R. Fearn1, D. Ziring2, K. Solis1, I. Dorofeeva1, C. Landon3
111 Health and Technologies Inc., Tustin, California; 2Cedars Sinai Hospital, Los Angeles, California; 3Ventura County Medical Centre, Ventura, California
Intestinal stomas are frequently constructed by colorectal surgeons either electively or emergently. Perioperative high stoma output is associated with increased length of stay, readmission and renal failure.
Leakage results in reduced quality of life and skin complications. Patients and healthcare professionals struggle to measure stoma output accurately.
Previously, we demonstrated that a flexible external sensor, wirelessly connected to a patient's smartphone can accurately and reproducibly measure stoma output, but was limited by a high noise to signal ratio in the real world.
Here we demonstrate and validate the proof of concept for using thermal and capacitive sensing to determine stoma output volume and leakage.
To investigate thermal sensing, a thermistor array was adhered to the outer layer of a standard ostomy bag using a wired temperature visualization.
Simulated effluent was infused in 50ml aliquots up to 350ml at 37oC. Here a graph demonstrating the temperature change at a single thermistor and a "heat-map" of all thermistors is seen.
Capacitive sensing is used widely in industry to measure fluid level due to changes in capacitance in the presence of water even if the sensor is not in direct contact with the fluid.
To investigate its use in this setting, A commercially available flexible capacitive sensing unit was mounted on the outer layer of a standard ostomy bag.
A capacitive sensor was able to respond to an incremental increase of volume in the bag in aliquots of 50 mL up to 350ml.
We investigated the use of skin temperature as a biomarker comparing thermal and optical images (FLIR E60). In this series of 33 peristomal thermographs, skin complaints were identified with 93% sensitivity.
We therefore developed a concentric ring array of thermistors attached to a hydrocolloid for use in a stoma wafer.
Simulated output was injected beneath the wafer worn on a healthy volunteer. Leakage under the base-plate was detected as a sharp increase of temperature along the flow-path.
On the basis of this data an array of thermistors and capacitive sensors was built into a flexible polymer to integrate with a stoma bag with an additional array of thermistors in two concentric circles to be embedded into the wafer.
The integrated final designs incorporated into a 450ml 2-piece SmartBag and wafer are seen here:
In order to integrate thermistor and capacitive data it was necessary to use a static feed-forward neural network. Resilient backpropagation algorithm (RPROP) was used for supervised training of the multilayer perceptron.
The multilayer perceptron architecture designed for volumetric prediction based on capacitive sensors signal.
In a real-world validation ostomates wore the SmartBag for a total of 51.6 hours. In this time 3139ml of stoma output was measured. The relationship between measured and predicted stoma output volumes can be seen.
Mean (standard deviation) stoma output across all participants was 53.3 (42.7) ml/h. The algorithm predicted output of 3332ml in the same time period, providing a predicted mean stoma output per participant of 56.3 (45.5) ml/h.
Predicted vs Measured Stoma Output in Real Ostomates
The cloud architecture was designed so that data from the sensors can be sent in real time to a patient’s or clinician’s mobile device. All these data can be accessed through a dedicated HIPAA compliant cloud service.
A companion smartphone software application was developed to support use of the integrated sensing device. This incorporates administrative tools, data visualization and access to educational and care resources. Alerts can be set allowing the patient, or the clinicians who are monitoring the patient, to take preventative action to avoid complications.
Conclusions: In this study of a novel connected "smart" stoma bag, we have demonstrated the proof of concept of using thermal and capacitive measurements and have derived and validated a neural network algorithm to accurately determine volume and leakage.
Complications in patients with intestinal stomas frequently result from challenges in monitoring this group. When combined with application-based resources and a telehealth communication link to the clinical team, remote monitoring provides the opportunity for early intervention and preventative education.
The evidence so far strongly suggests that this innovation, a remote monitoring system that can securely communicate instantaneous and cumulative output data to the patient and their clinical team should contribute to improved outcomes in this population.