artificial intelligence and machine learning – see …

24
T ORONTO – Michael Garron Hospi- tal (MGH) is a busy community fa- cility serving over 400,000 people in 22 neighbourhoods across East Toronto. The hospital is currently expanding its site with an eight-floor tower and an ad- ditional three-floor extension that will add 500,000 square feet of clinical space to the medical centre. With the opening of their new building (known as the Ken and Marilyn Thomson Patient Care Centre), expected to open sometime in late 2022, the hospital will ex- pand its use of Connexall’s enterprise plat- form. It’s the largest construction project in the hospital’s history, with an estimated cost of $560 million. At the same time, MGH is engaged in master planning via ‘Project Imagine’, an ef- fort to re-design healthcare and create an in- telligent, “connected” hospital. One of the hospital’s key challenges was to determine how to connect clinicians, staff, patients and equipment in real-time, mak- ing it possible to communicate instantly and locate people and resources as soon as possi- ble. To achieve this strategically, MGH chose to leverage its existing partnership with Connexall, an industry leader in clinical communication and collaboration. Simplified approach and system archi- tecture: Hospitals utilize a multitude of sys- tems and associated devices. Connexall in- tegrates its solutions in healthcare facilities without having to make any additional in- vestment to update their current infra- structure or replace their existing vendor relationships. The company’s sophisticated enterprise-grade platform and suite of so- lutions are vendor and device agnostic, supporting all types of systems and devices that hospitals use now and plan to use in the future. One such solution within Connexall’s portfolio is the MobileConnex application, an extension of its enterprise grade plat- form. The app will be a standard tool aid- ing clinical and corporate staff via inte- grating information from different systems into single mobile device interface from MGH leverages Connexall to enable SMART expansion CONTINUED ON PAGE 2 VoxNeuro, a Canadian neuroscience and health tech company, has developed a 30-minute, non-invasive Cognitive Health Assessment that records tens of thousands of data-points from a patient’s brain waves while they perform a series of tasks on a computer. The company’s med- ical software then rapidly analyzes, transforms and compares the patient data to a database of healthy cognitive function. SEE STORY ON PAGE 8. VoxNeuro’s data-driven future of brain health It’s the largest construction project in the hospital’s history, with an estimated cost of $560 million. PHOTO: STEVEN KIM INSIDE: Technology fuels burnout In an Ontario Medical Association survey, technology, and specifically issues involving tech usage, were at the top of the list of reasons physicians cited for burnout. Page 6 Improving access to care According to Statistics Canada, roughly 4.6 million Canadians in 2019 didn’t have access to a family healthcare provider, an issue the booming telehealth industry is trying to address. Page 10 Securing health information While we have seen impressive advancements in virtual care and digital health tools during the pandemic, a recent survey found that 74 per cent of Canadians are worried about the privacy and security of their personal health information. Page 12 Publications Mail Agreement #40018238 ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE PAGE 18 CANADA’S MAGAZINE FOR MANAGERS AND USERS OF INFORMATION SYSTEMS IN HEALTHCARE VOL. 26, NO. 8 NOVEMBER/DECEMBER 2021 DIAGNOSTIC IMAGING PAGE 14

Upload: others

Post on 07-Apr-2022

11 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

TORONTO – Michael Garron Hospi-tal (MGH) is a busy community fa-cility serving over 400,000 peoplein 22 neighbourhoods across East

Toronto. The hospital is currently expandingits site with an eight-floor tower and an ad-ditional three-floor extension that will add500,000 square feet of clinical space to themedical centre.

With the opening of their new building(known as the Ken and Marilyn ThomsonPatient Care Centre), expected to opensometime in late 2022, the hospital will ex-pand its use of Connexall’s enterprise plat-form. It’s the largest construction project inthe hospital’s history, with an estimated costof $560 million.

At the same time, MGH is engaged inmaster planning via ‘Project Imagine’, an ef-

fort to re-design healthcare and create an in-telligent, “connected” hospital.

One of the hospital’s key challenges wasto determine how to connect clinicians, staff,patients and equipment in real-time, mak-ing it possible to communicate instantly and

locate people and resources as soon as possi-ble. To achieve this strategically, MGH choseto leverage its existing partnership withConnexall, an industry leader in clinicalcommunication and collaboration.

Simplified approach and system archi-tecture: Hospitals utilize a multitude of sys-

tems and associated devices. Connexall in-tegrates its solutions in healthcare facilitieswithout having to make any additional in-vestment to update their current infra-structure or replace their existing vendorrelationships. The company’s sophisticatedenterprise-grade platform and suite of so-lutions are vendor and device agnostic,supporting all types of systems and devicesthat hospitals use now and plan to use inthe future.

One such solution within Connexall’sportfolio is the MobileConnex application,an extension of its enterprise grade plat-form. The app will be a standard tool aid-ing clinical and corporate staff via inte-grating information from different systemsinto single mobile device interface from

MGH leverages Connexall to enable SMART expansion

CONTINUED ON PAGE 2

VoxNeuro, a Canadian neuroscience and health tech company, has developed a 30-minute, non-invasive Cognitive Health Assessment thatrecords tens of thousands of data-points from a patient’s brain waves while they perform a series of tasks on a computer. The company’s med-ical software then rapidly analyzes, transforms and compares the patient data to a database of healthy cognitive function. SEE STORY ON PAGE 8.

VoxNeuro’s data-driven future of brain health

It’s the largest construction projectin the hospital’s history, with anestimated cost of $560 million.

PHO

TO:

STEV

EN K

IM

INSIDE:

Technology fuels burnoutIn an Ontario Medical Associationsurvey, technology, andspecifically issues involving techusage, were at the top of the listof reasons physicians cited forburnout. Page 6

Improving access to careAccording to Statistics Canada,roughly 4.6 million Canadians in2019 didn’t have access to afamily healthcare provider, anissue the booming telehealthindustry is trying to address. Page 10

Securing health informationWhile we have seen impressiveadvancements in virtual care anddigital health tools during thepandemic, a recent survey foundthat 74 per cent of Canadians areworried about the privacy andsecurity of their personal healthinformation. Page 12

Pub

lication

s Mail A

greem

ent #

40018238

ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE PAGE 18

CANADA’S MAGAZINE FOR MANAGERS AND USERS OF INFORMATION SYSTEMS IN HEALTHCARE VOL. 26, NO. 8 NOVEMBER/DECEMBER 2021

DIAGNOSTICIMAGING

PAGE 14

Page 2: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

which they can communicate and coor-dinate responses.

“It’s great that you’re not tied down to adesk to see what’s happening,” said NicoleGagne, telecommunications coordinatorat MGH. “Our staff can be at the bedsideand on the floors and still see all of the in-formation.”

Moreover, the combination of Connex-all’s enterprise platform and MobileCon-nex is helping caregivers handle multiplealerts and hand them off in order of im-portance to the most appropriate care-giver. If that caregiver is unavailable, it willescalate to the next logical person. Mobile-Connex will also ensure that the message isreceived and acted upon.

On MGH’s plans for the future, SandySaggar, SVP Innovation, Analytics and CIOat Connexall shared, “We are excited aboutMGH enabling their clinicians with our En-terprise Connected Hospital Solution, witha mobile-first strategy utilizing MobileCon-nex to prioritize all the alerts and messagesthey receive on any given shift. We need tohelp clinicians manage their workload, es-

pecially with burnout on the rise, with toolsto enable their day-to-day workflows inturn improving the patient experience.”

MGH has been using technologiesfrom Connexall since 2009, but the latestproject will substantially expand the scopeand capabilities.

“Since we began our partnership be-tween MGH and Connexall in 2009, wehave greatly benefited from the use of theConnexall products to enable streamlinedcommunications amongst our clinicalteams for patient care delivery,” com-mented Amelia Hoyt, CIO at Michael Gar-ron Hospital. “We continue to work to-gether to build additional communicationpathways with the use of Connexall prod-ucts, like MobileConnex, in support of ournew Ken and Marilyn Thomson PatientCare Centre opening in 2022.”

Previously, MGH used Connexall on itsinpatient floors with nurse call integrationto their mobile devices, allowing responseteams to manage critical events such ascode blue and code pinks. With the open-ing of their new tower, the hospital will ex-pand its use of the company’s enterpriseplatform. This will include providing sup-

port to outpatient clinics including daysurgery, complex continuing care and thematernal newborn areas.

Doing more with Connexall: Connex-all’s vast portfolio of solutions are inte-grated into critical systems such as Nurse

Call, Patient Monitoring, RTLS, EMR andmany more.

For instance, Connexall will receivealarms from MGH’s Nurse Call system anddeliver these alerts to the appropriate staffmembers via the MobileConnex app.MGH staff will have the ability to respondto events, including accepting and escalat-ing, and initiate a patient room call to clar-ify the situation promptly.

Similar to the Nurse Call solution, thePatient Monitoring solution will prioritizeand escalate alerts through Connexallwhenever the system receives a critical

alarm. In the new tower, Connexall willprocess the desired alarm types, based onstaff and clinician feedback and deliverthem to the appropriate team members viaMobileConnex. These alarms will be pri-oritized and communicated based on theworkflow designed for each alarm type.

An example of the RTLS integration isConnexall’s patient wandering solution,which “will be particularly helpful for pa-tients with dementia, with sensors and tagsenabling alerts – all showing up on Mobile-Connex for the closest and most appropriatestaff to be able to respond to,” said Gagne.

The deployment of MobileConnex willallow for meaningful utilization of infor-mation transmitted by various systemsand enable faster response to changes inpatient and system performance. The op-portunities to scale automation and en-hance data usage are promising for futureworkflow enhancements.

Greater insights and analytics: For hos-pital management, an important aspect oftheir quality improvement strategy will beits ability to automatically collect and pro-duce actionable analytics.

“We’re very excited about this for qual-ity improvement and reporting,” said LiniaShaji, manager of Health Records andTelecommunications. “Currently, we col-lect statistics manually which is labour in-tensive and can be inconsistent. Connexallwill automate all of this. It will be ideal forquickly assessing which units require themost staff attention for alerts, alarms androunds and how long staff members aretaking to respond to incidents.”

With real-time statistics, the hospitalwill be able to implement quality improve-ment programs in an efficient and effectivemanner. By automating the collection ofmetrics, the hospital will be able to createcustom analytics, with comparisons of dif-ferent units – such as medical, complexcare and palliative care – enabling it tomatch its resources with high-demand ar-eas better.

Digital resiliency: “The hospital is under-going substantial transformation of it digitalmaturity through this and other technologyupgrades which will better position us forthe future of care delivery,” said Hoyt.

Gagne also noted that there is a signifi-cant change management and trainingprogram slated to begin, as many new sys-tems will be deployed in the new towernext year. Hundreds of clinicians will bemoving over from the existing building tothe new tower. Connexall will be involvedin helping these efforts to ensure there isseamless user adoption. It’s all part of theshift towards modernized healthcare deliv-ery, which depends on fast and accurateinformation.

Art DirectorWalter [email protected]

Art AssistantJoanne [email protected]

Publisher & EditorJerry [email protected]

Office ManagerNeil [email protected]

Address all correspondence to Canadian Healthcare Technology, 1118 Centre Street, Suite 204, Thornhill ON L4J 7R9 Canada. Telephone: (905) 709-2330.Fax: (905) 709-2258. Internet: www.canhealth.com. E-mail: [email protected]. Canadian Healthcare Technology will publish eight issues in 2021. Featureschedule and advertising kits available upon request. Canadian Healthcare Technology is sent free of charge to physicians and managers in hospitals, clinicsand nursing homes. All others: $67.80 per year ($60 + $7.80 HST). Registration number 899059430 RT. ©2021 by Canadian Healthcare Technology. Thecontent of Canadian Healthcare Technology is subject to copyright. Reproduction in whole or in part without prior written permission is strictly prohibited.

Send all requests for permission to Jerry Zeidenberg, Publisher. Publications Mail Agreement No. 40018238. Returnundeliverable Canadian addresses to Canadian Healthcare Technology, 1118 Centre Street, Suite 204, Thornhill ON L4J 7R9.E-mail: [email protected]. ISSN 1486-7133.

CANADA’S MAGAZINE FOR MANAGERS AND USERS OF INFORMATION TECHNOLOGY IN HEALTHCARE

Volume 26, Number 8 Nov/Dec 2021

Contributing EditorsDianne [email protected]

Dianne [email protected]

Dr. Sunny MalhotraTwitter: @drsunnymalhotra

Norm [email protected]

Dave [email protected]

CONTINUED FROM PAGE 1

Digital Health Week is coming, Nov. 29-Dec. 5Join us as we connect with health care organizations, clinicians, government, industry and patients to celebrate, build relationships and show support for digital health. Share your story and add your voice to the conversation that takes place every year using #ThinkDigitalHealth.

Learn more about Digital Health Week: infoway-inforoute.ca/digital-health-week

MGH leverages Connexall to enable its SMART Hospital expansion

Connexall’s platform and suiteof solutions are vendor anddevice agnostic, supporting alltypes of systems and devices.

2 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1 h t t p : / / w w w . c a n h e a l t h . c o m

Page 3: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

ca.medical.canon

With a streamlined workflow in a single setting, you no longer need to transfer the patient between rooms, improving patient care and boosting productivity. Switch easily between CT and angiography systems to diagnose, treat or verify. The Alphenix 4D CT seamlessly integrates our flexible Alphenix interventional system with our advanced Aquilion CT imaging suite into one versatile solution.

Available now with our world first High-Definition Detector technology advancing interventional radiology, neurology, and cardiology procedures. Click to watch video.

Book your Virtual Experience Today [email protected]

One Room. One System. Many Possibilities.

A hybrid angiography CT solution.

Page 4: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

N E W S A N D T R E N D S

Rogers for Business, the last ofCanada’s big three telecommuni-cations companies to exit the pag-ing business, is now working with

Hypercare Inc. to promote the latter’s smart-phone-based communications solution tohealthcare systems across the country.

Pager subscriptions have fallen steadilyover the years, and as network technologyadvances with 5G, Rogers for Business sawan opportunity for the healthcare indus-try to leverage the benefits of this nextgeneration technology as they migratefrom the legacy paging platform. With thisnew agreement, Rogers and Hypercare of-fer pager customers an easy way to transi-tion to a more comprehensive communi-cation solution over smartphones and tobetter meet current and future needs ofmodern medicine.

