Abstract
Background and Objective:
To identify phenotypes of urinary urgency (URG) and urgency incontinence (UUI) from expert causal knowledge summarized as directed acyclic graphs (DAGs), and to determine whether they are discrete classes or probabilistic positions on a continuum.
Methods:
DAGs representing clinical phenotypes were constructed from knowledge elicited from urologists and urogynecologists. Ninety LURN II participants served as phenotype anchors; the remaining 539 were assigned to the nearest centroid. Compactness was each participant's distance to their centroid relative to median between-centroid distance. Validity was the prognostic information phenotypes added for symptom severity at 3 and 12 months beyond baseline severity.
Key Findings and Limitations:
629 participants (336 females, 293 males) were assigned to eight phenotypes: polyuria (n=265), nocturnal polyuria (n=129), metabolic (n=96), affective (n=41), bowel and bladder dysfunction (n=22), pelvic organ prolapse (n=19, females), UTI (n=4, females), and benign prostatic enlargement (n=53, males). Continuous position was more informative than categorical labels: adjusted for baseline severity, distances added 9–18% of new prognostic information beyond labels, labels only 3–9% beyond distances, across strata. No participant met the 0.20 compactness threshold (smallest ratio 0.73 in women, 0.68 in men); baseline severity alone remained more prognostic than the phenotypes.
Conclusions and Clinical Implications:
URG/UUI patients can be described by eight DAG-derived phenotypes, but these are probabilistic positions, not discrete classes. Continuous representations retained modestly more prognostic information than labels, though both added little beyond baseline severity. These phenotypes are interpretable but not yet suitable for clinical use.
Keywords: Urinary urgency, urgency incontinence, overactive bladder, causal inference, phenotyping
Introduction
The Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN II) is pursuing phenotyping of men and women with urinary urgency (URG) and urinary urgency incontinence (UUI). This effort aimed to investigate whether a multidimensional array of clinical and biological features naturally cluster into previously undescribed phenotypes.1 The overall clinical goal was to determine whether such groups exist and, if so, whether they are distinct and prognostically informative enough to be clinically useful.
Unsupervised statistical learning methods, such as clustering, are increasingly reported in the biomedical literature as a means of identifying novel phenotypes for heterogeneous conditions, including diabetes mellitus, acute kidney injury, and chronic obstructive pulmonary disease.2-4 However, recent systematic reviews have suggested that while clustering is a valuable method for detecting associations, there is a need for stronger statistical support regarding the clinical interpretability, distinctiveness, and validation of identified phenotypes.2,3 Specifically, researchers have noted that because clustering is a purely mathematical technique not driven by clinical knowledge, spurious groupings may be identified, resulting in uninterpretable phenotypes that lack clinical validity or usefulness.2-4 Recently, some authors have suggested that unsupervised learning approaches should draw substantially from existing causal theory to support the data-driven findings of clustering.5
This study proposes capturing experts’ knowledge of the causal factors of URG and UUI in the form of diagrams. These diagrams are not merely representations of known causal relations; they are intended to be a graphical tool for representing our qualitative expert knowledge and a priori assumptions about the causal structure of a system under study. They present knowledge and assumptions in an intuitive way and enhance communication among investigators.6
The primary aim of this study was to extract clinical knowledge and create expert knowledge-based diagrams to represent common phenotypes among patients seeking care for URG and UUI. We then asked whether participants grouped by these diagrams (directed acyclic graphs, DAGs) formed distinct classes or occupied positions along a continuum.
To create a phenotype driven by clinical intuition and knowledge, we used the information in the DAGs to classify participants with URG and UUI into their most likely phenotype. We examined whether there were expected differences between the phenotypes and determined whether these phenotypes were sufficiently distinct or compact to be used as forced-choice phenotypic labels. We hypothesized that these phenotypes would demonstrate interpretability, some distinctiveness, and, from a validity standpoint, contribute some prognostic information about symptom severity under usual care, beyond that provided by baseline severity alone.
Materials and Methods
Institutional Review Board approval was received at each participating institution, and informed consent was obtained from all participants. LURN is a multicenter, NIH/NIDDK-sponsored research network comprising six research sites and a data coordinating center. LURN II Urinary Urgency Phenotyping study enrolled adults seeking specialty care in the management of bothersome urinary urgency between March 2021 and June 2024. Consenting participants completed self-reported demographic and symptoms measures, a 3-day fluid intake and voiding diary, and a food irritant diary. Participants underwent a physical exam and biosample collection at baseline before undertaking any new LUTS treatment during the 12-month study period; established LUTS treatment was permitted at enrollment and was being taken by 131 of 336 women (39%) and 201 of 293 men (69%).1,7 Details on study design and data collection have been published previously.1,7
The current study was conducted in three phases.8 First, expert-derived information was collected and summarized through semi-structured interviews with practicing providers who routinely treated men and women seeking treatment for URG and UUI. Second, this information was used to generate DAGs through an iterative structured process to represent a set of phenotypes. Third, data from the LURN cohort were used to group participants into their most likely phenotype, and the distinctiveness, compactness, and validation of these phenotypes were assessed. For this study, phenotype distinctiveness was based on R. A. Fisher’s original concept that clustered groups should be compact sets, meaning there should be no significant heterogeneity within each cluster.9 Thus, before concluding that phenotype labels adequately summarize a patient, we assessed whether membership probabilities were near 0 or 1. Validity was assessed as whether the phenotype label provided prognostic information about symptom severity under usual care beyond baseline symptom severity.
DAGs are tools for identifying and guiding study design and variable adjustment strategies for empirical analyses.10 This study used a modified process, drawing on expert knowledge, and the terminology described by Rodrigues et al.8,11,12 This process was based on the premise that building DAGs with domain experts, as opposed to relying solely on researchers’ understanding of the literature, can promote a more robust and realistic representation of the model under study.
The research team solicited participation from the Society of Urodynamics, Female Pelvic Medicine, and Urogenital Reconstruction (SUFU) and the American Urogynecologic Society (AUGS) for physician membership. We invited clinicians treating predominantly men or predominantly women, in community-based or tertiary academic settings. Eligible clinicians were US-licensed urologists or urogynecologists (fellowship-trained in Female Pelvic Medicine and Reconstructive Surgery if treating women), reported that at least 25% of patients seen in the prior 6 months had URG or UUI, and were active SUFU or AUGS members. Clinicians in training, with under 5 or over 30 years in practice, or who were current LURN investigators were excluded.
After obtaining verbal consent, a 1-hour virtual interview was conducted following a semi-structured interview guide (Appendix A). A main interviewer experienced in qualitative research led each discussion, with a second clinical expert asking follow-up questions in real time; the interviewer and participant were matched on whether they primarily treated men or women. Interviews were conducted online with the guide and live notes shared on screen, and audio recordings were transcribed.
The interviews began with a patient vignette depicting common URG/UUI cases. Twelve total vignettes (Appendix B) were randomly assigned among clinician participants. Participants listed their top three causes of the patient’s condition, along with any supporting or contradicting information. They were asked about biological, social, or environmental mechanisms and potential treatments to understand responses based on their clinical experience.
Several strategies minimized interviewer bias. Interviewers received pilot training to align with each other, emphasizing a ‘facilitator role' rather than a teaching or evaluative role. After initial interviews, a pause was allowed for recording, audit, and feedback. At least two urogynecologic or urologic interviewers participated to reduce expertise bias.
We sought diversity in clinician sex, training, experience, and patient population, anticipating more interviews with clinicians treating women since both urologists and urogynecologists do so. We planned 20 to 30 interviews to reach data saturation.13
We conducted a systematic content analysis of each interview, following standard procedures.14 Analysis and data collection were conducted iteratively, with each informing the other. After 3-5 interviews, analysis began. The completed interview grid was then translated into a preliminary DAG, which was reviewed by an expert to ensure that clinician feedback was accurately incorporated. Separate DAGs were created for each case scenario, stratified by participant sex.
The goals of the second phase were to refine the initial DAGs. The research team determined whether any arrows could be omitted as non-causal given the temporal order depicted, and confirmed no further common causes of risk factors and outcome required addition. Experts then finalized DAGs for common pathways and themes, stratified by sex. Final DAGs were reviewed with the remaining LURN investigators from all sites.
The refined DAGs from each clinical case scenario were grouped by shared causal pathway into eleven condition-level phenotypes: eight not sex-specific (polyuria, metabolic, affective, nocturnal polyuria15, bowel and bladder dysfunction16, nervous system, pain, and tumor/implant) and three sex-specific (urinary tract infection and pelvic organ prolapse for females, and benign prostatic enlargement17 for males). Three of the eleven (nervous system, pain, and tumor/implant) were exclusion criteria for the LURN II cohort and were not analyzed further. For each of the remaining eight phenotypes, the variables and relationships depicted in that phenotype’s refined causal diagram were translated into an operational anchor definition (Table 1), which was then used to identify anchor participants and, subsequently, to classify the full cohort, as described below.
Table 1. Anchor participant definitions: variables, cohort percentile, and threshold value, by sex.
Anchor participants were used to compute each phenotype’s centroid; all remaining participants were assigned to the phenotype whose centroid was nearest. Thresholds were set iteratively within the LURN II cohort so that, where attainable, approximately 5 to 10 participants of each sex met the definition for each phenotype. The percentile therefore varies across phenotypes and between sexes, and we report both the percentile and the corresponding value in the cohort’s units. When several variables are listed, all conditions were required. All anchor-defining variables are baseline measurements obtained before any new LUTS treatment.
| Females | ||||
|---|---|---|---|---|
| Phenotype | Anchor variable | Percentile | Threshold (cohort units) |
Anchors, n |
| Bowel and bladder dysfunction | PROMIS GI constipation T-score21 | 98th | > 70.2 | 6 |
| Metabolic | Body mass index, kg/m2 | 80th | > 37.0 | 8 |
| Hemoglobin A1c, % | 80th | > 6.3 | ||
| Wake frequency, times per night | criterion | 3–4 | ||
| Polyuria (excess fluid intake) | Void volume, 3-day average, oz | 85th | > 87.3 | 7 |
| 24-hour intake volume, L | criterion | > 3.0 | ||
| Frequency predominates urgency * | criterion | yes | ||
| Nocturnal polyuria | Void volume during sleep, % of 24-hour volume | criterion | ≥ 25% | 8 |
| 24-hour void volume, L | criterion | normal, 0.5–3.0 | ||
| CASUS nocturia score19 | 90th | > 12 | ||
| Affective | PROMIS anxiety T-score22 | 90th | > 63.6 | 9 |
| PROMIS depression T-score22 | 90th | > 62.3 | ||
| Catastrophizing score †‡ | 90th | > 9 | ||
| Pelvic organ prolapse | POPDI-6§ | 90th | > 41.7 | 5 |
| max(POP-Q Ba, Bp, C), cm | 90th | > 2 | ||
| Urinary tract infection | Primary or secondary clinical diagnosis | criterion | present | 2 |
| Males | ||||
| Phenotype | Anchor variable | Percentile | Threshold (cohort units) |
Anchors, n |
| Bowel and bladder dysfunction | PROMIS GI constipation T-score21 | 97th | > 65.5 | 6 |
| Metabolic | Body mass index, kg/m2 | 75th | > 32.8 | 7 |
| Hemoglobin A1c, % | 75th | > 6.4 | ||
| Wake frequency, times per night | criterion | 3–4 | ||
| Polyuria (excess fluid intake) | Void volume, 3-day average, oz | 80th | > 83.8 | 4 |
| 24-hour intake volume, L | criterion | > 3.0 | ||
| Frequency predominates urgency * | criterion | yes | ||
| Nocturnal polyuria | Void volume during sleep, % of 24-hour volume | criterion | ≥ 25% | 12 |
| 24-hour void volume, L | criterion | normal, 0.5–3.0 | ||
| CASUS nocturia score19 | 80th | > 9 | ||
| Affective | PROMIS anxiety T-score22 | 90th | > 60.9 | 9 |
| PROMIS depression T-score22 | 90th | > 60.3 | ||
| Catastrophizing score †‡ | 90th | > 7.5 | ||
| Benign prostatic enlargement | Prostate size, digital rectal exam grade | criterion | 60 g (6+) or 80 g (8+) | 7 |
| Straining during urination | criterion | “sometimes” or more often | ||
Naming. Phenotypes were named after the mechanism each anchor definition operationalizes, using standard International Continence Society terminology where available. Nocturnal polyuria follows the ICS nocturnal polyuria index, the proportion of the 24-hour voided volume produced during sleep.15 Benign prostatic enlargement rather than benign prostatic hyperplasia or benign prostatic obstruction, since prostate size on digital rectal examination and a voiding symptom were available but neither histology nor a pressure-flow study was performed.17 Bowel and bladder dysfunction follows the International Children’s Continence Society standardization, applied here to adults.16 Metabolic and affective are mechanistic descriptors for which no ICS terminology exists; they denote, respectively, the adiposity and glycemia captured by the first anchor definition and the anxiety, depression, and catastrophizing captured by the second.
