PInChPitt Innovation Challenge 2020

Pitt Innovation Challenge 2020 Awardees

<< Back to competition home

LemurDx

Categories:
Mental health ADHD Diagnostics Machine learning Pediatrics Smart devices Software Wearables
Do you have any questions, feedback, or suggested contacts for the team?

LemurDx

A smartwatch activity monitor that combines sensors with machine learning algorithms to accurately measure hyperactivity associated with ADHD diagnosis.

Highlights

  • ADHD is the most common neurodevelopmental disorder of early childhood.

  • Diagnosing ADHD accurately is challenging due to a lack of objective measures.

  • LemurDx uses smartwatch sensors and machine learning algorithms to measure hyperactivity.

  • LemurDx has been developed and tested over the past three years and received NIH funding.

  • PInCh funds will allow us to collect additional data and refine our algorithms.

Problem

Attention-Deficit/Hyperactivity Disorder (ADHD) is the most common neurodevelopmental disorder of early childhood, affecting around 5% of children in the United States. Because there are no objective measures of hyperactivity (a core component of ADHD), diagnosing ADHD can be challenging and subjective.

Current standards for measuring hyperactivity are questionnaires completed by parents and teachers. It can be frustrating for families to be given a diagnosis based on subjective questionnaires. In addition, misdiagnoses (both overdiagnosis and underdiagnosis) can lead to unnecessarily medicating some children while others go untreated.

Solution

LemurDx is an application for smartwatches that uses sensor technology to measure hyperactivity, with the goal of improving the accuracy of ADHD diagnoses. The wide array of sensors embedded in wearable technology afford new opportunities to develop objective and accurate measures of hyperactivity.

LemurDx passively collects sensor data from built-in smartwatch sensors, including accelerometer and gyroscope. A child wears a smartwatch with the LemurDx application for up to one week, during which the data is transferred to a secure server. Machine learning algorithms are used to differentiate children with ADHD-hyperactive presentation or ADHD combined presentation from those with typical levels of activity.

LemurDx started as a collaboration between Pitt and CMU in 2017. We have since partnered with a local Pittsburgh tech company (NuRelm) and received early funding from the NIH. To date, we have collected data using LemurDx from 50 families, with promising results.

Competition

The competitive landscape analysis below summarizes key features of this solution, and current competitors working to solve similar healthcare problems.

Team

  • Oliver Lindhiem, PhD, Associate Professor of Psychiatry and Pediatrics, University of Pittsburgh. He has expertise in developing diagnostic tools for childhood behavioral health disorders, and has served as PI, Co-PI, or Co-I on seven NIH and foundation grants, including a Phase I SBIR and Phase I STTR.

  • Mayank Goel, PhD, Assistant Professor in the School of Computer Science, Carnegie Mellon University. He has extensive experience in designing, implementing, and testing new medical technologies.

  • Sam Shaaban, CEO. For the past 21 years at NuRelm, he has led the development and commercialization of numerous mobile and web-based products for commercial, non-profit, and university clients, as well as in-house software products.

Milestones

  • Milestone 1: We will recruit 100 new families (50 with ADHD; 50 without ADHD).

  • Milestone 2: We will collect sensor data from each child who will wear a smartwatch with the LemurDx app for one week.

  • Milestone 3: We will use our new data to refine our proprietary LemurDx algorithms.

Path to Impact Plan

The data generated from this PInCh project will be leveraged to prepare a competitive Phase II STTR application as our next step in the commercialization process. Our Phase II STTR application will include a complete commercialization plan.

IP protection will be sought as necessary, using the most appropriate protection mechanism. It is likely that the machine learning model parameters derived from future LemurDx work, and based on the training data which is collected, will fall under trade secret protection.

Dissemination strategies will include presentations at pediatrics conferences, networks of primary care providers, professional organizations, children’s hospitals, and health insurer networks. Focus groups will be conducted with providers and stakeholders at all key stages.

Eventually, LemurDx will be marketed to providers, and payers (insurance plans). Estimates on total annual ADHD spending range from $143 billion to $266 billion.

Frequently Asked Questions

What problem does LemurDx solve?

LemurDx provides an objective measure of hyperactivity, which is a core component of ADHD. Currently, providers rely largely on subjective questionnaires that are completed by parents and teachers.

How is LemurDx different from an actiwatch or FitBit?

One key difference is context. LemurDx goes beyond just measuring physical activity. It also classifies the environment/context within which the physical activity is happening.

Why is context so important?

Physical activity and movement are not enough to differentiate children with and without ADHD. Many very active children do not have ADHD.

Is LemurDx only useful for diagnostic purposes?

LemurDx also has the potential to measure response to medication and titration (figuring out optimal dosing).

Do you have any questions, feedback, or suggested contacts for the team?

*
*
*