of life and reductions in measures of impact or distress, compared to baseline and to placebo.

Findings reported in this review are consistent with three prior reviews.^{196, 270, 271} The most recent found: (1) antimuscarinics are efficacious compared to placebo, (2) mean decrease in UUI episodes per day ranged from 0.4 to 1.1, (3) mean decrease in the number of voids per day ranged from 0.5 to 1.3, (4) every treatment, with two exceptions, was associated with greater risk of adverse events compared to placebo, and (5) improvements were seen in quality of life.²⁷¹ This review added an additional 28 studies and incorporated evidence from study designs other than randomized clinical trials.

Table 31 below provides estimates of treatment effects for pharmacologic treatments represented by more than one trial arm. Some drugs and doses of drugs are not reported because the publications with trial arms for that treatment did not provide sufficient information to estimate variance in meta-analysis models. The models required that we have some estimate of the variance of the effect size such as standard deviation, standard error, or confidence bound, in order to achieve appropriate estimates.

Since baseline episodes of UUI per day ranged from 1.6 to 5.3, and voids per day from 7.2 to 13.7, these reductions (Table 31) reflect modest margins of benefit from baseline above placebo. Data was not consistently provided across studies to estimate the proportion of women who became symptom free.

Procedural and surgical treatments.

Methodologic consideration for procedural and surgical treatment studies. Studies in this treatment domain are of limited quality and predominantly case series in specialized treatment settings. Sacral neuromodulation has not had properly masked randomized clinical trials,¹⁹³ and botulinum toxin injections are promising but based on a small number of studies that identified urinary retention as a distinct risk factor that is self-resolving but troublesome.¹⁹³ Other procedures found no benefits or are no longer used in practice. Given consideration of 18 studies, of which 11 were fair quality and seven poor, with five RCTs, the strength of the evidence for managing OAB with procedural and surgical treatment is weak for all aspects of understanding outcomes of care.

Findings from procedural and surgical treatment studies. Among the trials of procedures and surgery, one demonstrated a statistically significant benefit of sacral neuromodulation over usual care for the reduction of episodes of urge urinary incontinence per day (average reduction of 7.1 compared to 2.1 increase with usual care) among subjects with OAB known to be refractory to medical therapy.¹²⁴ Enthusiasm is tempered primarily by reports from multiple case series that harms are not rare with these treatment approaches. Data that reflects newer sacral neuromodulation techniques in this refractory population are lacking.

Surgical and procedural treatments for OAB are not first line management due to the cost and risks of these types of procedures. Sacral neuromodulation decreases the number of urge incontinence episodes by at least 50 percent among patients refractory to conservative therapies. Moreover, the number of moderate-heavy incontinence episodes also decreases. These results persist even at five years. Sacral neuromodulation seems to have less of a benefit for urinary urgency (mixed results found in our review) and frequency (31 to 45 percent decrease), with benefits in frequency tapering off over time (23 percent reduction in daily voids at 5 years). These benefits come with a relatively high rate of adverse events. Early studies using older technology had more than one adverse event per subject, on average;^{115, 124} studies employing newer technology report lower rate, with events in 11 to 53 percent of subjects.^{112, 119} Nearly 40 percent describe pain or an unpleasant stimulation, 7 to 48 percent returned to the operating room (this increased to 67 percent at five years, but includes the need to change the implantable pulse generator battery) and between 2 to 6 percent have an infection (often requiring hospitalization, intravenous antibiotics or explantation). The overall explantation rate hovers around 10 percent.

Our review included only one study on peripheral neuromodulation, using an anal and/or vaginal probe. Benefits in frequency were unlikely to be clinically significant (decrease from 9 to 8 voids daily) and there was a high dropout rate due to pain and effects on the bowels. Other forms of peripheral neuromodulation such as posterior tibial stimulation were not reviewed

Electromagnetic stimulation with a portable unit was not found to be beneficial for OAB.

Of the drugs injected or instilled into the bladder, botulinum toxin and oxybutynin had the greatest benefit. One trial demonstrated benefit of instillation of oxybutynin compared to sterile water in the reduction of voids per day (average reduction of 6.8 compared to 2.4).¹²⁶ A trial we identified and the recent review by the Cochrane Collaboration found that small trials suggest benefit though researchers continue to evaluate means to decrease the risk of undesired side effects like urinary retention with botulinum toxin. Both botulinum toxin and instilled oxybutynin increase the postvoid residuals and the long term effects of higher residuals in terms of bladder infection and other risks is not known. Although evidence for these approaches is promising, the strength of the literature is inadequate to recommend any of these approaches for broader use in general practice. As of the date of the report, neither treatment is FDA approved for OAB.

