After institutional review board approval, we used a cross-sectional study design with sampling from regions surrounding 4 PLUS research centers (Washington University in St. Louis [201912088], Loyola University Medical Center [LU212615], University of Pennsylvania [protocol number 833718], and University of Alabama at Birmingham [IRB-300003879]) to maximize variability across key novel Where I Go variables. We purposely excluded the University of Michigan PLUS Center where the build and prior informal iterative usability testing was performed. Women were instructed to enter real-time data in the Where I Go app whenever they were thinking about their bladder or using the toilet over 2 days. In addition, ecological momentary assessment (EMA) data were collected 3 to 5 times per day to capture contextual factors around toileting including stress, mood, and fluid intake in addition to potentially missed toileting events. EMA data were collected via equally spaced prompts across participants’ usual waking hours as reported in the app set-up process at first log-in.

As published in our first report [11], we estimated that a minimal sample size of 40 women would be needed to evaluate an a priori cut-point score of 74 (representing the 70th percentile of SUS scores or a grade of B-) [13] on the SUS. Community-dwelling women were recruited through flyers, advertisements in local community centers, contacting participants from previous studies, social media, and word-of-mouth. Women were eligible if they (1) were 18 years and older; (2) were fluent in written and spoken English; (3) were able to stand or walk independently without assistance by another person (use of cane or walker allowed); (4) owned a smartphone (Android or iOS) on which calls could be made; (5) had downloaded at least 1 app in the past 6 months; (6) were willing to respond to Where I Go prompts or texts and input data about toileting behaviors for 48 consecutive hours, during which they were required to leave their home at least once, but not to change time zones; and (7) agreed to 2 in-person visits at the study site. Women were excluded if they had a physical or mental condition that prevented them from completing written questionnaires and interactions with Where I Go; lived in an institutional setting (ie, skilled nursing, long-term care, or rehabilitation center); were pregnant or had a neurogenic or congenital bladder condition; and were unable to independently use the toilet.

After consent, each participant downloaded and installed Where I Go onto her smartphone, and then entered basic demographic information and typical “wake” time and pinned her home and any other regularly frequented locations. Formal data collection began with a check-in notification the morning following the download, which was scheduled automatically based on the inputted “wake time.” Starting with this “wake time,” each woman was instructed to continue using Where I Go for 48 consecutive hours. Participants were provided with a modest honorarium for their involvement.

Where I Go was built at the University of Michigan’s Center for Health Communications Research (UM invention registration OTT Ref No: 2019‐292) using the Cordova dual platform (The Apache Software Foundation; simultaneous build for use on mobile phones of either iOS or Android operating systems).

As reported in detail elsewhere [11], Where I Go provides each woman with an opportunity for (1) real-time reporting and (2) EMA reporting of the last few hours. Real-time reporting allows the woman to report planning for toileting throughout her day, The app starts with an option to tap the opening screen of “I’m thinking about my bladder.” Once tapped, these screens follow: “How strong is your urge to pee?” (slider scale 0‐100), “When are you planning to pee?” (response options “now” or “later”), and if later, “Why aren’t you going to pee now?” (response options list “don’t have to pee that bad yet,” “no place to pee,” “work duties,” etc). The screen then changes to an option for tapping “Now I peed.” Once tapped, these screens follow: “When did you pee” (response items right now, 10, 20, or 30+ min ago), “Where did you pee?” (with tailored locations options, the last one being “new location”), “How strong was your urge to pee?” (slider scale), “Why did you pee” (eg, “had to go,” “no time to go later,” “wanted a break,” and “I leaked” ), and “If you had to wait before you peed, why did you wait?” (“work duties,” “errands or traveling,” “long line for toilet,” “no delay,” “peed when I needed to,” etc). Response option of “other - fill in” was also offered when applicable. An “other” response offered the option for typing a comment into the fill-in box to indicate the reason for delaying toileting. At each of these interactions with the app, geospatial data were gathered automatically.

