User login
Editor’s note: As of this writing, the following proposed health insurance policy from Blue Cross and Blue Shield of North Carolina is still active. The Coalition of State Rheumatology Organizations and other rheumatology advocacy groups are in ongoing discussions with the health insurer and hope to have major changes to this policy implemented.
While AI has been in our world for years, it is expanding by the minute, perhaps by the nanosecond, within the health care sector. The $6.7 billion dollar health care AI market in 2020 is expected to climb to more than $120 billion by 2028. There are many questions regarding the application of AI in our world. Is it a mere instructional algorithm that computes things in a much faster way, or does it create a new story based on the information it has access to? Does it engender excitement or fear ... or both? Remember HAL? As we have seen throughout history with new inventions and technologies, there are risks and rewards. Even the best can have harmful unintended consequences. AI is no different, particularly when it comes to health care. In this case, AI can get a bad name if it is utilized along with biased data input and bad policy.
Shared savings
Here is where “shared savings” comes into play. A shared savings program starts with a baseline cost analysis of a particular care plan and then tracks costs (performance) going forward after certain changes to the original care plan are instituted. If savings are accrued when compared with baseline spending, those savings are shared with the providers of the care. Depending on how the shared savings program is implemented, the optics can be very bad if it appears as though physicians are being paid to reduce care.
‘The volunteer opportunity’
Recently, Blue Cross and Blue Shield of North Carolina, in partnership with Outcomes Matter Innovations, a data analysis company that uses AI/machine-learning technology, offered rheumatologists a new voluntary shared savings, value-based care (VBC) “opportunity.” Rheumatologists would be able to “utilize a web-based machine-learning technology platform that suggests evidence-based care pathways” in the treatment of rheumatoid arthritis and psoriatic arthritis (PsA). The VBC/shared savings model uses the AI platform to propose two different pathways. One model would delay the start of biologics or Janus kinase inhibitors (JAKi), and the second model would taper and/or stop biologics or JAKi altogether.
Delaying the start of biologics/JAKi would be achieved through “methotrexate optimization” and/or the use of triple therapy with methotrexate, sulfasalazine, and hydroxychloroquine. The other model would recommend tapering biologic/JAKi dosing in patients in remission or low disease activity and might even suggest a “medication holiday.”
The intention of this 3-year VBC/shared savings program is to reduce costs and create savings by reducing the use of biologics or JAKi. A tangential question might be, “Reduce costs and create savings for whom?” Apparently, the patients will not reap any of the cost savings, as this is proposed to be a shared savings program with the savings going to the physicians and the insurance company. Perhaps the idea is that patients will benefit by reducing unneeded expensive medications.
How will it work?
A cost baseline will be established on biologic and JAKi use prior to the start of the program. Once started, there will be a calculation of savings based on biologic/JAKi use going forward. It was stated that physicians would receive 22% of the total costs saved. In one flyer, it was estimated that, with methotrexate optimization, rheumatologists could be paid an average of $1,527 a month per patient per month of delay before starting a biologic or JAKi.
The American College of Rheumatology has guidelines for the treatment of RA and PsA, and while optimizing methotrexate and triple therapy is mentioned, tapering or stopping treatment with biologics or JAKi is not. Additionally, after lack of response at 3 months, the standard of care is to change to a more effective treatment, which for most patients is a biologic disease-modifying antirheumatic drug (DMARD). It could be construed that rheumatologists are being monetarily incentivized to reduce the use of expensive medications through ways that are not included in ACR guidelines and are not standard of care.
What if after the medication holiday the patient cannot recapture control of their disease? Is there a liability concern? Remember, there is no institutional review board or informed patient consent for this VBC data gathering model.
How will a patient feel knowing that their physician was paid to withhold care, or even worse, if a patient is not told of this and then finds out later? Not only are the optics for this suboptimal (at best), where does the liability fall if the patient does not do well and it comes out that their rheumatologist was paid to reduce the care, particularly in a way that is not supported in the guideline. Clearly, this appears to be a clinical study without an institutional review board and without patient consent.