Resiniferatoxin injection was not beneficial in the study reviewed. Older treatment modalities such as prolonged bladder distention or bladder transection are no longer commonly used due to the morbidity of these procedures. The reviewed studies also lacked rigorous methods for evaluating treatment benefit.

Behavioral treatments.

Methodologic considerations in behavioral treatment studies. Most of the literature addressing behavioral interventions (with or without comparison to pharmacologic intervention) was of fair or poor quality. In general, studies of behavioral approaches rarely included a true and comparable placebo arm. Although it is well recognized that there are inherent challenges to developing placebos for behavioral techniques, a reasonably strong evidence base on means of doing so suggests that it is possible. Burgio and colleagues¹⁴³ worked to mitigate this issue by maintaining similar visit schedules and through the use of bladder diaries in all groups, which was considered adequate masking for quality grading purposes. Particularly in older studies (prior to 2002), the behavioral approach often is not fully described; and inconsistency in the language used to identify different approaches requires the reader to examine the manuscripts very carefully – multiple studies may have called their approach by the same name, but in fact be studying quite different interventions.

To mitigate against such confusion, we have attempted in this report to always describe the intervention along with the first description of results from a given study. Studies conducted prior to the ICS definition of OAB in 2002 tended to examine more limited approaches to bladder training, while later studies focused more on multicomponent approaches and delivery of the training in different ways. Prior treatment attempts are rarely documented in this work, which may make it difficult to compare treatment groups across studies. Finally, this body of literature includes very few studies that included similar combinations of intervention and comparator making it almost impossible to summarize across them. After review and analysis of 29 studies, of which 14 were fair quality and 15 poor, including 17 RCTs, the strength of the evidence for managing OAB with behavioral approaches treatment is moderate to weak for short term outcomes and weak for long term outcomes and harms.

Findings from behavioral treatment studies. Conclusions about the effectiveness of behavioral techniques for addressing the symptoms of overactive bladder are based on a total of 29 papers (27 studies) that encompass behavioral to behavioral comparisons as well as studies of combining behavioral approaches with pharmacologic treatment, and the reverse, combining pharmacologic approaches with behavioral ones. No two studies could be combined to produce summary data. Overall, behavioral approaches can be effective in reducing episodes of incontinence and daily voids. Multicomponent approaches are most effective, and they perform relatively equivalent to pharmacologic treatment. Generally speaking, reductions in symptoms were modest, with potential decreases in incontinence episodes of up to 1.9 per day, and reductions of up to about four voids per day. The addition of caffeine reduction to behavioral modification reduced frequency, but made no difference in reduction of incontinence episodes. There is no evidence the behavioral approaches enhance the effectiveness of pharmacologic therapy to reduce episodes of incontinence; and like pharmacologic approaches, there is no evidence for long term effectiveness beyond the period during which the intervention is being provided in the health care setting.

Complementary and alternative medicine treatments.

Findings from complementary and alternative therapy studies. We identified three studies that used complementary and alternative medicine therapies to treat OAB: a fair quality trial of acupuncture,¹³⁷ a fair quality trial of foot reflexology,¹³⁸ and a poor quality prospective case series of hypnotherapy. There is weak to no evidence for complementary and alternative approaches to managing OAB.

Outcomes of treatment. The small trial of acupuncture has intriguing results related to decreased frequency of voiding and reduced symptoms of urgency which are associated with changes in cystometrics related to improved bladder capacity that are logical intermediates of the improvement in symptoms. Women felt they were improved as measured by scales that capture bother and quality of life. Evidence is insufficient to support definitive choice of acupuncture but offers preliminary information that promises modest improvements similar to those reported in many pharmacologic trials.

Reflexology is represented by a small trial with unmasking of participants that could have biased the results. No evidence supports choice of this modality. Likewise, hypnotherapy is not supported by the scant information provided by one case series with little detail, patient reported outcomes, or statistical assessment. Given the scope of placebo effects demonstrated in other well-conducted studies of OAB treatment, it is difficult to know whether to attribute any effect to hypnotherapy.