Dr. Joseph Choi, co-founder and chiefmedical officer of Hypercare, says it’s longoverdue. “I can tell you real examples ofhow patients have been harmed due tomissed or delayed pages. It’s been a prob-lem for decades, and I’m excited thathealthcare is embracing these digital toolsthat we already use in our everyday lives. Ithink it will really improve the care we canprovide our patients.”

Hypercare has advantages over most al-ternatives to paging, said company CEO,Albert Tai. “Our solution isn’t just pagerreplacement. It’s a complete communica-tion solution for healthcare across differ-ent settings.

“Our core functionality is the ability tosend text messages, pictures, videos, andother multimedia in a fully encryptedHIPAA and PHIPA compliant manner.Our messaging feature allows providers toconvey the urgency of the message to en-sure the message gets through. Every mes-

sage has receipts, so the sender knows ex-actly when a message was read for maxi-mum situational awareness. We manageon-call schedules and allow clinicians toeasily connect to the on-call provider accu-rately and easily. Hypercare even allows forautomated escalation of care in critical sit-uations ensuring you get a timely re-sponse. We integrate into different IT sys-tems such as EHR and nursing call systemsto make an all-encompassing platform forclinical communication.”

Hypercare’s cloud-based solution is in-dependent of a hospital’s existing commu-nication networks, allowing for redun-dancy in case a customers’ network experi-ences issues such as downtime or a ran-somware attack. Moreover, it’s one of the

few solutions on the market that enablesproviders to connect across multiple caresettings for their patients. This capability isideal for Ontario Health Teams, which en-able hospitals, primary care physicians,skilled nursing facilities, home care andother healthcare organizations in the com-munity to coordinate care and streamlinepatient transitions between care settings.

Although transitioning to new technol-ogy can be daunting, the move to Hyper-care has been easy, said Tai. “Since mostusers are already familiar with instant mes-saging, it isn’t a very big jump. Wi-Fi infra-structure is decent in most hospitals, andsince it runs on mobile devices, it canswitch seamlessly to cellular data connec-tions when coverage is poor. Hypercare’s

encryption enables users to securely usetheir own device for clinical work.”

With Rogers and Hypercare working to-gether, the solution provides the hardware,network, and software, all in one package.

Dr. Choi is also optimistic about the fu-ture. “This is an exciting opportunity tohelp transition healthcare organizations tomore modern and effective communica-tion technologies. I truly believe it willhelp healthcare providers be more efficientand provide better care for patients. As weknow, the system is already stretched to itslimit and providers are becoming increas-ingly burnt out and frustrated with theirjobs. Anything to make the lives of health-care workers easier is desperately needed.”

Many forward-thinking hospitals are al-ready using Hypercare, including MichaelGarron Hospital, and Huron Perth Health-care Alliance consisting of four hospitalsacross Southwestern Ontario. The com-pany has also recently expanded into theUnited States with a hospital system in up-state New York.

“It’s been great to see the amount ofgrowth and adoption into different caresettings such as mental health, palliativecare, nursing homes, and hospices allow-ing them to quickly communicate inter-nally but also to be connected to vital re-sources such as specialists at the hospi-tal,” said Tai. “It’s really been an incredi-ble journey.”

Asked about the impact of the change,Serena Xie, Rogers director of WirelessProduct Management, replied “We’re veryexcited about this work. This solution willallow our customers to truly improvehealthcare outcomes for both patients andproviders. With the arrival of 5G technol-ogy, these types of improvements are onlygoing to accelerate.”

Rogers and Hypercare to offer digital solution to healthcare industry

Pictured, left to right: Albert Tai, chief executive officer of Hypercare; Dr. Joseph Choi, chief operating offi-cer and chief medical officer of Hypercare; Natalie Ellis, senior product manager at Rogers CommunicationsInc.; Serena Xie, director, Wireless Product Management at Rogers Communications Inc.

BY GARY BUFFETT

Kingston Health SciencesCentre (KHSC) has part-nered with the StreetHealth Centre (SHC) andthe University Hospitals

Kingston Foundation (UHKF) to bringinnovative ultrasound services to wherethey are most needed by vulnerable pop-ulations in Kingston.

SHC provides care to vulnerablepopulations by operating low-threshold,barrier-free access to primary care andaddiction services for those who arehomeless or precariously housed, mayhave been incarcerated, and may beaffected by or at risk of acquiringHepatitis C.

“One challenge faced by StreetHealth’s Hepatitis C program relates therequirement for clients to undergo diag-nostic liver ultrasound to support thedevelopment of a treatment plan,” saysBrittany Couto, manager of StreetHealth. “However, there is a very highno-show rate for this diagnostic proce-dure, with as many as eight out of ten

patients not showing up for their ap-pointment. This leads to delays with thetreatment plan and inefficient use ofhealthcare resources.”

SHC’s Hepatitis C program is one of20 provincially funded teams in Ontarioproviding access to education, testingand treatment for individuals who havedifficulty accessing main-stream healthcare services.KHSC partnered with theStreet Health Centre to de-liver the ultrasound imag-ing at its Barrack Street siterather than at the hospital.This innovative collabora-tion will bring the care tothe patient’s location forthe best outcome and bestuse of resources.

“We knew that weneeded to look at a newway to deliver care that wasmore patient centred forthis population,” said MikeMcDonald, vice president ofPatient Care and chief nurs-ing executive at KHSC.

“Partnerships such as this demonstrateKHSC’s ability to be a hospital beyondwalls that delivers complex, acute andspecialty care where and when it isneeded most. This important initiative isa great example of placing our expertsand services in the community, and ac-tively involving community partners as

members of our care teams to improveaccess to care and improve the health ofour community.

“As KHSC strives to improve thehealth of the community through in-novation, partnerships with our health-care partners, and donors from thecommunity through UHKF, we can

improve patient outcomes andmake our communities healthier,”said McDonald.As part of the Kingston Commu-nity Health Centre, the StreetHealth Centre is open 365 days ayear to deliver health servicesthrough a wrap-around caremodel for people who face barri-ers accessing mainstream healthservices. SHC began as a needleexchange program in 1992, andhas evolved into a multi-servicemodel, providing health, diseaseprevention, primary care, andtreatment services.

Gary Buffett is Strategic Communi-cations Advisor at Kingston HealthSciences Centre.

Hospital and community services partner to bring care to vulnerable

h t t p : / / w w w . c a n h e a l t h . c o m4 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Brittany Couto, manager of Street Health, and Mike McDonald, vicepresident of Patient Care and chief nursing executive at KHSC, led apartnership to bring innovative ultrasound services to where they are mostneeded by vulnerable populations in Kingston.

Page 5: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …
Page 6: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

N E W S A N D T R E N D S

BY DIANNE CRAIG

When we think of physi-cian burnout, the manychallenges relating todelivering and manag-ing care during the

COVID-19 pandemic seem to be the mostlikely reasons for the jump in reports ofburnout this year. To a large extent, theyare. Contributing significantly and addingfuel to the fire – according to a recent sur-vey by the Ontario Medical Association(OMA) – is technology, and specifically is-sues involving tech usage, which were atthe top of the list of reasons physicianscited for burnout.

These include stressors such as the timenecessary to complete large amounts of re-quired documentation associated with us-ing different digital health tools – some-times resulting in duplication – and thefact that many physicians have to spendmore time completing documents thancaring for patients. Also contributing toburnout are frustrations when differentsystems don’t work seamlessly together,and the multiple steps often required tolog in to reach key data, which can lead tofeelings of ‘click burnout’.

The OMA, which represents Ontario’s43,000 plus physicians, medical students,and retired physicians, revealed 72.9 percent of physicians surveyed said they expe-rienced some level of burnout in 2021, upfrom 66 per cent the previous year whenthey were polled on burnout in 2019. Justover one third – 34.6 per cent – reportedeither persistent feelings of burnout orfeeling completely burned out in 2021 – upfrom 29 per cent in 2020. Presumably,much of this jump is attributable to thedemands of treating a huge influx of pa-tients with COVID-19. Dr. Adam Kassam,president of the OMA, observes “We dobelieve this is highly attributable to thepandemic. All the physicians I’ve spoken tohave expressed burnout.”

The pandemic has been a key stressor.“The scale and severity of the pandemichas stressed healthcare workers in signifi-cant ways,” says Dr. Kassam, who indicatesthat the lack of respite time has made thischallenging situation even more stressfulfor physicians to cope with.

Asked how serious physician burnouthas become, Dr. Kassam said the situationis ‘increasingly dire’. “There are physicianson the cusp of retirement that may con-sider exiting the profession earlier than an-ticipated. Generally, the rising burnoutsentiment is a challenge that has been fu-eled by the pandemic,” says Dr. Kassam.

The issue of physician burnout is notnew. Just prior to the pandemic, the OMAestablished The Burnout Task Force with amandate to drive system-level changes andprovide support to members. Even then,nearly one third of Canadian physicianswere reporting high levels of burnout, ac-cording to the OMA, which acknowledgedmany other healthcare workers – especiallynurses – were also experiencing burnout.

Some physicians indicated they feltweighed down by the ‘burden of technol-ogy’: When he spoke with Ontario physi-cians, Dr. Kassam said he was made awareof the frustrations they were feeling and

the “burden of technology.” That burdenincludes issues such as the time and ef-fort needed to complete large, and insome cases, seemingly excessive amountsof documentation associated with usingtechnology tools and systems, and frus-trations with systems that don’t workseamlessly together.

Dr. Kassam reveals “Many physiciansare spending six hours a day working onelectronic charting – in addition to theirclinic hours.”

Following the survey, the OMA madefive recommendations designed to address

physician burnout, the first and third ofwhich relate to technology usage:

• Reducing and streamlining documentation;• More work/life balance through flexible

work arrangements;• Making digital health tools a seamless

part of physicians’ workflow, including byensuring different systems can speak toeach other;

• Support for physician wellness at theirworkplaces; and

• Fair, equitable compensation for allwork, including administrative work thatcannot be reduced.

The need to reduce and streamline doc-umentation has long been a big issue fornurses as well. According to Dr. Lynn Na-gle, the founding president of the Cana-dian Nursing Informatics Association, andeditor-in-chief of the Canadian Journal ofNursing Leadership, nurses can spend 20 to50 per cent of their time dealing with doc-umentation. “There is so much paper, andso much duplication,” she says.

Document duplication is increasinglyproblematic and can lead to inefficiencies:Dr. Kassam, who acknowledges some dupli-cation is necessary for legal requirements,also cites the amount of duplication as a keyproblem. “There is a growing concernabout duplication and carrying forward ofdocumentation. It would be good if wecould try and have physician and patient in-teraction more emphasized than the takingdown of information.”

The issue of duplication is increasinglyproblematic for nurses as well. While Dr.Nagle also notes sometimes “duplicativeeffort is required in an organization,” shesays that in some settings, “nurses are

asked to document in the organization’sinformation system and then to documentagain for the purposes of reporting to ex-ternal organizations.”

Nurses call for streamlining of informa-tion and more standardized approaches:“We need a unified approach to nursingdocumentation in this country. Even if weare using different vendors, the informa-tion should be streamlined, and includejust essential, relevant information in the(patient) care encounter,” says Dr. Nagle.

“If we’re going to computerize, take ad-vantage of that and don’t just replicate all

the old paper documents. The informationshould be pulled forward and updated,”she adds.

“In Toronto, 10 different teaching hos-pitals have different document practices.Some are still doing paper documenta-tion,” says Dr. Nagle. “We have a prepon-derance of hybrid records – a mix of paper

and electronic. So often things are missedbecause somebody forgets to look at eitherthe paper or the electronic document.

“Often, she adds, “we have to go lookingfor things.”

Some physicians are becoming weary ofwhat Dr. Kassam calls ‘click burnout’:Speaking from his own experience, Dr. Kas-sam said, “I work at a rehab hospital. Oftenwe don’t know what’s happened at anotherhospital when patients have had work doneat different hospitals.” He explained that of-ten multiple steps need to be taken, fromlogging in and finding the patient’s number,to getting to the key patient informationneeded. “I call that ‘click burnout,” he says.

Solutions should make digital healthtools a seamless part of physicians’ work-flow: Asked to elaborate on how the OMArecommendation for ‘making digitalhealth tools a seamless part of physicians’workflow, including by ensuring differentsystems can speak to each other’ might beachieved, Dr. Kassam offered the following

specific recommendations regarding theusage of digital health tools:

• Seamless access: Ensure physicians canseamlessly access patient records andshare patient health information amongcare providers, which would reduce theadministrative burden of obtaining pa-tient information;

• Inclusion in development: Involvephysicians as key partners from the start inthe procurement, design, implementationand ongoing optimization of digital healthtools to ensure usability;

• Ongoing training: Equip physicianswith comprehensive and ongoing trainingon digital health tools, beginning in med-ical school;

• Ongoing support: Provide physicianswith easily accessible and ongoing techni-cal support, which can free up time formore direct patient care; and

• Effective innovations: Explore techno-logical innovations including ‘virtualscribes’ and voice dictation that will reduceand simplify administrative demands, in-cluding billing administration.

Virtual scribes, which listen to a patientencounter virtually and give patients agreater sense of privacy by not havingsomeone else present in the room, have al-ready proven to be a valuable and neces-sary transition from in-person scribes forsome physicians during the pandemic.

Doctors and nurses want to be in-cluded at the table when new systems aredeveloped: Historically, observes Dr. Na-gle, “systems have been designed withoutthe aid of clinicians – up until the mid-2000s when that started to change.”

Dr. Kassam’s call for physicians’ ‘inclu-sion in development’ is echoed by Dr. Na-gle in her call for nurses to be included indevelopment of the systems they woulduse. “Nurses need to be involved in the de-velopment to ensure the systems work forthem. There’s a clear recognition now thatthat is important,” says Dr. Nagle. “For ex-ample,” she adds, “we haven’t askedenough nurses about what (information)gets captured in charts. Let’s capture theinformation once and use it many times.”

Dr. Nagle also suggests including somepatients ‘at the table’ for development ofsystems they would use.

