Two hundred registered nurses are planning to hold a march in Los Angeles on Aug. 29 to voice their concerns about the threat of artificial intelligence technologies and their ability to provide quality patient care.
The California Nurses Association and National Nurses United say nurses are alarmed about the increasing use of artificial intelligence technologies within patient care, stating that AI cannot replace hands-on care or solve healthcare's staffing crisis.
"Nurses across the country are demanding solutions like safe staffing, resources including personal protective equipment and adequate training, and union representation to advocate for better conditions so we can proudly deliver the best quality care to our patients," Sandy Reding, RN, president of CNA/NNU, said in an Aug. 28 news release. "In our fight to improve patient care, we will not cede our profession to the false solution of unchecked and unregulated technology."
These nurses are joining writers and actors who have been striking in Hollywood due to fear that movie studios would "start generating AI-produced scripts that would need only rewrites," eliminating the need for writers, according to The Washington Post.
Nurses will be carrying signs that read "Flip the script on AI" during the march.
Clínica Alemana Osorno, located in southern Chile, has introduced digital communication workflows that leverage connected care. Digital workflows are helping the clinic create a better patient experience, enhance workforce efficiency and increase access to care across its 46 medical specialties and four facilities.
With more than 300,000 outpatient appointments per year, Clínica Alemana Osorno is one of the largest healthcare providers in southern Chile. This private institution in the Los Lagos Region of northern Patagonia has been in operation for more than 50 years. Today, it has more than 200 medical specialists, 450 employees and a staff growth rate of 20% per year.
The clinic aims to provide high-quality healthcare by putting patients and their needs at the center of all healthcare activities.
The challenges: Competition in the healthcare sector and access to care
Private hospitals in Chile face intense competition for patients and qualified physicians, as patients are free to choose their hospital and doctors are free to choose where to work. Additionally, most healthcare providers are concentrated in the capital (Santiago), forcing patients in the south to travel long distance for specialized medical treatment.
To overcome these challenges and differentiate itself, Clínica Alemana Osorno aims to provide an exceptional experience that surpasses its regional competitors. The clinic prioritizes enhancing patient satisfaction and improving staff efficiency. It also seeks to address the geographic challenge by making southern Chile a more desirable place to live. This involves offering accessible and convenient high-quality care that meets the needs of the local community while ensuring workloads remain manageable for care teams and physicians.
The solution: Driving health information exchange and introducing connected-care applications
Clínica Alemana Osorno has responded to these issues with a comprehensive digital transformation strategy. It introduced solutions for care collaboration and patient engagement in 2022. This was preceded by a thorough preparatory phase, which began in 2021 and included migrating existing data, building an interoperability platform, integrating and aggregating data, and building digital pathways for patients and physicians on top of the platform.
The interoperability platform is crucial for a successful digital transformation. It is the foundation for all applications that drive collaboration between care teams and physicians, which allows them to engage with patients. It also enables applications to connect with and access data from all other systems in use at the clinic.
A highly efficient workforce
The interoperability platform and associated applications have streamlined administrative tasks for clinic physicians, reducing reliance on paper-based processes and the need to search for patient information across multiple systems. With a digital dashboard providing access to comprehensive patient data in one place, doctors can work more efficiently, prioritize patient needs and cultivate stronger patient relationships.
Scheduling was another issue that resulted in inefficiencies. Physicians often must manage multiple schedules for operations and consultations at different locations. Now that all their appointments and tasks are stored in one system, medical staff can plan their workdays much more efficiently and invest their time in high-value tasks.
“I work better now that I have the support of this system. It integrates all my tasks into my schedule, allowing me to better plan my days and be better prepared for all my activities. As a result, I feel I’m more productive and don’t waste any time on unnecessary organizational tasks.” Marko Gjuranovic, MD, Urologist, Clínica Alemana Osorno
Increased access to care
The new online system for appointment scheduling has significantly improved convenience and accessibility for the clinic’s patients. The ability to make, modify and cancel appointments at any time has reduced no-shows. This enables more efficient use of medical systems and treatment rooms, resulting in reduced waiting times and making it easier for the Osorno population to get the care they need.
The interoperability platform and applications have also improved access to care by centralizing healthcare data and enabling analysis through algorithms. This makes it possible to identify at-risk individuals and send automated invitations for checkups or preventive treatments, which contribute to improving overall community health.
“Managing patients’ health data in one system gives us the opportunity to really support patients to engage in their healthcare. This will have a positive impact on prevention, as patients can receive automated reminders for their medical checkup and other preventive offerings from the clinic.” Patricia Martinez, MD, Medical Director, Clínica Alemana Osorno
Furthermore, telehealth services significantly improve access to care by enabling virtual consultations, which eliminate or reduce the need for in-person appointments. Patients can receive healthcare at home through secure virtual connections, instead of traveling to and waiting in healthcare facilities. As a result, more people can benefit from more convenient healthcare services.
