A glimpse at AI’s near future in helping health care professionals make better decisions to improve care quality, patient safety, and efficiency.
NEJM Catalyst
With Eyal Zimlichman, MD, MSc & David Bates, MD, MSc
October 25, 2021
imagem:healthitanalytics
Summary
The Deputy Director General, Chief Medical Officer, and Chief Innovation Officer for Sheba Medical Center and the Chief of the Division of General Internal Medicine and Primary Care for Brigham and Women’s Hospital discuss the role of artificial intelligence on various aspects of health care delivery in 10 years and what organizations need to do to get ready for its use in decision support, digital pathology, control towers, innovation and implementation programs, sensing technologies, apps, precision medicine, and policy change.
Zimlichman describes how AI has already impacted the field of radiology and says that it will expand to other fields soon, helping to improve health care professionals’ decision-making.
“We’re going to be able to make better decisions to improve quality, to improve patient safety, and to improve efficiency,” he says.
We’ll also see AI help with automation.
But to prepare for the growth of AI in clinical care, we need to resolve a number of issues, particularly around workflow.
“We’re going to be able to make better decisions to improve quality, to improve patient safety, and to improve efficiency.” We’ll also see AI help with automation.
But to prepare for the growth of AI in clinical care, we need to resolve a number of issues, particularly around workflow.
Digital pathology is now where radiology was 5 or 10 years ago. Sheba Medical Center, for example, scans 1,500 slides per day.
The ability of AI to browse through pathology slides will be transformational and will improve patient safety in terms of making better decisions, not missing out on diagnoses, and in making new discoveries.
The ability of AI to browse through pathology slides will be transformational and will improve patient safety in terms of making better decisions, not missing out on diagnoses, and in making new discoveries.
Hospitals are running well below where they need to be in terms of efficiencies. But that load can be dealt with when you have the right operational tools, says Zimlichman.
Control towers will play a growing role in hospitals, helping us to greatly optimize patient flow and patient placement, and gain a better understanding of which patients are decompensating or deteriorating and may need to be moved quickly.
AI in these control towers will allow for much less manpower and more efficiency — an important transformation for smart hospitals.
Hospitals are running well below where they need to be in terms of efficiencies.
AI in these control towers will allow for much less manpower and more efficiency — an important transformation for smart hospitals.
Many organizations struggle with how to assimilate new approaches like AI, says Bates.
Something like Sheba’s Accelerate, Resign, Collaborate, or ARC, innovation program could help with this, where the goal is to accelerate innovation and redesign health care through collaborations by the year 2030.
To get to where we want to be with AI, Zimlichman says that we need robust programs that generate innovation both internally at academic medical centers and externally with startups, and that implement these innovations on a large scale, at multiple sites and even internationally.
“We seem to learn a lot when we implement together across institutions. What will not work in one institution might work in another, and we can learn a lot from that,” he says.
The ability to predict and prevent will play a huge role in driving down the costs of care and in the overall health of the population.
Sensing technologies, which can continuously measure several hundred parameters in an individual noninvasively, are another exciting area.
But while hundreds of parameters sound great, Zimlichman cautions that this is also a problem. How are we going to make sense of the data from hundreds of parameters coming from a patient walking around with a patch on their arm?
We need to determine what we need from that data, applying AI tools to sort and clean out the noise. Once we know which data points are useful, one area where sensing technologies could have particular value is in chronic disease management, to be able to see where the disease is progressing and react before hospitalization becomes necessary.
Coupling that sensor data with precision medicine will provide us the means of being able to predict and prevent. Another area of value for sensing technologies is acute care, especially for hospital and home, being able to monitor and receive smart alerts about patients.
The app marketplace is enormous. There are now thousands of health-related apps, yet as Bates says, many are hardly used. Many countries are now contemplating app regulation and recommending which apps to use. In addition to tougher regulation, we’ll also see more prescribing of apps.
Perhaps a clinician will prescribe a medication and pair it with a monitoring app relevant to that specific medication, or prescribe apps on a wearable sensor, or with medical devices, says Zimlichman.
