Saving babies: A new data capture project
A new data-capture project is helping Sick Kids care for its most vulnerable patients
By Trevor Marshall
October 1, 2009
October 1, 2009
In Greek mythology, Artemis played several important roles, including goddess of the hunt and goddess of fertility and childbirth. According to one story, her role as a guardian of young children and women in childbirth started the day she was born, when Artemis aided her mother Leto across the straits between Ortygia and Delos and then helped Leto give birth to her twin brother Apollo. Pretty impressive for a newborn.
Of course, human newborns are not so ready for the world, and that helplessness is all the more serious when babies are born prematurely. The health-care challenges for these patients are so severe that professionals in the Neonatal Intensive Care Unit (NICU) can use all the help they can get. So, how about a guardian to watch over their fragile charges?
That, put simply, is the idea behind Artemis, a collaborative project involving Toronto’s Hospital for Sick Children, the University of Ontario Institute of Technology (UOIT) in Oshawa, Ont., and IBM Canada. The goal of the project is to capture and analyze vast amounts of physiological data from premature babies and then present that information to physicians and nurses.
Too much information
The Artemis project addresses a couple of significant challenges caregivers face in NICU environments. Dr. Carolyn McGregor, Canada Research Chair (CRC) in Health Informatics at UOIT, said the first challenge is that monitoring machines generate much more data than a caregiver can absorb. “Data is coming out of those machines at a rate of a thousand readings per second, and we’re taking that down to make a note of a single reading every 30 or 60 minutes,” she said. “There’s an enormous amount of data loss.”
The second problem is near real-time monitoring. Even if humans could interpret that huge data stream, they’d rarely get the chance in the NICU environment. “There are frequent disruptions,” said Dr. Andrew James, associate professor in the Department of Paediatrics at the University of Toronto and a pediatric otorhinolaryngologist in the Neonatology Department at the Hospital for Sick Children. “We might be trying to make sense of a big chunk of information but then we’re called away to something urgent. That goes on continuously throughout the day.”
Canada calling
McGregor is a native of Australia and started her research there after experience working on informatics systems for the financial and retail sectors. Early on in her research, she met the founding dean of UOIT at a conference in Singapore, and soon found herself packing for a move to Ontario, lured by $500,000 in funding through the CRC program plus a $250,000 injection to establish her lab, courtesy of the Canada Foundation for Innovation, Ontario’s Ministry of Research and Innovation, and the university.
McGregor said the support has been invaluable. “The laboratory I’ve set up here is the first of its kind, anywhere in the world, where we have medical devices from all different makers that we can use for computer science and research,” she said, adding that work to digitize health records through the Canada Health Infoway program contributes to a remarkable environment for her research. “There are unique models of innovation in Canada in terms of the electronic health record, very strong support through the Canada Research Chair program and a very good team of people I’m collaborating with, as well as the University of Ontario, which has been incredibly supportive.”
IBM signs on
That collaboration includes a team from IBM, which came on board about two years ago following a presentation by McGregor at the company’s Thomas J. Watson Research Centre in Yorktown Heights, N.Y. “We were about three or four years into developing a whole new class of software that we referred to as System S, which handles streaming data,” said Don Aldridge, general manager of Research and Life Sciences at IBM Canada. “One thing we don’t have as a company is access to hospitals. We don’t have patients. We look to our partners for access to data in many cases and the combination of Dr. McGregor’s informatics skills and Dr. James, with access to the kids, seemed to be a really good marriage.”
Baby steps
As a first step, McGregor, James and the IBM team are exploring ways to use the wealth of data from patient monitoring machines to enable earlier detection of nosocomial infections in the NICU (see sidebar). This could help reduce infant mortality and even improve the overall ongoing health of the patient.