Anchors are markers, not definitions. Each anchor definition is a deliberately extreme marker used to seed a phenotype centroid. It is not a definition of the phenotype, which is specified by the corresponding causal diagram. The bowel and bladder dysfunction anchor, for example, uses a constipation score as a marker for a phenotype whose diagram describes bowel–bladder interaction.
Prostate size. Prostate size was recorded by the examining clinician on the physical examination form (“Genitourinary male abnormal findings — prostate and rectal exam”), with ordinal response options of 20 g (2+), 40 g (4+), 60 g (6+) and 80 g (8+). It is a digital rectal examination estimate in approximate 20 g increments, not a measured volume; no imaging-based volumetry was obtained. The benign prostatic enlargement anchor corresponds to the 60 g (6+) or 80 g (8+) categories.
Frequency predominates urgency = low urgency (rarely or sometimes had a sudden need to rush to urinate) AND high frequency (8 or more voids during waking hours on a typical day in the last month) AND typically 20 or more urination episodes per day.
Catastrophizing was scored using four items from the Cognitive Emotion Regulation Questionnaire: “I often think that what I have experienced is much worse than what others have experienced”; “I keep thinking about how terrible it is what I have experienced”; “I often think that what I have experienced is the worst that can happen to a person”; “I continually think how horrible the situation has been.” Response options: almost never, sometimes, regularly, often, almost always. The target of approximately 5 to 10 anchors per phenotype was not attainable for two phenotypes: two women met the urinary tract infection definition and twelve men met the nocturnal polyuria definition. Unlike the other anchor variables, both are defined by a fixed clinical criterion (diagnosis present; void volume during sleep ≥25% of the 24-hour total) rather than a cohort percentile chosen to hit the 5-to-10 target, so their counts were not iteratively tunable in the same way and were accepted as they stood.
Garnefski N, Kraaij V, Spinhoven P. Negative life events, cognitive emotion regulation and emotional problems. Personal Individ Differ. 2001;30(8):1311–1327.
Barber MD, Walters MD, Bump RC. Short forms of two condition-specific quality-of-life questionnaires for women with pelvic floor disorders (PFDI-20 and PFIQ-7). Am J Obstet Gynecol. 2005;193(1):103–113.
Phenotypes represent non-exclusive causal mechanisms rather than mutually exclusive diagnoses, and a participant could plausibly meet criteria consistent with more than one. We distinguish the conceptual phenotype, the causal diagram itself, from the operational assignment, the data-derived label each participant received. Assignment proceeded in four steps. (1) The refined diagrams defined eight conceptual phenotypes. (2) Each diagram was translated into one deliberately extreme anchor definition (Table 1); anchor participants meeting it were used only to compute that phenotype’s centroid. Thresholds were set iteratively so that roughly five to ten participants of each sex met each definition, where attainable (Table 1). (3) Every remaining participant was assigned to the phenotype whose centroid was nearest in multidimensional space, following scaling, weighting, and distance methodology previously described for LURN subtyping analyses.1,18 Continuous variables were scaled using the distribution of an independent control sample (71 females, 55 males) who did not meet criteria for urinary urgency or urgency incontinence; these participants were used only for scaling and were not among the 629 phenotyped participants described below. (4) The resulting groups were then evaluated both as categorical labels and as continuous positions. Meeting an anchor definition was not required for membership: most members of a phenotype did not meet its anchor criteria. For example, a woman meeting no anchor definition was assigned to the nearest centroid across all measured dimensions, with a median probability of belonging to that phenotype near 0.2. Critically, the Table 1 criteria used objective, directly measured baseline variables only; none of the interview material used to construct the DAGs, including experts' ratings of likely treatment response (Appendix A), entered assignment. The prognostic-validity outcome (LURN SI-29 total score at 3 and 12 months) was likewise a directly measured, independently collected patient-reported outcome. Phenotype assignment and the validation outcome therefore shared no common data source. Postvoid residual volume was measured in all participants but was not used for anchoring, as it is a downstream consequence of several distinct mechanisms; bladder outlet obstruction was a cohort exclusion criterion.
The phenotypes were assessed in three ways: First, through descriptive pairwise comparisons of demographics, clinical and physical characteristics, and patient-reported outcomes, including the LURN Comprehensive Assessment of Self-Reported Urinary Symptoms (CASUS), LURN Symptom Index-29 (SI-29), and PROMIS Gastrointestinal and PROMIS Depression and Anxiety scores;19-22 Second, we assessed cluster compactness using the ratio , where (unitless) from participant to the centroid of their assigned phenotype and = median pairwise distance between all phenotype centers (computed separately by sex). Third, validation was assessed by quantifying whether distances or logits of probabilities to all phenotype centers added additional prognostic (i.e. predictive) information beyond the cluster labels using likelihood ratio and 1- adequacy index.23,24 Because several variables in these comparisons also define anchors (Table 1), differences on them are expected by construction and appear in Tables 2 and 3 as verification, not findings. Pairwise comparisons were descriptive and unadjusted for multiplicity; effect estimates and p-values appear in Tables 2 and 3.
Table 2.
Significant pairwise differences in female participant characteristics among phenotypes
| Characteristics | Metabolic (n = 53) |
Nocturnal polyuria (n = 38) |
Polyuria (n =189) |
Affective (n = 21) |
Bowel
and bladder dysfunction (n = 12) |
UTI (n = 4) |
Prolapse (n = 19) |
% Missing |
Pairwise p < 0.05 |
|---|---|---|---|---|---|---|---|---|---|
| Demographics | |||||||||
| Age, years | 62.5 (11.9) | 58.6 (13.4) | 60.5 (14.4) | 54.0 (12.7) | 54.1 (14.7) | 78.0 (11.6) | 51.4 (14.8) | 0% | 1v4, 1v6, 1v7, 2v6, 3v4, 3v6, 3v7, 4v6, 5v6, 6v7 |
| Case type | 0% | 1v3, 1v7, 2v3, 2v7, 3v4, 3v5, 4v7 | |||||||
| Urinary urgency | 1.9% | 0.0% | 36.5% | 4.8% | 0.0% | 25.0% | 31.6% | ||
| Urinary urgency with incontinence | 98.1% | 100.0% | 63.5% | 95.2% | 100.0% | 75.0% | 68.4% | ||
| Comorbidities | |||||||||
| Diabetes mellitus | 35.8% | 18.4% | 9.0% | 19.0% | 25.0% | 0.0% | 5.3% | 0% | 1v3, 1v7 |
| Depression | 25.0% | 55.3% | 32.6% | 66.7% | 66.7% | 0.0% | 15.8% | <1% | 1v2, 1v4, 1v5, 2v3, 2v7, 3v4, 3v5, 4v6, 4v7, 5v7 |
| Anxiety or panic disorder | 26.4% | 52.6% | 34.6% | 52.4% | 50.0% | 0.0% | 31.6% | <1% | 1v2, 1v4, 2v3 |
| Urinalysis | |||||||||
| Urine specific gravity | 1.01 (0.01) | 1.02 (0.01) | 1.01 (0.01) | 1.01 (0.01) | 1.02 (0.01) | 1.01 (0.01) | 1.01 (0.01) | 3% | 1v6, 2v3, 2v6, 5v6 |
| Labs | |||||||||
| Serum creatinine | 0.85 (0.16) | 0.85 (0.14) | 0.84 (0.44) | 0.83 (0.21) | 0.83 (0.10) | 0.88 (0.32) | 0.76 (0.13) | 9% | 2v3, 2v7 |
| HbA1c, % | 6.42 (1.22) | 6.14 (1.00) | 5.68 (0.66) | 6.03 (0.95) | 5.63 (0.66) | 6.20 (−) | 5.60 (0.49) | 51% | 1v3, 1v7 |
| HDL | 57.6 (17.7) | 60.9 (17.9) | 62.9 (15.6) | 49.2 (12.7) | 63.8 (16.4) | 63.0 (11.3) | 66.0 (17.5) | 12% | 1v3, 2v4, 3v4, 4v5, 4v7 |
| Triglycerides | 149 (78.3) | 137 (78.4) | 114 (58.7) | 171 (95.0) | 101 (48.3) | 146 (9.19) | 91.4 (32.4) | 13% | 1v3, 1v5, 1v7, 3v4, 4v5, 4v7 |
| Patient-reported nocturia | |||||||||
| Times typically wake up due to need to urinate | 2.25 (1.14) | 2.29 (1.29) | 1.77 (1.07) | 2.24 (1.26) | 2.33 (0.89) | 2.75 (0.50) | 1.84 (1.30) | 0% | 1v3, 2v3, 3v5, 3v6 |
| CASUS nocturia score (sum of 5 questions) | 9.62 (3.36) | 12.4 (3.47) | 6.58 (2.82) | 10.0 (4.11) | 10.5 (3.83) | 12.5 (2.89) | 5.74 (2.84) | 1% | 1v2, 1v3, 1v7, 2v3, 2v4, 2v7, 3v4, 3v5, 3v6, 4v7, 5v7, 6v7 |