Reducing the backlog of delayed ap-pointments, procedures is a current OMApriority: “Right now, we’re trying to navi-gate through the backlog,” says Dr. Kas-sam, referring to the massive backlog ofappointments and procedures delayed bythe pandemic. “In Ontario,” he said, “16million points of care have been delayed –things like mammograms, cancer surgeryand diagnostics, addiction screens andmental health screenings.”

The OMA survey has put a spotlight onthe dramatic increase in physician burnoutin the past year. While Ontario physiciansfocus on tackling the backlog of appoint-ments and procedures delayed during thepandemic, Dr. Kassam hopes these surveyresults and recommendations will moti-vate government, healthcare organiza-tions, and other healthcare stakeholders todrive implementation of new solutions.

“We need more integration. We want tomake physicians’ tools more integrated,more seamless,” says Dr. Kassam.

Dr. Adam Kassam Dr. Lynn Nagle

Technology use fuels rising burnout among physicians and nurses

Stressors include the timenecessary to complete thedocumentation associated withusing different digital tools.

h t t p : / / w w w . c a n h e a l t h . c o m6 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 7: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

SPONSORED CONTENTI

NN

OV

AT

IO

N

BY NORM TOLLINSKY

IllumiSonics, a medical device company withoffices in Toronto and R&D facilities at theUniversity of Waterloo’s PhotoMedicine Labs,has selected Microsoft Azure’s cloud servicesto accelerate the commercialization of itsground-breaking optical imaging technology.

illumiSonics’ Photoacoustic Remote Sensing(PARS) system uses lasers and the physics of light-matter interaction to identify the optical absorptionrates of different tissues within the body.

“The incredibly high-resolution images requirecloud capabilities for processing and storage,” saidLisa Carroll, Microsoft Canada’s National Public Sec-tor Lead. “This innovation in imaging needs a plat-form that delivers high availability, scalability and in-teroperability to enable rapid deployment, makingillumiSonics an ideal candidate for Azure.”

While attempting to source a powerful GPU-pow-ered server for its research, the company realized thatno amount of onsite hardware would accommodate itsrapidly scaling requirements. A cloud computing solu-tion using Microsoft Azure with its machine learningand assorted capabilities was the ideal solution.

“As we began to transition from a technology de-velopment company to a products and partneringcompany, our needs for a scalable platform that reg-ularly uses the latest technological advances in hard-ware without requiring regular updates or mainte-nance on our end became paramount,” said illu-miSonics Chief Executive Officer Rocky Ganske. “Weneeded a GPU-inclusive platform capable of scalingwith our business and delivering the real-time resultsof PARS microscopy to what will soon be a steadilygrowing number of customers.”

Real-time, non-invasive pathology is one of themost exciting potential applications for PARS, notesGanske. “Surgical excision with the goal of removingall of the cancer is an integral part of the treatmentfor most solid tumors. Deciding how much tissuearound the margin of a tumor to remove, however, isseldom obvious because differentiating healthy frommalignant tissue by eye or palpation can be highlyinaccurate. The current gold standard – post-opera-tive histological analysis – can take anywhere fromfive working days to several weeks for results to be re-turned after interpretation by a pathologist.”

If the post-operative analysis reveals unexcised can-cer tissue, follow-up surgery may be required. How-ever, using PARS imaging while the surgeon is still inthe operating room would provide the analysis in sec-onds and allow for the complete removal of a tumor.

“PARS technology can also be used as an ad-vanced ophthalmology tool capable of real-time,non-invasive eye imaging,” said Ganske. “An esti-mated 1.5 million Canadians suffer from vision lossand an estimated six million more have been diag-nosed with sight-threatening eye disease. Canadaspends $23.5 billion on vision loss every year eventhough approximately three-quarters of cases are ei-ther preventable or treatable.

“After cataracts, the three leading causes of blind-ness in Canada are age-related macular degenera-tion, diabetic retinopathy and glaucoma. These dis-eases have no cure and because they have no symp-toms and no technology can accurately screen forthem at their earliest stages, they are usually diag-nosed only after vision has been irreversibly affected.

“However, all three diseases have metabolic foun-

dations that with the right imaging tool could be de-tected pre-symptomatically through abnormal reti-nal blood oxygen saturation, metabolic rate of oxy-gen and melanin loss in the retinal pigment epithe-lium allowing early interventions to prevent or sig-nificantly slow vision loss. Currently, there is no clin-ical tool capable of providing these crucial details, sothere is a desperate need for a real-time, non-contact,safe, and accurate ophthalmology tool capable ofstructural and functional imaging.”

The selection of Azure cloud services made sensebecause of its wide adoption in hospital and researchcommunities and the familiarity of illumiSonics’ ul-timate customer base with its user interface.

Security was also an important consideration be-cause the patient data collected by PARS mi-

croscopy must comply with health information pri-vacy standards.

“The goal has been to adopt a platform that grantsus knowledge and control over which of our imagescan be accessed by a customer,” said Dr. Haji Reza, il-lumiSonics’ Founder and Chief Technology Officer.“We also need those images, which can be up to 70gigabytes in size, to be viewable without requiringdownload.”

“Building and validating HIPAA-compliant plat-forms takes a lot of investment, both in terms ofmoney and energy,” added Ganske. “That compliantmodules are readily available as part of Azure was abig plus.”

illumiSonics is looking forward to using AzureMachine Learning to accurately colourize image gen-eration, identify cancer cells and build libraries ofdifferent cancer images.

“As more people begin using PARS images for di-agnostics, Azure AI will play a major role in deliver-ing fast, accurate results,” predicted Ganske.

Another important component of Azure cloudservices is its Research Centre of Excellence on Cloud(RCEC) service, which helps institutions seamlesslymigrate their research workloads to the cloud, scaletheir research, collaborate with partner organiza-tions and simplify their computing needs.

“At its core, cloud storage enables users to storeand share data with ease and security, but Azure ismore than that,” said Microsoft’s Lisa Carroll. “Mi-crosoft’s Cloud for Healthcare provides trusted and

integrated capabilities that deliver automation forhigh-value workflows as well as deep data analysis forboth structured and unstructured data that enablehealthcare organizations to turn insight into action.

“There has been a tremendous increase in theamount of data collected in the healthcare sectorwith fragmentation and disparate data formatsacross providers and care teams. The global adoptionof Fast Healthcare Interoperability Resources (FHIR)has transformed the healthcare industry by enablingthe rapid exchange of medical information electron-ically between healthcare providers across a single,simplified data management solution.

“This summer, we introduced Azure API forFHIR, which offers a turnkey platform for a cloud-based FHIR service where healthcare providers can

quickly connect existing data sources such as elec-tronic health record systems and research databases.”

The Research Centre of Excellence on Cloud of-fering will leverage Azure’s capabilities to enable ac-celerated research outcomes in a secure and trustedenvironment.

“This research offering will allow our customersto easily collaborate and jointly share their experi-ences,” said Ganske.

Azure will also enable illumiSonics to make the appsand infrastructure it’s building available to customersthrough a subscription-based model. The first of theseapps, created with Microsoft’s help and using a prebuiltcustom vision API from Azure Cognitive Services, al-lows remote clinicians to view PARS images and auto-matically identify multiple types of cancer cells.

“Microsoft’s Web App development toolkit al-lowed us to easily deploy a fully functional medicalimaging viewer for our clinician collaborators, en-abling them to securely access large images from anydevice, zoom in and zoom out, and validate the qual-ity of our PARS images in comparison with state-of-the-art imaging modalities,” said Ganske.

“The hands-on support we received from differ-ent teams at Microsoft was unmatched and helped usmake the transition from no experience on Azure tobeing capable of running workloads and deployingapplications.”

illumiSonics anticipates commercialization of aclinical product in two or three years, but may beready to introduce a pre-clinical product even sooner.

illumiSonics’ PARS systems needed a solution delivering high availability, scalability and interoperability.

Microsoft cloud provides ideal platform for ground-breaking imaging technology

h t t p : / / w w w . c a n h e a l t h . c o m N O V E M B E R / D E C E M B E R 2 0 2 1 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y 7

Page 8: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

VoxNeuro technology pointing to the data-driven future of brain health

N E W S A N D T R E N D S

VoxNeuro is ready to changethe way brain health ismanaged and treatedacross the globe. Today’scognitive assessments arereliant on subjective mea-

sures like patient self-reporting and ob-served behaviour, opening the door forfaulty results. The types of brain health as-sessments using objective measures thatare available today, such as MRI or CT,only scan for structural issues (e.g. lesionsin brain tissue, brain bleeds, or damage tothe skull) and cannot provide an under-standing of a brain’s cognitive health.

Enter VoxNeuro. The Canadian neuro-science and health tech company has de-veloped a 30-minute, non-invasive Cog-nitive Health Assessment that uses elec-troencephalography (EEG) to record tensof thousands of data-points from a pa-tient’s brain waves while they perform aseries of tasks on a computer. The com-pany’s medical software, the CognitiveHealth Assessment Management Plat-form (CHAMP), then rapidly analyzes,transforms and compares the patient datato a database of healthy cognitive func-tion. This is a procedure similar to thatused with MRI and CT scans, which com-pares a patient’s brain structure to ahealthy brain structure database. The re-sult of VoxNeuro’s assessment is a reportthat objectively scores the patient’s cogni-tive function in the areas of attention andconcentration, information processing,and memory.

When used as part of existing protocolsto assess brain health, clinicians are nowarmed with a technology that providesthem with a reliable report to supporttheir clinical findings, decision making,and treatment of a patient’s condition. Theaddition of VoxNeuro’s technology makesan immediate impact, as clinicians are nolonger forced to leverage patient self-re-ported symptoms or subjective measuresof other observable behaviours in order todiagnose an issue. Instead, they can nowaccurately measure the neurophysiologicalhealth of a patient’s brain.

Dr. Abdel Kaleel is a neurologist and theowner of HeadworX Neurology Clinic inKitchener, Ontario, one of five sites acrosssouthern Ontario where VoxNeuro’s tech-nology is currently available to patients.

“VoxNeuro rounds out my practice verywell and acts as an extension to our clinicalexam, allowing us to develop a more real-istic and complete picture of the patient’sroad to recovery,” says Dr. Kaleel. “With theaddition of VoxNeuro, our patients canfeel very comfortable that we have themost current means of giving them a trueunderstanding of their condition and away to ensure their treatment is working.”He concludes, “With the objective dataprovided by VoxNeuro, we will be betterable to determine if a patient’s cognitiveconcerns warrant further investigations ormanagement options.”

With its success in southern Ontario,VoxNeuro now seeks to continue scalingacross North America. This fall, the com-pany launched in the USA through a net-work of diagnostic imaging centers in NewYork state. It will also soon be deploying atthe first of many neurology clinics in the

south-eastern United States where its as-sessment will become available in the next12 months. VoxNeuro plans to be opera-tional in approximately 100 clinics inNorth America by the end of 2022.

As VoxNeuro continues to grow, so doits ambitions to improve cognitive healthby taking brain healthcare into the future.Data-driven at its core, the company seeksto build the largest database in the worldon neurophysiological measures of cogni-tive function. The database will fuel un-precedented insights that can be leveragedto improve patient diagnosis and treat-ment – a process that has already begun.

With the use of machine learning,VoxNeuro’s database has enabled the iden-tification of neurophysiological biomark-ers that reflect whether a patient who hassuffered a concussion is experiencing alevel of cognitive dysfunction that requires

therapeutic rehabilitation. As the Cogni-tive Health Assessment becomes morewidely used, its normative database willgrow ever-larger. Machine learning willcontinue to be used to parcel out the keyfactors that affect brain health for certainconditions, such as dementia orAlzheimer's, and to establish biomarkersthat can help with early detection of theseconditions.

VoxNeuro's goal is to get to a pointwhere its normative database reaches ascale that would enable personalizedtreatment, informed by a data set of com-parable circumstances. The future ofVoxNeuro’s work is to be able to identifythe optimal treatment for individuals, de-pending on their specific situation andconditions, through continual collabora-tion with healthcare providers and sys-tems and the use of machine learning.

With the use of artificial intelligence, thecompany believes it may even be possibleto administer therapeutics at specific mo-ments where they’ll be most effective. Forexample, VoxNeuro’s scientists envisionthis being useful in identifying comatosepatients on the verge of emerging. In thiscase, notifications could be sent to at-tending healthcare personnel that a par-ticular patient’s prognosis is improving.Even information relevant to a patient’schances of being rehabilitated after theiremergence based on their cognitive statecould be sent.

As brain health becomes more empha-sized in healthcare, VoxNeuro aims to be-come an integral part of people’s lives. Thecompany sees a future where its CognitiveHealth Assessment can be incorporatedinto annual checkups to promote long-term well-being and prolonged quality oflife, be integrated into pharmaceutical tri-als to add an objective data point andprove effectiveness, and be providedthrough remote healthcare services for en-hanced and more convenient monitoringof brain health.

VoxNeuro is backed by 30+ years ofglobally celebrated, peer-reviewed scien-tific research, which has shown time andagain that their technology has the speed,accuracy and objectivity to offer nearlylimitless clinical applications. As of thissummer, VoxNeuro is an FDA RegisteredMedical Device Establishment, while theirmedical software, CHAMP, is listed as aClass II Exempt Medical Device in the USAand a Class II Medical Device in Canada.

VoxNeuro has developed a 30-minute, non-invasiveCognitive Health Assessment that uses electroen-cephalography (EEG) to record tens of thousands ofdata-points from a patient’s brain waves while theyperform a series of tasks on a computer.

PHO

TOS:

STE

VEN

KIM

h t t p : / / w w w . c a n h e a l t h . c o m8 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 9: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

SPONSORED CONTENTD

IA

GN

OS

TI

C

IM

AG

IN

G

GEHealthcare has developed a next-generation, cloud-based PictureArchiving and CommunicationSystem (PACS) that aims to help

healthcare organizations keep current with rapidlyevolving technology, relieve the pressure on capitalbudgets and free up IT resources. The subscription-based diagnostic imaging solution called Edison TruePACS encompasses diagnostic reading, exam work-flow, AI Applications, 3D post processing, enterprisevisualization and archiving in a single platform.

Currently available in the United States, GE Health-care is preparing for a Canadian launch in 2022.