The impact of connected systems and a holistic patient view
Clínica Alemana Osorno has benefited greatly from introducing a health information exchange platform and patient engagement applications. The clinic has seen an increase in medical consultations, lab tests and imaging studies. It has become a more popular employer, attracting more physicians. Notably, patient experience has greatly improved, with a Net Promoter Score that is 3.7 times higher than before the transformation. Other achievements include a no-show rate of 9.2%, 39% reduction in waiting times, 15% increase in workforce efficiency and 84% patient satisfaction.
Outlook
The clinic will continue developing the platform and introducing new functionalities for the applications. CEO Leonidas Rosas also expects the clinic to meet its strategic goal of doubling the number of treatments offered by 2028, with the help of the platform and its applications. This is a long-term project that significantly impacts physicians, patients and the entire community. The clinic is well on its way to achieving its aim of making southern Chile a better place to live.
To learn more about successful connected-care projects visit here.
Elizabeth Amirault had never heard of a Narx Score. But she said she learned last year the tool had been used to track her medication use.
During an August 2022 visit to a hospital in Fort Wayne, Indiana, Amirault told a nurse practitioner she was in severe pain, she said. She received a puzzling response.
“Your Narx Score is so high, I can’t give you any narcotics,” she recalled the man saying, as she waited for an MRI before a hip replacement.
Tools like Narx Scores are used to help medical providers review controlled substance prescriptions. They influence, and can limit, the prescribing of painkillers, similar to a credit score influencing the terms of a loan. Narx Scores and an algorithm-generated overdose risk rating are produced by health care technology company Bamboo Health (formerly Appriss Health) in its NarxCare platform.
Such systems are designed to fight the nation’s opioid epidemic, which has led to an alarming number of overdose deaths. The platforms draw on data about prescriptions for controlled substances that states collect to identify patterns of potential problems involving patients and physicians. State and federal health agencies, law enforcement officials, and health care providers have enlisted these tools, but the mechanics behind the formulas used are generally not shared with the public.
Artificial intelligence is working its way into more parts of American life. As AI spreads within the health care landscape, it brings familiar concerns of bias and accuracy and whether government regulation can keep up with rapidly advancing technology.
The use of systems to analyze opioid-prescribing data has sparked questions over whether they have undergone enough independent testing outside of the companies that developed them, making it hard to know how they work.
Lacking the ability to see inside these systems leaves only clues to their potential impact. Some patients say they have been cut off from needed care. Some doctors say their ability to practice medicine has been unfairly threatened. Researchers warn that such technology — despite its benefits — can have unforeseen consequences if it improperly flags patients or doctors.
“We need to see what’s going on to make sure we’re not doing more harm than good,” said Jason Gibbons, a health economist at the Colorado School of Public Health at the University of Colorado’s Anschutz Medical Campus. “We’re concerned that it’s not working as intended, and it’s harming patients.”
Amirault, 34, said she has dealt for years with chronic pain from health conditions such as sciatica, degenerative disc disease, and avascular necrosis, which results from restricted blood supply to the bones.
Subscribe to KFF Health News' free Weekly Edition.
The opioid Percocet offers her some relief. She’d been denied the medication before, but never had been told anything about a Narx Score, she said.
In a chronic pain support group on Facebook, she found others posting about NarxCare, which scores patients based on their supposed risk of prescription drug misuse. She’s convinced her ratings negatively influenced her care.
“Apparently being sick and having a bunch of surgeries and different doctors, all of that goes against me,” Amirault said.
Database-driven tracking has been linked to a decline in opioid prescriptions, but evidence is mixed on its impact on curbing the epidemic. Overdose deaths continue to plague the country, and patients like Amirault have said the monitoring systems leave them feeling stigmatized as well as cut off from pain relief.
The last state to adopt a program, Missouri, is still getting it up and running.
More than 40 states and territories use the technology from Bamboo Health to run PDMPs. That data can be fed into NarxCare, a separate suite of tools to help medical professionals make decisions. Hundreds of health care facilities and five of the top six major pharmacy retailers also use NarxCare, the company said.
The platform generates three Narx Scores based on a patient’s prescription activity involving narcotics, sedatives, and stimulants. A peer-reviewed study showed the “Narx Score metric could serve as a useful initial universal prescription opioid-risk screener.”
NarxCare’s algorithm-generated “Overdose Risk Score” draws on a patient’s medication information from PDMPs — such as the number of doctors writing prescriptions, the number of pharmacies used, and drug dosage — to help medical providers assess a patient’s risk of opioid overdose.
Bamboo Health did not share the specific formula behind the algorithm or address questions about the accuracy of its Overdose Risk Score but said it continues to review and validate the algorithm behind it, based on current overdose trends.