Bates adds that most people rely on star ratings to pick an app, and yet those star ratings have little relevance to how the app actually performs, and so having apps prescribed will be a more objective use of them.
Precision medicine may feel like a disappointment; we haven’t seen this field advance as much as we thought we would after the human genome was first sequenced.
But people need to realize that precision medicine goes well beyond genomics. “We need to start looking at multi-omics,” Zimlichman says, including genomics, metabolomics, proteomics, microbiomics, and so on.
It goes beyond genomics to digital pathology, digital radiology, and other fields as well.
Having AI sort through the different multi-omics to find distinctions on predicting which drugs to prescribe and which patients will develop which complications is going to be central in the future.
“The ability to predict and prevent will play a huge role in driving down the costs of care and in the overall health of the population,” Zimlichman adds.
There are so many solutions available, and yet, as Bates notes, many get stuck in niches. We need new policies to help organizations adopt these solutions and change care delivery.
One critical issue Zimlichman calls out is addressing payments through policy. He points to how changes in telemedicine policy helped ramp up those services during Covid-19. Coming up with the right incentives can push innovation to change health care in many ways.
We also need regulation on AI, making sure we use the right tools and not those that cause damage. And we need financial incentives for precision medicine.
Video
The Coming AI Transformation of Care Delivery
From the NEJM Catalyst event Technology & Data in 2030 , sponsored by Optum, September 9, 2021. Eyal Zimlichman, MD…catalyst.nejm.org
About the authors
Eyal Zimlichman, MD, MSc &
Deputy Director General, Chief Medical Officer, and Chief Innovation Officer, Sheba Medical Center, Ramat Gan, Israel
David Bates, MD, MSc
Chief of the Division of General Internal Medicine and Primary Care, Brigham and Women’s Hospital, Boston, Massachusetts, USA; Medical Director, Clinical and Quality Analysis, Information Systems, Mass General Brigham, Inc., Boston, Massachusetts, USA
Originally published at https://catalyst.nejm.org on October 25, 2021.
From the NEJM Catalyst event Technology & Data in 2030, sponsored by Optum, September 9, 2021.
A glimpse at AI’s near future in helping health care professionals make better decisions to improve care quality, patient safety, and efficiency.
NEJM Catalyst
With Eyal Zimlichman, MD, MSc & David Bates, MD, MSc
October 25, 2021
imagem:healthitanalytics
Summary
The Deputy Director General, Chief Medical Officer, and Chief Innovation Officer for Sheba Medical Center and the Chief of the Division of General Internal Medicine and Primary Care for Brigham and Women’s Hospital discuss the role of artificial intelligence on various aspects of health care delivery in 10 years and what organizations need to do to get ready for its use in decision support, digital pathology, control towers, innovation and implementation programs, sensing technologies, apps, precision medicine, and policy change.
Zimlichman describes how AI has already impacted the field of radiology and says that it will expand to other fields soon, helping to improve health care professionals’ decision-making.
“We’re going to be able to make better decisions to improve quality, to improve patient safety, and to improve efficiency,” he says.
We’ll also see AI help with automation.
But to prepare for the growth of AI in clinical care, we need to resolve a number of issues, particularly around workflow.
“We’re going to be able to make better decisions to improve quality, to improve patient safety, and to improve efficiency.” We’ll also see AI help with automation.
But to prepare for the growth of AI in clinical care, we need to resolve a number of issues, particularly around workflow.
Digital pathology is now where radiology was 5 or 10 years ago. Sheba Medical Center, for example, scans 1,500 slides per day.
The ability of AI to browse through pathology slides will be transformational and will improve patient safety in terms of making better decisions, not missing out on diagnoses, and in making new discoveries.
The ability of AI to browse through pathology slides will be transformational and will improve patient safety in terms of making better decisions, not missing out on diagnoses, and in making new discoveries.
Hospitals are running well below where they need to be in terms of efficiencies. But that load can be dealt with when you have the right operational tools, says Zimlichman.
Control towers will play a growing role in hospitals, helping us to greatly optimize patient flow and patient placement, and gain a better understanding of which patients are decompensating or deteriorating and may need to be moved quickly.