In addition to real-time reporting, Where I Go has built-in EMA, which the woman sees as pushed “Check-In” notifications. Push notifications for additional assessments offer users the opportunity to update information in the app if the recording of a toileting event was missed in real time. The screen series starts with automatic feedback to the woman about her prior real time reports, “These are the times you said you peed. Are there more you want to report?” and she can input additional toilet events and approximate time they occurred.

Within the EMA section of the app, additional screens with questions are “During this time period, I felt…” with slider scale response options anchored at each end with calm or stressed (0=calm to 10=stressed) and similarly not busy to busy.

EMA check-ins offered reporting during the look-back period of “Any pain with peeing,” and “Any leakage with peeing.” These screens were followed by “How much liquid did you drink?” with response options of selecting drink portion-size pictorials to add, and finally a screen “Did any of these drinks have these ingredients?” with response options “caffeine,” “artificial sweeteners,” “alcohol,” and “none of these.” Push notifications offered women the opportunity to record information on variables that are easier or more accurately reported through short look-back periods over the interval of time since the last notification rather than in real time. Geospatial data were also gathered automatically at each one of these EMAs.

The International Consultation on Incontinence Questionnaire (ICIQ) Overactive Bladder Module [14], a validated measure assessing overactive bladder and its impact on quality of life, was administered to participants at the end of the study. Responses to novel measures collected by Where I Go were compared with responses to these validated survey items on frequency, urgency, and incontinence.

Given the novel nature of the app and the exploratory nature of this report, there were no a priori rules for evaluating participant response rate to prompts. To assess the accuracy of the location-pinning function (home or away from home), we compared participants’ GPS longitude and latitude pinned location of “home,” set at baseline, to the longitude and latitude automatically captured when a participant interacted with the app to record a real-time toileting event at home. For participants with a missing or inaccurate home location pinned at baseline, their home location was replaced with the longitude and latitude of their first toileting event when they indicated that they were at home.

To illustrate if plausible values were captured across the wide variety of types of data collected by Where I Go, descriptive statistics were calculated for the daily number of toileting events, average urinary urge ratings, number of locations reported, locations of each toileting event, and reasons for delaying toileting. In addition, descriptive statistics were calculated on the EMA component of the app (check-ins), comparing a number of locations, stress ratings, busy ratings, leakage episodes, toileting locations, fluid intake volume and ingredients, and management strategies across day 1 and day 2.

Descriptive statistics were calculated to portray the frequency or lack of use of real-time and EMA survey elements. The ability to record toileting events in real time was assessed by comparing real-time recordings to the number of missed toileting events documented in the EMA section. We evaluated beverage intake response rates to prompts as at least 1 fluid intake recorded in a day, making the logical assumption that all women would have consumed at least a minimal amount of fluid on each day. Finally, the relationships between quantities measured within the app and the ICIQ items were assessed.

Descriptive statistics were calculated separately for day 1 and day 2 and reported as frequency and percentage for categorical variables and median and IQR for continuous variables. For comparisons of continuous variables (eg, daily toileting frequency and daily intake volume), Pearson correlations were calculated. For comparisons of a continuous variable across levels of categorical variables (eg, urge ratings at home vs away), box plots were used for qualitative visual assessment. All analyses were completed in the SAS Enterprise Guide (version 9.3; SAS Institute).

A total of 44 women completed the baseline survey with 48 hours of interaction with Where I Go. Participants had a mean age of 44 (range 21‐85) years, with 25/44 (56%) of participants identifying as White, 30/44 (68%) working one or more jobs, and 32/44 (73%) having a college degree or higher. Most participants had 2 or 3 toilets in the home (32/44, 73%), and most used an Apple phone (33/44, 75%) to interact with the app. Additional details on demographic information are provided in Table 1.

