closing the loop on medical device systems simulation

19-20 April, 2018 Twin Cities, MinnesotaHow Systems Engineering Can Reduce Cost & Improve Quality #hwgsec

Closing the Loop on Medical Device Systems Simulation

Marc Horner, Ph.D.

Technical Lead, Healthcare

ANSYS, Inc.


A. Healthcare Industry Overview• Systems Engineering for Medical Devices

• Digital Systems Prototyping

• Regulatory Update

B. Insulin Pump Example• Background

• Drug Delivery Sub-system Model• Kink Detection Modeling

• Virtual Patient Modeling

C. Conclude

Outline


Healthcare Industry Overview


Electric Smarter Connected

* Cogizant, How the IOT is Transformng Medical Devices, May 2016

Today’s Medical Devices are Increasingly


Electronic Control

Embedded Software

Sensors

Actuators

Components


Electronic Control

Embedded Software

Sensors Operating Conditions

Safety Requirements

Operational Profiles

Actuators

…are ultimately part of a system


Electronic Control

Embedded Software

Sensors Operating Conditions

Safety Requirements

Operational Profiles

Actuators

…with complex interactions.


Challenge: System Complexity

• Understand and optimize performance

• Eliminate late-stage integration failures

• Improve collaboration among design disciplines

• Enhance or reduce physical testing

• Accelerate innovation

Digital System Prototyping


Design Still Happens in Silos

Each discipline has its’ own set of tools, processes, and expertise.


Systems Engineering: A Unifying Approach

Reduced Order

Model Creation

System Model

Interoperability Embedded Software

IntegrationCo-simulation

with 3D Physics

Language-Based

Modeling

Multi-Domain

Model Libraries


Reduced Order

Model Creation

System Model

Interoperability Embedded Software

IntegrationCo-simulation

with 3D Physics

Language-Based

Modeling

Multi-Domain

Model Libraries

Thousands of Built-in

Component Models

Industry Standards

for System Modeling

Interfaces with all ANSYS 3D Physics 3rd Party System

Modeling Tools

ANSYS SCADE

and More

Systems Modeling– Model Flexibility, Reusability, and Interoperability


System Simulation & Digital Twins

Simplorer

Model-Based Software Engineering

Model-Based Systems Engineering

System/SW Architecture

System Safety Analysis

System Architecture

The ANSYS Portfolio

3D Physics Simulation

RO

M


FDA Analyses of Product Recalls

“failures in product design and manufacturing process control caused more than half of all product recalls”

1. from FDA Report “Understanding Barriers to Medical Device Quality” (2011)

“The most frequent causes for recalls are related to device design,software, and non-conforming material or component issues.”

http://www.fda.gov/downloads/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDRH/CDRHTransparency/UCM388442.pdf

http://www.fda.gov/downloads/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDRH/CDRHReports/UCM277323.pdf

19-20 April, 2018 Twin Cities, MinnesotaHow Systems Engineering Can Reduce Cost & Improve Quality #hwgsechttp://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm209042.htm

“many of the reported problems appear to be related to

deficiencies in device design and engineering”

Infusion Pump Safety

http://diyabetimben.com/diyabet-muzesi/

1963

http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm209042.htm


Insulin Pump Model


What is Diabetes?

diabetes image from https://i.ytimg.com/vi/SCCb5GqhnrI/maxresdefault.jpgPump image from http://www.medtronicdiabetes.com/products/minimed-530g-diabetes-system-with-enlite

• Insulin is a hormone created by the pancreas. It is required for sugar molecules (from the food you eat) to move inside cells. Patients with diabetes either do not produce insulin (Type 1) or do not use insulin the right way (Type 2).

• Insulin pumps replace the function of the pancreas by injecting insulin under the skin throughout the day.

https://i.ytimg.com/vi/SCCb5GqhnrI/maxresdefault.jpg


HBM EM

Catheter

Manufacture

Antenna

Design/Optimization

Electro-

Mechanical

Design

Drug

Delivery

Catheter Kink

Testing

Catheter Pressure

TestingCatheter

Occlusion ROM

Embedded Software -

Display

Drop Test

Embedded Software

- Controller

Components, Components, Components


Drug Delivery Sub-System


EMBEDDED SOFTWARE

ELECTRICAL

FLUIDIC+MECHANICAL

Drug Delivery Sub-SystemM

OD

EL D

OM

AIN

S


display controller

power

electronicshydraulics

kin

k d

ete

ctio

n

volu

me

de

live

red

INFO

RM

ATI

ON

FLO

WDrug Delivery Sub-System


Pressure-Flow Analysis of Tube Bending

ROM

Family of Structural Fluid Simulations


Q

bend

DPkink

REQUIREMENT:if Psensor > Pthreshold

then (imotor = 0;warn patient; )

Psensor = DPcath + DPcannula + DPkink

Insulin Pump – ROM Behavior

DPcath = 𝟏𝟐𝟖𝝁𝑸𝑳𝒄𝒂𝒕𝒉

𝝅𝑫𝒄𝒂𝒕𝒉𝟒

DPcann = 𝟏𝟐𝟖𝝁𝑸𝑳𝒄𝒂𝒏𝒏

𝝅𝑫𝒄𝒂𝒏𝒏𝟒


Drug Delivery Sub-SystemV

IRTU

AL

PATI

ENT

Couple the device model to a patient predictive model

of glucose metabolism.


Virtual Patient ModelTwo-compartment insulin model

Two-compartment glucose model

Insulin effectiveness

• The patient model requires a mathematical representation of the relevant physics.

• The model should capture insulin metabolism as well as the ability of insulin to effect glucose uptake into cells.

• Researchers and industry typically rely on pharmacokinetic/pharmacodynamics (PK/PD) modeling to represent these processes.

*Kanderian et al., Identification of Intraday Metabolic Profiles during Closed-Loop Glucose Control in Individuals with Type 1 Diabetes, J Diabetes Sci and Tech , Vol. 3 (2009).


Virtual Patient ModelTwo-compartment insulin model

Two-compartment glucose model

Insulin effectiveness

*Kanderian et al., Identification of Intraday Metabolic Profiles during Closed-Loop Glucose Control in Individuals with Type 1 Diabetes, J Diabetes Sci and Tech , Vol. 3 (2009).

CLINICAL DATA


Dig

ital

Tw

in

Execution Layer

S

Monitor & Visualization

Ph

ysic

al A

sse

tDigital Twin Predictive Platform

MIoT Connector

S

ROM Builder Data Layer - Device

Simulation DataReal Test DataPLM/ERP/etc.S

ROM Builder Data Layer - Patient

Simulation DataElectronic Medical RecordPatient Physiology


▪Chronic diseases and the aging population are placing significant strain on healthcare systems, motivating the need for more effective medical technologies.

▪ The risk (and failure) of medical devices has increased since incorporating new technologies and functionality, much of which is related to embedded software.

▪Systems modeling can improve the robustness and safety of today’s medical devices.

▪Digital twins for implanted devices that include models of human physiology (enabled by computer modeling) can improve treatment outcomes.

Conclusions

closing the loop on medical device systems simulation

Documents