What distinguishes Edison True PACS from tradi-tional PACS systems is how it offers radiology de-partments access to innovative artificial intelligencetools, claims Peter Eggleston, GE Healthcare’s GlobalProduct Marketing director for radiology. “We seethe cloud as a way to democratize access to AI be-cause the reality now is that the AI market is frag-mented. There are a lot of different vendors focusingon very narrow areas and the cost to an organizationto work with all these vendors is very high.”

GE Healthcare simplifies the adoption of AIthrough what it calls the Edison ecosystem. “We takeaway a lot of the work of finding, acquiring andstanding up the AI,” said Eggleston. “In the cloud, it’svery easy for us to provision new services, so we cangreatly shorten the amount of time it takes for an or-ganization to adopt AI. And, it’s only one vendor, GE,that you have to work with because we serve as a dis-tributor of AI tools through resell and distributionagreements we have with third parties.”

Currently, GE Healthcare offers AI algorithmsthat are available through Edison True PACS, but it’sworking with more and more companies that are de-veloping algorithms for a whole suite of AI use cases,noted Matthew Collingridge, General Manager ofGE Healthcare’s Digital Solutions Group. They in-clude for example, algorithms being developed tohelp identify and prioritize likely cases of pneumoth-orax, stroke and cancer, speed workflow and help en-hance diagnostic confidence.

“AI is gaining confidence among users,” saidEggleston. Aside from prioritizing studies that aretruly critical, “it’s good at doing things like measur-ing and quantifying physiological changes comparedto previous readings and identifying things thataren’t visually perceptible to humans because thecomparative data is so low or it’s so small. “Accep-tance of AI is pretty high because radiologists are al-ways concerned that they’re going to miss some-thing,” he added. “They’ll tell us, for example, thatthey feel like they’ve got a second set of eyes and thatsomebody is watching their back.”

A study by the American College of Radiologyfound that clinical adoption of AI by radiologistswent from virtually zero to 30 per cent from 2015 to2020. Other studies cited by GE Healthcare havefound that read times using AI applications are 26 percent2 faster for evaluating lung nodules and up to 60per cent faster for evaluating multiple sclerosis cases3.

Radiologists see the value of AI, “but haven’t hadan easy way of using it. That’s what the Edisonecosystem does,” said Eggleston.

“We’re excited about the next 24 months,” com-mented Collingridge. “We see a number of new tech-nologies emerging that have had early clinical valida-tion and we now need to push into broader access thatwe think will benefit not only the large institutions,

but also many of the rural and regional areas. We’reembarking on an initiative in Quebec, for example, tobring AI access to radiologists serving rural areas.”

GE Healthcare’s Edison True PACS is deployed onthe Amazon Web Services in the United States to en-sure a high level of security and data protection, sohospital IT departments don’t have to worry aboutmaintaining all of the security and intrusion proto-cols themselves. It also takes care of software up-grades, freeing IT staff to spend their time on otherenterprise challenges.

Neuro Imaging Winter Park, an imaging centre inFlorida, is one GE Healthcare customer that saw thevalue of a cloud-based PACS during their evaluationof the cloud based solution.

“With storage getting so expensive and constantlyhaving to switch out drives, maintain them and makesure the RAID arrays are in good shape, it just makessense to have it hosted offsite,” said Richard Duemm-ling, Neuro Imaging’s Chief of Business Operations.“If we have a fire or there’s a cyber attack on our net-work, we are likely in a better position with our dataon the cloud-based PACS. It’s also easier and less ex-pensive to deploy a new PACS because we don’t haveto buy new server hardware. We don’t have to main-tain it, or worry about operating software updatesand patches. There are a lot of hours involved inmaintaining an onsite server and when you go with acloud-based solution, IT infrastructure is effectivelymanaged as a Service. If you don’t have a big IT teamthat’s constantly dedicated to making sure every-thing is up to date, then you’re at risk.” A cloud-basedPACS is also ideal for hospitals forced to leverageoutsourced or distributed reading due to increasingworkloads and the shrinking radiologist workforce,said Eggleston because “if you’re initiating a relation-ship with a third party and everything is in the cloud,

it’s much easier to connect and scale that infrastruc-ture to another organization.”

Edison True PACS is available through software asa service (SaaS) or on premise in three subscription-based models: Edison True PACS Technologist, EdisonTrue Essentials and Edison True Professional. Hospi-tals opting for the on premise solution look after theirown storage. For the SaaS solution, “we manage theservers and everything that goes along with that suchas the third-party software, and since we managethem, we’re also responsible for keeping them up-dated, so the customer doesn’t have to do that,” saidEggleston. “When you’re on premise, software updatesoften incur charges or disruption. Usually, you have tofind a server to stand it up and test it because youdon’t want to just take it and put it into production, sothere’s time and cost there. With the SaaS version, youdon’t need to do any of that.” The on premise offering“is a toe in the water to help healthcare providers andhospitals move to a more native cloud environmentwhere things can be more seamless and secure, partic-ularly from the threat of cyber risks, and enjoy the ad-vantage of continuous improvement and upgradingof functionality and features,” said Collingridge.

Edison True PACS Technologist accommodatesthe QA or technologist workflow and is designed fororganizations that just acquire the images and out-source the reading. Edison True Essentials includes adiagnostic viewer and the basic requirements for di-agnostic reading. Edison True Professional is for or-ganizations that want all the latest technology, in-cluding GE Healthcare’s Open AI Orchestrator thatcan integrate AI tools into the radiologist’s workflow.

Cloud technology, interoperability and AI have allreached a state of maturity that was required for thelaunch of a cloud-based picture archiving and com-munication system, said Collingridge.

“Edison True PACS,” he predicted, “lays thegroundwork for what we think will be an excitingtwo or three years in Canada for machine learning.”

Edison True PACS1 aims to help healthcare organizations keep current with rapidly evolving technology.

GE Healthcare offers radiology departmentsaccess to innovative AI tools

Matthew Collingridge Pete Eggleston

h t t p : / / w w w . c a n h e a l t h . c o m N O V E M B E R / D E C E M B E R 2 0 2 1 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y 9

1. Edison True PACS Solution consists of the following products:Universal Viewer, Enterprise Archive, Centricity UniversalViewer Zero Footprint, Edison Open AI Orchestrator and 3rd partyAI applications

2. Source: Riverain Technologies Medical DeltaView FDA 510(K)Reader Study Results, 2011

3. Source: Lo, S. B., Freedman, M. T., Gillis, L. B., White, C. S., & Mun,S. K. (2018). JOURNAL CLUB: Computer-Aided Detection of Lung Nodules on CT With a Computerized Pulmonary Vessel Suppressed Function. American Journal of Roentgenology, 210(3), 480–488. doi: 10.2214/ajr.17.18718

Page 10: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

OAKVILLE, ONT. – SyngoCarbon is Health Canadalicensed and is now com-mercially available inCanada. The new platform

has been designed by Siemens Healthi-neers to provide clinicians with easy ac-cess to all relevant data generated in theprocesses of imaging and reporting.

While in theory, companies and re-searchers have been saying for years thatradiological and other imaging data havebeen unified, in practice, clinicians havefound obstacles to accessing the infor-mation they need.

Syngo Carbon has been created toquickly and easily reach the images anddata contained in various silos and tomake it easier for different areas towork together.

“With our daily clinical environment

being so highly fragmented, we face a de-mand to have clinical images and datasetsmeticulously prepared and enhanced us-ing AI, and to ensure knowledge can beshared,” said Niles Geminiuc, director ofDigital Health and Business Develop-ment, Siemens Healthineers Canada.

“Many clinical departments have dif-ferent image management systems forarchiving and drawing up findings. Thatmeans reports can differ, and the data is-n’t always available. Syngo Carbon en-ables our customers to break down barri-ers in Enterprise Imaging and Reporting.It provides holistic insights across depart-ments and responsibilities to meet certaindecisions along the patient pathway.”

Syngo Carbon protects existing in-vestments by integrating and continuingto use existing technologies and data. Italso incorporates existing Syngo solu-

tions conveniently and seamlessly. Itsopen data concept is another key aspectin making it possible to incorporatethird-party solutions, enabling datamanagement and archiving systems

from various departments within a hos-pital to be combined and consolidated.

Integrated AI tools help with the effi-cient performance of image-based diag-nostics. Unlike a traditional PictureArchiving and Communication System(PACS), the focus in this case extends be-yond only DICOM-based medical images.

Syngo Carbon works with all image-related data of relevance for diagnosticsand decision-making. Examples includeimages from pathology, endoscopy, andcardiology, and information generated aspart of a longer process, e.g. camera im-ages from surgery to document the con-dition of a wound.

Syngo Carbon is said to be the start-ing point for a new type of organization-wide system for imaging and reporting,which combines existing technologies toform a unified solution. Refinements aremade in close collaboration with users.Thanks to its open architecture, it canadapt to constantly changing conditions,modules and functions can be expanded,and the system can be scaled to suit cus-tomer needs.

For more information, visit siemens-healthineers.ca/syngo-carbon

Syngo Carbon provides enterprise-wide image reading and reporting

Syngo Carbon protects existinginvestments by integrating and continuing to use existingtechnologies and data.

N E W S A N D T R E N D S

The most recent Physicians inCanada report, commissioned bythe Canadian Institute for HealthInformation (CIHI), showed that

for the last 13 years, the growth rate of thephysician population has outpaced that ofthe general population. However, accord-ing to Statistics Canada, roughly 4.6 mil-lion Canadians in 2019 didn’t have accessto a family healthcare provider.

Canadians needing immediate health-care support, whether it’s because they livein rural or remote communities, have noluck on waiting lists, or don’t have timelyaccess, either delay care or end up in busyemergency rooms and healthcare clinics.Though the impact is not always obvious,putting off healthcare can have long-termeffects, especially to those with pre-exist-ing health conditions.

These are issues the booming telehealthindustry is trying to address. For example,Maple, a leading virtual care platform, hasseen tremendous growth and adoptionduring the pandemic, eliminating barriersto accessing healthcare for Canadians whodidn’t have a family healthcare providerprior to the pandemic and for those wor-ried about leaving home during a publichealth crisis.

Founded in 2015 with a vision to im-prove healthcare accessibility and relievepressure from hospital systems, the Mapleplatform has since supported long-termcare homes and independent clinics, im-plemented virtual emergency room sup-port for busy hospitals, and provided carefor patients without primary care doctorsin various provinces.

“After years of practice, the strains onour healthcare system became clear to me,”said Dr. Brett Belchetz, CEO and co-founder at Maple, and practicing physi-cian. “I knew there was a role that technol-ogy could play in improving the sustain-ability of our system and people’s health-care experiences across the country.”

Prior to the pandemic, there wereroughly 47,000 people on the Need a Fam-ily Practice Registry in Nova Scotia, a

provincial list of those looking for a pri-mary care provider. Last fall, the numberof people adding their name to the registrybegan to rise.

Dr. Gail Tomblin Murphy, vice presi-dent of Research, Innovation and Discov-ery and chief nurse executive at Nova Sco-tia Health, recognized an opportunity toleverage the adoption of virtual care thathad been occurring in the province duringthe first two waves of the pandemic andapply it to those on the registry.

In May 2021, Nova Scotia Healthlaunched VirtualCareNS, a proof-of-con-cept initiative where local family doctors

and nurse practitioners provided care de-livered through the Maple platform. Thisstarted with a few communities that hadthe highest number of unattached resi-dents, with access expanded over the sum-mer to communities in the northern andwestern parts of the province. A gradualexpansion is planned for the rest of theprovince beginning in December.

“Our Research, Innovation, and Dis-covery team has worked closely with theDepartment of Health and Wellness, Pri-mary Health Care and other health systempartners to find a solution to provide vir-tual care for those who have been on the

Need a Family Practice Registry thelongest,” said Dr. Tomblin Murphy. “Virtu-alCareNS is based on best evidence and ex-amples in other jurisdictions and providesanother option to access care for thosewho are without a family doctor or nursepractitioner.”

Thanks to the successful implementa-tion of VirtualCareNS, unattached pa-tients in Nova Scotia are now able to getvirtual diagnoses, referrals to specialists,prescriptions, and lab requisition forms allwithin the Maple platform, which can beaccessed from a smartphone, tablet, orcomputer.

So far, VirtualCareNS has seen promis-ing results, with over 1,600 consultationsover a four-month period.

One reason for this success can be at-tributed to Maple’s easy-to-use platform.“As evidenced by the thousands of Canadi-ans accessing care through our platformevery day, Maple was built with the patientand provider experience at the forefront,”says Peter Forte, director of Public SectorBusiness Development at Maple. “A big fo-cus of ours is to be seen as a partner andally in furthering the modernization ofCanadian healthcare, versus a standalonesolution that exists in parallel to the tradi-tional healthcare system.”

Hospitals, clinics, and long-term carehomes have started taking notice of thebenefits of a virtual care offering becauseof its ability to enable a more modern andtailored care experience for patients.Maple’s technology platform has beenused to take pressure off of hospital emer-gency departments, reduce wait times forspecialty clinics, and help with staff short-ages across care systems. In addition,long-term care homes have benefitedfrom this technology, leveraging Maple tofacilitate services like remote bedside con-sultations and provider-to-provider carecollaboration.

To learn more about Maple’s virtualcare technology solution for hospitals andother healthcare systems, visit their web-site at www.getmaple.ca

App helps Canadians access services while waiting for a family doctor

h t t p : / / w w w . c a n h e a l t h . c o m1 0 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 11: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

Our enterprise platform provides hospital-wide integration and interoperability and converts disparate environments - leveraging existing technologies - into a connected healthcare ecosystem that enables

To learn more, go to connexall.com/cahealth

Page 12: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

N E W S A N D T R E N D S

BY ABIGAIL CARTER-LANGFORD

While we have seen impressiveadvancements in virtualcare and digital health toolsduring the pandemic, a re-

cent survey found that 74 per cent ofCanadians are worried about the privacyand security of their personal health infor-mation. This is because cyber attacks thataffect the integrity of healthcare quite lit-erally hit Canadians where it hurts most.The risks associated with these attacks,such as the loss of access to or misuse ofone’s critical healthcare information, arepain points for patients across the country.