Guidance from the CDC advised clinicians to consult PDMP data before prescribing pain medications. But the agency warned that “special attention should be paid to ensure that PDMP information is not used in a way that is harmful to patients.”
This prescription-drug data has led patients to be dismissed from clinician practices, the CDC said, which could leave patients at risk of being untreated or undertreated for pain. The agency further warned that risk scores may be generated by “proprietary algorithms that are not publicly available” and could lead to biased results.
Bamboo Health said that NarxCare can show providers all of a patient’s scores on one screen, but that these tools should never replace decisions made by physicians.
Some patients say the tools have had an outsize impact on their treatment.
Bev Schechtman, 47, who lives in North Carolina, said she has occasionally used opioids to manage pain flare-ups from Crohn’s disease. As vice president of the Doctor Patient Forum, a chronic pain patient advocacy group, she said she has heard from others reporting medication access problems, many of which she worries are caused by red flags from databases.
“There’s a lot of patients cut off without medication,” according to Schechtman, who said some have turned to illicit sources when they can’t get their prescriptions. “Some patients say to us, ‘It’s either suicide or the streets.’”
Elizabeth Amirault of Indiana has dealt with chronic pain for years. She believes a tool that tracks her prescription drug use negatively influenced her ability to get the medication she needs. (Nicholas Amirault)
The stakes are high for pain patients. Research shows rapid dose changes can increase the risk of withdrawal, depression, anxiety, and even suicide.
Some doctors who treat chronic pain patients say they, too, have been flagged by data systems and then lost their license to practice and were prosecuted.
Lesly Pompy, a pain medicine and addiction specialist in Monroe, Michigan, believes such systems were involved in a legal case against him.
His medical office was raided by a mix of local and federal law enforcement agencies in 2016 because of his patterns in prescribing pain medicine. A year after the raid, Pompy’s medical license was suspended. In 2018, he was indicted on charges of illegally distributing opioid pain medication and health care fraud.
“I knew I was taking care of patients in good faith,” he said. A federal jury in January acquitted him of all charges. He said he’s working to have his license restored.
One firm, Qlarant, a Maryland-based technology company, said it has developed algorithms “to identify questionable behavior patterns and interactions for controlled substances, and for opioids in particular,” involving medical providers.
The company, in an online brochure, said its “extensive government work” includes partnerships with state and federal enforcement entities such as the Department of Health and Human Services’ Office of Inspector General, the FBI, and the Drug Enforcement Administration.
In a promotional video, the company said its algorithms can “analyze a wide variety of data sources,” including court records, insurance claims, drug monitoring data, property records, and incarceration data to flag providers.
William Mapp, the company’s chief technology officer, stressed the final decision about what to do with that information is left up to people — not the algorithms.
Mapp said that “Qlarant’s algorithms are considered proprietary and our intellectual property” and that they have not been independently peer-reviewed.
“We do know that there’s going to be some percentage of error, and we try to let our customers know,” Mapp said. “It sucks when we get it wrong. But we’re constantly trying to get to that point where there are fewer things that are wrong.”
Prosecutions against doctors through the use of prescribing data have attracted the attention of the American Medical Association.
“These unknown and unreviewed algorithms have resulted in physicians having their prescribing privileges immediately suspended without due process or review by a state licensing board — often harming patients in pain because of delays and denials of care,” said Bobby Mukkamala, chair of the AMA’s Substance Use and Pain Care Task Force.
Even critics of drug-tracking systems and algorithms say there is a place for data and artificial intelligence systems in reducing the harms of the opioid crisis.
“It’s just a matter of making sure that the technology is working as intended,” said health economist Gibbons.
Google unveiled a wide array of new generative AI-powered services at its Google Cloud Next 2023 conference in San Francisco on August 29. At the pre-briefing, we got an early look at Google’s new Cloud TPU, A4 virtual machines powered by NVIDIA H100 GPUs and more.
Jump to:
Vertex AI increases capacity, adds other improvements
June Yang, vice president of cloud AI and industry solutions at Google Cloud, announced improvements to Vertex AI, the company’s generative AI platform that helps enterprises train their own AI and machine learning models.
Customers have asked for the ability to input larger amounts of content into PaLM, a foundation model under the Vertex AI platform, Yang said, which led Google to increase its capacity from 4,000 tokens to 32,000 tokens.
Customers have also asked for more languages to be supported in Vertex AI. At the Next ’23 conference, Yang announced PaLM, which resides within the Vertex AI platform, is now available in Arabic, Chinese, Japanese, German, Spanish and more. That’s a total of 38 languages for public use; 100 additional languages are now options in private preview.
Vertex AI Search, which lets users create a search engine inside their AI-powered apps, is available today. “Think about this like Google Search for your business data,” Yang said.
Also available today is Vertex AI Conversation, which is a tool for building chatbots. Search and Conversion were previously available under different product names in Google’s Generative AI App Builder.