There are also the data that are collected from this voluntary “opportunity.” A valid question would be, “What kind of data will this produce if rheumatologists are paid to delay, reduce, or stop the use of biologics/JAKi?” Is it possible that physicians may subconsciously delay putting patients on a biologic and taper more rapidly because of the reimbursement? This could lead to faulty, biased, AI-generated data that erroneously show this type of care is working. It would not be unheard of to wonder whether this once-voluntary opportunity might evolve into mandatory policy because now, they have “data to prove it.” … only this time there is no shared savings.
Low disease activity results in long-term savings
This is not meant to be an indictment of AI in health care, value-based care, or shared savings programs. In reality, AI had very little to do with how poorly this program was presented. Hopefully, it will bring about further discussions on how to achieve savings without sacrificing care. In fact, optimal care in RA and PsA is probably one of the best ways to save money in the long run. Nowhere in this program is there any mention of the high cost associated with uncontrolled disease activity in patients with RA or PsA. The downstream costs can be enormous when long- and short-term sequelae are taken into consideration: joint replacements, cardiovascular disease, certain kinds of malignancies, and all the side effects of increased steroid usage are just a few of the consequences we see with uncontrolled disease activity. It is only recently that we have been able to achieve low disease activity and remission in our patients. The rush to get patients off these medications is not the answer to achieving long-term savings. In addition to the very bad optics of paying rheumatologists to delay, taper, or stop using expensive mediations in their patients, the ultimate data achieved will be biased, and the only real winner will be the health insurance company.
Again, AI machine-learning and shared saving programs are not the guilty parties here. In fact, AI may be helpful in coming up with solutions to long-term health care costs, whether in the realm of economics or scientific research. CSRO and our state member organizations continue to educate the health insurance company on the significant drawbacks to this “volunteer opportunity.” Let’s hope a more reasonable program is put forward with AI-generated data that can be trusted. Hopefully not with a platform named “HAL,” for those of you old enough to remember “2001: A Space Odyssey.”
Dr. Feldman is a rheumatologist in private practice with The Rheumatology Group in New Orleans. She is the CSRO’s vice president of advocacy and government affairs and its immediate past president, as well as past chair of the Alliance for Safe Biologic Medicines and a past member of the American College of Rheumatology insurance subcommittee. You can reach her at rhnews@mdedge.com.
Editor’s note: As of this writing, the following proposed health insurance policy from Blue Cross and Blue Shield of North Carolina is still active. The Coalition of State Rheumatology Organizations and other rheumatology advocacy groups are in ongoing discussions with the health insurer and hope to have major changes to this policy implemented.
While AI has been in our world for years, it is expanding by the minute, perhaps by the nanosecond, within the health care sector. The $6.7 billion dollar health care AI market in 2020 is expected to climb to more than $120 billion by 2028. There are many questions regarding the application of AI in our world. Is it a mere instructional algorithm that computes things in a much faster way, or does it create a new story based on the information it has access to? Does it engender excitement or fear ... or both? Remember HAL? As we have seen throughout history with new inventions and technologies, there are risks and rewards. Even the best can have harmful unintended consequences. AI is no different, particularly when it comes to health care. In this case, AI can get a bad name if it is utilized along with biased data input and bad policy.
Shared savings
Here is where “shared savings” comes into play. A shared savings program starts with a baseline cost analysis of a particular care plan and then tracks costs (performance) going forward after certain changes to the original care plan are instituted. If savings are accrued when compared with baseline spending, those savings are shared with the providers of the care. Depending on how the shared savings program is implemented, the optics can be very bad if it appears as though physicians are being paid to reduce care.
‘The volunteer opportunity’
Recently, Blue Cross and Blue Shield of North Carolina, in partnership with Outcomes Matter Innovations, a data analysis company that uses AI/machine-learning technology, offered rheumatologists a new voluntary shared savings, value-based care (VBC) “opportunity.” Rheumatologists would be able to “utilize a web-based machine-learning technology platform that suggests evidence-based care pathways” in the treatment of rheumatoid arthritis and psoriatic arthritis (PsA). The VBC/shared savings model uses the AI platform to propose two different pathways. One model would delay the start of biologics or Janus kinase inhibitors (JAKi), and the second model would taper and/or stop biologics or JAKi altogether.