Both before and during the COVID-19pandemic, healthcare information hasbeen a focus of attack among bad actors.However, it is also clear that digital healthsolutions and virtual care, which is any in-teraction between a patient and a health-care provider that doesn’t involve directcontact, such as a video or phone appoint-ment, are here to stay. According to aCanada Health Infoway survey, 51 per centof patient-reported visits are now virtual,over double pre-pandemic levels.

This data reveals that while Canadianshave concerns about privacy and cybersecu-rity when it comes to their health informa-tion, they still want virtual care and digitalhealth services to exist. Someone who is cau-tious about privacy is not inherently anti-digital, but it does mean they expect their in-formation to be privacy protected. That iswhy it is more critical than ever for physi-cians, pharmacists and patients to choosetrusted providers within the digital carespace that prioritize privacy and security.

Take the prescription process, for in-

stance. e-Prescribing services like Pre-scribeIT modernize the prescriptionprocess by enabling physicians to electron-ically transmit a prescription directly froman Electronic Medical Record (EMR) tothe pharmacy management system of a pa-tient’s pharmacy of choice. e-Prescribingcan be a significant asset when in-personvisits with healthcare professionals are notpossible or recommended, which has cer-tainly been the case over the course of thepandemic. Prescribers and pharmacies canalso send secure clinical communicationsto each other through an integrated mes-saging tool, permitting them to quicklyalign on an appropriate course of actionand provide the best approach to medica-tion for their patients.

Of equal importance, e-prescribing de-creases the risk of privacy breaches due tofax transmission issues or unsecuredemail. When we compare the old paperworld to the new digital world, we oftenconsider the risks associated with the lattermuch more critically than the former. Faxis the most common transmission toolused within healthcare, yet fax machinesdo not have any security or protection fea-tures to protect patient data. There is norecord or accountability to show who hasseen the information that comes through afax machine, nor is there a way of guaran-teeing that data ends up in the right hands.

Conversely, digital tools like e-prescrib-ing allow physicians and pharmacists toknow where the information on a prescrip-tion has been, who has seen it, and whetherit has arrived at the intended destination.Before a care provider can use PrescribeIT,their identities and credentials are verified.The credentials they use must be unique,

and they must use two-factor authentica-tion to place an order. Security controls arein place to ensure that the informationshared is protected from inappropriate dis-closure and from tampering.

In the same way Canadians have cometo expect that their banking informationwill be kept safe, patients need the assur-ance that their healthcare information willbe kept safe end-to-end. The beauty ofsomething like e-prescribing is that data isencrypted, in transit and at rest, using in-

dustry standard cryp-tographic algorithmsto significantly im-prove data integrity,with antivirus andfirewall protectionsalso in effect to com-bat the risk of dataloss and manipula-tion or cyberattack.It is also importantthat a patient feelsconfident their in-

formation is being used only for the bene-fit of their own healthcare. Concerns thatthe information will be sold or otherwiseexploited for profit are real among Canadi-ans, making the choice of trusted digitalhealth providers and tools all the more im-portant. Services like PrescribeIT, for in-stance, have data protection woven intotheir DNA, and work closely with min-istries and oversight bodies to prioritizetransparency and the protection of patientdata from commercial exploitation.

While there are risks associated withany form of aggregated electronic infor-mation, we must ask ourselves if thoserisks outweigh the rewards, and secure dig-

ital health tools have proven themselves tobe fundamental to receiving good qualitycare. In the same way that someonechooses to get into a car – despite the risks– to perform essential tasks, digital healthtools play an essential role in improving apatient’s healthcare journey. To someonewho has experienced the benefits of virtualcare first hand during the pandemic, thenotion of having to go somewhere and ex-perience the potential of additional risk, inorder to get a prescription that could havebeen readily, safely and securely sent totheir pharmacy of choice using e-prescrib-ing, seems counterintuitive.

As we look beyond the pandemic, thequestion will no longer be whether wecontinue to implement electronic toolslike e-prescribing into our healthcare sys-tem. Canadian patients, physicians andpharmacists across the country have al-ready answered that with a resounding yes.The question will be how to implementthese tools safely, ensuring that their bene-fits outweigh their risks. When it comes toe-prescribing, this means investing in thelatest security measures and equippingphysicians and pharmacists with the toolsand information they need to prioritizeprivacy and cybersecurity throughout theprescription process.

Abigail Carter-Langford is Chief Privacy &Security Officer and Executive Vice President,Governance, Risk & Compliance at CanadaHealth Infoway. Since 2017, she has workedto promote Infoway’s commitment to ad-dressing health information governance, in-cluding the privacy and security of personalhealth information, throughout the digitalhealth industry and within Infoway.

How can we improve the security of Canadians’ healthcare information?

BY DANIEL MARTZ

The COVID-19 pandemic hasbeen a catalyst for innovation,and has forced Canada’s healthecosystem to reimagine its ca-pabilities. In the past year

alone, we’ve learned a lot about the shiftinghealthcare needs of Canadians, includingthe importance of adaptability and flexibil-ity when it comes to how we access care.

Prior to COVID-19, timely access toprimary care support was a challenge formany people in Canada, and the pan-demic exacerbated these issues. ManyCanadians have continued to delay careout of fear of being exposed to the virusin clinics and hospitals, or have had caredelayed as a result of the pandemic re-sponse. While Canada’s healthcare sys-tem still faces significant COVID-relatedhurdles, solutions such as virtual carehave shown us that digital services arenow an essential component of ourhealthcare system.

According to the Canadian MedicalAssociation Journal, by June 2020, vir-tual care represented more than 70 percent of ambulatory care provided byhospitals and doctors’ offices. As of

March 2021, more than five millionvirtual care appointments had beenconducted in Canada. TELUS Healthmet the urgent demand by leveragingits suite of existing digital health tools,already used by Canadians long beforethe pandemic hit, and new digitalhealth tools. The accelerated focus thatthe pandemic placed on virtual care in-spired us to quickly bring new servicesinto the hands of Canadians whoneeded them most, across a variety ofhealthcare issues.

One of the most pressing issues facedby Canadians throughout the pandemichas been mental health challenges. Ac-cording to Statistics Canada, since theonset of the pandemic, 55 per cent ofCanadians reported having good or ex-cellent mental health; this is a 13 percent drop from 2019. In response togrowing mental health concerns, TELUSHealth expanded its virtual care servicesbeyond primary care to include access tomental health counsellors and otherwell-being services, to ensure Canadiansand their families felt supported in everyaspect of their healthcare journey – bothphysical and mental. Understanding thata one-size-fits-all approach to care never

works, the company also launched Espriby TELUS Health, a mobile app specifi-cally designed to deliver mental healthresources that support the unique, ongo-ing and ever-changing challenges facedby frontline workers.

In addition to providing a variety ofservices and support to patients, virtualcare has also enabled Canadians to re-think how they’re able to access care.For example, the ability to access care

from the safety of their homes hasforced Canadians to rethink what theywant from their careers, employers andhealth benefits plans, and has themlooking to their employers to supportthem with flexibility, including both in-person and virtual clinical services.TELUS Health Care Centres enablesCanadians to do just that, offering a va-riety of services from preventive healthassessments, wellness coaching and pri-

mary care, to nutrition, mental healthsupport, and occupational health.

The end of the pandemic won’t meanthe end of shifting health needs and ex-pectations in Canada. Looking ahead,2022 needs to be about strengthening asystem that will support advancementssuch as virtual care in the long-term andacross the entirety of the care contin-uum. Technology enables access to ser-vices and interventions that address notonly current challenges and changes, butalso those that will continue to impactCanadians long after the pandemic. Inresponse to these growing obstacles,TELUS Health diversified its offering tomeet the unique needs of many diversedemographics, providing support whereand when it is needed.

Moving forward, we must ensure in-novative solutions are designed so theycan be adapted to meet the unique,constantly evolving needs of all Canadi-ans, throughout the entirety of the pa-tient journey, and foster a healthcareecosystem that supports sustainabilityinto the future.

Daniel Martz is vice president, VirtualCare, TELUS Health.

Sustaining the strides recently made in digital health

One of the most pressingissues faced by Canadiansthroughout the pandemic hasbeen mental health challenges.

Abigail Carter-Langford

h t t p : / / w w w . c a n h e a l t h . c o m1 2 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 13: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

siemens-healthineers.ca/syngo-carbon

The connecting element that gives rise to knowledgeAre you looking for an all-in-one solution for your imaging and data management?

Syngo Carbon connects what is not connected yet. It brings all imaging disciplines into one and enables access to all relevant data – across departments and responsibilities to meet certain decisions along the patient pathway.

If you know better, you can do better. That’s how you can achieve better patient care. Syngo Carbon gives rise to knowledge by providing a highly efficient, high quality and optimized clinical environment, opening up a wide range of possibilities in image interpretation, reporting, AI implementation, data management, archiving and migration, including unique access to innovation platforms, open source and third party functionality and the most up-to-date software available.

Syngo Carbon is a new species that goes far beyond PACS systems.

Talk to our digital health team about your needs and find out more about our digital health portfolio.

Email: [email protected]

View our digital health solutions: siemens-healthinees.ca/digitalhealth

Syngo Carbonenabled by teamplay digital health platform

Page 14: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

D I A G N O S T I C I M A G I N G

BY IRENE DUBE

SIOUX LOOKOUT, ONT. – For years,patients with breast cancer whorequired a lumpectomy have un-dergone a multi-step process onthe days leading up to theirsurgery. The process is often un-

comfortable for patients and can sometimescause delays and added stress on patients.But now, here at Sioux Lookout Meno YaWin Health Centre (SLMHC), we arethrilled to announce the purchase and use ofthe Endomag system – a game-changer forbreast cancer patients.

“Compared to a conventional proce-dure known as wire localization beforebreast surgery, our newly purchased En-domag system allows the diagnosticimaging department to precisely markthe sites of lesions and tumors in a highlyaccurate method that can occur at thetime of biopsy, which eliminates the needfor additional procedures to occur im-mediately before the surgical interven-tion,” explains Dr. Neety Panu, lead radi-ologist at SLMHC.

Not only is this cutting-edge technol-ogy making it easier for the SLMHChealth team and our patients, but this is animportant step forward for rural hospi-tals, as SLMHC is now the first hospital innorthwestern Ontario, and one of onlytwo rural hospitals in Canada, to adoptthis technology and demonstrate that thiskind of technology can be accessible tomore Canadians.

“We could be at the forefront here withthis technology, as many smaller hospitalslike ours experience logistical challengesfor breast cancer patients,” explains Dr.Justin Poling, surgeon at SLMHC. “Thereare hundreds of hospitals like ours who Ithink we could be a really good examplefor, and this technology has potential to begame changing for Canada’s geographi-cally-distributed population.”

The Endomag system process startswith a tiny Magseed marker, that is easilyplaced by the radiologist before thesurgery. Ideally, it is placed at the time ofbiopsy so that the patient requires onlyone visit/procedure before surgery in or-der to mark the exact site of breast cancer.Once placed, it can’t be broken or dis-lodged in any way, and it is detectable us-ing a Sentimag probe. The use of Magseedmarkers at SLMHC provides patients witha more accurate placement of a detectorpre-surgically, and enables a more seam-less transition to their surgical care and abetter outcome post-surgically.

Prior to this, a wire would have been in-serted into a patient’s breast on the sameday as the surgery. This meant that pa-tients would first be scheduled with the ra-diologist in the diagnostic imaging depart-ment, and then make their way to theirscheduled surgery. Coordinating the carebetween two different departments createsa logistical challenge, and also requires atwo-step process for the patient (the firststep at the time of biopsy and the secondbeing a very early start on the day of thesurgical excision).

“In the past, if a lesion or area of con-cern was spotted and surgery was recom-mended, a patient would be booked for

when the radiologist would be on-site forthe localizer wire insertion. Because thewire could come loose or be dislodged, thewire insertion and surgery needed to be thesame day,” explains DeAnna Lance, medicalradiation technologist at SLMHC. “Thismeant that all our breast surgery patientswould be booked ahead to meet the sched-ule requiring coordination between twovery busy radiology and surgical depart-ments. Now, with the Magseed technology,the radiologist can implant the Magseed atany time and it will stay there until it is sur-gically removed. This means we can have

patients have their biopsies/Magseed place-ment independent of department sched-ules, and this removes a big hurdle in logis-tical care for the patient.”

“It really frees us up with regards to alot of scheduling challenges,” adds Pol-ing. “It makes us less dependent on thetime frame. Not only that, there are somany other benefits to our patients, in-cluding a lot less travel and stress. A pa-tient might have had to drive to ThunderBay to get a radioactive tracker injectedat the end of the day, then once injectedthey’d have to drive back right away sincewe would only have 12-18 hours to oper-ate after that injection. In the middle ofwinter, driving on a dark highway atnight and worrying about your surgerythe next morning… this was obviouslystressful for our patients.”

Now, our team at SLMHC can inject a

patient with Magtrace and that need forradioactive tracer injection and back-and-forth travel to Thunder Bay is eliminated.

The use of Magseed markers at SLMHCwas made possible through the generoussupport of the SLMHC Foundation, whichprovided approximately $95,000 in fund-ing after the SLMHC team sent their pro-posal in for the equipment request thispast summer.

“Our rep at Endomag said this was oneof the quickest turnarounds for the timebetween the presentation of the technol-ogy to the processing of the purchase… weare thrilled that the SLMHC Foundationunderstood how valuable this equipmentpurchase was to our patients and team andthat they acted so fast to make this dreama reality,” says Poling.

“The breast imaging department andpatients are truly grateful for the SLMHCFoundation in bringing this new technol-ogy to our region,” adds Lance. “We areable to improve the efficiency and decreasethe time delays of patients requiring surgi-cal removal of suspicious or cancerousbreast lesions.”

For other rural hospitals looking at thepurchase of the Endomag system, Polingadds that while it does cost a lot up front,it does save the healthcare system moreover time.

“With a huge reduction in travel costsrelated to breast cancer treatment andsurgery, this new technology saves thehealthcare system a lot of money. Thereare a lot of rural hospitals across Canadathat have well-trained breast surgeons notable to do breast procedures because ofsignificant logistical challenges. This newtechnology is going to open doors for a lotof hospitals to reinitiate breast cancer pro-grams and provide better care for patients.

I think we are setting a really good exam-ple,” adds Poling.