| Bladder diaries | |||||||||
| Voids per day, 3-day avg | 8.04 (2.19) | 9.46 (3.47) | 9.20 (3.21) | 6.67 (4.02) | 10.1 (3.92) | 11.5 (6.84) | 10.7 (6.09) | 29% | 1v2, 1v3, 3v4, 4v7 |
| SI-29 summary variables | |||||||||
| SI-29 Incontinence Score (range 0 to 100) | 22.8 (11.5) | 51.6 (18.3) | 9.56 (10.8) | 42.5 (20.9) | 43.8 (22.1) | 32.3 (18.8) | 9.21 (10.5) | <1% | 1v2, 1v3, 1v4, 1v5, 1v7, 2v3, 2v7, 3v4, 3v5, 3v6, 4v7, 5v7, 6v7 |
| SI-29 Pain Score (range 0 to 100) | 8.96 (9.69) | 21.9 (21.0) | 10.3 (12.5) | 28.3 (29.0) | 45.3 (20.7) | 20.3 (12.9) | 25.7 (13.0) | 0% | 1v2, 1v4, 1v5, 1v7, 2v3, 2v5, 3v4, 3v5, 3v7, 4v5, 5v6, 5v7 |
| SI-29 Voiding Difficulty Score (range 0 to 100) | 19.2 (18.0) | 15.6 (14.0) | 18.4 (15.5) | 28.2 (16.1) | 30.0 (17.8) | 20.0 (5.77) | 33.3 (18.8) | 4% | 1v4, 1v5, 1v7, 2v4, 2v5, 2v7, 3v4, 3v5, 3v7 |
| SI-29 Urgency Score (range 0 to 100) | 61.6 (17.2) | 80.5 (14.4) | 39.6 (19.0) | 69.6 (21.6) | 75.0 (21.9) | 81.3 (12.5) | 39.0 (13.6) | 0% | 1v2, 1v3, 1v5, 1v6, 1v7, 2v3, 2v7, 3v4, 3v5, 3v6, 4v7, 5v7, 6v7 |
| SI-29 Nocturia Score (range 0 to 100) | 66.6 (28.1) | 73.7 (28.1) | 61.1 (28.8) | 65.3 (30.8) | 65.5 (28.2) | 85.7 (0.00) | 54.9 (34.1) | <1% | 2v3, 2v7, 3v6 |
| LURN SI-29 Total Score (range 0 to 100) | 32.4 (7.46) | 46.4 (8.24) | 24.9 (7.84) | 45.8 (16.9) | 49.6 (7.57) | 41.9 (3.48) | 31.9 (6.43) | 0% | 1v2, 1v3, 1v4, 1v5, 1v6, 2v3, 2v7, 3v4, 3v5, 3v6, 3v7, 4v7, 5v6, 5v7, 6v7 |
| E. PROMIS GI constipation 9a (past 7 days) | |||||||||
| Constipation T-Score | 52.4 (7.47) | 53.2 (7.32) | 51.8 (7.39) | 58.6 (9.30) | 61.9 (11.6) | 52.4 (8.64) | 52.5 (8.55) | <1% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v7, 5v7 |
| E1:How often did you pass very hard or lumpy stools? (If Never, go to #3) | 1.89 (0.82) | 1.47 (0.65) | 1.82 (1.02) | 2.25 (1.21) | 2.17 (1.19) | 2.00 (1.15) | 1.74 (1.10) | 1% | 1v2, 2v4, 2v5 |
| E2:How much did hard or lumpy stools bother you? | 2.59 (1.07) | 2.69 (1.01) | 2.46 (1.16) | 3.92 (1.31) | 3.13 (1.36) | 3.00 (1.41) | 2.88 (1.46) | 50% | 1v4, 2v4, 3v4 |
| E3:How often did you strain while trying to have bowel movements? (If Never, go to #6) | 2.49 (1.01) | 2.53 (1.11) | 2.45 (1.00) | 3.29 (1.27) | 3.58 (1.51) | 2.50 (1.29) | 2.53 (1.12) | 2% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 5v7 |
| E4:How much did you usually strain while trying to have a bowel movement? | 2.47 (1.01) | 2.68 (0.83) | 2.42 (0.93) | 3.21 (1.23) | 4.10 (0.99) | 2.67 (1.53) | 2.73 (1.10) | 16% | 1v4, 1v5, 2v5, 3v4, 3v5, 5v7 |
| E5:How much did straining during bowel movements bother you? | 2.33 (1.28) | 2.60 (1.19) | 2.28 (1.11) | 3.58 (1.50) | 3.90 (1.29) | 2.33 (1.53) | 2.60 (1.30) | 17% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v7, 5v7 |
| E6:How often did you feel pain in your rectum or anus while trying to have bowel movements? (If Never, go to #8) | 1.77 (1.03) | 1.92 (1.02) | 1.69 (0.86) | 2.55 (1.19) | 3.33 (1.50) | 1.50 (0.58) | 1.68 (0.95) | 3% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v7, 5v6, 5v7 |
| E7:At its worst, how would you rate the pain in your rectum or anus during bowel movements? | 2.61 (1.20) | 2.29 (0.85) | 2.11 (1.11) | 2.94 (0.93) | 3.60 (1.17) | 1.00 (0.00) | 2.63 (1.19) | 50% | 1v3, 1v5, 2v4, 2v5, 2v6, 3v4, 3v5, 4v6 |
| E8:How often after a bowel movement did you feel unfinished - that is, that you had not passed all your stool? | 2.75 (1.10) | 2.89 (1.09) | 2.46 (0.96) | 3.15 (1.39) | 3.50 (1.24) | 2.50 (1.29) | 2.58 (1.12) | 1% | 2v3, 3v4, 3v5 |
| E9:How often did you use your finger or toilet paper to get out a stool? | 1.62 (0.86) | 1.61 (1.05) | 1.41 (0.74) | 1.65 (0.99) | 2.33 (1.44) | 2.25 (1.26) | 1.42 (0.61) | <1% | 3v5, 3v6 |
| F. PROMIS GI diarrhea 6a (past 7 days) | |||||||||
| Diarrhea T-Score | 52.4 (9.50) | 56.1 (10.5) | 48.5 (8.06) | 54.2 (11.0) | 53.2 (10.6) | 55.6 (10.6) | 50.5 (9.64) | <1% | 1v3, 2v3, 3v4 |
| Diarrhea SE | 3.57 (2.41) | 3.07 (1.96) | 4.21 (2.43) | 3.60 (2.33) | 3.73 (2.52) | 2.93 (2.78) | 3.90 (2.38) | <1% | 2v3 |
| F1:How many days did you have loose or watery stools? (If No Days, go to #4) | 2.37 (1.43) | 2.42 (1.43) | 1.87 (1.16) | 1.84 (1.12) | 1.83 (1.11) | 2.00 (0.82) | 1.84 (1.21) | 4% | 1v3, 2v3 |
| F2:How much did having loose or watery stools interfere with your day-to-day activities? | 2.27 (1.31) | 3.08 (1.38) | 1.91 (1.07) | 2.89 (1.27) | 2.33 (1.21) | 3.00 (1.00) | 2.25 (1.04) | 51% | 1v2, 2v3, 3v4 |
| F3:How much did having loose or watery stools bother you? | 2.97 (1.35) | 3.54 (1.22) | 2.28 (1.15) | 3.00 (1.58) | 2.67 (1.21) | 3.67 (0.58) | 2.63 (1.51) | 52% | 1v3, 2v3, 3v6 |
| F4:How often did you feel like you needed to empty your bowels right away or else you would have an accident? (If Never, you are finished) | 2.00 (1.03) | 2.39 (1.26) | 1.63 (0.77) | 2.19 (1.12) | 1.67 (0.65) | 3.00 (1.41) | 1.68 (0.75) | 1% | 1v3, 2v3, 2v7, 3v4, 3v6 |
| F5:How much did feeling you needed to empty your bowels right away interfere with your day-to-day activities? | 2.58 (1.25) | 2.83 (1.34) | 2.09 (0.88) | 2.93 (1.39) | 3.14 (1.07) | 2.67 (0.58) | 2.45 (1.37) | 44% | 2v3, 3v4, 3v5 |
| F6:How much did feeling you needed to empty your bowels right away bother you? | 2.88 (1.22) | 3.14 (1.46) | 2.56 (1.04) | 3.73 (1.33) | 3.86 (0.90) | 3.67 (0.58) | 3.10 (1.45) | 44% | 1v4, 3v4, 3v5 |
| G. PROMIS GI bowel incontinence 4a (past 7 days) | |||||||||
| GI Bowel Incontinence Score | 5.37 (2.10) | 6.97 (3.56) | 4.79 (1.53) | 6.86 (3.40) | 4.83 (1.75) | 7.75 (2.87) | 4.89 (2.00) | 3% | 1v2, 2v3, 2v5, 2v7, 3v4, 3v6, 4v5, 4v7, 6v7 |
| G1:How often did you have bowel incontinence - that is, have an accident because you could not make it to the bathroom in time? | 1.23 (0.47) | 1.68 (1.03) | 1.11 (0.35) | 1.52 (0.87) | 1.08 (0.29) | 1.50 (0.58) | 1.11 (0.46) | <1% | 1v2, 1v3, 2v3, 2v5, 2v7, 3v4, 3v6, 4v7, 6v7 |
| G2:How often did you soil or dirty your underwear before getting to a bathroom? | 1.37 (0.66) | 1.63 (0.91) | 1.14 (0.40) | 1.62 (0.86) | 1.17 (0.39) | 1.75 (0.50) | 1.11 (0.46) | 1% | 1v3, 1v7, 2v3, 2v7, 3v4, 3v6, 4v7, 5v6, 6v7 |
| G3:How often did you leak stool or soil your underwear? | 1.37 (0.72) | 1.76 (1.00) | 1.20 (0.53) | 1.62 (0.92) | 1.25 (0.45) | 1.75 (0.96) | 1.17 (0.51) | 2% | 1v2, 2v3, 2v7, 3v4 |
| G4:How often did you think you were going to pass gas, but stool or liquid came out instead? | 1.44 (0.67) | 1.92 (0.94) | 1.33 (0.60) | 2.10 (1.09) | 1.33 (0.78) | 2.75 (1.26) | 1.50 (0.86) | 1% | 1v2, 1v4, 1v6, 2v3, 2v5, 3v4, 3v6, 4v5, 5v6, 6v7 |
| Q: PROMIS depression short form 8a (past 7 days) | |||||||||
| Depression T-Score | 47.7 (8.11) | 49.4 (9.71) | 48.4 (8.55) | 64.5 (8.38) | 53.3 (6.94) | 45.7 (10.7) | 49.3 (8.15) | 0% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v5, 4v6, 4v7 |
| Depression SE | 3.33 (1.65) | 3.23 (1.83) | 3.22 (1.65) | 1.95 (0.35) | 2.28 (1.17) | 4.08 (1.95) | 3.01 (1.51) | 0% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v6, 4v7, 5v7 |
| Q1:I felt worthless | 1.45 (0.85) | 1.45 (0.83) | 1.41 (0.75) | 2.71 (1.49) | 1.67 (0.78) | 1.25 (0.50) | 1.26 (0.73) | <1% | 1v4, 2v4, 3v4, 4v7 |
| Q2:I felt helpless | 1.38 (0.79) | 1.61 (0.97) | 1.48 (0.80) | 3.33 (1.20) | 1.75 (0.87) | 1.50 (0.58) | 1.32 (0.58) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6, 4v7 |
| Q3:I felt depressed | 1.77 (0.98) | 2.08 (1.10) | 1.86 (1.04) | 3.76 (0.94) | 2.82 (0.87) | 1.50 (1.00) | 1.78 (1.00) | <1% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v5, 4v6, 4v7, 5v7 |
| Q4:I felt hopeless | 1.30 (0.72) | 1.63 (0.97) | 1.44 (0.87) | 3.20 (1.28) | 1.58 (0.79) | 1.67 (1.15) | 1.74 (1.05) | 1% | 1v2, 1v4, 1v7, 2v4, 3v4, 4v5, 4v7 |
| Q5:I felt like a failure | 1.47 (0.82) | 1.71 (1.11) | 1.47 (0.90) | 3.52 (1.36) | 1.83 (0.83) | 1.25 (0.50) | 1.58 (1.07) | 0% | 1v4, 2v4, 3v4, 3v5, 4v5, 4v6, 4v7 |
| Q6:I felt unhappy | 2.06 (0.93) | 2.13 (1.04) | 2.01 (0.94) | 3.45 (0.94) | 2.33 (0.89) | 1.75 (1.50) | 2.26 (0.99) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6, 4v7 |
| Q7:I felt that I had nothing to look forward to | 1.45 (0.75) | 1.59 (0.96) | 1.51 (0.93) | 3.05 (1.53) | 1.83 (0.83) | 1.75 (1.50) | 1.47 (0.90) | <1% | 1v4, 2v4, 3v4, 4v5, 4v7 |
| Q8:I felt that nothing could cheer me up | 1.28 (0.57) | 1.47 (0.76) | 1.41 (0.75) | 2.86 (1.39) | 1.75 (1.06) | 1.50 (1.00) | 1.47 (0.96) | 0% | 1v4, 2v4, 3v4, 4v5, 4v7 |