Delaying the start of biologics/JAKi would be achieved through “methotrexate optimization” and/or the use of triple therapy with methotrexate, sulfasalazine, and hydroxychloroquine. The other model would recommend tapering biologic/JAKi dosing in patients in remission or low disease activity and might even suggest a “medication holiday.”
The intention of this 3-year VBC/shared savings program is to reduce costs and create savings by reducing the use of biologics or JAKi. A tangential question might be, “Reduce costs and create savings for whom?” Apparently, the patients will not reap any of the cost savings, as this is proposed to be a shared savings program with the savings going to the physicians and the insurance company. Perhaps the idea is that patients will benefit by reducing unneeded expensive medications.
How will it work?
A cost baseline will be established on biologic and JAKi use prior to the start of the program. Once started, there will be a calculation of savings based on biologic/JAKi use going forward. It was stated that physicians would receive 22% of the total costs saved. In one flyer, it was estimated that, with methotrexate optimization, rheumatologists could be paid an average of $1,527 a month per patient per month of delay before starting a biologic or JAKi.
The American College of Rheumatology has guidelines for the treatment of RA and PsA, and while optimizing methotrexate and triple therapy is mentioned, tapering or stopping treatment with biologics or JAKi is not. Additionally, after lack of response at 3 months, the standard of care is to change to a more effective treatment, which for most patients is a biologic disease-modifying antirheumatic drug (DMARD). It could be construed that rheumatologists are being monetarily incentivized to reduce the use of expensive medications through ways that are not included in ACR guidelines and are not standard of care.
What if after the medication holiday the patient cannot recapture control of their disease? Is there a liability concern? Remember, there is no institutional review board or informed patient consent for this VBC data gathering model.
How will a patient feel knowing that their physician was paid to withhold care, or even worse, if a patient is not told of this and then finds out later? Not only are the optics for this suboptimal (at best), where does the liability fall if the patient does not do well and it comes out that their rheumatologist was paid to reduce the care, particularly in a way that is not supported in the guideline. Clearly, this appears to be a clinical study without an institutional review board and without patient consent.
There are also the data that are collected from this voluntary “opportunity.” A valid question would be, “What kind of data will this produce if rheumatologists are paid to delay, reduce, or stop the use of biologics/JAKi?” Is it possible that physicians may subconsciously delay putting patients on a biologic and taper more rapidly because of the reimbursement? This could lead to faulty, biased, AI-generated data that erroneously show this type of care is working. It would not be unheard of to wonder whether this once-voluntary opportunity might evolve into mandatory policy because now, they have “data to prove it.” … only this time there is no shared savings.
Low disease activity results in long-term savings
This is not meant to be an indictment of AI in health care, value-based care, or shared savings programs. In reality, AI had very little to do with how poorly this program was presented. Hopefully, it will bring about further discussions on how to achieve savings without sacrificing care. In fact, optimal care in RA and PsA is probably one of the best ways to save money in the long run. Nowhere in this program is there any mention of the high cost associated with uncontrolled disease activity in patients with RA or PsA. The downstream costs can be enormous when long- and short-term sequelae are taken into consideration: joint replacements, cardiovascular disease, certain kinds of malignancies, and all the side effects of increased steroid usage are just a few of the consequences we see with uncontrolled disease activity. It is only recently that we have been able to achieve low disease activity and remission in our patients. The rush to get patients off these medications is not the answer to achieving long-term savings. In addition to the very bad optics of paying rheumatologists to delay, taper, or stop using expensive mediations in their patients, the ultimate data achieved will be biased, and the only real winner will be the health insurance company.
Again, AI machine-learning and shared saving programs are not the guilty parties here. In fact, AI may be helpful in coming up with solutions to long-term health care costs, whether in the realm of economics or scientific research. CSRO and our state member organizations continue to educate the health insurance company on the significant drawbacks to this “volunteer opportunity.” Let’s hope a more reasonable program is put forward with AI-generated data that can be trusted. Hopefully not with a platform named “HAL,” for those of you old enough to remember “2001: A Space Odyssey.”