“In the era of COVID-19 and through-out this pandemic, there have been manyhardships in healthcare across the coun-try. However, despite all of that, to see aproject like this lift off quickly and effi-ciently speaks volumes to the level of col-laboration between our physicians andadministration here at SLMHC,” saysPanu. “We can’t thank the Foundationenough for seeing the value in this, how itwill positively impact so many patientsand their families.”

“SLMHC is thankful for the continuoussupport of the SLMHC Foundation, andthe tireless work of volunteers and com-munity members, who ensure crucialfunding is provided for necessary equip-ment and upgrades at SLMHC,” addsHeather Lee, president and CEO ofSLMHC. “We’re excited to be a part ofsomething so innovative, and to be provid-ing better care closer to home.”

After training and completing trials ofthe technology at SLMHC, the surgicaland diagnostic imaging team is excited toannounce that the technology is here andalready in use for breast cancer patients.SLMHC is proud to continue growing ourcare capacity as we strive to provide crucialservices to keep quality healthcare.

To learn more about the importantwork the Foundation does, visit the Foun-dation website at www.slmhc.on.ca/foun-dation. To learn more about SLMHC’smammography services and breast care inour diagnostic imaging department here atSLMHC, visit slmhc.on.ca/di.

Irene Dube is corporate communicationsmanager at Sioux Lookout Meno Ya WinHealth Centre.

New technology at SLMHC brings more precise breast cancer surgery

In September 2021, Sioux Lookout Meno Ya Win Health Centre’s surgical team received the new Sentimag Localization System that will be used for breast proce-dures. The Sentimag system uses magnetic sensing to detect magnetic markers using the Magseed or Magtrace method. Pictured, left to right: Jeremy Doke, Sen-timag representative, Dr. Eric Touzin, Dr. Justin Poling, and Dr. Matthew Parkinson.

SLMHC is now the first hospitalin northwestern Ontario, and oneof only two rural hospitals inCanada, to adopt this technology.

h t t p : / / w w w . c a n h e a l t h . c o m1 4 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 15: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

SAS and Microsoft Accelerating the Power of AI for Everyone.Joining forces to define the future of analytics in the cloud. Now you can run workloads that work for you and everyone is empowered with insights that enable more trusted decisions.

Learn how at SAS.com/ca/microsoft-sas

Page 16: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

BY DAVID KOFF, MD

With the ongoing pandemic,and no prospect to return toan in-person meeting in 2021,MIIT has again been deliveredonline this year, but with adifferent format. Instead of a

full day conference, the program committee decidedto divide the conference talks into a session in Springand one in Fall, as well as a keynote talk and severalpanel discussions.

Lectures were uploaded and made available forparticipants to view at their own pace, followed 10days later by a roundtable discussion with the speak-ers. In an environment where many webinars are of-fered to our public of professionals, we wanted ourconference to be as flexible as possible.

We’ve been privileged that Dr. Rasu Shresthaagreed to deliver the keynote address at the kick-offsession. Dr. Shrestha is chief strategy and transfor-mation officer at Atrium Health, one of the largestnot-for-profit health systems in the United States. Hetold us about the new VUCA (volatility, uncertainty,complexity, ambiguity) environment we are living inand how it brings us unique opportunities to re-imagine and re-build.

Other speakers gave us their perspective on anumber of topics related to medical imaging and itsfuture, with of course a share for artificial intelli-gence and a special focus on the human impact ofour initiatives. As I cannot quote them all, I’d like tohighlight three of the talks.

Dr. George Shih, an associate professor of Radiol-ogy and Vice-Chair Informatics at Weill Cornell Uni-versity in New York, told us about “Bias and Fairnessin Artificial Intelligence”, a major issue with the ex-ploding development of AI-based solutions designedto assist radiologists, where AI detects abnormalitiesand generates reports.

But AI is not always accurate. A good example isfacial recognition where algorithms don’t perform aswell on women and people with dark skin, with one

highlighted case leading to jail time due to incorrectidentification. This unfortunately also applies tohealthcare where AI would not discriminate basedon attributes, and will not take into considerationminority populations, minority diseases, and minor-ity manifestation of diseases.

The issue of Bias and Data was raised in a recentjoint statement between the American College or Ra-diology, European Society of Radiology and Cana-dian Association of Radiologists, titled “Ethics of AIin Radiology” and published in the journals of all so-cieties in 2019.

The document, which calls for the implementa-tion of a Code of Ethics for the use of AI, states that“ethical use of AI in radiologyshould promote well-being andminimize harm resulting frompotential pitfalls and inherentbiases” (source ACR, Oct 1,2019). Bias is a systemic devia-tion from the truth, which mayhappen when the sampled datado not represent the popula-tion accurately or when the re-ported data does not representthe real world.

Aiden Kivisto, is a Trans Ed-ucator, developing organizational policies for health-care institutions, and providing workshops for non-profit, social welfare sector organizations on genderand sexuality. His talk addressed “The Patient Per-spective on HL7 Updates,” and he also told us abouthis personal story of gender transition, helping himbreaking down social barriers for staff with little ex-perience of the trans community.

Aiden participates in the Infoway Sex and GenderWorking Group whose goal is to develop and imple-ment a strategy to modernize Gender, Sex, and Sex-ual Orientation (GSSO) documentation in CanadianElectronic Health Records.

The focus is on policy and practice to integratethe best GSSO information with organizationalstructures, policies, and workflow processes. The

“HL7 Gender Harmony project” aims to use modelinformation (data elements, value sets, code sys-tems) to define and harmonize some aspects repre-senting sex and gender. The project takes into ac-count user needs including the LGBTQ communityand patients.

Ameena Elahi is a senior technical analyst at Uni-versity of Pennsylvania, the international director ofInformatics Operations, RAD-AID informatics and aboard member of SIIM, the Society for Imaging In-formatics in Medicine. She told us about “ImagingInformatics and Global Health Outreach,” and hersuccessful experience implementing a PACS systemin a University Hospital in Nigeria.

RAD-AID is an amazing charitable organization,started in 2008 at Johns Hopkins, and whose missionis to improve access to radiology services in the low-and middle-income countries (LMIC) as well as inunderserved areas of high-income countries (HIC)like Canada. With more than 14,000 volunteers from140 countries, they are present in over 38 countries.Ameena told us about their challenges, and how theyfollowed a rigorous Assessment-Planning-Imple-mentation-Training-Analysis strategy to achievetheir goal. She also told us how artificial intelligencecan improve radiology access in LMIC, and how withthe support of a Google Foundation Grant, they de-veloped ways to detect and correct racial bias in AI,to be used in resource-poor hospitals.

These talks gave us good examples of all the workperformed to understand and correct biases and dis-crimination, being artificial intelligence dealing withminority populations, minority diseases and minor-ity manifestation of diseases, Electronic HealthRecords integrating Sex and Gender considerationsor divide between high and low/middle incomecountries access to radiology. Of course, there ismore to do, but all the hard work solving inequitiesmust be acknowledged and recognized.

David Koff is a Professor Emeritus of radiology at Mc-Master University, and past Chair, Department of Ra-diology, McMaster University.

VI

EW

PO

IN

T

The first steps are to be aware of the biases being generated by technologies, and of systemic social biases.

MIIT 2021: How to correct biases andinclusion issues in radiology and medicine?

TORONTO – The pandemic hasserved as a powerful opportu-nity to accelerate digital health

capabilities across the healthcare sys-tem. Whether through virtual medi-cine or large-scale use of technolo-gies, over the last 19 months we haveseen incredible innovation andadoption of digital solutions to meetthe needs of the more than 400,000people served by North York GeneralHospital (NYGH).

NYGH’s Digital Strategy 2020-2025 is a key driver of the seven-siteorganization’s new Strategic Plan.The Digital Strategy is being deliv-ered jointly by the hospital and morethan 20 community and primarycare partners that make up the area’s

Ontario Health Team, North YorkToronto Health Partners. United bya common vision of creating a user-friendly, connected, patient-centredcommunity health system, NorthYork providers are using digital tech-nology to achieve a new level ofhealth service integration.

The latest initiative in service ofthis vision is the introduction of theNorth York digital eReferral plat-form. The platform facilitates aseamless, simple process for primarycare providers to securely refer pa-tients for common hospital services.

“The traditional referral system offaxing and phone calls is an anti-quated multi-step process that doesn’twork well for patients. They don’t

know where they are in the referralprocess and if the appointment theyreceive doesn’t work, making a changebecomes a process for them,” says Dr.

David Eisen, chief of Family andCommunity Medicine, NYGH.

eReferral enables everyone to seethe same information at the sametime which speeds up and simplifiesthe referral process.

“Expediting, simplifying, and

making referrals a transparent anddirect process reduces patient anxietyand empowers them. It’s exactly thiskind of digital platform that enablesan integrated health system, bringingto reality the patient’s medical neigh-bourhood as a real supportive systemin their care,” adds Dr. Eisen.

NYGH is among the first GTAhospitals to offer physicians and pa-tients eReferral, using the well-estab-lished Ocean platform. Ocean eRe-ferral senders include primary carephysicians or specialists in the com-munity referring patients to the hos-pital’s specialist services.

Ocean eReferral receivers areNYGH physicians and clinicians who

COVID crisis prompts NYGH to accelerate its progress in e-referral

Dr. David Koff

eReferral enables everyoneto see the same informationat the same time, speedingup the referral process.

CONTINUED ON PAGE 22

h t t p : / / w w w . c a n h e a l t h . c o m1 6 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 17: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

V I E W P O I N T

BY DERRICK CHOW

The promise of artificial intelli-gence (AI) in healthcare – rang-ing from operational efficienciesto patient outcomes – is vast. Still,

many organizations struggle to turn thatpotential into actual value to make it moreefficient and collaborative, and most im-portantly, to improve care. One of the mostsignificant obstacles to unlocking AI’s po-tential is the quality of the data itself.

The healthcare industry creates signifi-cant data, but the insight AI provides islimited when data sets are incomplete.While the problem is easy enough to un-derstand, addressing it is challenging – butnot impossible – to overcome. Two thingshealthcare organizations can do are to pro-vide structured, coded healthcare data andimprove the patient consent process.

In the US, the 21st Century Cures Actwill accelerate digi-tal access for pa-tients to their healthrecords in a ma-chine-readable, in-teroperable form.This will open thedoor for AI develop-ers who need accessto greater amountsof personalized datato provide precisionhealthcare insights.

While Canada doesn’t have a similarregulation, healthcare organizations shouldstill be working toward more structureddata to keep pace with patients who preferproviders that can offer data portability totheir consumer apps.

Many offices and health systems relyheavily on faxing and mailing informationto one another, which seems antiquatedwhen contrasted against how patients usetechnology via wearables and apps tomanage their health.

By digitizing health data, providers canshare and combine data from other de-partments within, or external to, the orga-nization to make their operations more ef-ficient. And when the historical patientdata is combined with patient-generateddata from their apps and wearable devices,care can be dramatically improved.

Additionally, healthcare organizationsare better equipped to address publichealth emergencies like COVID-19 whenhealth data is available to help them makecritical, real-time decisions.

An often-overlooked, but critical, compo-nent to freeing data for AI insights is the pa-tient consent process, which can be tediousand unintuitive for patients. When access isstuck at consent, it is difficult to generate thebenefits AI can offer. Healthcare organiza-tions can expedite the process by implement-ing intuitive methods for patients to give theirconsent. Here are some things to consider:

• Device integration: Digital consentforms should be compatible with whateverdevice a patient uses, whether it’s a desktopcomputer, phone, or tablet. The experienceshould also integrate with whatever brandof device a patient uses.

• Transparency: Patients won’t want tospend a lot of time giving consent, but atthe same time, they need to know what

they’re consenting to. So, it is essential toinclude as much information as necessaryby law while quickly getting to the specificreason you’re asking for their consent.

• Reward: It’s also important to share withpatients the benefits of giving their con-sent. In some cases, it may be used for vital

medical research, for example. In other in-stances, the data may be shared with thepatient’s attorney or insurance companythat needs it to represent them in a lawsuitor complete a life insurance submission.

Obtaining access to data is about morethan technology – trust is also essential. One

way to gain trust is to ensure that the data apatient shares with you is secure. Once dataprivacy is breached, a slew of issues can oc-cur that may offset the value AI can bring.

Derrick Chow is co-founder and COO ofMedchart.

Structured health data and patient consent can unlock the value of AI

Revolutionize care delivery with Canada’s top virtual care platformHealthcare organizations across Canada leverage our cutting-edge technology to power virtual programs

Learn more at getmaple.ca/care-delivery

h t t p : / / w w w . c a n h e a l t h . c o m N O V E M B E R / D E C E M B E R 2 0 2 1 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y 1 7

Derrick Chow

Page 18: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

FE

AT

UR

E

RE

PO

RT

BY DIANNE DANIEL

If health data is the next frontier in personal-ized medicine, the learning health systemcould be the vehicle to transport us there.That’s the vision of recent collaborationsacross Canada bringing together academics,clinicians, businesses and funding partners to

unlock critical health data found in physician notes,imaging databases, electronic medical records andother digital health systems. Not only are the groupsenabling researchers and clinicians to access high-quality pools of patient information that have histor-ically been difficult to apply to decision making,they’re also working to build a healthcare system thatcontinuously learns about the population it serves.

“What we’re doing is building a system so thatwe’ll be prospective, by which I mean thatevery day the system will be newly updated,”said Jeremy Petch, PhD, director of DigitalHealth Innovation at Hamilton Health Sci-ences (HHS) in Ontario and founder of thehospital’s Centre for Data Science and DigitalHealth (CREATE). “As every new patientcomes in and is diagnosed, we’ll be able to ac-cess their data,” said Dr. Petch.

In collaboration with Dr. Mark Levine, theBuffet Taylor chair in breast cancer research atMcMaster University, the team at EscarpmentCancer Research Institute (ECRI) and theToronto-based artificial intelligence (AI) firmPentavere, CREATE – which is staffed by ex-perts in artificial intelligence, data sciencesand software engineering – is working to dig-itally recreate patient ‘stories’ for breast cancerpatients currently being treated at the HHSJuravinski Cancer Centre. The platformbuilds on an earlier pilot of 50 patients thatused IBM’s Watson Health. Moving forward,the initiative is relying on Pentavere’s propri-etary natural language processing (NLP) en-gine called DARWEN to extract real worldpatient evidence and insights from vastamounts of unstructured clinical data.