| R. PROMIS anxiety short form 8a (past 7 days) | |||||||||
| Anxiety T-Score | 48.8 (8.21) | 51.2 (9.62) | 49.4 (9.22) | 63.8 (9.67) | 53.9 (6.86) | 45.8 (3.06) | 52.8 (8.26) | 0% | 1v4, 1v7, 2v4, 3v4, 4v5, 4v6, 4v7 |
| Anxiety SE | 3.01 (1.30) | 2.75 (1.18) | 3.04 (1.39) | 2.15 (0.32) | 2.22 (0.48) | 2.83 (0.55) | 2.58 (1.32) | 0% | 1v4, 1v5, 1v7, 3v4, 3v5, 3v7, 4v6 |
| R1:I felt fearful | 1.33 (0.62) | 1.55 (0.86) | 1.42 (0.72) | 2.38 (1.53) | 1.45 (0.69) | 1.00 (0.00) | 1.58 (0.77) | 1% | 1v4, 3v4 |
| R2:I found it hard to focus on anything other than my anxiety | 1.31 (0.65) | 1.66 (1.07) | 1.49 (0.83) | 3.10 (1.26) | 1.67 (0.89) | 1.25 (0.50) | 1.53 (0.90) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6, 4v7 |
| R3:My worries overwhelmed me | 1.64 (0.83) | 1.89 (1.03) | 1.65 (0.95) | 3.29 (1.19) | 2.25 (1.14) | 1.00 (0.00) | 2.00 (0.82) | 0% | 1v4, 2v4, 3v4, 3v5, 3v7, 4v5, 4v6, 4v7, 5v6, 6v7 |
| R4:I felt uneasy | 1.64 (0.86) | 1.76 (1.02) | 1.75 (0.93) | 3.14 (1.31) | 1.92 (0.90) | 1.00 (0.00) | 2.21 (0.79) | <1% | 1v4, 1v7, 2v4, 2v7, 3v4, 3v7, 4v5, 4v6, 4v7, 6v7 |
| R5:I felt nervous | 1.79 (0.91) | 2.05 (1.01) | 1.90 (1.02) | 3.19 (1.33) | 2.17 (0.94) | 1.50 (0.58) | 2.16 (0.83) | 1% | 1v4, 2v4, 3v4, 4v5, 4v6, 4v7 |
| R6:I felt like I needed help for my anxiety | 1.50 (0.90) | 1.66 (1.05) | 1.59 (0.97) | 3.15 (1.09) | 2.00 (0.74) | 1.50 (1.00) | 1.83 (0.86) | 2% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v5, 4v6, 4v7 |
| R7:I felt anxious | 1.81 (0.91) | 2.18 (1.11) | 1.95 (1.03) | 3.14 (1.15) | 2.17 (0.94) | 1.50 (0.58) | 2.37 (1.01) | <1% | 1v4, 1v7, 2v4, 3v4, 4v5, 4v6, 4v7 |
| R8:I felt tense | 2.21 (1.04) | 2.18 (1.16) | 2.12 (1.09) | 3.57 (1.25) | 2.50 (0.80) | 1.50 (0.58) | 2.53 (1.07) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6, 4v7, 5v6 |
| T. PROMIS sleep disturbance short form 8a (past 7 days) | |||||||||
| Sleep Disturbance T-Score | 52.3 (4.32) | 52.7 (4.34) | 52.7 (3.52) | 55.9 (4.88) | 54.7 (3.34) | 50.6 (3.88) | 51.2 (5.14) | 0% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v6, 4v7 |
| Sleep Disturbance SE | 2.85 (0.52) | 3.07 (0.64) | 2.79 (0.46) | 3.44 (0.99) | 2.85 (0.67) | 2.53 (0.26) | 2.96 (0.56) | 0% | 1v4, 2v3, 3v4 |
| T1:My sleep quality was | 3.17 (1.03) | 3.37 (1.15) | 2.71 (1.07) | 3.95 (1.02) | 3.25 (0.97) | 3.50 (0.58) | 3.21 (0.98) | 0% | 1v3, 1v4, 2v3, 3v4, 4v7 |
| T2:My sleep was refreshing | 3.36 (1.18) | 3.63 (1.17) | 2.93 (1.08) | 4.20 (1.01) | 3.50 (0.80) | 3.25 (0.96) | 3.68 (1.20) | <1% | 1v3, 1v4, 2v3, 3v4, 3v7, 4v5 |
| T3:I had a problem with my sleep | 3.25 (1.22) | 3.34 (1.32) | 2.66 (1.18) | 3.81 (1.25) | 3.25 (1.06) | 3.50 (1.00) | 3.26 (1.24) | 0% | 1v3, 2v3, 3v4, 3v7 |
| T4:I had difficulty falling asleep | 2.26 (1.26) | 2.39 (1.44) | 2.14 (1.13) | 3.40 (1.31) | 2.83 (1.19) | 1.75 (0.96) | 2.21 (1.13) | <1% | 1v4, 2v4, 3v4, 3v5, 4v6, 4v7 |
| T5:My sleep was restless | 2.62 (1.11) | 2.89 (1.23) | 2.32 (1.16) | 3.71 (1.35) | 3.00 (1.21) | 2.00 (0.82) | 2.58 (1.17) | 0% | 1v4, 2v3, 2v4, 3v4, 3v5, 4v6, 4v7 |
| T6:I tried hard to get to sleep | 2.21 (1.19) | 2.58 (1.52) | 2.13 (1.15) | 3.62 (1.40) | 2.92 (1.38) | 2.25 (0.96) | 2.32 (1.34) | 1% | 1v4, 2v4, 3v4, 3v5, 4v7 |
| T7:I was worried about not being able to fall asleep | 1.88 (1.18) | 2.00 (1.34) | 1.80 (1.03) | 3.29 (1.31) | 2.58 (1.08) | 1.50 (0.58) | 1.79 (1.18) | <1% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v6, 4v7, 5v7 |
| T8:I was satisfied with my sleep | 3.38 (1.24) | 3.76 (1.24) | 2.99 (1.23) | 4.29 (1.15) | 3.50 (1.17) | 3.50 (0.58) | 3.84 (1.17) | 0% | 1v3, 1v4, 2v3, 3v4, 3v7, 4v5 |
| X. PROMIS physical function short form 10b | |||||||||
| X1:Are you able to do chores such as vacuuming or yard work? | 3.94 (1.24) | 3.89 (1.18) | 4.37 (0.98) | 3.43 (1.33) | 4.08 (1.38) | 4.00 (2.00) | 4.68 (0.58) | <1% | 1v3, 1v7, 2v3, 2v7, 3v4, 4v7 |
| X2:Are you able to get in and out of a car? | 4.46 (0.75) | 4.47 (0.83) | 4.85 (0.44) | 4.24 (1.09) | 4.50 (1.00) | 4.50 (1.00) | 4.84 (0.37) | <1% | 1v3, 2v3, 3v4 |
| X3:Are you able to go up and down stairs at a normal pace? | 3.85 (1.32) | 3.71 (1.21) | 4.47 (0.83) | 3.05 (1.16) | 4.00 (1.41) | 3.50 (0.58) | 4.47 (0.70) | <1% | 1v3, 1v4, 2v3, 2v4, 2v7, 3v4, 3v6, 4v5, 4v7, 6v7 |
| X4:Are you able to run errands and shop? | 4.27 (1.15) | 4.16 (0.99) | 4.72 (0.65) | 3.71 (1.31) | 4.25 (1.14) | 5.00 (0.00) | 4.89 (0.32) | 1% | 1v3, 1v7, 2v3, 2v7, 3v4, 3v5, 4v7, 5v7 |
| X5:Are you able to bend down and pick up clothing from the floor? | 4.35 (1.03) | 4.11 (1.16) | 4.65 (0.73) | 3.29 (1.38) | 4.45 (1.04) | 3.25 (2.06) | 4.84 (0.37) | 1% | 1v3, 1v4, 2v3, 2v4, 2v7, 3v4, 4v5, 4v7 |
| X6:Are you able to lift 10 pounds (5 kg) above your shoulder? | 3.87 (1.39) | 3.84 (1.28) | 4.24 (1.14) | 2.90 (1.29) | 3.64 (1.63) | 3.75 (1.89) | 4.68 (0.48) | 1% | 1v4, 1v7, 2v4, 2v7, 3v4, 4v7 |
| X7:Does your health now limit you in doing vigorous activities, such as running, lifting heavy objects, participating in strenuous sports? | 2.58 (1.37) | 2.53 (1.47) | 3.41 (1.38) | 2.19 (1.33) | 3.08 (1.51) | 2.25 (1.89) | 2.84 (0.83) | 1% | 1v3, 2v3, 3v4, 4v7 |
| X8:Does your health now limit you in bathing or dressing yourself? | 4.78 (0.55) | 4.58 (0.79) | 4.87 (0.43) | 4.19 (1.03) | 4.58 (1.00) | 4.50 (1.00) | 4.89 (0.32) | 1% | 1v4, 2v3, 3v4, 4v7 |
| X9:Does your health now limit you in putting a trash bag outside? | 4.51 (0.95) | 4.55 (0.76) | 4.80 (0.63) | 4.14 (1.20) | 4.45 (1.29) | 5.00 (0.00) | 5.00 (0.00) | 1% | 1v3, 1v7, 2v3, 2v7, 3v4, 4v7 |
| X10:Does your health now limit you in doing moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf? | 3.92 (1.41) | 3.84 (1.31) | 4.42 (1.02) | 3.05 (1.36) | 4.33 (1.37) | 4.00 (2.00) | 4.79 (0.54) | <1% | 1v3, 1v4, 1v7, 2v3, 2v4, 2v7, 3v4, 4v5, 4v7 |
| J. Physical Exam Findings | |||||||||
| J2:Systolic Blood Pressure | 136 (20.1) | 130 (18.3) | 128 (18.2) | 127 (16.7) | 134 (16.8) | 130 (13.9) | 117 (21.7) | <1% | 1v3, 1v7, 2v7, 3v7, 5v7 |
| J3:Diastolic Blood Pressure | 78.4 (8.00) | 75.8 (10.2) | 74.7 (10.1) | 76.4 (11.3) | 79.5 (10.7) | 74.8 (6.95) | 72.4 (12.9) | <1% | 1v3 |
| J4:Weight, kg | 91.4 (21.6) | 89.8 (20.3) | 74.8 (18.5) | 89.3 (24.1) | 95.2 (20.4) | 73.1 (3.73) | 76.1 (18.4) | 1% | 1v3, 1v6, 1v7, 2v3, 2v7, 3v4, 3v5, 5v6, 5v7 |
| Body Mass Index | 34.0 (8.98) | 33.7 (7.48) | 28.0 (6.87) | 33.5 (7.58) | 33.8 (5.09) | 26.2 (2.07) | 28.8 (7.72) | 1% | 1v3, 1v6, 1v7, 2v3, 2v6, 2v7, 3v4, 3v5, 4v6, 4v7, 5v6, 5v7 |
| J6:Waist measurement | 57.6 (32.5) | 52.9 (24.5) | 49.5 (26.5) | 52.5 (25.7) | 61.5 (35.7) | 34.7 (4.93) | 52.2 (38.2) | 21% | 1v3, 2v3, 2v6 |
| J7:Ambulatory aids | 9.4% | 2.6% | 3.3% | 23.8% | 25.0% | 25.0% | 5.3% | 2% | 2v4, 2v5, 3v4, 3v5 |
| J18:Pelvic Organ Prolapse Quantification System (POP-Q) - Overall Stage | 18.0 (37.3) | 18.9 (38.4) | 13.5 (32.8) | 14.5 (35.3) | 8.92 (28.4) | 49.3 (56.3) | 21.9 (40.9) | <1% | 4v7 |
| J19:POP-QAa result | −1.79 (1.44) | −1.84 (1.49) | −1.96 (1.38) | −2.25 (1.13) | −2.35 (0.75) | −2.50 (0.50) | −0.90 (2.16) | 16% | 3v7, 4v7 |
| J20:POP-QBa result | −1.80 (1.48) | −2.02 (1.33) | −1.89 (1.68) | −1.88 (1.71) | −2.45 (0.76) | −2.83 (0.29) | −0.63 (2.38) | 16% | 2v7, 3v7, 5v7 |
| J21:POP-QC result | −5.88 (3.20) | −4.79 (4.73) | −6.21 (3.09) | −5.86 (4.50) | −6.70 (2.45) | −6.75 (0.35) | −4.30 (3.76) | 21% | 3v7 |
| J22:POP-QGH result | 3.48 (1.23) | 3.30 (0.95) | 3.01 (1.16) | 3.44 (1.01) | 3.40 (1.26) | 2.50 (0.50) | 3.60 (1.54) | 17% | 1v3 |
| J23:POP-QPB result | 3.21 (0.93) | 3.15 (0.98) | 3.15 (0.88) | 2.81 (1.00) | 2.35 (1.76) | 3.83 (1.04) | 3.03 (0.74) | 17% | 3v5 |
| J24:POP-QTVL result | 8.76 (2.01) | 8.58 (1.59) | 8.87 (1.40) | 8.44 (1.21) | 8.20 (1.75) | 7.00 (0.00) | 8.50 (1.94) | 17% | 3v6 |
Results shown as mean (standard deviation) or prevalence; pairwise p-values calculated via Wilcoxon rank sum test for continuous variables or chi-square test for categorical variables
Phenotypes are numbered in column order: 1 = metabolic, 2 = nocturnal polyuria, 3 = polyuria, 4 = affective, 5 = bowel and bladder dysfunction, 6 = urinary tract infection, 7 = pelvic organ prolapse.
Tables 2 and 3 report summary scores. Item-level CASUS comparisons are described in the Results but are not tabulated.
Table 3.