On both days 1 and 2, the median number of toileting events recorded in real time was 5 (IQR 3‐7 on day 1 and IQR 3‐8 on day 2) and most toileting events were recorded at the time of the event (112/140, 80%, on day 1 and 136/171, 80%, on day 2, Table 2). The most common toileting locations were home (85/140, 61%, on day 1 and 129/171, 75%, on day 2) and workplace (34/140, 24%, on day 1 and 25/171, 15%, on day 2). Most participants reported toileting due to a need to go (114/140, 66%, on day 1 and 153/171, 74%, on day 2), with work duties and errands or traveling as common reasons for delaying toileting as well as selection of “other” both when reporting toileting now and later (Table 2).

Real-time urge ratings were similar across day 1 (median rating 64, IQR 52‐81) and day 2 (median rating 62, IQR 46‐85) when participants reported toileting now (Table 2). Ratings were lower but similarly consistent across days when participants reported toileting later (day 1 median 54.5, IQR 34-74; day 2 median 55, IQR 38-71).

A total of 6 out of 44 participants (14%) on day 1 reported toileting via the EMA survey only and did not use the real-time function. Similarly, on day 2, four out of 44 participants (9%) used the EMA survey only. In contrast, 3 participants out of 44 (7%) on day 1 and 2 participants out of 44 (5%) on day 2 did not respond to any EMA notifications. The average number of responses to the EMA survey was higher on day 2 than on day 1 (day 1=3.2, day 2=4.3; P<.001). On the EMA survey, participants reported a median of 1 (IQR 1‐2) additional toileting event on day 1 and 2 (IQR 1‐2) additional toileting events on day 2 (Table 3).

A total of 15 out of 44 participants (34%) had a missing or incorrect pinned home location at baseline. Two participants out of 44 (5%) did not have a pinned home location at baseline and 13 out of 44 (30%) had a pinned home location that was set incorrectly. Of all toileting events that had location automatically pinned by the app as “Home,” 97% (103/106) of those events were self-reported as having occurred at home. However, for toileting events where participants’ location was automatically pinned as “Away,” only 60% (85/141) were self-reported by those participants as being away from home, potentially due to an incorrectly pinned home location.

Median number of beverage servings put into the EMA beverage section ranged from 3 to 4 per day, and the approximate median daily intake volume was 32 (IQR20-40) to 41 (IQR 20-60) ounces. While most participants did not report using strategies to manage their toileting events, those that did most reported toileting at certain times of the day due to their schedule (32/178, 18%, on day 1 and 21/228, 9%, on day 2) and before certain activities (26/178, 15%, on day 1 and 32/228, 14%, on day 2). Most participants reported low stress generally (median scores 1‐2 on a scale from 0=calm to 10=stressed) during the look-back periods.

When comparing urge ratings between “Home” and “Away,” urge ratings tended to be lower on average when participants were at home (Figure 1). This relationship was more pronounced for the “Now I Peed” survey using participants’ reported location (median rating 72, IQR 56-98, when away vs 61, IQR 41-84, when at home) and the “Thinking About My Bladder” survey using participants’ pinned location (median rating 58, IQR 41.5-74.5, when away vs 48, IQR 24-67, when at home). Daily fluid intake had a small to medium positive correlation with daily toileting frequency (day 1 r=0.3, day 2 r=0.24, Figure 2). In addition, participants typically reported higher urge ratings when they reported they would “Pee Now” versus “Pee Later,” (median 64, IQR 52-81 vs 54.5, IQR 34-74, on day 1; median 62, IQR 46-85 vs 55, IQR 38-71, on day 2; Figure 3).

Participants’ daily toileting frequency reported in Where I Go showed a small positive correlation with the daily urinary frequency item on the ICIQ (r=0.31 with day 1 toileting frequency and r=0.21 with day 2 toileting frequency, Figure 4). Most participants reported urination frequencies of 1 to 6 times and 7 or 8 times on the ICIQ frequency item (37/44, 84%); therefore, the relationship between toileting events recorded in Where I Go and ICIQ responses cannot be assessed precisely for those reporting urination frequency greater than 8 times per day on the ICIQ.