Smart Toilet: A Disruptive Technology to Improve Health and Wellness (2020-2021)

Background

Precision medicine is an emerging approach aimed at tailoring medical care to the individual to achieve efficacious treatments and long-term wellness. As opposed to a snapshot of a visit to a doctor’s office, precision medicine is empowered by time-dense high-quality data and digital technologies.

Achieving the goals of health and wellness requires novel approaches to daily monitoring of analytical biomarkers from physiological fluids, which currently cannot be achieved using noninvasive approaches. Our daily excreta – feces and urine – are rich in latent data. In addition to biochemical markers, stool physical characteristics can contribute to the diagnosis and management of many acute and chronic GI conditions. Gastrointestinal (GI) diseases burden the U.S. healthcare system more than heart disease and significantly impact morbidity and mortality as well as quality of life. The current approach of patient self-reporting on bowel movement is limited by subjectivity and is particularly unreliable in vulnerable populations.

In 2019-2020, this project team created the “Smart Sampling Toilet” – a novel device that has the potential to transform healthcare as a noninvasive source of individualized biological data. This device employs hands-off extraction to collect samples of human excreta that can later be used for diagnosis of wellness and disease.

In its first year, the team focused on the engineering development, quality control, refinement, business and regulatory strategy of the Smart Sampling Toilet platform. This year, the team will build upon their technical achievements to investigate the physical characteristics of stool samples within the same platform. This analysis will contribute to real-time diagnosis and management of GI diseases and help inform the business strategy for this new technology. 

Project Description

Building off the work of last year’s team, this project team will develop integrated inline sensors for real-time categorization of bowel movements and further leverage the Smart Sampling Toilet platform.

A fundamental advantage of the Smart Sampling Toilet design is the immobilization of the fecal specimen in the toilet plumbing environment, allowing for physical characterization of fecal specimens outside the purview of the user. The integration of sensors and machine learning with this platform will facilitate development of a novel tool for precision medicine and transform the sampling toilet into a health analytics platform.

This proposed sensor-enabled Smart Sampling Toilet will ultimately enable the establishment of wellness baselines for individuals with deviations, triggering health interventions and specimen extraction for separate biochemical assays at specific time intervals with no user intervention nor privacy concerns associated with a camera in a bathroom.

To achieve these goals, team members will participate in three main activities, organized in distinct but interrelated thrusts:

  1. Engineering integration of sensors and data collection with fecal specimens: Team members will select, implement and collect data from commercial devices (e.g., endoscopic camera, optical transmission system, volume measurement devices) with both surrogates and fecal specimens on a testbed developed at the Duke University Center for WaSH-AID. Biochemical analysis will be implemented to ensure these modifications do not impact the quality of the specimen extraction.
  2. Development of machine learning and algorithms for classification of bowel movements: Students will analyze data from the sensors and aim at the automatic extraction of information with respect to the type and nature of an excretion event.  Image analysis classification and time series analysis will be implemented.
  3. Development of a business strategy for the technology: Students will advance the development of a business case for the technology. The product will include both hardware and data service models. This thrust will define the product description and will research the market requirements, in particular potential first adopters, such as long-term care facilities and Continuing Care Retirement communities.

These three thrusts are interdependent and will require team members to adjust the analytics and machine learning approach based on the engineering data and the requirement and constraints identified by the use cases.

Anticipated Outputs

In-lab data collection and analysis; oral presentations; report

Timing

Fall 2020 – Summer 2021

  • Fall 2020: Begin reading phase; start weekly meetings; begin research implementation and data collection
  • Spring 2021: Continue data collection and analysis; develop a business model for product description and cost margin; submit final presentations and reports to the team
  • Summer 2021 (optional): Continue activities on tasks to be defined by the outcomes of previous two semesters

See earlier related team: Smart Toilet: A Disruptive Technology to Improve Health and Wellness (2019-2020).

 

Image: Katie Sellgren and colleagues, courtesy of Center for WaSH-AID

Katie Sellgren.

Team Leaders

  • Geoffrey Ginsburg, School of Medicine-Medicine: Cardiology
  • Sonia Grego, Pratt School of Engineering-Electrical & Computer Engineering
  • Katelyn Sellgren, Pratt School of Engineering-Electrical & Computer Engineering

/graduate Team Members

  • Sandhya Kal, Master of Engineering Mgmt-MEG

/undergraduate Team Members

  • Nicolas Decapite, Electrical & Computer Egr(BSE), Computer Science (BSE2)
  • Shehzan Maredia
  • Jackson McNabb, Electrical & Computer Egr(BSE), Computer Science (BS2)
  • Matthew Murray, Economics (BS), Statistical Science (BS2)
  • Alina Suarez
  • Bryent Takayama

/yfaculty/staff Team Members

  • Krishnendu Chakrabarty, Pratt School of Engineering-Electrical & Computer Engineering|Arts & Sciences-Computer Science
  • Holly Dressman, School of Medicine-Molecular Genetics and Microbiology
  • Jeffrey Glass, Pratt School of Engineering-Electrical & Computer Engineering
  • Brian Stoner, Pratt School of Engineering-Electrical & Computer Engineering