KQ3: Comparisons of Treatments

Methodologic issues in studies that compare treatments. Evidence for direct comparisons of treatments was based on 19 studies: 12 were fair and 7 poor. Of these, 14 were RCTS. Evidence is currently weak to absent for choosing one therapy over another.

Pharmacologic. A number of pharmaceutical agents have been studied in direct comparison. Nine of these comparisons explicitly examined outcomes for urge urinary incontinence episodes and voids per day. Fewer report on the symptom of urgency. Five of these studies are comparability studies in which a new drug aimed to establish equivalence to oxybutynin, the first drug FDA approved for OAB. One study tested the hypothesis of whether the ER formulation of tolterodine was superior to the IR formulation and the remaining studies were “challenges” of newer drugs to tolterodine, which was the second drug to receive FDA approval. In this context lack of statistical differences between active drugs is somewhat uninformative as trials were often powered for the comparison of the drugs individually to placebo and in many cases statistical testing of the outcomes of drug-to-drug comparison are not provided. Where studies reported being powered to detect differences between drugs, the postulated differences were unlikely and/or the withdrawals from protocol prevented meaningful effectiveness analysis beyond ITT. In a health services context, ideal comparisons would be of extended durations, since OAB is a chronic condition, and would report on all primary outcomes of relevance including differences in medication adherence, satisfaction with treatment, and quality of life.

Procedures. Participants in these studies were often not exclusively those with OAB. Strict application of inclusion criteria for this review would have eliminated virtually all studies of procedures for this reason. Studies of OAB included those with urinary retention and at times neurologic conditions as the indication. Masking, though challenging, was approximated by insertion of leads for sacral neuromodulation without activation; however given that a test period is done to establish efficacy before implantation of the permanent device, individuals may have been aware of their status and assessment of unmasking is not provided. Developing sham procedure methods may be of special importance in this area.

Behavioral. Variations in the behavioral approaches used and methods for teaching them, as well as differences in the duration and intensity of treatment make comparisons challenging. As a category the methods were generally strong with the continued challenge of developing attention control comparison methods and documenting testing of the degree to which individuals believed they knew their treatment status.

Findings of direct comparisons of treatments.

Pharmacologic. Among 14 pharmacologic RCTs that made ten unique comparisons among pharmacologic agents, the majority did not report statistical tests that showed one drug to be superior to another. The exceptions were from three RCTs. Both oxybutynin and tolterodine in extended release form were superior to tolterodine in immediate release forms.^{83, 140} One trial reported oxybutynin ER superior to tolterodine ER for reducing voids per day.⁸⁴ Given heterogeneity of participant populations and study designs, this limited number of studies is insufficient for any drug to be considered definitively superior.

Procedures. For procedures, sacral neuromodulation was compared to wait list participants on medications. It is important to note that failure of prior medical management was a criteria for entry into the study; those waiting had worsening of many symptoms. This is in contradistinction to improvements noted in virtually all other comparison groups and likely reflects bias from the knowledge that they were awaiting what they considered more definitive treatment for severe OAB. No conclusions can be reached with this data and future research should address the differences in risk profile of sacral neuromodulation versus medications.¹²⁴

Behavioral. Seven studies compared behavioral to pharmacologic treatments. One study in this group reported significant reductions in incontinence episodes with a multi-component intervention;¹⁴³ no study found differences in reductions in voids per day. Participants in one study who were queried reported a preference for behavioral treatment over pharmacologic.¹⁴³ Adding behavioral treatments to pharmacologic treatments did not improve outcomes for incontinence episodes or voids per day above pharmacologic treatment alone.

KQ4: Modifiers of Treatment Outcomes

Methodologic issues for study of modifiers. Individual characteristics of participants in OAB studies were highly varied on characteristics that are plausibly related to treatment outcomes such as: age, menopausal status, prior treatment, prior refractory symptoms during treatment, prior incontinence or gynecologic surgeries, severity of OAB at baseline, and presence and type of incontinence. Higher quality studies reported on these characteristics and either found them comparable across trial arms or adjusted for baseline differences in their analyses. However, few studies indicated a priori goals of conducting sub-analyses in order to better understand treatment response. Among publications that did report on predictors of treatment response, the majority were under-powered to detect differences so the resulting claims of comparability are of limited value from an individual study. We found cross-cutting similarities for several of these characteristics such as age, severity, and prior treatment and have compiled those to present the limited picture that is coming into view. Overall this treatment literature is at an early stage of development in which the primary objective has been documenting superiority of the treatment to placebo. Population-based cohorts, such as the few provided by national and payor databases and larger clinical trials designed explicitly to more closely examine treatment response patterns and long-term effectiveness and tolerability will be required to have definitive information that can be used clinically with confidence.