The impetus for the project came from Dr.Levine, who, as he was winding down hishighly recognized career as a medical oncologist atHHS, had the notion that AI could solve a funda-mental challenge in healthcare – how to obtain aclear picture of a patient population before, duringand after treatment.

“Thirty-five years ago if a patient asked me,‘How many patients with breast cancer do you seewith the same problem as me every year and howdo they do?’ ... I would either do a chart review my-self or ask a resident to do it,” said Dr. Levine. “Doyou know what we do now, 35 years later? The samething, and with all of the technology now available,that’s unbelievable.”

The group’s initial pilot demonstrated the abilityof AI to accurately identify key information frommultiple digital sources for 50 stage three breast can-cer patients treated at Juravinski. The current effort,funded in part by Roche Pharmaceuticals and nowscaled to include 3,000 breast cancer patients, is us-ing Pentavere’s technology to generate similaranonymized information that includes the date ofdiagnosis, how the cancer was diagnosed, the pathol-ogy of the tumour, the data and type of surgery,

whether or not chemotherapy or radiation were ad-ministered, and the outcome.

“All of this data is very siloed, spread over multi-ple different source systems, so really the task ini-tially is to bring that data all together in its raw un-structured form, into a data lake – it’s still messy, it’sstill unstructured, it still has conflicting data in it,”explained Pentavere medical director Dr. Christo-pher Pettengell. “Our technology can then be de-ployed within that environment and structure it tomake sense of it.”

The idea is to build an automated function intothe HHS health system workflow so that disparatehealth data can be grouped into high-quality rowand data sets, preparing it for analysis. Machinelearning algorithms are then trained to generate ac-curate, de-identified patient information from the

data, giving clinicians a first-of-its-kind view intooutcome data specific to Juravinski cancer patients.

“The analytics and the statistics on the data –that’s something we know how to do, but we’ve justnever really been able to access the data before be-cause it’s been locked away, and there’s been no effi-cient way to get that data out and make it available tothe statisticians and analysts,” said Petch.

Though he’s careful to avoid the ‘hype’ around thepromise of what AI can deliver, Petch said the overallambition is to better understand the cancer clinic’spatient population so that the hospital can move to-wards a personalized care model, and ultimatelymake quality improvements. One possible deliver-able is a learning health system dashboard utility thatwould enable treating clinicians to access real-worldevidence about similar patient journeys within HHS,as well as recommendations based on leading-edgeresearch or clinical trials, when deciding on the bestcourse of treatment for the patient in front of them.

“What’s unique about this is that a clinician canget a picture of their complete practice, and get aprofile of their practice over all of the patients

they’ve treated over many years, as well as the pa-tients that their colleagues have treated, to maybe seepatterns that we haven’t seen before,” said Petch. Theapproach also supports the hospital’s mission to ad-dress health equity, because it analyses outcomes forentire patient populations regardless of socioeco-nomic status or ethnicity, removing the bias often as-sociated with clinical trials, he added.

As work on the learning health system moves for-ward, the next step is to validate that the informationgenerated by the AI model matches the informationin the actual patient records. Two leading breast can-cers experts, independent of the collaboration, arecurrently comparing key patient attributes such ashistology, stage, estrogen receptor or type of surgery,and the machine learning algorithms will be refinedas needed, until the match is 95 per cent or higher,

explained Levine.Another Canadian team working to‘unlock’ the benefit of digital healthdata involves the University of Al-berta Faculty of Medicine and Den-tistry, Alberta Health Services (AHS)and Indianapolis-based Pinnacle So-lutions Inc., a long-time SAS resellerthat entered the Canadian analyticsmarket in 2019. This past summer,the faculty launched the Data Ana-lytics Research Core (DARC), a se-cure high-performance computingand on-premise data storage envi-ronment it built in partnership withPinnacle Solutions to facilitate healthdata sharing between AHS and fac-ulty researchers.The goal is to grow the university’scapacity to perform big data analysisand AI, advance e-health innovationssuch as personalized care, and pro-vide student training in health dataanalytics. At the same time, DARC,which is supported with fundingfrom the AHS Chair in Health Infor-matics Research, is also available todata analysts funded by the AlbertaStrategy for Patient Oriented Re-

search Support Unit who are invested in building aprovincial learning health system.

“As medicine really becomes a data science, we’regoing to need to train a next generation of clinicianswho know data, as well as data scientists who knowclinical medicine,” said vice-dean of researchLawrence Richer, a professor in the faculty’s depart-ment of pediatrics who led the project. “DARC pro-vides an environment where that can start to happen.”

Historically, if a faculty researcher wanted to workwith sensitive health data, they would make an indi-vidual request to AHS, the provincial custodian ofpatient information, and agree to work within speci-fied security parameters. Richer found that as morerequests were being made, and the sensitivity of thedata increased, the custodian’s willingness to sharedecreased. At the same time, researchers interested inapplying AI and machine learning require access tomuch higher computer capacity than is typicallyavailable to them in a small server or desktop envi-ronment. DARC solves both challenges, he said.

Architected by Pinnacle Solutions, the integratedplatform sits behind the AHS firewall at a local hospi-

Groups are working to build a healthcare system that continuously learns about the population it serves.

Artificial intelligence making progress in collaborative projects across Canada

ILLU

STRA

TIO

N:

LIN

DA

WEI

SS

h t t p : / / w w w . c a n h e a l t h . c o m1 8 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 19: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

A I A N D M A C H I N E L E A R N I N G

tal, where it shares the hospital’s network in-frastructure and security protocols. In addi-tion to featuring SAS Viya, a visual AI, ana-lytic and data management tool that makesdata analysis as easy as ‘dragging and drop-ping,’ it also supports open-source softwareand other free tools popular among stu-dents. The benefit is that both ‘clickers andcoders’ are supported, said Richer.

“Not everyone can be that expertcoder, and we can’t continuously hire thatcoder every time we want access to a dataset,” he said.

Richer experienced the platform’s easeof use firsthand when he decided to test amodel he had previously requested from adata analyst to predict which children ar-riving in the emergency department arebest served by brain imaging. “What weknow is many kids present to emergencywith worrisome symptoms, but only asmall fraction actually have a worrisomedisorder like a stroke, brain bleed or braintumour,” he said. “My hope was that wecould build an algorithm to take clinicalfeatures and then give the point-of-careclinician a risk score of ‘should I, or shouldI not’ do a brain scan.”

When DARC went live, Richer was ableto test six different iterations of the modelin one hour, even though he had neverused SAS Viya prior. “I didn’t need to waitfor an analyst to tell me which modelworks best. I already have an idea of whichis best – I took that as proof that we’vedone the right thing here,” said Richer.

As of October 2021, roughly 60 usershad been granted access to DARC, andRicher expects the number to grow. Ap-proximately one-quarter of the faculty’s750 members are involved in some aspectof health data research.

One example is a big data initiativeled by Dr. Daniel Baumgart, pro-fessor and director of the univer-

sity’s Division of Gastroenterology, who isworking in partnership with Crohn’s andColitis Canada. The project aims to useDARC to deliver personalized therapies toCanadians living with inflammatory boweldisease (IBD), and is currently analysinghealth data from 60,000 Albertans who livewith the disease to identify patterns in howthey’re doing over the long term and whatcontributes to their outcomes.

The hope is that the AI models will beable to predict which medication will bemost effective and at what dose, who is morelikely to experience complications related toIBD, and who is at greater risk for futurehospitalizations. With AHS recently trans-ferring terabytes of diagnostic imaging datato DARC – representing a first for re-searchers – the researchers will also work torefine diagnosis with imaging techniques, toassist in more precisely detecting the disease.

“Right now we’re limiting access toDARC. We have a select number of teamsand the demand is growing,” said Richer. “Idon’t want to sit with a big white elephantthat I can’t sustain over time, but at thesame time, I want to maintain what thecore needs are for our core researchers whoare truly innovating in the health space.”

A third collaboration working to ad-vance the concept of a learning health sys-tem in Canada is the new ‘junior-senior’ Re-search Chair in Digital Health announcedby the Université de Sherbrooke in Quebecand funded by the Ministère de l’Économie

et de l’Innovation (MEI). Co-directed byprofessor Anita Burgun, a French expert inbiomedical informatics with ties to the Uni-versity of Paris, and Christina Khnaisser, ahealth informatics researcher at Sher-brooke, the chair is overseeing an ambitiousresearch program based on a vast networkof France-Quebec collaborations includingpartnerships between pediatric hospitalsand research centres.

The program is using a digital platformcalled PARS3, developed at the universityby the Interdisciplinary Research Group inHealth Informatics (GRIIS), to accesshealth data in situ, meaning it doesn’t haveto be moved to a shared location, alleviat-ing issues related to privacy. The goal is tobuild a system that can “securely and ethi-cally analyze the data that accumulatesduring the patient-care trajectory,” said

Burgun, as they work to better understand,diagnose and treat rare diseases.

For example, medical teams will beable to securely analyze clinical records,medical images and physiological infor-mation as it is collected throughout pa-tients’ daily lives. Doctors will then beable to compare medical profiles in orderto speed up diagnostic decisions and op-timize treatment.

Canadian Healthcare Technology offers

the country’s leading healthcare I.T.

publications. Are you getting all the news?

One-stopshopping.

www.canhealth.com

Canadian Healthcare Technology magazineCanadian Healthcare Technology breaks the news about importantprojects, programs and technologies, and provides hospital executives andsenior managers with an excellent source of information for improvingthe delivery of healthcare. It’s sent to over 5,200 readers in print formatand to 6,800 opt-in subscribers as a digital edition.

For advertising or editorial inquiries, contact Jerry Zeidenberg, Publisher, [email protected]

eMessenger newsletterCanadian Healthcare Technology’s e-Messenger contains breaking newsabout important deals, installations and developments. Four blasts aresent each month, via e-mail, to over 7,000 senior managers and executivesin hospitals, clinics and health regions.

White PapersCanadian Healthcare Technology’s White Papers are sent out once amonth, via e-mail, to over 5,000 senior managers and executives inCanadian hospitals and health regions. The monthly blast containssummaries and links to White Papers issued by various organizations,providing cutting-edge information about topics of interest tohealthcare decision-makers.

h t t p : / / w w w . c a n h e a l t h . c o m N O V E M B E R / D E C E M B E R 2 0 2 1 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y 1 9

Page 20: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

A I A N D M A C H I N E L E A R N I N G

TORONTO – Liver diseases areone of the fastest growingcauses of morbidity and mor-tality in the world. Over 2 bil-lion people globally are living

with or are at risk of liver disease, and asthis disease is mostly asymptomatic untilvery advanced stages, most don’t evenknow it.

This surge in disease is being acceler-ated by the alarming rise of the obesityepidemic and type 2 diabetes, which canboth cause non-alcoholic fatty liver disease(NAFLD) and its most advanced and dan-gerous form, non-alcoholic steatohepati-tis, or NASH.

While traditional approaches to screenfor liver disease require referrals for highend imaging or highly invasive biopsy, On-coustics, a Toronto-based AI diagnosticscompany, has developed a better way to di-agnose and catch liver disease early – whentreatment options are broader and chancesfor reversal and full remission are best.

Oncoustics, founded in 2018 andfunded by the Creative Destruction Lab, iscreating and deploying multiple advancedAI solutions based on ultrasound for lowcost, non-invasive surveillance, diagnos-tics, and treatment monitoring of diseaseswith high unmet clinical need.

Unlike other players in the space, theyare not doing image recognition. Instead,they apply AI to raw sound signals fromreadily available handheld ultrasound de-

vices to rapidly differentiate healthy fromdiseased tissues.

There’s a wealth of information in theseraw signals and their patented approachreveals novel biomarkers that can bealigned with existing standards and cate-gorization systems. Oncoustics is a hard-ware agnostic, software only solution, buttheir enhanced value comes via low-costpoint-of-care ultrasound that are nowcoming onto the market.

These systems that now cost under$5,000 are exciting, but they’re limited bypoor image quality and lack of medical ap-plications. But the Oncoustics approachdoesn’t require images and their quantifiedsystems don’t require expert sonographersor radiologists either.

The approach at Oncoustics makes theexpensive and difficult process of diagnos-ing liver disease seamless, painless and donein five minutes – right at the point-of-care.

“The idea came out of my PhD work atthe Sunnybrook Research Institute inToronto. Part of my work looked at ultra-sound raw data and using statistical ap-proaches of extracting parameters relatedto tumor cell death or tumor response todifferent cancer therapies,” said Ahmed ElKaffas, PhD and a founder of Oncoustics.“Playing around with the technology, wesaw a lot of potential for prevention andearly detection of diseases as well, whichwe soon narrowed down to the liver mar-ket as a first target for our startup; liver

disease is a major global health problem.” Their first product, the OnX Liver Fi-

brosis Categorization Solution, has nowachieved over 95 percent AUROC com-pared to big and expensive systems foundin hospitals like shear wave elastographyand transient elastography and they arebeginning their regulatory filings.

Locally in Toronto, Oncoustics is begin-ning deeper integration and data collec-tion work with hepatologists at UniversityHealth Network (UHN), the largest healthresearch organization in North America.

With this and other clinical partner-ships in the USA and in Asia, they are

looking to go beyond replacing existingimaging and elastography devices and areaiming at replacing painful and invasivebiopsies and to build an entire pipeline ofsolutions for the liver and beyond.

Oncoustics has also gotten the attentionof provider networks, pharmaceuticalcompanies and industry organizations,and recently announced that they were se-lected by Mayo Clinic’s Innovation Ex-change program and they are moving for-ward on a research program with RocheDiagnostics as well.

“Liver disease disproportionately affectspeople of color, lower socio-economic sta-tus and native peoples, and these samepeople are less likely to have access to ex-pert clinicians and the high-end diagnos-tics tests that may be available,” said BethRogozinski, CEO of Oncoustics. “Our ap-proach democratizes access to early diag-nosis by making it much less expensive,less painful and invasive, and readily avail-able at the point-of-care.”