Improvements to the Codey foundation model
Codey, the text-to-code model inside Vertex AI, is getting an upgrade. Although details on this upgrade are sparse, Yang said developers should be able to work more efficiently on code generation and code chat.
“Leveraging our Codey foundation model, partners like GitLab are helping developers to stay in the flow by predicting and completing lines of code, generating test cases, explaining code and many more use cases,” Yang noted.
Match your business’ art style with text-to-image AI
Vertex’s text-to-image model will now be able to perform style tuning, or matching a company’s brand and creative guidelines. Organizations need to provide just 10 reference images for Vertex to begin to work within their house style.
New additions to Model Garden, Vertex AI’s model library
Google Cloud has added Meta’s Llama 2 and Anthropic’s Claude 2 to Vertex AI’s model library. The decision to add Llama 2 and Claude 2 to the Google Cloud AI Model Garden is “in line with our commitment to foster an open ecosystem,” Yang said.
“With these additions compared with other hyperscalers, Google Cloud now provides the widest variety of models to choose from, with our first-party Google models, third-party models from partners, as well as open source models on a single platform,” Yang said. “With access to over 100 curated models on Vertex AI, customers can now choose models based on modality, size, performance latency and cost considerations.”
BigQuery and AlloyDB upgrades are ready for preview
Google’s BigQuery Studio — which is a workbench platform for users who work with data and AI — and AlloyDB both have upgrades now available in preview.
BigQuery Studio added to cloud data warehouse preview
BigQuery Studio will be rolled out to Google’s BigQuery cloud data warehouse in preview this week. BigQuery Studio assists with analyzing and exploring data and integrates with Vertex AI. BigQuery Studio is designed to bring data engineering, analytics and predictive analysis together, reducing the time data analytics professionals need to spend switching between tools.
Users of BigQuery can also add Duet AI, Google’s AI assistant, starting now.
AlloyDB enhanced with generative AI
Andy Goodman, vice president and general manager for databases at Google, announced the addition of generative AI capabilities to AlloyDB — Google’s PostgreSQL-compatible database for high-end enterprise workloads — at the pre-brief. AlloyDB includes capabilities for organizations building enterprise AI applications, such as vector search capabilities up to 10 times faster than standard PostgreSQL, Goodman said. Developers can generate vector embeddings within the database to streamline their work. AlloyDB AI integrates with Vertex AI and open source tool ecosystems such as LangChain.
“Databases are at the heart of gen AI innovation, as they help bridge the gap between LLMs and enterprise gen AI apps to deliver accurate, up to date and contextual experiences,” Goodman said.
AlloyDB AI is now available in preview through AlloyDB Omni.
A3 virtual machine supercomputing with NVIDIA for AI training revealed
General availability of the A3 virtual machines running on NVIDIA H100 GPU as a GPU supercomputer will open next month, announced Mark Lohmeyer, vice president general manager for compute and machine learning infrastructure at Google Cloud, during the pre-brief.
The A3 supercomputers’ custom-made 200 Gbps virtual machine infrastructure has GPU-to-GPU data transfers, enabling it to bypass the CPU host. The GPU-to-GPU data transfers power AI training, tuning and scaling with up to 10 times more bandwidth than the previous generation, A2. The training will be three times faster, Lohmeyer said.
NVIDIA “enables us to offer the most comprehensive AI infrastructure portfolio of any cloud,” said Lohmeyer.
Cloud TPU v5e is optimized for generative AI inferencing
Google introduced Cloud TPU v5e, the fifth generation of cloud TPUs optimized for generative AI inferencing. A TPU, or Tensor Processing Unit, is a machine learning accelerator hosted on Google Cloud. The TPU handles the massive amounts of data needed for inferencing, which is a logical process that helps artificial intelligence systems make predictions.
Cloud TPU v5e boasts two times faster performance per dollar for training and 2.5 times better performance per dollar for inferencing compared to the previous-generation TPU, Lohmeyer said.
“(With) the magic of that software and hardware working together with new software technologies like multi-slice, we’re enabling our customers to easily scale their [generative] AI models beyond the physical boundaries of a single TPU pod or a single TPU cluster,” said Lohmeyer. “In other words, a single large AI workload can now span multiple physical TPU clusters, scaling to literally tens of thousands of chips and doing so very cost effectively.”
The new TPU is generally available in preview starting this week.
Introducing Google Kubernetes Engine Enterprise edition
Google Kubernetes Engineer, which many customers use for AI workloads, is getting a boost. The GKE Enterprise edition will include muti-cluster horizontal scaling and GKE’s existing services running across both cloud GPUs and cloud TPUs. Early reports from customers have shown productivity gains of up to 45%, Google said, and reduced software deployment times by more than 70%.
GKE Enterprise Edition will be available in September.
