Online research communities are becoming increasingly common and provide a low-cost resource to researchers and low burden opportunity for volunteers. These platforms are prime for studies utilizing self-reported measures to explore novel, potential risk and protective factors and can be particularly useful to understand the usability and performance of newly developed research instruments in additional populations. The objectives of this online bladder health survey were to (1) assess the response rate and feasibility of using ResearchMatch, a national online volunteer registry established by the National Institute of Health for the recruitment of study participants, (2) compare the demographics, health characteristics, and lower urinary tract symptoms (LUTS) of survey respondents to published prevalence estimates from population-based surveys, and (3) pilot the newly developed Lower Urinary Tract Research Network-Symptom Index (LURN-SI 29) in women, including its ability to detect known associations between demographics, health characteristics, and LUTS (i.e., convergent validity).  Data obtained in this study will be used to inform future population based studies and intervention trials.

A cross-sectional survey of women recruited through ResearchMatch was performed to assess bladder health/LUTS, known LUTS-related risk and protective factors, as well as other health characteristics and behaviors. At the time of study launch, ResearchMatch included 98,300 women 18 years of age and older (480 women 80-100 years-old) and 630 transgender participants. ResearchMatch permits contact of 1,500 participants via email at one time with additional contacts made until sample size is reached. 

The web-based survey was developed by content experts in the field of bladder health and LUTS, and edited and revised by a team of investigators. LUTS were assessed using the new Lower Urinary Tract Research Network-Symptom Index (LURN-SI 29).[1] Additional validated questionnaires were used when possible but several understudied areas lacked validated items. New questions were drafted and revised by our research team taking into account the available published and unpublished literature in the field. We then tested these questions for 6th grade readability using MS Word and assessed content validity. The survey was entered into the secure web platform REDCap and logic was used to minimize participant burden with irrelevant questions. The survey consisted of 100 questions and took approximately 14 minutes to complete. A REDCap link was provided for survey completion.

Participant demographics, health characteristics and LUTS were summarized and compared to estimates from the Epidemiology of LUTS (EpiLUTS) study [2], which used representative population-based sampling of men and women >40 years from an internet based panel of US, UK, and Swedish residents (allowing for multiple contacts and incentives to improve response rates). Participant LURN-SI total scores and sub-scores were tabulated and compared by demographics and health characteristics. Total scores and sub-scores ranged from 0 (least severe) to 100 (most severe); p-values were calculated by the non-parametric Kruskal-Wallis test. Free text entry for some fields was used to “hear” participants’ voices on understudied topics (e.g., preferred terminology, effects of exercise).  Free text entries were analyzed using qualitative research methods, grouping of common language and meaning, and identification of themes.

A random sample of 18,609 of the 99,920 registered female and transgender volunteers between the ages of 18-100 received the request to participate in the survey. Of 18,609 volunteers emailed, 2,144 agreed to participate and 1,747 (ages 18-88 years) completed the survey for a response rate of 9.4% of those contacted, and a completion rate of 81.5% of those who agreed. The response rate to EpiLUTS was 59.6%. 

Table 1 shows participant demographics and LUTS characteristics among ResearchMatch volunteers compared with participants in EpiLUTS. ResearchMatch participants had a mean age of 44 years, most had completed some post-high school education, and were White/Caucasian, making the sample younger, more highly-educated with similar racial/ethnic diversity to the EpiLUTS population. Self-reported health, distribution of comorbid conditions and body mass index (BMI) were similar between the two studies while menopausal status and parity were not. ResearchMatch respondents were less likely to be postmenopausal and had lower parity consistent with their younger age. ResearchMatch respondents reported a high prevalence and bother from LUTS similar to EpiLUTS. 

Table 2 shows bivariate analysis of participant characteristics and LURN-SI total scores and sub-scores.  LURN-SI 29 total scores ranged from 0 to 77; median score (25th%, 75th%) was 17 (11, 26).  Notably, the Nocturia sub-score had the highest median score of 29 (14, 57).  LURN-SI 29 sub-scores as well as total score increased with age, BMI, medical comorbidity, parity, menopausal status, urinary symptom bother and each decrement in self-reported health, and varied with education and race, thus providing evidence of convergent validity.