Findings about modifiers of treatment outcomes. Advancing age was associated with more severe symptoms at baseline and with potentially observed attenuated treatment effects. Nonetheless, the majority of studies that reported on the effect of age in relationship to treatment outcomes found significant improvements in the older groups when active treatment was compared to placebo. No studies reported lack of efficacy among older participants. Older individuals do benefit from treatment. Race and ethnicity were not associated with outcomes in two analyses addressing this topic.

Presence of UUI, or urge-dominant mixed urinary incontinence, was not associated with treatment failure. Neither was severity of symptoms; in some cases those with the most severe symptoms, including more severe UUI, achieved the greatest treatment gains but were less likely to be symptom free.

Urodynamic findings do not provide consistent information about likelihood of treatment benefit or failure. A single study noted, among women who all had documented detrusor overactivity at enrollment, that those who had detrusor response to provocative maneuvers such as running water and washing hands were more likely to fail treatment. Further investigation of this finding from a small study would be intriguing as would examining the overall self-reported relationship between cues, urgency, UUI and treatment response. Another small study reported that anterior vaginal wall prolapse was a strong predictor of non-response to treatment. It is important to note that while these factors were associated with lower likelihood of treatment response, the majority of those treated with these characteristics did see improvement in symptoms.

Gender, while not a focus of this report, is important in interpreting findings with caution. Multiple research teams noted outcomes were not as favorable in men as in women. This likely reflects different pathophysiology behind the symptoms. Because some studies in this report included men (up to 25 percent of participants), results should be viewed in light of this potential bias.

KQ5: Costs of OAB Treatment

Methodologic considerations about cost studies. Studies that use administrative data are limited in their ability to adjust by clinical comorbidities and concomitant conditions, although they benefit from large enough numbers to provide a reasonable global estimate. Conversely studies that survey patients on their own health and health care may obtain more detailed information, but suffer from recall bias on the part of the respondents. None of the studies reports on their funding source; nor do they provide any information on investigator conflict of interest.

Findings about costs of treatment. Total direct health care costs for women with OAB in 2000 were estimated at $6.9 billion, of which $1.1 billion was for pharmacologic treatment and $550 million for surgical treatment. The rest was estimated for “consequences” costs, which would include things like falls, longer hospital LOS and skin conditions. Medication costs for OAB with the two most commonly used drugs (oxybutynin and tolterodine) range from $56 to $360 over a twelve month period for newly diagnosed patients. However, overall health care costs were highest for patients who take oxybutynin, relative to tolterodine in any formulation, with costs lowest for patients on tolterodine ER. No study adequately determines why the observed differences exist, in particular whether they are actually a reflection of the differences in the populations who are prescribed the various medications. The one study that measures baseline costs found that patients whose incident OAB prescription was for tolterodine ER had lower costs in the six months prior to prescription than did patients whose incident prescription was either oxybutynin ER or oxybutynin IR. None of the studies conducted a cost effectiveness analysis, although cost-utility analyses have been conducted in Europe (which would be difficult given the low effectiveness of any of the drugs, and short followup of almost all efficacy studies). Nor does any study specifically assess the effect of medication on peripheral outcomes of OAB such as falls.

In studies of adherence and persistence, most of which are conducted in managed care populations, fewer than half of patients ever refill their prescriptions for OAB medications. Average quit time is about a month; among those who persist, adherence is best among patients taking extended release versus immediate release formulations. Even in this group, adherence is low; the highest medication possession ratio noted in the studies we identified was about 36 percent.^272–275

State of the Literature

The study of OAB as a syndrome is entering its second decade. As is typical of advancing areas of research, publications based on case series are giving way to observational cohorts. Trials, beyond those required for FDA approval of indication for OAB, are appearing in the literature and health services researchers are investigating population-level factors such as cost of care and risk of rare but serious side effects of treatments.

The 2002 ICS standardization of terminology¹⁸⁰ was associated with a productive trend toward greater attention to and clarity of operational definitions in research. Documentation of inclusion and exclusion criteria, baseline characteristics, and change in symptom profiles have become more detailed and nuanced in the last five to seven years. Improved clarity about research definitions for conducting the study and analyzing data was the case even when authors departed from ICS definitions. Simultaneously important research gains have been made in crafting, refining, and validating questionnaire and interview instruments for classifying symptoms, assessing severity of symptoms, describing impact of OAB on quality of life, and measuring satisfaction with outcomes.