Two hundred registered nurses are planning to hold a march in Los Angeles on Aug. 29 to voice their concerns about the threat of artificial intelligence technologies and their ability to provide quality patient care.
The California Nurses Association and National Nurses United say nurses are alarmed about the increasing use of artificial intelligence technologies within patient care, stating that AI cannot replace hands-on care or solve healthcare's staffing crisis.
"Nurses across the country are demanding solutions like safe staffing, resources including personal protective equipment and adequate training, and union representation to advocate for better conditions so we can proudly deliver the best quality care to our patients," Sandy Reding, RN, president of CNA/NNU, said in an Aug. 28 news release. "In our fight to improve patient care, we will not cede our profession to the false solution of unchecked and unregulated technology."
These nurses are joining writers and actors who have been striking in Hollywood due to fear that movie studios would "start generating AI-produced scripts that would need only rewrites," eliminating the need for writers, according to The Washington Post.
Nurses will be carrying signs that read "Flip the script on AI" during the march.
Clínica Alemana Osorno, located in southern Chile, has introduced digital communication workflows that leverage connected care. Digital workflows are helping the clinic create a better patient experience, enhance workforce efficiency and increase access to care across its 46 medical specialties and four facilities.
With more than 300,000 outpatient appointments per year, Clínica Alemana Osorno is one of the largest healthcare providers in southern Chile. This private institution in the Los Lagos Region of northern Patagonia has been in operation for more than 50 years. Today, it has more than 200 medical specialists, 450 employees and a staff growth rate of 20% per year.
The clinic aims to provide high-quality healthcare by putting patients and their needs at the center of all healthcare activities.
The challenges: Competition in the healthcare sector and access to care
Private hospitals in Chile face intense competition for patients and qualified physicians, as patients are free to choose their hospital and doctors are free to choose where to work. Additionally, most healthcare providers are concentrated in the capital (Santiago), forcing patients in the south to travel long distance for specialized medical treatment.
To overcome these challenges and differentiate itself, Clínica Alemana Osorno aims to provide an exceptional experience that surpasses its regional competitors. The clinic prioritizes enhancing patient satisfaction and improving staff efficiency. It also seeks to address the geographic challenge by making southern Chile a more desirable place to live. This involves offering accessible and convenient high-quality care that meets the needs of the local community while ensuring workloads remain manageable for care teams and physicians.
The solution: Driving health information exchange and introducing connected-care applications
Clínica Alemana Osorno has responded to these issues with a comprehensive digital transformation strategy. It introduced solutions for care collaboration and patient engagement in 2022. This was preceded by a thorough preparatory phase, which began in 2021 and included migrating existing data, building an interoperability platform, integrating and aggregating data, and building digital pathways for patients and physicians on top of the platform.
The interoperability platform is crucial for a successful digital transformation. It is the foundation for all applications that drive collaboration between care teams and physicians, which allows them to engage with patients. It also enables applications to connect with and access data from all other systems in use at the clinic.
A highly efficient workforce
The interoperability platform and associated applications have streamlined administrative tasks for clinic physicians, reducing reliance on paper-based processes and the need to search for patient information across multiple systems. With a digital dashboard providing access to comprehensive patient data in one place, doctors can work more efficiently, prioritize patient needs and cultivate stronger patient relationships.
Scheduling was another issue that resulted in inefficiencies. Physicians often must manage multiple schedules for operations and consultations at different locations. Now that all their appointments and tasks are stored in one system, medical staff can plan their workdays much more efficiently and invest their time in high-value tasks.
“I work better now that I have the support of this system. It integrates all my tasks into my schedule, allowing me to better plan my days and be better prepared for all my activities. As a result, I feel I’m more productive and don’t waste any time on unnecessary organizational tasks.” Marko Gjuranovic, MD, Urologist, Clínica Alemana Osorno
Increased access to care
The new online system for appointment scheduling has significantly improved convenience and accessibility for the clinic’s patients. The ability to make, modify and cancel appointments at any time has reduced no-shows. This enables more efficient use of medical systems and treatment rooms, resulting in reduced waiting times and making it easier for the Osorno population to get the care they need.
The interoperability platform and applications have also improved access to care by centralizing healthcare data and enabling analysis through algorithms. This makes it possible to identify at-risk individuals and send automated invitations for checkups or preventive treatments, which contribute to improving overall community health.
“Managing patients’ health data in one system gives us the opportunity to really support patients to engage in their healthcare. This will have a positive impact on prevention, as patients can receive automated reminders for their medical checkup and other preventive offerings from the clinic.” Patricia Martinez, MD, Medical Director, Clínica Alemana Osorno
Furthermore, telehealth services significantly improve access to care by enabling virtual consultations, which eliminate or reduce the need for in-person appointments. Patients can receive healthcare at home through secure virtual connections, instead of traveling to and waiting in healthcare facilities. As a result, more people can benefit from more convenient healthcare services.
