Three different electronic pumps were evaluated for their fluid flow capabilities: a motor pipette pump, a peristaltic pump, and a diaphragm pump. The pumps were evaluated based on 6 different criteria: rate of emptying, control over pressure, ability to provide both suction and irrigation pressure, reusability, portability, and cost. Using a Pugh matrix, an engineering decision-making tool,⁷ the peristaltic pump was determined to be the most effective pump for our device since it can generate both suction and irrigation pressure while also preventing pump contact with fluid by use of replaceable and easily cleanable tubing.

To perform initial verification of the device’s functions, tests were performed against a phantom bladder model. The phantom bladder is modeled by an emptied-out stress ball, which we have previously shown to have high compliance with filling under low pressures,⁸ as do most medically and surgically optimized neurogenic bladders. The phantom bladder was placed under 5 cm H₂O of external pressure to mimic resting abdominal wall pressures, and a catheter was then tunneled and sealed through the phantom bladder 0.5 cm from the bottom to mimic catheterizable channel entry site at the base of the bladder. Phantom stoma site was 16 cm from entry into the bladder.

We constructed models for 10, 12, and 14 Fr catheters. Bladder emptying rates and remaining volumes of fluid were tested against phantom bladders filled with 150, 300, or 600 mL of water. Comparisons were made via the catheter either by passive drainage or with aid of the pump. Irrigation was tested by instilling empty bladders with 120 mL of water using a 14 Fr catheter, mixing fluid back and forth 5 times, and then draining the bladder using either a syringe or the pump. Five independent tests were performed for each condition, and Student’s t tests were used to calculate significance between conditions. Emptying efficiency was defined as the fractional percentage of drained volume by starting volume. Preliminary usability testing was also performed to simulate performance of the device in a CIC environment. Eight college student users attached the device to a pre-filled phantom bladder model, the time for suction and irrigation was measured, and user feedback was collected, comparing the performance of the pump and syringe. In the feedback survey, the users were asked to rate the ease of use of the pump compared to the syringe on a 1 to 5 scale with 1 being very easy to use and 5 being very difficult.

The mechanical components of the device were assembled using a 3D-printed housing, an ANKO A200 peristaltic pump (ANKO Products, Bradenton, FL) with ¼ in internal diameter × ⅜ in outer diameter latex tubing. The ANKO A200 pump was selected due to its high flow rates up to 10 mL/s and easy attachment of tubing. The 3D-printed box was designed using SolidWorks (Dassault Systèmes, Vélizy-Villacoublay, France) with a center hole for the pump motor and threaded M4 screw holes to secure the peristaltic pump head to the outside of the box. Four pegs were positioned at the base of the box for the solderable breadboard and circuit components to be secured and grounded to the plastic housing. The box was printed using a Form 3 SLA printer and clear resin (Formlabs, Somerville, MA). When fully assembled, the dimensions were 14.22 × 8.89 × 6.71 cm, as seen in Figure 1.

An Arduino Nano (Arduino, Turin, Italy) was used to control the main components of our system. The Nano powered and controlled 2 LEDs and a buzzer for patient feedback on the device operation status and received input from 2 rocker switches that the patient uses to control pump direction. A L293D H-bridge controller (PIXNOR, Shenzhen City, China) was used to control the direction and speed of the pump. The Nano also monitored the current through the pump and was capable of automatic shutdown if the current passed a threshold current of 390 mA, which is set to 30% above the expected load on the pump. A high current load suggested (although not formally tested) the pump was under high pressure, either due to a kink in the tubing or due to sucking on the phantom bladder wall. Three 9V rechargeable lithium batteries were used to power the system, and the pump could be operated continuously for roughly one hour before recharging was required.

Testing results showed that for all bladder volumes and catheter sizes, the pump had >98% emptying efficiency and drained between 1.6 and 6.3 times faster than passive drainage. In general, the pump emptying rate continually increased with catheter gauge, and emptying rate increased as initial bladder volume increased, both as expected (Figure 2). The largest improvements in drainage rate between pump drainage and passive drainage were found in 10 Fr catheters and between small bladder volumes. Across all catheter sizes, the pump vastly outperforms passive drainage with regard to the amount of fluid able to be drained from the bladder. With regard to irrigation testing, instilling 120 mL into bladders, performing 5 back-and-forth mixes, and subsequent draining took a mean time of 76.2 ± 3.56 seconds using the pump, compared to 107.4 ± 4.83 seconds using a syringe (p<0.001). During verification testing, unintended automatic shutdown secondary to high current load did not occur.

Usability testing indicated that the pump had a significant decrease in the total time on task when compared to the syringe irrigation method. Users were able to complete the usability tasks 38% faster from initiation of first catheter connection to final disconnection and storage of the pump compared to the syringe technique (pump: 138 ± 32 seconds; syringe: 224 ± 51 seconds; p=0.001). The users rated the pump as 3 points easier to use on a 5-point scale, where 1 is the easiest to use and 5 is the most difficult, compared to the syringe (median score: pump 1; syringe 4).