Significant pairwise differences in male participant characteristics among phenotypes
| Characteristics | Metabolic (n = 43) |
Nocturnal polyuria (n = 91) |
Polyuria (n =76) |
Affective (n = 20) |
Bowel
and bladder dysfunction (n = 10) |
BPE (n = 53) |
% Missing |
Pairwise p < 0.05 |
|---|---|---|---|---|---|---|---|---|
| Demographics | ||||||||
| Age, years | 71.0 (8.89) | 66.1 (12.7) | 65.6 (10.5) | 57.1 (11.2) | 59.7 (11.4) | 62.4 (12.8) | 0% | 1v2, 1v3, 1v4, 1v5, 1v6, 2v4, 2v6, 3v4, 4v6 |
| Case type | 0% | 1v3, 1v6, 2v3, 2v6, 3v4, 3v5, 4v6, 5v6 | ||||||
| Urinary urgency | 16.3% | 27.5% | 77.6% | 15.0% | 20.0% | 66.0% | ||
| Urinary urgency with incontinence | 83.7% | 72.5% | 22.4% | 85.0% | 80.0% | 34.0% | ||
| Comorbidities | ||||||||
| Diabetes mellitus | 34.9% | 17.6% | 10.5% | 35.0% | 30.0% | 17.0% | 0% | 1v2, 1v3, 1v6, 3v4 |
| Congestive heart failure | 9.3% | 5.6% | 1.4% | 10.0% | 10.0% | 0.0% | 1% | 1v6 |
| Depression | 19.0% | 26.1% | 14.9% | 75.0% | 11.1% | 13.7% | 3% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Anxiety or panic disorder | 14.3% | 19.5% | 13.9% | 57.9% | 10.0% | 18.0% | 4% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Urinalysis | ||||||||
| Blood/Hemoglobin/Red Blood Cells level | 0.02 (0.16) | 0.09 (0.36) | 0.05 (0.23) | 0.30 (0.66) | 0.00 (0.00) | 0.10 (0.57) | 3% | 1v4, 3v4, 4v6 |
| White Blood Cells positive | 2.4% | 3.4% | 2.7% | 0.0% | 0.0% | 0.0% | 3% | |
| Glucose positive | 17.1% | 3.4% | 5.4% | 5.0% | 20.0% | 5.8% | 3% | 1v2 |
| Urine specific gravity | 1.02 (0.01) | 1.01 (0.01) | 1.01 (0.01) | 1.02 (0.01) | 1.02 (0.01) | 1.01 (0.01) | 3% | 3v4, 4v6 |
| Labs | ||||||||
| Total Cholesterol | 150 (27.5) | 161 (34.5) | 165 (35.4) | 166 (35.8) | 199 (69.4) | 168 (41.0) | 5% | 1v3, 1v5 |
| LDL | 80.9 (22.5) | 89.2 (29.9) | 95.0 (31.2) | 85.8 (37.1) | 115 (49.1) | 92.4 (32.8) | 6% | 1v3, 1v5, 2v5 |
| Triglycerides | 124 (67.9) | 110 (63.9) | 108 (63.2) | 155 (108) | 211 (227) | 136 (72.2) | 5% | 2v6, 3v4, 3v6 |
| Patient-reported nocturia | ||||||||
| Times typically wake up due to need to urinate | 2.47 (1.26) | 2.37 (1.16) | 2.05 (1.06) | 2.50 (1.15) | 1.80 (1.48) | 2.15 (1.03) | 0% | 2v3 |
| CASUS nocturia score (sum of 5 questions) | 9.10 (2.89) | 7.55 (2.88) | 6.48 (2.43) | 8.85 (3.33) | 6.70 (3.89) | 7.29 (2.38) | 2% | 1v2, 1v3, 1v6, 2v3, 3v4, 3v6 |
| Bladder diaries | ||||||||
| Intake volume, 3 day avg oz | 62.7 (20.8) | 59.0 (23.9) | 69.0 (24.2) | 58.1 (27.7) | 79.1 (37.4) | 68.0 (23.1) | 32% | 2v3 |
| Void volume, 3 day avg oz | 62.9 (23.5) | 56.8 (25.9) | 66.9 (26.1) | 60.6 (37.0) | 92.3 (7.58) | 63.9 (22.7) | 31% | 1v5, 2v3, 2v5, 5v6 |
| Voids per day, 3 day avg | 10.0 (4.25) | 9.07 (2.34) | 9.27 (2.96) | 7.15 (3.03) | 12.8 (3.85) | 10.1 (2.84) | 31% | 1v4, 2v4, 3v4, 4v6 |
| SI-29 summary variables | ||||||||
| SI-29 Incontinence Score (range 0 to 100) | 12.4 (10.2) | 5.74 (6.69) | 1.51 (3.74) | 11.6 (11.9) | 6.25 (6.29) | 2.12 (4.74) | <1% | 1v2, 1v3, 1v6, 2v3, 2v6, 3v4, 3v5, 4v6, 5v6 |
| SI-29 Pain Score (range 0 to 100) | 6.83 (8.61) | 7.81 (9.78) | 6.00 (9.27) | 28.4 (17.2) | 18.1 (14.3) | 22.3 (15.6) | 0% | 1v4, 1v5, 1v6, 2v4, 2v5, 2v6, 3v4, 3v5, 3v6 |
| SI-29 Voiding Difficulty Score (range 0 to 100) | 25.8 (18.4) | 28.8 (16.8) | 20.3 (15.7) | 42.1 (18.9) | 53.1 (17.7) | 40.6 (21.8) | 9% | 1v4, 1v5, 1v6, 2v3, 2v4, 2v5, 2v6, 3v4, 3v5, 3v6 |
| SI-29 Urgency Score (range 0 to 100) | 62.2 (20.9) | 47.2 (18.2) | 27.7 (13.5) | 58.7 (27.0) | 60.8 (17.6) | 38.5 (18.5) | 0% | 1v2, 1v3, 1v6, 2v3, 2v5, 2v6, 3v4, 3v5, 3v6, 4v6, 5v6 |
| SI-29 Nocturia Score (range 0 to 100) | 77.4 (22.1) | 69.0 (28.8) | 67.3 (27.1) | 71.4 (25.4) | 55.7 (41.2) | 70.4 (27.3) | <1% | 1v3 |
| LURN SI-29 Total Score (range 0 to 100) | 32.7 (6.52) | 28.1 (6.49) | 21.0 (6.33) | 41.3 (10.0) | 38.2 (8.83) | 32.5 (8.81) | 0% | 1v2, 1v3, 1v4, 1v5, 2v3, 2v4, 2v5, 2v6, 3v4, 3v5, 3v6, 4v6 |
| E. PROMIS GI constipation 9a (past 7 days) | ||||||||
| Constipation T-Score | 52.6 (7.14) | 50.7 (7.08) | 49.1 (7.22) | 58.1 (7.09) | 62.3 (5.38) | 50.7 (6.51) | <1% | 1v3, 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| Constipation SE | 2.50 (1.14) | 2.65 (1.46) | 2.83 (1.44) | 2.13 (0.95) | 1.93 (0.23) | 2.51 (1.25) | <1% | 3v4, 3v5 |
| E1:How often did you pass very hard or lumpy stools? (If Never, go to #3) | 1.83 (0.99) | 1.74 (0.96) | 1.47 (0.72) | 2.21 (1.40) | 3.00 (1.25) | 1.61 (0.90) | 3% | 1v5, 2v5, 3v4, 3v5, 5v6 |
| E2:How much did hard or lumpy stools bother you? | 2.65 (0.81) | 2.41 (0.91) | 2.08 (1.09) | 3.11 (1.27) | 3.89 (1.27) | 2.27 (1.08) | 57% | 1v3, 1v5, 2v5, 3v4, 3v5, 5v6 |
| E3:How often did you strain while trying to have bowel movements? (If Never, go to #6) | 2.47 (1.03) | 2.44 (0.93) | 2.15 (0.88) | 3.15 (0.81) | 3.80 (0.79) | 2.38 (0.89) | 2% | 1v4, 1v5, 2v3, 2v4, 2v5, 3v4, 3v5, 4v5, 4v6, 5v6 |
| E4:How much did you usually strain while trying to have a bowel movement? | 2.61 (1.03) | 2.41 (0.89) | 2.24 (0.88) | 3.05 (0.97) | 3.30 (1.25) | 2.40 (0.73) | 19% | 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| E5:How much did straining during bowel movements bother you? | 2.62 (1.04) | 2.04 (0.91) | 2.00 (1.08) | 3.21 (1.18) | 3.90 (1.20) | 2.00 (0.87) | 18% | 1v2, 1v3, 1v5, 1v6, 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| E6:How often did you feel pain in your rectum or anus while trying to have bowel movements? (If Never, go to #8) | 1.56 (0.73) | 1.57 (0.74) | 1.49 (0.69) | 2.55 (1.00) | 3.00 (1.12) | 1.58 (0.70) | 2% | 1v4, 1v5, 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| E7:At its worst, how would you rate the pain in your rectum or anus during bowel movements? | 2.00 (0.84) | 1.82 (0.76) | 1.84 (0.78) | 2.81 (1.11) | 2.60 (1.35) | 1.77 (0.76) | 52% | 1v4, 2v4, 3v4, 4v6 |
| E8:How often after a bowel movement did you feel unfinished - that is, that you had not passed all your stool? | 2.53 (0.98) | 2.14 (0.91) | 2.03 (0.99) | 3.00 (1.12) | 3.20 (1.03) | 2.27 (0.91) | <1% | 1v2, 1v3, 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| E9:How often did you use your finger or toilet paper to get out a stool? | 1.63 (0.93) | 1.38 (0.66) | 1.28 (0.69) | 1.95 (0.94) | 1.50 (0.85) | 1.21 (0.46) | <1% | 1v3, 1v6, 2v4, 3v4, 4v6 |
| F. PROMIS GI diarrhea 6a (past 7 days) | ||||||||
| Diarrhea T-Score | 51.5 (8.57) | 46.0 (7.06) | 44.9 (6.50) | 56.0 (9.49) | 48.4 (7.67) | 48.4 (7.49) | <1% | 1v2, 1v3, 2v4, 3v4, 3v6, 4v5, 4v6 |
| Diarrhea SE | 3.49 (2.26) | 4.85 (2.42) | 5.18 (2.38) | 2.95 (2.16) | 3.93 (2.74) | 4.06 (2.38) | <1% | 1v2, 1v3, 2v4, 2v6, 3v4, 3v6, 4v6 |
| F1:How many days did you have loose or watery stools? (If No Days, go to #4) | 2.07 (1.24) | 1.61 (1.03) | 1.45 (0.93) | 2.42 (1.26) | 1.70 (0.82) | 1.96 (1.22) | 3% | 1v2, 1v3, 2v4, 3v4, 3v6 |
| F2:How much did having loose or watery stools interfere with your day-to-day activities? | 2.13 (0.97) | 1.47 (0.68) | 1.63 (0.83) | 2.77 (1.30) | 1.60 (0.89) | 1.63 (0.71) | 61% | 1v2, 2v4, 3v4, 4v6 |
| F3:How much did having loose or watery stools bother you? | 2.43 (0.95) | 1.90 (0.88) | 1.95 (0.91) | 3.38 (1.19) | 3.00 (1.22) | 2.04 (1.00) | 61% | 1v2, 1v4, 2v4, 2v5, 3v4, 4v6 |
| F4:How often did you feel like you needed to empty your bowels right away or else you would have an accident? (If Never, you are finished) | 1.88 (1.00) | 1.48 (0.70) | 1.47 (0.73) | 2.63 (1.12) | 1.80 (0.79) | 1.73 (0.95) | 1% | 1v2, 1v3, 1v4, 2v4, 3v4, 4v5, 4v6 |
| F5:How much did feeling you needed to empty your bowels right away interfere with your day-to-day activities? | 2.44 (1.19) | 1.82 (0.76) | 1.71 (0.60) | 2.88 (1.15) | 1.67 (0.52) | 2.08 (0.69) | 54% | 1v2, 1v3, 2v4, 3v4, 4v5, 4v6 |
| F6:How much did feeling you needed to empty your bowels right away bother you? | 2.88 (1.36) | 2.38 (0.95) | 1.86 (0.71) | 3.50 (1.26) | 2.33 (0.52) | 2.38 (1.06) | 54% | 1v3, 2v3, 2v4, 3v4, 4v6 |
| G. PROMIS GI bowel incontinence 4a (past 7 days) | ||||||||
| GI Bowel Incontinence Score | 5.50 (2.02) | 5.11 (1.92) | 4.47 (0.94) | 6.15 (2.58) | 5.44 (1.81) | 4.81 (1.30) | 2% | 1v3, 2v3, 2v4, 3v4, 3v5, 4v6 |
| G1:How often did you have bowel incontinence - that is, have an accident because you could not make it to the bathroom in time? | 1.21 (0.52) | 1.07 (0.25) | 1.05 (0.36) | 1.35 (0.99) | 1.10 (0.32) | 1.10 (0.30) | 1% | 1v3, 3v4 |
| G2:How often did you soil or dirty your underwear before getting to a bathroom? | 1.31 (0.60) | 1.24 (0.68) | 1.04 (0.20) | 1.35 (0.67) | 1.20 (0.63) | 1.08 (0.27) | 1% | 1v3, 1v6, 2v3, 3v4, 4v6 |
| G3:How often did you leak stool or soil your underwear? | 1.33 (0.65) | 1.31 (0.71) | 1.07 (0.30) | 1.40 (0.68) | 1.22 (0.44) | 1.19 (0.44) | 2% | 1v3, 2v3, 3v4, 3v6 |
| G4:How often did you think you were going to pass gas, but stool or liquid came out instead? | 1.64 (0.88) | 1.49 (0.71) | 1.30 (0.54) | 2.05 (0.89) | 1.90 (0.57) | 1.44 (0.64) | 1% | 1v3, 2v4, 2v5, 3v4, 3v5, 4v6, 5v6 |
| Q: PROMIS depression short form 8a (past 7 days) | ||||||||