Concerning deficits. As a body of literature, a number of concerning deficits were common. Fewer than seven percent of included studies met criteria for good quality. For example, for clinical trials, this meant that publications lacked one or more of the following:

• Any description of randomization method
• Masking of participants and assessors to treatment assignment
• Description of participant and selection process sufficient to understand generalizability
• Details of intervention provided to subjects sufficient to replicate
• Followup of treatment effects for 12 weeks or longer
• Loss to followup less than 10 percent
• Drop out less than 10 percent
• Power calculations (preferably for two-sided tests)
• Use of appropriate statistical comparisons and tests
• Clear description of methods used to measure outcomes
• Description of validity or reliability of outcome measures for primary outcomes

Each of these criteria is fundamental to the conduct and reporting of research of sufficient quality to build knowledge and inform care. The treatment literature is currently hindered by critical flaws that must be eliminated. These include use of data from only those who completed the whole treatment course and not intention-to-treat analyses. Likewise authors frequently noted covariates that were associated with either baseline severity or outcomes and did not adjust for these factors in analysis or conduct stratified analyses. This was often the case because the size of the study would not support modeling or sub-analyses. Conclusions often over-reached findings in ways that in some instances were blatantly biased in favor of a newer treatment.

The large magnitude of placebo effect in OAB studies deserves special note. The fact of robust placebo response implies that uncontrolled studies will be notably biased. Indeed in this literature observational studies, with rare exceptions, overestimated treatment benefits when compared to trials. High quality trials and innovations in masking treatment group (especially for procedural and behavioral studies) will be essential to firmly establishing treatment effects.

Conflict of Interest. Trends in increasing transparency about sources of funding and potential conflicts of interest have been steadily positive over the past two decades (Table 32). However, a fundamental mismatch exists between the initial research needed to obtain regulatory approval and broader, longer term research needs to assess a wider range of questions about outcomes of care in typical practice settings. For a condition as common as OAB, the funding and conflict of interest picture that emerges suggests a research area that is urgently in need of additional sources of independent funding for the next wave of clinical effectiveness research.

Future Research Directions

Momentum in the direction of higher quality and more informative research will follow from attention to:

• Reporting greater information about key characteristics of populations studied in order to allow assessment of comparability of study populations and applicability of findings. This also facilitates understanding of candidate confounders and variations in findings across studies and study types.
• Conducting studies of sufficient size to conduct hypothesis testing or assess treatment effects. Small studies preclude meaningful descriptive analysis of modifiers and appropriate adjustment of confounders. Inclusion of small numbers of men in much larger studies was a recurrent example of a modifier of treatment outcomes that was noted or ignored without sufficient study size to meaningfully interpret.
• Continuing the expansion of standardized nomenclature and use of validated measures. The current literature is challenging to synthesize and interpret because outcomes measured are varied, not cross-cutting, and measured on different time scales (e.g., episodes per day, per week or per other unit of time). Use of validated measures is improving but measures are perhaps too numerous to help bring results into focus. Networks of researchers or those working in common areas would benefit from a rigorous, evidence-based consensus process to prioritize tools to use across studies to improve comparability of measures of outcomes like severity and quality of life.

Content priorities. Well-conducted larger studies with study populations that reflect the severity of conditions seen in both primary care and specialty practice settings are critical. It is imperative that the time window of followup be extended. The literature suggests that treatment effects can be achieved in the early weeks of treatment with pharmacologic, procedural, behavioral, and complementary and alternative therapies. However this does not mean that the duration of study should be truncated. To the degree that long-term efficacy and effectiveness is poorly documented, the resolution or worsening of side effects is poorly characterized, and satisfaction and continuation of treatment is not assessed over extended periods of time, the literature is not relevant for informing care for this chronic condition.

Each area that was a focus of this report would benefit from continued study of:

• Etiology and natural history of disease, risk factors and potential preventive measures
• Novel treatments
• Properly powered investigations of direct comparisons of existing treatments
• Longer term investigation of benefits and side effects of treatment
• Continued study of combinations of treatments
• Further investigation of complementary and alternative treatments
• Investigation of predictors of both good and poor treatment responses across treatment modalities.
• Selection of treatment options after prior treatment failure