The impact of connected systems and a holistic patient view
Clínica Alemana Osorno has benefited greatly from introducing a health information exchange platform and patient engagement applications. The clinic has seen an increase in medical consultations, lab tests and imaging studies. It has become a more popular employer, attracting more physicians. Notably, patient experience has greatly improved, with a Net Promoter Score that is 3.7 times higher than before the transformation. Other achievements include a no-show rate of 9.2%, 39% reduction in waiting times, 15% increase in workforce efficiency and 84% patient satisfaction.
Outlook
The clinic will continue developing the platform and introducing new functionalities for the applications. CEO Leonidas Rosas also expects the clinic to meet its strategic goal of doubling the number of treatments offered by 2028, with the help of the platform and its applications. This is a long-term project that significantly impacts physicians, patients and the entire community. The clinic is well on its way to achieving its aim of making southern Chile a better place to live.
To learn more about successful connected-care projects visit here.
Elizabeth Amirault had never heard of a Narx Score. But she said she learned last year the tool had been used to track her medication use.
During an August 2022 visit to a hospital in Fort Wayne, Indiana, Amirault told a nurse practitioner she was in severe pain, she said. She received a puzzling response.
“Your Narx Score is so high, I can’t give you any narcotics,” she recalled the man saying, as she waited for an MRI before a hip replacement.
Tools like Narx Scores are used to help medical providers review controlled substance prescriptions. They influence, and can limit, the prescribing of painkillers, similar to a credit score influencing the terms of a loan. Narx Scores and an algorithm-generated overdose risk rating are produced by health care technology company Bamboo Health (formerly Appriss Health) in its NarxCare platform.
Such systems are designed to fight the nation’s opioid epidemic, which has led to an alarming number of overdose deaths. The platforms draw on data about prescriptions for controlled substances that states collect to identify patterns of potential problems involving patients and physicians. State and federal health agencies, law enforcement officials, and health care providers have enlisted these tools, but the mechanics behind the formulas used are generally not shared with the public.
Artificial intelligence is working its way into more parts of American life. As AI spreads within the health care landscape, it brings familiar concerns of bias and accuracy and whether government regulation can keep up with rapidly advancing technology.
The use of systems to analyze opioid-prescribing data has sparked questions over whether they have undergone enough independent testing outside of the companies that developed them, making it hard to know how they work.
Lacking the ability to see inside these systems leaves only clues to their potential impact. Some patients say they have been cut off from needed care. Some doctors say their ability to practice medicine has been unfairly threatened. Researchers warn that such technology — despite its benefits — can have unforeseen consequences if it improperly flags patients or doctors.
“We need to see what’s going on to make sure we’re not doing more harm than good,” said Jason Gibbons, a health economist at the Colorado School of Public Health at the University of Colorado’s Anschutz Medical Campus. “We’re concerned that it’s not working as intended, and it’s harming patients.”
Amirault, 34, said she has dealt for years with chronic pain from health conditions such as sciatica, degenerative disc disease, and avascular necrosis, which results from restricted blood supply to the bones.
Subscribe to KFF Health News' free Weekly Edition.
The opioid Percocet offers her some relief. She’d been denied the medication before, but never had been told anything about a Narx Score, she said.
In a chronic pain support group on Facebook, she found others posting about NarxCare, which scores patients based on their supposed risk of prescription drug misuse. She’s convinced her ratings negatively influenced her care.
“Apparently being sick and having a bunch of surgeries and different doctors, all of that goes against me,” Amirault said.
Database-driven tracking has been linked to a decline in opioid prescriptions, but evidence is mixed on its impact on curbing the epidemic. Overdose deaths continue to plague the country, and patients like Amirault have said the monitoring systems leave them feeling stigmatized as well as cut off from pain relief.
The last state to adopt a program, Missouri, is still getting it up and running.
More than 40 states and territories use the technology from Bamboo Health to run PDMPs. That data can be fed into NarxCare, a separate suite of tools to help medical professionals make decisions. Hundreds of health care facilities and five of the top six major pharmacy retailers also use NarxCare, the company said.
The platform generates three Narx Scores based on a patient’s prescription activity involving narcotics, sedatives, and stimulants. A peer-reviewed study showed the “Narx Score metric could serve as a useful initial universal prescription opioid-risk screener.”
NarxCare’s algorithm-generated “Overdose Risk Score” draws on a patient’s medication information from PDMPs — such as the number of doctors writing prescriptions, the number of pharmacies used, and drug dosage — to help medical providers assess a patient’s risk of opioid overdose.
Bamboo Health did not share the specific formula behind the algorithm or address questions about the accuracy of its Overdose Risk Score but said it continues to review and validate the algorithm behind it, based on current overdose trends.