| Depression T-Score | 46.4 (7.48) | 47.7 (8.71) | 45.5 (6.82) | 62.2 (6.60) | 48.9 (6.08) | 45.4 (6.99) | 0% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Depression SE | 3.52 (1.78) | 3.38 (1.71) | 3.64 (1.67) | 1.78 (0.13) | 2.73 (1.59) | 3.70 (1.66) | 0% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q1:I felt worthless | 1.31 (0.68) | 1.32 (0.72) | 1.18 (0.53) | 2.40 (1.19) | 1.00 (0.00) | 1.15 (0.50) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q2:I felt helpless | 1.33 (0.64) | 1.40 (0.85) | 1.22 (0.60) | 2.79 (0.85) | 1.50 (0.85) | 1.21 (0.57) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q3:I felt depressed | 1.40 (0.66) | 1.73 (0.99) | 1.53 (0.82) | 3.40 (0.75) | 1.80 (0.79) | 1.53 (0.85) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q4:I felt hopeless | 1.19 (0.45) | 1.37 (0.77) | 1.21 (0.55) | 3.05 (0.89) | 1.33 (0.50) | 1.19 (0.48) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q5:I felt like a failure | 1.48 (0.77) | 1.52 (0.88) | 1.17 (0.44) | 2.95 (1.15) | 1.60 (0.70) | 1.30 (0.57) | <1% | 1v3, 1v4, 2v3, 2v4, 3v4, 3v5, 4v5, 4v6 |
| Q6:I felt unhappy | 1.90 (0.85) | 1.96 (0.99) | 1.80 (0.74) | 3.60 (0.94) | 2.10 (0.74) | 1.83 (0.91) | 1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| Q7:I felt that I had nothing to look forward to | 1.42 (0.66) | 1.53 (0.84) | 1.29 (0.63) | 2.95 (1.35) | 1.30 (0.48) | 1.23 (0.51) | 1% | 1v4, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| Q8:I felt that nothing could cheer me up | 1.28 (0.59) | 1.51 (0.91) | 1.20 (0.52) | 2.95 (1.32) | 1.20 (0.42) | 1.19 (0.44) | <1% | 1v4, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| R. PROMIS anxiety short form 8a (past 7 days) | ||||||||
| Anxiety T-Score | 44.0 (7.41) | 47.6 (8.98) | 45.6 (7.85) | 62.6 (8.15) | 51.7 (8.31) | 47.7 (8.16) | 0% | 1v2, 1v4, 1v5, 1v6, 2v4, 3v4, 3v5, 4v5, 4v6 |
| Anxiety SE | 3.84 (1.56) | 3.26 (1.49) | 3.52 (1.54) | 2.01 (0.21) | 2.63 (1.18) | 3.16 (1.53) | 0% | 1v2, 1v4, 1v5, 1v6, 2v4, 3v4, 4v6 |
| R1:I felt fearful | 1.09 (0.29) | 1.31 (0.69) | 1.20 (0.52) | 2.42 (1.12) | 1.50 (0.71) | 1.30 (0.54) | 3% | 1v4, 1v5, 1v6, 2v4, 3v4, 4v5, 4v6 |
| R2:I found it hard to focus on anything other than my anxiety | 1.16 (0.48) | 1.37 (0.82) | 1.30 (0.70) | 2.80 (1.24) | 1.40 (0.84) | 1.19 (0.48) | 1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| R3:My worries overwhelmed me | 1.35 (0.65) | 1.48 (0.87) | 1.27 (0.60) | 2.90 (1.17) | 1.60 (0.97) | 1.42 (0.78) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| R4:I felt uneasy | 1.40 (0.69) | 1.67 (0.82) | 1.51 (0.70) | 3.10 (1.07) | 2.00 (1.05) | 1.74 (0.76) | <1% | 1v2, 1v4, 1v6, 2v4, 3v4, 4v5, 4v6 |
| R5:I felt nervous | 1.42 (0.66) | 1.70 (0.84) | 1.54 (0.77) | 3.10 (0.97) | 2.40 (1.07) | 1.81 (0.79) | 1% | 1v4, 1v5, 1v6, 2v4, 2v5, 3v4, 3v5, 3v6, 4v6 |
| R6:I felt like I needed help for my anxiety | 1.20 (0.52) | 1.50 (0.96) | 1.24 (0.62) | 3.25 (1.33) | 1.60 (0.84) | 1.40 (0.77) | 3% | 1v4, 2v3, 2v4, 3v4, 4v5, 4v6 |
| R7:I felt anxious | 1.40 (0.69) | 1.78 (0.96) | 1.58 (0.80) | 3.26 (1.15) | 2.33 (1.00) | 1.89 (0.99) | <1% | 1v2, 1v4, 1v5, 1v6, 2v4, 3v4, 3v5, 4v5, 4v6 |
| R8:I felt tense | 1.72 (0.88) | 1.96 (1.04) | 1.79 (0.88) | 3.40 (0.82) | 2.40 (0.97) | 1.89 (0.89) | 0% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v5, 4v6 |
| T. PROMIS sleep disturbance short form 8a (past 7 days) | ||||||||
| Sleep Disturbance T-Score | 51.2 (4.96) | 52.0 (3.55) | 52.0 (2.86) | 54.2 (4.27) | 53.8 (4.35) | 51.3 (3.43) | <1% | 3v4, 4v6 |
| Sleep Disturbance SE | 2.77 (0.47) | 2.72 (0.50) | 2.68 (0.36) | 3.18 (0.71) | 2.88 (0.47) | 2.75 (0.44) | <1% | 1v4, 2v4, 3v4, 4v6 |
| T1:My sleep quality was | 2.65 (1.11) | 3.02 (0.86) | 2.64 (0.90) | 3.85 (0.93) | 2.78 (1.20) | 2.87 (0.94) | <1% | 1v2, 1v4, 2v3, 2v4, 3v4, 4v5, 4v6 |
| T2:My sleep was refreshing | 2.93 (1.12) | 3.09 (0.96) | 2.67 (0.98) | 4.35 (0.93) | 2.89 (0.78) | 2.89 (1.09) | <1% | 1v4, 2v3, 2v4, 3v4, 4v5, 4v6 |
| T3:I had a problem with my sleep | 2.60 (1.16) | 2.82 (1.15) | 2.50 (1.15) | 3.80 (1.20) | 2.44 (1.51) | 2.46 (1.18) | <1% | 1v4, 2v4, 3v4, 4v5, 4v6 |
| T4:I had difficulty falling asleep | 1.79 (1.08) | 2.08 (1.10) | 1.75 (0.87) | 3.30 (1.22) | 2.78 (1.48) | 1.83 (0.90) | 2% | 1v4, 1v5, 2v4, 3v4, 3v5, 4v6, 5v6 |
| T5:My sleep was restless | 2.28 (1.22) | 2.30 (1.05) | 2.03 (0.88) | 3.40 (1.31) | 2.56 (1.24) | 2.31 (1.02) | 2% | 1v4, 2v4, 3v4, 4v6 |
| T6:I tried hard to get to sleep | 1.74 (1.05) | 2.03 (1.03) | 1.70 (0.78) | 3.25 (1.16) | 2.00 (1.41) | 1.71 (0.87) | 1% | 1v4, 2v3, 2v4, 3v4, 4v5, 4v6 |
| T7:I was worried about not being able to fall asleep | 1.40 (0.79) | 1.59 (0.89) | 1.53 (0.76) | 2.80 (1.32) | 2.11 (1.45) | 1.53 (0.87) | <1% | 1v4, 2v4, 3v4, 4v6 |
| T8:I was satisfied with my sleep | 3.02 (1.14) | 3.10 (1.11) | 2.71 (1.07) | 4.15 (0.88) | 2.56 (1.01) | 2.94 (1.18) | <1% | 1v4, 2v3, 2v4, 3v4, 4v5, 4v6 |
| X. PROMIS physical function short form 10b | ||||||||
| X1:Are you able to do chores such as vacuuming or yard work? | 3.93 (1.18) | 4.44 (1.01) | 4.82 (0.45) | 3.53 (1.22) | 4.30 (1.25) | 4.70 (0.67) | <1% | 1v2, 1v3, 1v6, 2v3, 2v4, 3v4, 4v6 |
| X2:Are you able to get in and out of a car? | 4.19 (0.91) | 4.59 (0.68) | 4.89 (0.31) | 4.05 (0.76) | 4.80 (0.42) | 4.88 (0.38) | <1% | 1v2, 1v3, 1v5, 1v6, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| X3:Are you able to go up and down stairs at a normal pace? | 4.05 (0.84) | 4.33 (0.97) | 4.71 (0.61) | 3.45 (1.36) | 4.70 (0.48) | 4.79 (0.49) | 0% | 1v2, 1v3, 1v5, 1v6, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| X4:Are you able to run errands and shop? | 4.33 (1.08) | 4.62 (0.77) | 4.96 (0.20) | 4.00 (1.08) | 4.90 (0.32) | 4.88 (0.32) | <1% | 1v3, 1v6, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| X5:Are you able to bend down and pick up clothing from the floor? | 4.21 (0.99) | 4.41 (0.88) | 4.73 (0.53) | 3.65 (1.14) | 4.22 (0.67) | 4.74 (0.45) | 1% | 1v3, 1v4, 1v6, 2v3, 2v4, 2v6, 3v4, 3v5, 4v6, 5v6 |
| X6:Are you able to lift 10 pounds (5 kg) above your shoulder? | 4.40 (0.94) | 4.58 (0.82) | 4.88 (0.43) | 4.25 (1.07) | 4.90 (0.32) | 4.89 (0.32) | <1% | 1v3, 1v6, 2v3, 2v6, 3v4, 4v6 |
| X7:Does your health now limit you in doing vigorous activities, such as running, lifting heavy objects, participating in strenuous sports? | 2.79 (1.20) | 3.27 (1.37) | 3.99 (1.16) | 2.65 (1.42) | 3.60 (1.43) | 4.15 (0.97) | <1% | 1v2, 1v3, 1v6, 2v3, 2v6, 3v4, 4v6 |
| X8:Does your health now limit you in bathing or dressing yourself? | 4.60 (0.76) | 4.71 (0.71) | 4.97 (0.16) | 4.25 (0.79) | 5.00 (0.00) | 4.98 (0.14) | <1% | 1v3, 1v4, 1v6, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| X9:Does your health now limit you in putting a trash bag outside? | 4.70 (0.64) | 4.79 (0.59) | 4.97 (0.16) | 4.25 (0.97) | 4.80 (0.63) | 4.98 (0.14) | <1% | 1v3, 1v4, 1v6, 2v3, 2v4, 2v6, 3v4, 4v6 |
| X10:Does your health now limit you in doing moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf? | 4.09 (1.27) | 4.45 (0.93) | 4.83 (0.50) | 3.50 (1.36) | 4.40 (1.35) | 4.83 (0.47) | <1% | 1v3, 1v6, 2v3, 2v4, 2v6, 3v4, 4v5, 4v6 |
| J. Physical Exam Findings | ||||||||
| J4:Weight, kg | 98.2 (21.1) | 95.6 (24.0) | 89.4 (14.3) | 106 (23.3) | 101 (36.8) | 85.0 (14.7) | 2% | 1v6, 2v6, 3v4, 3v6, 4v6 |
| Body Mass Index | 31.7 (6.41) | 30.3 (7.25) | 28.1 (4.20) | 33.2 (6.78) | 32.4 (10.4) | 27.2 (4.39) | 2% | 1v3, 1v6, 2v6, 3v4, 4v6 |
| J6:Waist measurement | 40.4 (8.28) | 42.1 (15.0) | 40.4 (16.4) | 48.3 (20.2) | 40.1 (6.55) | 40.4 (18.0) | 8% | 1v3, 1v6, 2v6, 3v4, 4v6 |
| J7:Ambulatory aids | 9.3% | 6.9% | 1.4% | 21.1% | 0.0% | 0.0% | 3% | 1v6, 3v4, 4v6 |
Results shown as mean (standard deviation) or prevalence; pairwise p-values calculated via Wilcoxon rank sum test for continuous variables or chi-square test for categorical variables
Phenotypes are numbered in column order: 1 = metabolic, 2 = nocturnal polyuria, 3 = polyuria, 4 = affective, 5 = bowel and bladder dysfunction, 6 = benign prostatic enlargement.
Tables 2 and 3 report summary scores. Item-level CASUS comparisons are described in the Results but are not tabulated.
Results
Two hundred and twenty-eight clinicians responded to the eligibility screening; 74 were eligible, including 45 who provided their contact information. Nine of the 45 clinicians predominantly treated men, and 36 predominantly treated women. Invitations were sent to all nine clinicians who treated men and 22 clinicians who treated women. Twenty clinicians scheduled and completed the interview, including eight who predominantly treated men and 12 who predominantly treated women (Supplemental Table 1).
Eleven expert knowledge-based phenotypes were created based on a review of the refined scenario DAGs, including eight that were not sex-specific: polyuria, metabolic, affective, nocturnal polyuria, bowel and bladder dysfunction, nervous system, pain, and tumor/implant (e.g. mesh) and three that were sex-specific, including UTI and pelvic organ prolapse for females, and benign prostatic enlargement (BPE) for males. Figure 1 provides an example of one DAG, and the remaining are maintained on a GitLab site (https://lurn.pages.oit.duke.edu/dags/).25 Nervous system, pain, and tumor/implant DAGs were included for completeness’ sake; however, these DAGs were not used in the analysis, as these were exclusion criteria for participants in the LURN II cohort.