These diagnostic tests, which will beeasier to order and perform than a bloodtest, are gaining the interest of clinicians aswell. Primary care physician MatthewSakumoto, MD at UCSF added, “As a pri-mary care physician, I love that the OnXwould allow me to screen and counsel thepatient within a single encounter.” Thispromise of AI and of Oncoustics to decen-tralize and democratize healthcare, are ex-actly what the doctor ordered.

BY JONATHAN WOODS

Undetected patient deteriora-tion is a major reason for un-planned ICU admissions,which are serious adverse

events correlated with higher patientmortality and morbidity. Early warningsystems exist, but they’re reactive; if analarm sounds, it means deterioration isalready underway.

St. Michael’s Hospital of Unity HealthToronto took a proactive approach to re-ducing unplanned ICU transfers. Insteadof spotting deterioration early, they’reusing artificial intelligence (AI) to pre-dict it. They do this using CHARTwatch,an AI system trained on 10 years’ worthof hospital encounters in order to antici-pate patient decline and signal the needfor early intervention care.

During its pilot in the hospital’s Gen-eral Internal Medicine unit, CHART-watch provided a near 1.5-day time in-terval between alarm and event, a marginof safety that translated into a 20 percentreduction in mortality among high-riskpatients according to early results. Thesystem is now in regular use in the unit.

CHARTwatch is a Pathfinder Project,one of a series of health AI pilots led bythe Vector Institute for Artificial Intelli-gence and Ontario health organizations.Pathfinder Projects are small-scale im-plementations designed to turn promis-ing AI research into clinical benefits.

These pilots create pathways for AI up-take by addressing known adoptionchallenges in healthcare settings and bybridging the gap between AI researchfunding from the federal Pan-CanadianAI Strategy and AI application in aprovincial healthcare system.

“Low acceptance of machine-learn[ing] solutions by clinician users,and uncertainty about how to evaluatethem” are key reasons for slow AI uptakein healthcare, according to Implement-ing Machine Learning in Medicine, a re-cent paper published in the CanadianMedical Association Journal. The paperreads: “One common barrier to theadoption of machine-learn[ing] technol-ogy is whether clinicians trust themodel’s output.” When real health out-comes are on the line, concerns about AImodel robustness, complexity, and opac-ity create trepidation.

Pathfinder Projects are designed tobuild the trust required for AI accep-tance. This is done through a focus oncontinuous learning, close work with theinstitutions implementing the systems,and applications that complement,rather than replace, human judgment.

This is exemplified by another VectorPathfinder Project, an AI update to CoralReview. Coral Review is an existing tool,developed by the University Health Net-work (UHN), that enables imaging-re-lated peer review and learning.

The project – a collaboration between

Vector, UHN , and the KIMIA Lab at theUniversity of Waterloo – involves addingan AI element to the tool: the ability toidentify X-ray images that include pneu-mothorax, pinpoint abnormalitieswithin them, and then search throughthousands of historical images to iden-tify others that share characteristics.

The project has the potential to speedup X-ray review and reduce time totreatment for patients, but, importantly,it’s also designed to instill trust. A paperon Coral Review describes how the up-

grade “enable[s] a virtual ‘second opin-ion’ for diagnostic purposes and pro-vide[s] computerized explanations forthe decision support.”

The model supports physician deci-sions. Physicians aren’t asked to deferjudgment when patient health is at stakeand they can get a sense of how themodel comes to its conclusions. At thispoint in the pilot, the search functionhas produced an 80 to 90 percent accu-racy rate. The detection function – theability to pinpoint abnormalities – iscurrently being tested.

These pilots and their focus on trust

align with the concept of Responsible AI,a set of principles and practices that aimto ensure AI serves, and doesn’t harm, itsstakeholders. Because of AI’s novelty,complexity, and rapid change, standardsfor managing fairness, risks, and trade-offs between performance and trans-parency are still unsettled.

Pathfinders’ small-scale deploymentshelp fill in the picture about these con-cerns on a case-by-case basis, clarifyingeach model’s role in a clinical processand growing acceptance.

Pathfinder Projects also highlight therole that the Vector Institute has in con-necting federal funding for advanced AIresearch with practical implementationopportunities in Ontario organizations.“Health innovation in Canada is split,”says Roxana Sultan, vice president ofHealth at Vector. “While most fundamen-tal health research is funded federally, thedeployment of health AI in practice hap-pens provincially in the health system.”

While Pathfinder Projects receivefunding from the Ontario government,“many of the researchers driving theseprojects are funded by CIFAR Canada AIChairs,” Sultan says. CIFAR Canada AIChairs are appointments fundedthrough the federal government’s PanCanadian AI Strategy. “Building thatbridge between research and its imple-mentation at actual sites is the role thatVector and Pathfinder Projects are play-ing,” says Sultan.

Vector demonstrates quick success in bringing AI into clinical usage

Pathfinder Projects are small-scale implementationsdesigned to turn promising AIresearch into clinical benefits.

Oncoustics AI democratizes diagnostics, beginning with liver disease

h t t p : / / w w w . c a n h e a l t h . c o m2 0 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 21: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

Caring for the health of Canadians, every step of the way.Together, let’s create a healthier future.

telushealth.com/forall

Page 22: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

A I A N D M A C H I N E L E A R N I N G

TORONTO – In one of the latest ex-amples of artificial intelligence(AI) and machine learning (ML)leading to impressive gains, a num-

ber of hospitals across Canada have de-ployed Semantic Health’s AI-powered med-ical coding and data quality auditing soft-ware, which uses AI to improve the effi-ciency of manual coding, abstracting, andclinical documentation improvement (CDI)processes. Hospitals are reporting initial ef-ficiency gains of up to 25 per cent, and ex-pect this number to increase over time as Se-mantic Health’s engine continues to learnhow the process is completed at each site.

As more Canadian hospitals consideradopting AI solutions, Dr. Nicola Sahar,CEO at Semantic Health, suggests that asuccessful implementation starts by asking,“who are we trying to help and can AI ad-dress their challenges?” Some problems donot need to be solved with AI. Problems thatare inherently data-driven or involve largeamounts of data, however, can be solvedmore efficiently and accurately with AI.

There are three primary types of AI so-lutions for health systems:

• clinical facing solutions that help im-prove care delivery;

• operational solutions that help improvehospital management and decision making;

• and financial or administrative solu-tions, which support coding, abstracting,and data quality improvement processes.

For Dr. Sahar, administrative solutionsare an attractive place for Canadian hospitalsto start with AI because, “the departments incharge of coding and auditing usually haveall the data – structured or unstructured –flowing through them, and are currentlymanually processing all of the data.”

One of the challenges with AI adoptionat hospitals is that it often has an air ofmystery around it about how it works. Atits heart, AI is the process by which we canteach computers to learn insights fromdata, and to make predictions about thatdata. A common example of the applica-

tion of AI is the prediction of the length ofstay for a particular patient. In that case,the AI can be “taught” to consider variousclinical data points, including the patient’sclinical notes and medical issues, to gener-ate an accurate prediction for the patient’slength of stay. It does this by analyzingthousands, and sometimes millions, ofprevious examples of patient data. As theAI sees more examples of patient data, andthe corresponding length of stay for thatpatient, it gradually learns what to pay at-tention to in the data in order to make amore accurate prediction of the patient’slength of stay.

With their ability to continuously learnfrom clinical data, AI systems can delivermeasurable and continuously improvingresults for hospitals. But, according to Dr.Sahar, the biggest factor for success in im-plementing AI is understanding the typeand volume of data to use.

If a hospital has just implemented anEMR, for example, now would be the per-fect time to explore AI use cases that can bebuilt on top of digitized EMR data streams.

However, even if a hospital does nothave a fully digitized data ecosystem, an AIuse case can often be framed in a way thatcan provide a solution with less data. Inthese cases, it is important to partner withexperts and vendors in AI. “There will al-ways be pockets of areas where you canstart. It all begins by focusing on the userand the use case,” Dr. Sahar adds.

As health systems seek to adopt AI, Dr.Sahar advises leaders to start with problem-atic processes that would significantly ben-efit from increased efficiency or accuracy. “Aplace where processes are slow and manual,and results are not fully accurate or reliable,is a very good starting point. These are theareas where you will see the best results onteam operations and return on investment.”

“In Semantic Health’s case, we have foundthat when it comes to coding, auditing, anddata quality, there is a lot of room for AI tohelp existing processes.” Semantic Health’s

product supports coding and auditing teamsby predicting the most relevant medicalcodes within a record, and auditing alreadyassigned codes to ensure that they are accu-rate. “That is really transformational because

the coders and audi-tors we help have toread tens, if not hun-dreds, of pages tofind key informationabout the patient.The AI engine can dothe heavy lifting anddirect their attentionwhere needed.” The impact of in-vesting in founda-tional AI use cases

can be significant. With Semantic Health,by improving the speed and accuracy ofmanual coding processes, hospitals alsobuild a layer of insight on top of unstruc-tured data. These insights can be used todo other things, like provide feedback on

operations, generate analytics, help withdecision support, and future-proof codingand auditing workflows. As a result, in thelong term, the software facilitates more AIuse cases, such as clinical decision support,clinical documentation improvement, andfacilitating access for researchers to real-world evidence generation.

It is also important to invest in AI sys-tems that are designed for clinical datastreams. Semantic Health’s approach, for ex-ample, uses a model that was trained on realclinical data, and is uniquely designed to un-derstand the nuances of clinical text.

“With Semantic Health’s approach,” Dr.Sahar adds, “it is easier to learn how patientpopulations differ at different hospitals, howdifferent treatments and conditions are be-ing referred to in a given text, and to moreeffectively identify the information codersneeded to make them more accurate.”

To learn more about AI for healthcareor Semantic Health’s product, please visittheir website at www.semantichealth.ai

review patient referrals for specialist ser-vices and programs. If more information isneeded before moving ahead with a refer-ral, they can communicate with thesenders, ultimately improving the processand patient experience.

“The eReferral platform is an easy-to-use system that integrates into primarycare office electronic medical records(EMR),” said Dr. Ronik Kanani, chief ofPaediatrics, NYGH. “It is a convenient wayto send referrals to our specialists in thehospital and allows us to email and com-municate directly with families. The long-term goal is to eliminate the need for fax-ing paper referrals to our clinic and to fullyintegrate the system into our hospital

EMR. We often have incomplete informa-tion on paper referrals and they are some-times illegible. If we have a more completereferral form, we can triage the referral ina more efficient way.”

The shift from fax-based and phone-based referrals to a digital system has sig-nificant benefits for providers and patientsalike; expanded access to specialists, fasterreferrals, better communication and infor-mation sharing, and the overall ability totrack and support patients throughouttheir care journey.

Currently, NYGH patients with a Fitpositive test who require a colonoscopy,those referred to the Complex Centre forDiabetes Care (CCDC), and paediatricspatients are benefiting from the platform.Soon Paediatric and Adult Mental Health,Diagnostic Imaging, and other high-vol-ume programs will offer digital referralsthrough this initiative.

“The eReferral process allows patientsand families to follow along as various spe-cialists are contacted and appointmentsscheduled. The communications are farsuperior and timely, reducing stress, elimi-nating the need for phone calls and followups,” said Bruce LaRue, Patient and Familyadvisor. “We are hoping to implement textmessaging as well as email notifications aswe progress through implementation.”

By using eReferral, patients, familiesand caregivers are better informed andsupported throughout the care process.

“COVID has been incredibly difficultfor so many people while also showing usthat we can work quickly in healthcare torespond to the needs of patients, theirfamilies and healthcare providers. This in-cludes pushing the health system closer towhere it should be in terms of providingmore digital options,” says Duska Kennedy,chief digital officer, NYGH. “eReferral isone such gain and another step in ourcommitment to provide patients with dig-ital health tools that help empower themaround tracking and monitoring informa-tion that supports their health.”

CONTINUED FROM PAGE 16

NYGH accelerates its progress in e-referral

Doctor and founder offers hospitals advice on achieving ROI with AI

Explore TryCycle’sDigital Compassionate Tether.

* in New England, USA (January 2019–October 2021)

[email protected]

Zero overdose deaths*

Zero deaths by suicide*

Dr. Nicola Sahar

h t t p : / / w w w . c a n h e a l t h . c o m2 2 C A N A D I A N H E A L T H C A R E T E C H N O L O G Y N O V E M B E R / D E C E M B E R 2 0 2 1

Page 23: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

[email protected]

New solution streamlines

communication among

radiologists and other

providers, improving

clinical collaboration

The integration of Agfa HealthCare’s XERO®

diagnostic imaging viewer with MicrosoftTeams allows for easy sharing of imagesamong groups of healthcare professionals.Save time tracking down colleagues in thehospital when a review of images is needed—instead, images can be sent quickly andsecurely in a way that is already used bymany physicians and allied professionals.

Physicians requesting a consult can tagspecific members of the channel to review an image. If they fail to respond, the requestcan be escalated via email and repeatednotifications.

Physicians participating in a consult can viewthe images and communicate with each otherusing audio, video and chat. Also available isa markup tool allowing clinicians to interactwith the images using their cursor and toshare the markups in real-time.

The solution can be customized to meet theneeds of specific hospitals or clinicians. Forexample, a COVID button can be added tothe navigation bar in the XERO viewer andprogrammed to transmit images to a‘channel’ of predetermined specialists,including pulmonologists and infectiousdisease experts. Channels can be added forcritical care and cardiology specialists,ophthalmology, dermatology, and others.

The XERO/Microsoft Teams app can savevaluable time over the course of a week,month or year. Agfa HealthCare estimatesthat with a time saving of 10 minutes perconsult, an average hospital could save 75

days of productive time per year. The app canalso be life-saving, if a patient has COVID-19and needs to be placed in quarantine beforeinfecting someone else.

Installation is via a simple plug-in with nodowntime or interruption to viewer use.Following successful implementations in theUK, the companies are now offering thesolution to North American customers.

©2021 Agfa HealthCare Inc.

Integration of XERO® with Teams enablesclinicians and specialists to share images

www.agfahealthcare.com

Page 24: ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING – SEE …

Introducing

Edison True PACSIntelligent diagnostic reading and workflow to allow radiologists to work smarter, not harder

Working smarter - not harderHelps enable fast report turn-around times

Intelligent Workflows

Helps enhance productivity and diagnostic accuracy1

Future-proof your investment

AI Integration

1 Via integration with GE Healthcare or 3rd party AI applications.