Guidance from the CDC advised clinicians to consult PDMP data before prescribing pain medications. But the agency warned that “special attention should be paid to ensure that PDMP information is not used in a way that is harmful to patients.”
This prescription-drug data has led patients to be dismissed from clinician practices, the CDC said, which could leave patients at risk of being untreated or undertreated for pain. The agency further warned that risk scores may be generated by “proprietary algorithms that are not publicly available” and could lead to biased results.
Bamboo Health said that NarxCare can show providers all of a patient’s scores on one screen, but that these tools should never replace decisions made by physicians.
Some patients say the tools have had an outsize impact on their treatment.
Bev Schechtman, 47, who lives in North Carolina, said she has occasionally used opioids to manage pain flare-ups from Crohn’s disease. As vice president of the Doctor Patient Forum, a chronic pain patient advocacy group, she said she has heard from others reporting medication access problems, many of which she worries are caused by red flags from databases.
“There’s a lot of patients cut off without medication,” according to Schechtman, who said some have turned to illicit sources when they can’t get their prescriptions. “Some patients say to us, ‘It’s either suicide or the streets.’”
Elizabeth Amirault of Indiana has dealt with chronic pain for years. She believes a tool that tracks her prescription drug use negatively influenced her ability to get the medication she needs. (Nicholas Amirault)
The stakes are high for pain patients. Research shows rapid dose changes can increase the risk of withdrawal, depression, anxiety, and even suicide.
Some doctors who treat chronic pain patients say they, too, have been flagged by data systems and then lost their license to practice and were prosecuted.
Lesly Pompy, a pain medicine and addiction specialist in Monroe, Michigan, believes such systems were involved in a legal case against him.
His medical office was raided by a mix of local and federal law enforcement agencies in 2016 because of his patterns in prescribing pain medicine. A year after the raid, Pompy’s medical license was suspended. In 2018, he was indicted on charges of illegally distributing opioid pain medication and health care fraud.
“I knew I was taking care of patients in good faith,” he said. A federal jury in January acquitted him of all charges. He said he’s working to have his license restored.
One firm, Qlarant, a Maryland-based technology company, said it has developed algorithms “to identify questionable behavior patterns and interactions for controlled substances, and for opioids in particular,” involving medical providers.
The company, in an online brochure, said its “extensive government work” includes partnerships with state and federal enforcement entities such as the Department of Health and Human Services’ Office of Inspector General, the FBI, and the Drug Enforcement Administration.
In a promotional video, the company said its algorithms can “analyze a wide variety of data sources,” including court records, insurance claims, drug monitoring data, property records, and incarceration data to flag providers.
William Mapp, the company’s chief technology officer, stressed the final decision about what to do with that information is left up to people — not the algorithms.
Mapp said that “Qlarant’s algorithms are considered proprietary and our intellectual property” and that they have not been independently peer-reviewed.
“We do know that there’s going to be some percentage of error, and we try to let our customers know,” Mapp said. “It sucks when we get it wrong. But we’re constantly trying to get to that point where there are fewer things that are wrong.”
Prosecutions against doctors through the use of prescribing data have attracted the attention of the American Medical Association.
“These unknown and unreviewed algorithms have resulted in physicians having their prescribing privileges immediately suspended without due process or review by a state licensing board — often harming patients in pain because of delays and denials of care,” said Bobby Mukkamala, chair of the AMA’s Substance Use and Pain Care Task Force.
Even critics of drug-tracking systems and algorithms say there is a place for data and artificial intelligence systems in reducing the harms of the opioid crisis.
“It’s just a matter of making sure that the technology is working as intended,” said health economist Gibbons.
Google unveiled a wide array of new generative AI-powered services at its Google Cloud Next 2023 conference in San Francisco on August 29. At the pre-briefing, we got an early look at Google’s new Cloud TPU, A4 virtual machines powered by NVIDIA H100 GPUs and more.
Jump to:
Vertex AI increases capacity, adds other improvements
June Yang, vice president of cloud AI and industry solutions at Google Cloud, announced improvements to Vertex AI, the company’s generative AI platform that helps enterprises train their own AI and machine learning models.
Customers have asked for the ability to input larger amounts of content into PaLM, a foundation model under the Vertex AI platform, Yang said, which led Google to increase its capacity from 4,000 tokens to 32,000 tokens.
Customers have also asked for more languages to be supported in Vertex AI. At the Next ’23 conference, Yang announced PaLM, which resides within the Vertex AI platform, is now available in Arabic, Chinese, Japanese, German, Spanish and more. That’s a total of 38 languages for public use; 100 additional languages are now options in private preview.
Vertex AI Search, which lets users create a search engine inside their AI-powered apps, is available today. “Think about this like Google Search for your business data,” Yang said.
Also available today is Vertex AI Conversation, which is a tool for building chatbots. Search and Conversion were previously available under different product names in Google’s Generative AI App Builder.