Figure 1 -.

Directed acyclic graph for the metabolic phenotype
Six hundred and twenty-nine LURN participants (336 females and 293 males) were assigned to eight phenotypes: polyuria (n=265), nocturnal polyuria (n=129), metabolic (n=96), affective (n=41), bowel and bladder dysfunction (n=22), pelvic organ prolapse (n=19 females), UTI (n=4 females), and benign prostatic enlargement (n=53, males).
Pairwise differences between the phenotypes are presented in Tables 2 and 3 for females and males, respectively. Demographic characteristics and self-reported comorbidities differed by age and by urinary urgency alone versus urgency with incontinence. Comparisons involving the UTI phenotype (n=4) are unstable and are not interpreted. Differences on the variables used to define anchors (Table 1) — PROMIS GI constipation, body mass index, wake frequency, PROMIS depression and anxiety, CASUS nocturia, POPDI-6, and straining — follow by construction and are not restated below; those reported here arose on variables used in no anchor definition.
In females, the metabolic phenotype was more prevalent in older individuals, and incontinence severity was lowest in the polyuria and pelvic organ prolapse phenotypes. Congestive heart failure did not differ significantly across the female phenotypes (smallest pairwise p = 0.33) and is not tabulated.
Female CASUS responses showed that bowel/bladder phenotype participants reported more bloating, pain, and aching during urination. Affective phenotype participants experienced more frequent urination needs after emptying, a sensation of incomplete bladder emptying, and dribbling. Those in the affective phenotype reported greater difficulty with voiding.
Female LURN SI-29 responses showed incontinence scores highest in the nocturnal polyuria phenotype and lowest in polyuria and prolapse phenotypes. Pain scores were highest with bowel and bladder dysfunction and lowest with the metabolic and polyuria phenotypes. Urgency scores peaked in the nocturnal polyuria phenotype and were lowest in prolapse.
Female PROMIS responses demonstrated that constipation scores were significantly higher in the affective phenotype. Diarrhea scores were significantly lower in the polyuria phenotype. On physical examination, pelvic floor contraction strength, measured using the Oxford scale26, was numerically highest in the affective phenotype (2.89, SD 1.32) but did not differ significantly across phenotypes (smallest pairwise p = 0.054) and is therefore not shown in Table 2.
In males, participants with the polyuria phenotype generally had fewer comorbidities compared to those with other phenotypes. Participants in the affective phenotype reported a wider variety of sensations with and after urination, pain with urination, and voiding difficulty. The metabolic phenotype reported the highest incontinence and urgency scores, and pain scores were highest in the affective phenotype. PROMIS diarrhea and bowel incontinence scores were highest in the affective phenotype. On physical exam, the BPE phenotype had the lowest BMI of all phenotypes. Demographic and clinical characteristics for participants assigned to the anchor groups are presented in Supplemental Tables 2 and 3, respectively, for females and males.
When evaluating compactness of the phenotypes, the median pairwise distance (unitless) between all phenotype centroids was 19.4 for females and 17.4 for males. Figure 2 shows the distribution of the cluster compactness ratio, . Because we used the recommendation that (for forced-choice labeling to be credible) >80% of patients should have ,9 our phenotypes did not meet compactness requirements; thus, phenotype labels alone were unlikely to be clinically discriminating. No participant of either sex had a compactness ratio below 0.20; the smallest observed value was 0.73 in women and 0.68 in men, more than three times the threshold (Figure 2). Ambiguity of assignment, the difference between a participant’s distances to the two nearest centroids divided by the distance to the nearest, had a median of 0.08 in women and 0.07 in men: for a typical participant the second most likely phenotype was less than 10% farther away. As an internal consistency check, phenotype assignment was recomputed for the anchor participants who originally defined each phenotype centroid; 69 of 90 anchors (77%) returned to their own phenotype, 33 of 45 (73%) in women and 36 of 45 (80%) in men, ranging from 50% (female metabolic) to 100% (polyuria, both sexes). Median membership probability in the assigned phenotype was similarly low relative to certainty: the median probability of belonging to the assigned phenotype ranged from 0.18 to 0.25 across the seven female phenotypes against 0.14 expected by chance, and from 0.21 to 0.33 across the six male phenotypes against 0.17.
Figure 2 – Phenotype compactness diagnostics.

(A) Pairwise distances between all female phenotype centroids (median = 19.4); (B) Pairwise distances between all male phenotype centroids (median = 17.4); (C) Distribution of the ratio of each woman’s distance to her assigned phenotype centroid divided by the median inter-centroid distance; (D) Same as (C) for men. The 0.20 threshold is that recommended in reference 9.
For both sexes, nearly all ratios clustered around 1.0, with none <0.20 — the threshold Harrell recommends for compact clusters. This indicates that no participants were sufficiently close to their assigned centroid relative to the spacing of centroids, suggesting phenotype labels alone do not yield groupings that are distinct enough for the label alone to be of clinical use.
Supplemental Table 4 summarizes the incremental prognostic value provided by phenotype labels and their continuous representations (centroid distances, logits of membership probability) for LURN SI-29 total score at 3 and 12 months, calculated using the likelihood ratio and 1-adequacy index, separately by sex. Unadjusted, distances added 29% to 47% of new information beyond labels across the four sex-by-time strata, and phenotype labels added 3% to 12% beyond distances. After adjusting for baseline SI-29 score, a known prognostic factor, these contributions were substantially smaller and more consistent: distances added 9% to 18% of new information beyond labels, and labels added 3% to 9% beyond distances, across sex and follow-up time. Baseline SI-29 score alone (likelihood ratio χ2 = 190.0, females, 3 months, from the original phenotyping analysis) provided more prognostic information than the entire phenotype apparatus of labels and distances combined (likelihood ratio χ2 = 151.6, Supplemental Table 4); across the four sex-by-time strata, baseline severity alone accounted for 65% to 84% of the predictive information in the full model containing baseline severity, labels and distances. Phenotype information is therefore best interpreted as a modest supplement to, rather than a replacement for, baseline symptom severity in predicting prognosis under usual care. The asymmetry was consistent in direction across both sexes and both follow-up intervals.
Discussion
This study demonstrated a method for eliciting and summarizing clinical knowledge from clinicians treating URG and UUI. That continuous distances outperformed categorical labels is expected rather than surprising: a label is a coarsening of the geometry from which it is derived, and coarsening cannot add information. The contribution here is therefore the magnitude of that loss in a clinical dataset (distances carried 9% to 18% of new information beyond labels, labels only 3% to 9% beyond distances), together with the more consequential observation that even the continuous representation added little once baseline symptom severity was accounted for.
Anchors were selected as the clearest examples of each phenotype and defined by extreme percentile thresholds (Table 1), yet lay farther from their centroids than typical participants: median distance 27.2 versus 21.5 in females and 23.9 versus 19.7 in males. This is largely mechanical: each centroid is dominated by the non-anchor participants assigned to it by proximity, so anchors at the extreme of the distribution are necessarily displaced from it. Whether anything beyond this artifact is present, for example, whether clinicians recognize phenotypes by archetype rather than by average, or whether anchor displacement is radially outward from the global centroid and therefore tracks overall severity, was not tested here and cannot be inferred from these data.
This study has several implications for clinical phenotyping research. While unsupervised machine learning is sometimes referred to as “data-driven,” it is also exploratory by nature, as subjective decisions shape which variables are included and how resulting groups are compared. This study has demonstrated that clinical knowledge and intuition about causation can be effectively introduced into the process using DAGs, making the variable selection process more efficient while maintaining the biological plausibility of the resulting phenotypes. 18 Finally, DAGs could ensure that all important variables are included in data collection, measurements, and clustering, and that none of the clinically or physiologically important variables are missed. Once the inclusive list of variables is created, proper scaling and weighting can be achieved by comparing their distributions with those of controls, using weights based on the correlation matrices of the variables as described.2
This study's strengths include providing expert-derived causal diagrams (DAGs) for URG and UUI. It extracts beliefs from subject-matter knowledge rather than statistical data analysis. Unlike methods that rely on literature estimates, it explains how subject-matter knowledge informed DAG creation.27 Using a multicenter cohort, it captures a broad range of URG/UUI phenotypes across settings, reflecting their heterogeneity. Notably, several differences arose in variables used in no anchor definition, providing independent corroboration that the phenotypes capture coherent structure: PROMIS diarrhea and bowel incontinence were highest in males with the affective phenotype, and CASUS item-level responses differed significantly by phenotype in both sexes. Congestive heart failure and pelvic floor contraction strength were numerically consistent with the phenotypes but did not reach significance. Clinically, these results do not support using phenotype labels to select treatment. The diagrams remain useful as a structured prompt for inexpensive assessments that change management: a voiding diary separates global from nocturnal polyuria and from frequency at normal volumes, body mass index and hemoglobin A1c identify a metabolic contribution, and validated affective and constipation measures identify contributions that are independently treatable.
The study has several limitations. It covers eight of 11 phenotypes, as three (nervous system, pain, tumor/implant) were excluded from the LURN cohort. The phenotypes were not compact enough for labels alone to be useful in clinical care, highlighting the challenge of subgrouping URG/UUI patients. The 0.20 threshold is a proposed heuristic rather than a validated standard, and has not, to our knowledge, been applied to other phenotyping schemes in this population; our finding that these phenotypes fail it should therefore not be read as evidence that they are less compact than data-driven alternatives. Several DAGs describe multiple pathways a single anchor cannot capture; the nocturia DAG included both polyuria-mediated and sleep-disorder-mediated pathways, and only the former was operationalized (Table 1). Relatedly, the ICS nocturnal polyuria index carries age-dependent cutoffs (>20% in younger adults, 20-33% in middle age, >33% in older adults); a fixed 25% threshold was applied here in a cohort with a median age near 60, where the ICS-recommended cutoff would be higher, so the nocturnal polyuria anchor definition (Table 1) is a simplification of the full ICS operationalization. The clinician panel, while balanced by the sex of patients predominantly treated (8 clinicians treating predominantly men, 12 predominantly women), was limited in racial and ethnic diversity: with no Black, Native Hawaiian, or Hispanic clinicians among those treating predominantly men (Supplemental Table 1). The elicited causal beliefs may therefore not generalize to clinicians of different backgrounds or practice settings, particularly for care of male patients. The elicitation process was also resource-intensive: 228 clinicians were screened to yield 20 completed one-hour, two-interviewer sessions. Broader, more diverse, and more efficiently recruited panels would strengthen transportability and feasibility. Phenotype sizes were highly unequal, with polyuria accounting for 265 of 629 participants and urinary tract infection for four. Participants in the polyuria phenotype reported the lowest incontinence severity, fewest comorbidities, and lowest pain scores, consistent with this phenotype functioning partly as a default assignment for participants without a distinctive profile. Between-phenotype comparisons were descriptive and unadjusted for multiplicity across the many variables and phenotype pairs examined, and should be regarded as hypothesis-generating. Each participant was assigned a single phenotype rather than a score for each; given the low membership probabilities, a profile across phenotypes would represent these data better and is a priority for future work. Anchor variables were not varied in the sensitivity analysis, and prostate size was graded by digital rectal examination in approximate 20 g increments, which coarsely estimates gland size and would blur that centroid. Most participants, including 69% of men, were receiving LUTS treatment at baseline; because treatment attenuates the very symptoms that distinguish these phenotypes, assignment reflects treated rather than untreated state and would bias toward the null. Finally, the polyuria anchor definition combined void volume above the 85th percentile with a fluid-intake criterion rather than the ICS threshold of 40 mL/kg/24 h, so this label should not be read as ICS-defined polyuria.
Conclusions
This study showed a method for eliciting causal understanding from clinicians treating URG and UUI and integrating it into DAGs representing phenotypes. Cohort participants fell into eight such phenotypes, but these behave as probabilistic positions rather than discrete classes: median membership probability in the assigned phenotype ranged only from 0.18 to 0.33, and continuous representations retained modestly more prognostic information than labels. Because both added little beyond baseline symptom severity, these phenotypes are not yet suitable for clinical classification in either form; their present value lies in providing interpretable, modifiable causal hypotheses for future URG/UUI research.
Supplementary Material
Acknowledgments
This is publication number 54 of the Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN).
This study is supported by the National Institute of Diabetes & Digestive & Kidney Diseases through cooperative agreements (grants DK097780, DK097772, DK097779, DK099932, DK100011, DK100017, DK099879).