Several novel areas of bladder health were evaluated with our survey and may be important in future studies; these include:

(1)	Respondents used over 30 different terms to describe urine leakage. Preferred terms included: “Had an accident” (24.0%), “Peed on myself” (22.3%) and “Urinary Incontinence” (19.7%). Less preferred terms: “Urine accident” (3.0%) and “Accidental urine loss” (5.6%). 4% of participants used “Other” terms (e.g. piddle, dribble, drips, drops, sneeze pees, weak bladder, old lady bladder, mommy bladder). 
(2)	24.7% of women reported that they experienced bladder pressure during certain activities (10.4% during jumping, 8.5% during running, 5.7% during fast walking, 5% during yoga, 4.6% during hard landings). 
(3)	12.9% of participants avoided physical activity and 15.2% interrupted a workout due to LUTS.  
(4)	Adaptive behaviors were employed during exercise by 32.7% of participants (pad use 18.7%, dark clothing 8.6%, frequent toilet use 16.3%, sitting down to exercise 3.4%).

Response rates among self-identified volunteers in the ResearchMatch web-based platform for the study of bladder health and LUTS (9.4%) were similar to representative sampling of the population using postal and telephone recruitment [3] but were less than internet based incentivized sampling in EpiLUTS (59.6%). ResearchMatch participants were younger, more educated, and less parous than EpiLUTS participants. Despite this, the spectrum of LUTS was similar indicating this platform is an appropriate resource to pilot and inform the development of novel study instruments in the area of bladder health and LUTS. Respondents report a high prevalence of bladder symptoms consistent with other LUTS measures used in population-based studies. Furthermore, the novel LURN-SI 29 was capable of detecting expected associations with demographics (i.e., age) and health characteristics (i.e., comorbidity and BMI), supporting its use in future etiologic and health outcomes studies.

Internet-based recruitment for a bladder health survey is feasible, with similar response rates as population based postal and telephone recruitment studies. Participant demographic characteristics varied but the prevalence and spectrum of LUTS were comparable to other population-based samples. For future studies, selective sampling is feasible in ResearchMatch to allow a more nationally representative racial/educational distribution although its use is still limited in terms of achieving diversity in socioeconomic status and related domains affecting internet access and computer and health literacy. Internet-based population use of the novel LURN-SI 29 provided data on its performance outside of LUTS patients, demonstrating its ability to detect expected associations with demographics and health characteristics. For future studies refining and validating the LURN-SI 29, use of patient preferred language and bladder symptom assessment in the context of activity may improve performance of the measure.

Cella D, Smith AR, Griffith JW, Flynn KE, Bradley CS, Gillespie BW, Kirkali Z, Talaty P, Jelovsek JE, Helfand BT, Weinfurt KP; LURN Study Group. A new outcome measure for LUTS: Symptoms of Lower Urinary Tract Dysfunction Research Network Symptom Index-29 (LURN SI-29) questionnaire. Neurourol Urodyn. 2019 Aug;38(6):1751-1759. doi: 10.1002/nau.24067. Epub 2019 Jun 21.
Coyne KS, Sexton CC, Thompson CL, Milsom I, Irwin D, Kopp ZS, Chapple CR, Kaplan S, Tubaro A, Aiyer LP, Wein AJ. The prevalence of lower urinary tract symptoms (LUTS) in the USA, the UK and Sweden: results from the Epidemiology of LUTS (EpiLUTS) study. BJU Internat. 2009 Aug;104(3):352-60
Czajka JL, Beyler A. Background Paper. Declining Response Rates in Federal Surveys: Trends and Implications. Mathematica Policy Research Report. June 15, 2016. https://www.mathematica.org/our-publications-and-findings/publications/declining-response-rates-in-federal-surveys-trends-and-implications-background-paper