# Development of a personalized infusion failure detection algorithm combining tissue counter pressure and blood glucose data for closed-loop diabetes management

> **NIH NIH R43** · DIATECH DIABETES, INC. · 2021 · $299,989

## Abstract

PROJECT SUMMARY/ABSTRACT
Advances in diabetes care technology over the past several decades, including improvements in continuous
subcutaneous insulin infusion (CSII) and continuous glucose monitoring (CGM), have led to more streamlined
treatments and reduced burden for patients with type 1 diabetes (T1D). As the field moves toward increasingly
autonomous, closed-loop systems for management of T1D, infusion set failure (ISF) remains a health risk to
patients and a barrier to the development of artificial pancreas systems. ISF, the disruption of fluid flow from
insulin pump to patient causing loss of glycemic control, affects an estimated 50% of pump wearers, placing
them at risk for hyperglycemia and life-threatening complications such as diabetic ketoacidosis. Because modern
insulin pumps are not equipped with a mechanism for monitoring infusion performance, patients may not know
an ISF has occurred until experiencing symptoms of hyperglycemia. There is no existing technology capable of
accurately identifying ISF in real-time, before dysregulation of blood glucose (BG).
Diatech Diabetes is addressing this unmet need with SmartFusion, an AI-based platform to monitor infusion
performance, immediately alert patients when ISF occurs, and help users infuse confidently and safely. Diatech’s
novel algorithm leverages tissue counter pressure (TCP) and CGM data to offer a superior method for ISF
detection. The company’s long-term vision is that by detecting ISF in real time, SmartFusion will improve patients’
glycemic control, prevent complications of hyperglycemia, and reduce excess medical costs. Diatech has already
collected preliminary data to show that TCP waveforms can feasibly be leveraged to differentiate healthy and
malfunctioned infusions. The goal of this Phase I SBIR proposal is to collect increasingly complex and
representative preclinical data to train and optimize the TCP-CGM algorithm and demonstrate proof-of-concept
that it can accurately detect ISF. Diatech will pursue this goal through the following aims: 1) collect and
characterize labeled in vivo TCP and x-ray imaging data of typical and malfunctioned infusions, 2) collect and
characterize labeled in vivo TCP and BG/CGM data of failure modes consistent with ISF in immobilized diabetic
swine, and 3) collect 3-day TCP and BG/CGM data from ambulatory diabetic swine with failure modes consistent
with ISF.
Successful completion of the project will result in a novel TCP-CGM algorithm to accurately detect ISF in closed-
loop systems. The algorithm will be further optimized through collection and integration of clinical data in Phase
II. SmartFusion will ultimately be integrated into insulin pumps and diabetes management platforms to provide
real-time ISF detection and personalized recommendations for ISF prevention. Further, by enabling new
avenues for control of closed-loop systems, SmartFusion can offer a significant leap forward for implementation
of artificial pancreas technology, leadi...

## Key facts

- **NIH application ID:** 10296116
- **Project number:** 1R43DK130036-01
- **Recipient organization:** DIATECH DIABETES, INC.
- **Principal Investigator:** John Clark Gray
- **Activity code:** R43 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2021
- **Award amount:** $299,989
- **Award type:** 1
- **Project period:** 2021-09-01 → 2022-12-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10296116

## Citation

> US National Institutes of Health, RePORTER application 10296116, Development of a personalized infusion failure detection algorithm combining tissue counter pressure and blood glucose data for closed-loop diabetes management (1R43DK130036-01). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/10296116. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*