Research reported in this publication was supported at Northwestern University, in part, by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number UL1TR001422. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The following individuals were instrumental in the planning and conduct of this study at each of the participating institutions:
Duke University, Durham, North Carolina (DK097780): PIs: Cindy Amundsen, MD, J. Eric Jelovsek, MD; Co-Is: Kathryn Flynn, PhD, Jim Hokanson, PhD, Aaron Lentz, MD, David Page, PhD, Nazema Siddiqui, MD, , Chad Gridley, MD, Jordan Foreman, MD, Tara Morgan, MD, Study Coordinators: Magaly Guerrero, BSc, Stephanie Yu.
University of Iowa, Iowa City, Iowa (DK097772): PIs: Catherine S Bradley, MD, MSCE, Karl Kreder, MD, MBA; Co-Is: Bradley A. Erickson, MD, MS, Daniel Fick, MD, Vince Magnotta, PhD, Philip Polgreen, MD, MPH; Study Coordinator: Nancy Hollenbeck.
University of Chicago, Chicago, Illinois (DK097779): PIs: C. Emi Bretschneider, MD, James W Griffith, PhD, Kimberly Kenton, MD, MS, Brian Helfand, MD, PhD; Co-Is: David Cella, PhD, Julia Geynisman-Tan, MD, Alex Glaser, MD, Margaret Mueller, MD, Francesca Farina, PhD, Richard Fantus, MD, Devin Boehm, BS; Study Coordinators: Melissa Marquez, MBA, Malgorzata Antoniak, PhD, Pooja Talaty, MS, Karen John, Jinxuan Shi, MA, Tara Samsel, BS.
Dr. Helfand, Dr. Glaser, Dr. Antoniak, and Ms. Talaty are at Endeavor Health. Dr. Bretschneider, Dr. Cella, Dr. Geynisman-Tan, Mr. Boehm, Ms. Marquez, and Ms. John are at Northwestern Medicine.
University of Michigan Health System, Ann Arbor, Michigan (DK099932): PI: J Quentin Clemens, MD, FACS, MSCI; Co-Is: John DeLancey, MD, Dee Fenner, MD, Rick Harris, MD, Steve Harte, PhD, Anne P. Cameron, MD, Aruna Sarma, PhD, Giulia Ippolito, MD, Priyanka Gupta, MD, Whitney Horner, MD; Jannah Thompson, MD; Payton Schmidt, MD; Study Coordinators: Greg Mowatt, BA, Sarah Richardson, BS, Sneha Mathai, BS, Syedah Mubeenah, MBBS.
University of Washington, Seattle, Washington (DK100011): PI: Claire Yang, MD; Co-Is: Anna Kirby, MD, Swati Rane Levendovszky, PhD, Sreya Gutta; Study Coordinators: Brenda Vicars, RN, Holly Covert.
Washington University in St. Louis, St. Louis, Missouri (DK100017): PI: H. Henry Lai, MD; Co-Is: Joshua Shimony, MD, PhD, Fuhai Li, PhD; Study Coordinators: Patricia Hayden, BSN, Aleksandra Klim, RN, MHS, CCRC
Arbor Research Collaborative for Health, Data Coordinating Center (DK099879): PI: John Graff, PhD, MS; Co-Is: Victor Andreev, PhD, DSc, Brenda Gillespie, PhD; Project Manager: Jessica Durkin, M.Ed, MBA; Clinical Monitors: Melissa Sexton, BA, CCRP, Julia Nashif, BA, CCRP; Research Analysts: Brian Bieber, MPH, Nathan Goodrich, MS, Calvin Andrews, MS, Jon Wiseman, MS. Project Associate: Naveen Fawaz.
National Institute of Diabetes and Digestive and Kidney Diseases, Division of Kidney, Urology, and Hematology, Bethesda, Maryland: Project Scientist: Ziya Kirkali, MD; Project Officer: Christopher Mullins, PhD.
Funding:
National Institute of Diabetes & Digestive & Kidney Diseases through cooperative agreements (grants DK097780, DK097772, DK097779, DK099932, DK100011, DK100017, DK099879). DK097780 funded J. Eric Jelovsek’s efforts.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declaration of Generative AI and AI-assisted Technologies in the Writing Process
During the preparation of this work, the authors used Claude (Anthropic), a generative AI assistant, to support manuscript editing, to audit numerical, tabular, and citation consistency across the text, and to help prepare ancillary submission materials. The authors directed and reviewed all AI-assisted work, verified all data, results, and citations against source materials, and take full responsibility for the accuracy and content of this publication. This tool is not an author and did not contribute to study conception, data analysis, or interpretation of results.
Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
J. Eric Jelovsek: reports royalties from UpToDate and participation on the Collamedix advisory board.
Cindy L. Amundsen: reports consulting fees, honoraria for lectures/presentations, and travel support from Bluewind, and unpaid participation on the Data Safety Monitoring Board/advisory board of Bluewind and on the PCORI advisory board.
Victor P. Andreev: holds a prior patent (US 7,409,298 B2, assigned to Northeastern University) unrelated to urology.
M. Alan Brookhart: serves on the Scientific Advisory Board of AbbVie, Accompany Health, the American Academy of Allergy, Asthma & Immunology, Amgen, Axsome Therapeutics, Brigham and Women's Hospital, Gilead/Kite, Merck, NIDDK, Regeneron, and Target RWE, and holds equity in Accompany Health and VitriVax.
Catherine S. Bradley: reports royalties from Elsevier textbooks; serves unpaid on the American Urogynecologic Society (AUGS) Executive Committee/Board; receives honoraria as an American Board of Obstetrics and Gynecology (ABOG) Urogynecology Division member and oral examiner; and serves as an Elsevier journal editor, with salary support paid to her institution.
Anna Kirby: reports honoraria from FemSelect for two teaching sessions on prolapse surgery at AUGS.
H. Henry Lai: reports institutional payment from Medtronic for a post-marketing product registry, and consulting/leadership payments from the American Urological Association.
Tara Nikonow Morgan, Aaron Lentz, Brian Bieber, Nathan P. Goodrich, Ziya Kirkali, and Kathryn E. Flynn declare no competing interests.
Informed consent and patient details
Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed.
Studies in Human
This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place.
This study was approved by the Institutional Review Board.
(Approval No. Pro00106706)
Data Availability Statement
De-identified LURN II Urinary Urgency Phenotyping cohort data will be made available through the NIDDK Central Repository (https://repository.niddk.nih.gov/), per NIDDK’s data-sharing policy. DAGs and analytic code used to construct and evaluate the phenotypes are maintained on the GitLab site (https://lurn.pages.oit.duke.edu/dags/).
References
- 1.Cameron AP, Yang CC, Bradley CS, et al. : Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN): An introduction to the Urinary Urgency Phenotyping Protocol LURN II. Neurourol. Urodyn. 2024; 43: 1800–1808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nikolaou V, Massaro S, Fakhimi M, et al. : COPD phenotypes and machine learning cluster analysis: A systematic review and future research agenda. Respir. Med. 2020; 171: 106093. [DOI] [PubMed] [Google Scholar]
- 3.Shah AP, Snead W, Daga A, et al. : Methods for phenotyping adult patients with acute kidney injury: a systematic review. J. Nephrol. 2024; 38: 75–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ahlqvist E, Storm P, Karajamaki A, et al. : Novel subgroups of adult-onset diabetes and their association with outcomes: a data-driven cluster analysis of six variables. Lancet Diabetes Endocrinol 2018; 6: 361–369. [DOI] [PubMed] [Google Scholar]
- 5.Hennig C, Meila M, Murtagh F, et al. eds: Handbook of Cluster Analysis. 0 ed. Chapman and Hall/CRC; 2015. Available at: https://www.taylorfrancis.com/books/9781466551893, accessed July 28, 2025. [Google Scholar]
- 6.Hernán M and Robins JM: Causal inference: what if. First edition. Boca Raton: Taylor and Francis; 2023. [Google Scholar]
- 7.Ippolito GM, Kenton K, Bradley CS, et al. : Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN II) Urinary Urgency Phenotyping Study: Methods and Baseline Urinary Symptoms by Age and Sex. Neurourol. Urodyn. 2025; 44: 1007–1021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rodrigues D, Kreif N, Lawrence-Jones A, et al. : Reflection on modern methods: constructing directed acyclic graphs (DAGs) with domain experts for health services research. Int J Epidemiol 2022; 51: 1339–1348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Harrell F: The Burden of Demonstrating Statistical Validity of Clusters. Stat. Think. 2024. Available at: https://www.fharrell.com/post/cluster/, accessed June 3, 2026. [Google Scholar]
- 10.Pearl J: Causal Diagrams for Empirical Research. Biometrika 1995; 82: 669–688. [Google Scholar]
- 11.Tennant PWG, Murray EJ, Arnold KF, et al. : Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int. J. Epidemiol. 2021; 50: 620–632. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Greenland S, Pearl J and Robins JM: Causal diagrams for epidemiologic research. Epidemiology 1999; 10: 37–48. [PubMed] [Google Scholar]
- 13.Saunders B, Sim J, Kingstone T, et al. : Saturation in qualitative research: exploring its conceptualization and operationalization. Qual. Quant. 2018; 52: 1893–1907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Miles MB and Huberman AM: Qualitative data analysis: an expanded sourcebook. 2nd ed. Thousand Oaks: Sage Publications; 1994. [Google Scholar]
- 15.Hashim H, Blanker MH, Drake MJ, et al. : International Continence Society (ICS) report on the terminology for nocturia and nocturnal lower urinary tract function. Neurourol. Urodyn. 2019; 38: 499–508. [DOI] [PubMed] [Google Scholar]
- 16.Austin PF, Bauer SB, Bower W, et al. : The standardization of terminology of lower urinary tract function in children and adolescents: Update report from the standardization committee of the International Children’s Continence Society. Neurourol. Urodyn. 2016; 35: 471–481. [DOI] [PubMed] [Google Scholar]
- 17.D’Ancona C, Haylen B, Oelke M, et al. : The International Continence Society (ICS) report on the terminology for adult male lower urinary tract and pelvic floor symptoms and dysfunction. Neurourol. Urodyn. 2019; 38: 433–477. [DOI] [PubMed] [Google Scholar]
- 18.Andreev VP, Helmuth ME, Liu G, et al. : Subtyping of common complex diseases and disorders by integrating heterogeneous data. Identifying clusters among women with lower urinary tract symptoms in the LURN study. PloS One 2022; 17: e0268547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Weinfurt KP, Griffith JW, Flynn KE, et al. : The Comprehensive Assessment of Self-Reported Urinary Symptoms: A New Tool for Research on Subtypes of Patients with Lower Urinary Tract Symptoms. J Urol 2019; 201: 1177–1183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cella D, Smith AR, Griffith JW, et al. : A new outcome measure for LUTS: Symptoms of Lower Urinary Tract Dysfunction Research Network Symptom Index-29 (LURN SI-29) questionnaire. Neurourol. Urodyn. 2019; 38: 1751–1759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Spiegel BMR, Hays RD, Bolus R, et al. : Development of the NIH Patient-Reported Outcomes Measurement Information System (PROMIS) Gastrointestinal Symptom Scales. Am. J. Gastroenterol. 2014; 109: 1804–1814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pilkonis PA, Choi SW, Reise SP, et al. : Item banks for measuring emotional distress from the Patient-Reported Outcomes Measurement Information System (PROMIS®): depression, anxiety, and anger. Assessment 2011; 18: 263–283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Harrell F: Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. 2nd ed. 2015. Cham: Springer; 2015. [Google Scholar]
- 24.Harrell FE: Statistically Efficient Ways to Quantify Added Predictive Value of New Measurements. Stat. Think. 2020. [Google Scholar]
- 25.Jelovsek JE, Amundsen C, Brookhart A, et al. : Developing a causal framework for identifying urinary urgency and urgency incontinence phenotypes – Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN). 2024. Available at: https://lurn.pages.oit.duke.edu/dags/refined-conditions.html#sec-ref-cond, accessed August 7, 2025. [Google Scholar]
- 26.Schüssler B ed: Clinical evaluation of the pelvic floor. In: Pelvic floor re-education: principles and practice. London Berlin Heidelberg: Springer; 1997; pp 42–48. [Google Scholar]
- 27.Hernan MA, Hernandez-Diaz S, Werler MM, et al. : Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology. Am J Epidemiol 2002; 155: 176–84. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
De-identified LURN II Urinary Urgency Phenotyping cohort data will be made available through the NIDDK Central Repository (https://repository.niddk.nih.gov/), per NIDDK’s data-sharing policy. DAGs and analytic code used to construct and evaluate the phenotypes are maintained on the GitLab site (https://lurn.pages.oit.duke.edu/dags/).