AI in these control towers will allow for much less manpower and more efficiency — an important transformation for smart hospitals.
Hospitals are running well below where they need to be in terms of efficiencies.
AI in these control towers will allow for much less manpower and more efficiency — an important transformation for smart hospitals.
Many organizations struggle with how to assimilate new approaches like AI, says Bates.
Something like Sheba’s Accelerate, Resign, Collaborate, or ARC, innovation program could help with this, where the goal is to accelerate innovation and redesign health care through collaborations by the year 2030.
To get to where we want to be with AI, Zimlichman says that we need robust programs that generate innovation both internally at academic medical centers and externally with startups, and that implement these innovations on a large scale, at multiple sites and even internationally.
“We seem to learn a lot when we implement together across institutions. What will not work in one institution might work in another, and we can learn a lot from that,” he says.
The ability to predict and prevent will play a huge role in driving down the costs of care and in the overall health of the population.
Sensing technologies, which can continuously measure several hundred parameters in an individual noninvasively, are another exciting area.
But while hundreds of parameters sound great, Zimlichman cautions that this is also a problem. How are we going to make sense of the data from hundreds of parameters coming from a patient walking around with a patch on their arm?
We need to determine what we need from that data, applying AI tools to sort and clean out the noise. Once we know which data points are useful, one area where sensing technologies could have particular value is in chronic disease management, to be able to see where the disease is progressing and react before hospitalization becomes necessary.
Coupling that sensor data with precision medicine will provide us the means of being able to predict and prevent. Another area of value for sensing technologies is acute care, especially for hospital and home, being able to monitor and receive smart alerts about patients.
The app marketplace is enormous. There are now thousands of health-related apps, yet as Bates says, many are hardly used. Many countries are now contemplating app regulation and recommending which apps to use. In addition to tougher regulation, we’ll also see more prescribing of apps.
Perhaps a clinician will prescribe a medication and pair it with a monitoring app relevant to that specific medication, or prescribe apps on a wearable sensor, or with medical devices, says Zimlichman.
Bates adds that most people rely on star ratings to pick an app, and yet those star ratings have little relevance to how the app actually performs, and so having apps prescribed will be a more objective use of them.
Precision medicine may feel like a disappointment; we haven’t seen this field advance as much as we thought we would after the human genome was first sequenced.
But people need to realize that precision medicine goes well beyond genomics. “We need to start looking at multi-omics,” Zimlichman says, including genomics, metabolomics, proteomics, microbiomics, and so on.
It goes beyond genomics to digital pathology, digital radiology, and other fields as well.
Having AI sort through the different multi-omics to find distinctions on predicting which drugs to prescribe and which patients will develop which complications is going to be central in the future.
“The ability to predict and prevent will play a huge role in driving down the costs of care and in the overall health of the population,” Zimlichman adds.
There are so many solutions available, and yet, as Bates notes, many get stuck in niches. We need new policies to help organizations adopt these solutions and change care delivery.
One critical issue Zimlichman calls out is addressing payments through policy. He points to how changes in telemedicine policy helped ramp up those services during Covid-19. Coming up with the right incentives can push innovation to change health care in many ways.
We also need regulation on AI, making sure we use the right tools and not those that cause damage. And we need financial incentives for precision medicine.
Video
The Coming AI Transformation of Care Delivery
From the NEJM Catalyst event Technology & Data in 2030 , sponsored by Optum, September 9, 2021. Eyal Zimlichman, MD…catalyst.nejm.org
About the authors
Eyal Zimlichman, MD, MSc &
Deputy Director General, Chief Medical Officer, and Chief Innovation Officer, Sheba Medical Center, Ramat Gan, Israel
David Bates, MD, MSc
Chief of the Division of General Internal Medicine and Primary Care, Brigham and Women’s Hospital, Boston, Massachusetts, USA; Medical Director, Clinical and Quality Analysis, Information Systems, Mass General Brigham, Inc., Boston, Massachusetts, USA
Originally published at https://catalyst.nejm.org on October 25, 2021.
From the NEJM Catalyst event Technology & Data in 2030, sponsored by Optum, September 9, 2021.