The Artemis project is still at a relatively early stage. Over the past year, the team has accomplished fundamental development work, from fleshing out the theory to building the technical framework necessary to capture and analyze data, and creating some rules to get the team started. The team has run data from another study through the system, offline, to test and refine these rules, worked with the hospital’s IT department to ensure its gear won’t interfere with other critical information systems, and applied to the hospital’s research ethics board for permission to proceed with data collection in near real-time from actual patients. This phase, which is about to proceed, allows researchers to confirm data is being streamed without distortion and to further refine the rules used to search for nosocomial infections.
It will be a few more years before all clinical trials have been completed and the system is ready to be put to regular use, but James is confident that their work is going to give clinicians a new tool to hunt down and eradicate infections. “The direction of this research is incredibly exciting and this collaborative partnership is phenomenal,” he said.
As an engineer, McGregor is equally excited by the potential for those with technology skills to make a difference. “It’s not only about saving lives, but also about giving these babies the best, healthiest start we can,” she said. “This is a very good example of where—rather than the one-to-one relationship that doctors have—we have the potential to help thousands.”
Why nosocomial infections?
Nosocomial infections are more commonly known as hospital-acquired infections, and they are a serious issue for the health-care sector. Patients in a neonatal intensive care unit (NICU) are at particular risk for nosocomial infections for a number of reasons. First, many have tubes entering their bodies in several locations, each presenting a potential gateway for infection. Second, premature babies have immune systems that are both immature and inexperienced. In other words, not only are the body’s defence mechanisms still developing, they’re also learning what they need to defend against, and how.
Dr. Andrew James describes a model for the progress of nosocomial infections this way:
Phase zero: there is no infection
Phase one: the infection is present, but there are currently no means of detecting it
Phase two: the infection is present, as are some very soft, subtle, non-specific clinical signs of infection that could mean anything
Phase three: the infection is present and the signs are more obvious
Phase four: the infection is present and the signs are very obvious
James admits that even in phase two, the signs are so subtle that currently it’s difficult to diagnose an infection. He hopes Artemis will enable NICU clinicians to do much better, and even identify infections in phase one. “If we can detect changes in heart rate, breathing and blood pressure in that phase then we can act earlier,” he said. “If our research yields positive results then the benefit to our patient population could be huge.”
Combined with clinical investigation and treatment regimens, James hopes Artemis will allow caregivers to meet nosocomial infections head on. “We hope we could either prevent the infection from becoming obvious or reduce the severity of it,” he said. “That would be good for the kids.”
Of course, human newborns are not so ready for the world, and that helplessness is all the more serious when babies are born prematurely. The health-care challenges for these patients are so severe that professionals in the Neonatal Intensive Care Unit (NICU) can use all the help they can get. So, how about a guardian to watch over their fragile charges?
That, put simply, is the idea behind Artemis, a collaborative project involving Toronto’s Hospital for Sick Children, the University of Ontario Institute of Technology (UOIT) in Oshawa, Ont., and IBM Canada. The goal of the project is to capture and analyze vast amounts of physiological data from premature babies and then present that information to physicians and nurses.
Too much information
The Artemis project addresses a couple of significant challenges caregivers face in NICU environments. Dr. Carolyn McGregor, Canada Research Chair (CRC) in Health Informatics at UOIT, said the first challenge is that monitoring machines generate much more data than a caregiver can absorb. “Data is coming out of those machines at a rate of a thousand readings per second, and we’re taking that down to make a note of a single reading every 30 or 60 minutes,” she said. “There’s an enormous amount of data loss.”
The second problem is near real-time monitoring. Even if humans could interpret that huge data stream, they’d rarely get the chance in the NICU environment. “There are frequent disruptions,” said Dr. Andrew James, associate professor in the Department of Paediatrics at the University of Toronto and a pediatric otorhinolaryngologist in the Neonatology Department at the Hospital for Sick Children. “We might be trying to make sense of a big chunk of information but then we’re called away to something urgent. That goes on continuously throughout the day.”
Canada calling
McGregor is a native of Australia and started her research there after experience working on informatics systems for the financial and retail sectors. Early on in her research, she met the founding dean of UOIT at a conference in Singapore, and soon found herself packing for a move to Ontario, lured by $500,000 in funding through the CRC program plus a $250,000 injection to establish her lab, courtesy of the Canada Foundation for Innovation, Ontario’s Ministry of Research and Innovation, and the university.