Improvements to the Codey foundation model
Codey, the text-to-code model inside Vertex AI, is getting an upgrade. Although details on this upgrade are sparse, Yang said developers should be able to work more efficiently on code generation and code chat.
“Leveraging our Codey foundation model, partners like GitLab are helping developers to stay in the flow by predicting and completing lines of code, generating test cases, explaining code and many more use cases,” Yang noted.
Match your business’ art style with text-to-image AI
Vertex’s text-to-image model will now be able to perform style tuning, or matching a company’s brand and creative guidelines. Organizations need to provide just 10 reference images for Vertex to begin to work within their house style.
New additions to Model Garden, Vertex AI’s model library
Google Cloud has added Meta’s Llama 2 and Anthropic’s Claude 2 to Vertex AI’s model library. The decision to add Llama 2 and Claude 2 to the Google Cloud AI Model Garden is “in line with our commitment to foster an open ecosystem,” Yang said.
“With these additions compared with other hyperscalers, Google Cloud now provides the widest variety of models to choose from, with our first-party Google models, third-party models from partners, as well as open source models on a single platform,” Yang said. “With access to over 100 curated models on Vertex AI, customers can now choose models based on modality, size, performance latency and cost considerations.”
BigQuery and AlloyDB upgrades are ready for preview
Google’s BigQuery Studio — which is a workbench platform for users who work with data and AI — and AlloyDB both have upgrades now available in preview.
BigQuery Studio added to cloud data warehouse preview
BigQuery Studio will be rolled out to Google’s BigQuery cloud data warehouse in preview this week. BigQuery Studio assists with analyzing and exploring data and integrates with Vertex AI. BigQuery Studio is designed to bring data engineering, analytics and predictive analysis together, reducing the time data analytics professionals need to spend switching between tools.
Users of BigQuery can also add Duet AI, Google’s AI assistant, starting now.
AlloyDB enhanced with generative AI
Andy Goodman, vice president and general manager for databases at Google, announced the addition of generative AI capabilities to AlloyDB — Google’s PostgreSQL-compatible database for high-end enterprise workloads — at the pre-brief. AlloyDB includes capabilities for organizations building enterprise AI applications, such as vector search capabilities up to 10 times faster than standard PostgreSQL, Goodman said. Developers can generate vector embeddings within the database to streamline their work. AlloyDB AI integrates with Vertex AI and open source tool ecosystems such as LangChain.
“Databases are at the heart of gen AI innovation, as they help bridge the gap between LLMs and enterprise gen AI apps to deliver accurate, up to date and contextual experiences,” Goodman said.
AlloyDB AI is now available in preview through AlloyDB Omni.
A3 virtual machine supercomputing with NVIDIA for AI training revealed
General availability of the A3 virtual machines running on NVIDIA H100 GPU as a GPU supercomputer will open next month, announced Mark Lohmeyer, vice president general manager for compute and machine learning infrastructure at Google Cloud, during the pre-brief.
The A3 supercomputers’ custom-made 200 Gbps virtual machine infrastructure has GPU-to-GPU data transfers, enabling it to bypass the CPU host. The GPU-to-GPU data transfers power AI training, tuning and scaling with up to 10 times more bandwidth than the previous generation, A2. The training will be three times faster, Lohmeyer said.
NVIDIA “enables us to offer the most comprehensive AI infrastructure portfolio of any cloud,” said Lohmeyer.
Cloud TPU v5e is optimized for generative AI inferencing
Google introduced Cloud TPU v5e, the fifth generation of cloud TPUs optimized for generative AI inferencing. A TPU, or Tensor Processing Unit, is a machine learning accelerator hosted on Google Cloud. The TPU handles the massive amounts of data needed for inferencing, which is a logical process that helps artificial intelligence systems make predictions.
Cloud TPU v5e boasts two times faster performance per dollar for training and 2.5 times better performance per dollar for inferencing compared to the previous-generation TPU, Lohmeyer said.
“(With) the magic of that software and hardware working together with new software technologies like multi-slice, we’re enabling our customers to easily scale their [generative] AI models beyond the physical boundaries of a single TPU pod or a single TPU cluster,” said Lohmeyer. “In other words, a single large AI workload can now span multiple physical TPU clusters, scaling to literally tens of thousands of chips and doing so very cost effectively.”
The new TPU is generally available in preview starting this week.
Introducing Google Kubernetes Engine Enterprise edition
Google Kubernetes Engineer, which many customers use for AI workloads, is getting a boost. The GKE Enterprise edition will include muti-cluster horizontal scaling and GKE’s existing services running across both cloud GPUs and cloud TPUs. Early reports from customers have shown productivity gains of up to 45%, Google said, and reduced software deployment times by more than 70%.
GKE Enterprise Edition will be available in September.