McGregor said the support has been invaluable. “The laboratory I’ve set up here is the first of its kind, anywhere in the world, where we have medical devices from all different makers that we can use for computer science and research,” she said, adding that work to digitize health records through the Canada Health Infoway program contributes to a remarkable environment for her research. “There are unique models of innovation in Canada in terms of the electronic health record, very strong support through the Canada Research Chair program and a very good team of people I’m collaborating with, as well as the University of Ontario, which has been incredibly supportive.”
IBM signs on
That collaboration includes a team from IBM, which came on board about two years ago following a presentation by McGregor at the company’s Thomas J. Watson Research Centre in Yorktown Heights, N.Y. “We were about three or four years into developing a whole new class of software that we referred to as System S, which handles streaming data,” said Don Aldridge, general manager of Research and Life Sciences at IBM Canada. “One thing we don’t have as a company is access to hospitals. We don’t have patients. We look to our partners for access to data in many cases and the combination of Dr. McGregor’s informatics skills and Dr. James, with access to the kids, seemed to be a really good marriage.”
Baby steps
As a first step, McGregor, James and the IBM team are exploring ways to use the wealth of data from patient monitoring machines to enable earlier detection of nosocomial infections in the NICU (see sidebar). This could help reduce infant mortality and even improve the overall ongoing health of the patient.
The Artemis project is still at a relatively early stage. Over the past year, the team has accomplished fundamental development work, from fleshing out the theory to building the technical framework necessary to capture and analyze data, and creating some rules to get the team started. The team has run data from another study through the system, offline, to test and refine these rules, worked with the hospital’s IT department to ensure its gear won’t interfere with other critical information systems, and applied to the hospital’s research ethics board for permission to proceed with data collection in near real-time from actual patients. This phase, which is about to proceed, allows researchers to confirm data is being streamed without distortion and to further refine the rules used to search for nosocomial infections.
It will be a few more years before all clinical trials have been completed and the system is ready to be put to regular use, but James is confident that their work is going to give clinicians a new tool to hunt down and eradicate infections. “The direction of this research is incredibly exciting and this collaborative partnership is phenomenal,” he said.
As an engineer, McGregor is equally excited by the potential for those with technology skills to make a difference. “It’s not only about saving lives, but also about giving these babies the best, healthiest start we can,” she said. “This is a very good example of where—rather than the one-to-one relationship that doctors have—we have the potential to help thousands.”
Why nosocomial infections?
Nosocomial infections are more commonly known as hospital-acquired infections, and they are a serious issue for the health-care sector. Patients in a neonatal intensive care unit (NICU) are at particular risk for nosocomial infections for a number of reasons. First, many have tubes entering their bodies in several locations, each presenting a potential gateway for infection. Second, premature babies have immune systems that are both immature and inexperienced. In other words, not only are the body’s defence mechanisms still developing, they’re also learning what they need to defend against, and how.
Dr. Andrew James describes a model for the progress of nosocomial infections this way:
Phase zero: there is no infection
Phase one: the infection is present, but there are currently no means of detecting it
Phase two: the infection is present, as are some very soft, subtle, non-specific clinical signs of infection that could mean anything
Phase three: the infection is present and the signs are more obvious
Phase four: the infection is present and the signs are very obvious
James admits that even in phase two, the signs are so subtle that currently it’s difficult to diagnose an infection. He hopes Artemis will enable NICU clinicians to do much better, and even identify infections in phase one. “If we can detect changes in heart rate, breathing and blood pressure in that phase then we can act earlier,” he said. “If our research yields positive results then the benefit to our patient population could be huge.”
Combined with clinical investigation and treatment regimens, James hopes Artemis will allow caregivers to meet nosocomial infections head on. “We hope we could either prevent the infection from becoming obvious or reduce the severity of it,” he said. “That would be good for the kids.”