Dr. Feldman is a rheumatologist in private practice with The Rheumatology Group in New Orleans. She is the CSRO’s vice president of advocacy and government affairs and its immediate past president, as well as past chair of the Alliance for Safe Biologic Medicines and a past member of the American College of Rheumatology insurance subcommittee. You can reach her at rhnews@mdedge.com.
Editor’s note: As of this writing, the following proposed health insurance policy from Blue Cross and Blue Shield of North Carolina is still active. The Coalition of State Rheumatology Organizations and other rheumatology advocacy groups are in ongoing discussions with the health insurer and hope to have major changes to this policy implemented.
While AI has been in our world for years, it is expanding by the minute, perhaps by the nanosecond, within the health care sector. The $6.7 billion dollar health care AI market in 2020 is expected to climb to more than $120 billion by 2028. There are many questions regarding the application of AI in our world. Is it a mere instructional algorithm that computes things in a much faster way, or does it create a new story based on the information it has access to? Does it engender excitement or fear ... or both? Remember HAL? As we have seen throughout history with new inventions and technologies, there are risks and rewards. Even the best can have harmful unintended consequences. AI is no different, particularly when it comes to health care. In this case, AI can get a bad name if it is utilized along with biased data input and bad policy.
Shared savings
Here is where “shared savings” comes into play. A shared savings program starts with a baseline cost analysis of a particular care plan and then tracks costs (performance) going forward after certain changes to the original care plan are instituted. If savings are accrued when compared with baseline spending, those savings are shared with the providers of the care. Depending on how the shared savings program is implemented, the optics can be very bad if it appears as though physicians are being paid to reduce care.
‘The volunteer opportunity’
Recently, Blue Cross and Blue Shield of North Carolina, in partnership with Outcomes Matter Innovations, a data analysis company that uses AI/machine-learning technology, offered rheumatologists a new voluntary shared savings, value-based care (VBC) “opportunity.” Rheumatologists would be able to “utilize a web-based machine-learning technology platform that suggests evidence-based care pathways” in the treatment of rheumatoid arthritis and psoriatic arthritis (PsA). The VBC/shared savings model uses the AI platform to propose two different pathways. One model would delay the start of biologics or Janus kinase inhibitors (JAKi), and the second model would taper and/or stop biologics or JAKi altogether.
Delaying the start of biologics/JAKi would be achieved through “methotrexate optimization” and/or the use of triple therapy with methotrexate, sulfasalazine, and hydroxychloroquine. The other model would recommend tapering biologic/JAKi dosing in patients in remission or low disease activity and might even suggest a “medication holiday.”
The intention of this 3-year VBC/shared savings program is to reduce costs and create savings by reducing the use of biologics or JAKi. A tangential question might be, “Reduce costs and create savings for whom?” Apparently, the patients will not reap any of the cost savings, as this is proposed to be a shared savings program with the savings going to the physicians and the insurance company. Perhaps the idea is that patients will benefit by reducing unneeded expensive medications.
How will it work?
A cost baseline will be established on biologic and JAKi use prior to the start of the program. Once started, there will be a calculation of savings based on biologic/JAKi use going forward. It was stated that physicians would receive 22% of the total costs saved. In one flyer, it was estimated that, with methotrexate optimization, rheumatologists could be paid an average of $1,527 a month per patient per month of delay before starting a biologic or JAKi.
The American College of Rheumatology has guidelines for the treatment of RA and PsA, and while optimizing methotrexate and triple therapy is mentioned, tapering or stopping treatment with biologics or JAKi is not. Additionally, after lack of response at 3 months, the standard of care is to change to a more effective treatment, which for most patients is a biologic disease-modifying antirheumatic drug (DMARD). It could be construed that rheumatologists are being monetarily incentivized to reduce the use of expensive medications through ways that are not included in ACR guidelines and are not standard of care.
What if after the medication holiday the patient cannot recapture control of their disease? Is there a liability concern? Remember, there is no institutional review board or informed patient consent for this VBC data gathering model.
How will a patient feel knowing that their physician was paid to withhold care, or even worse, if a patient is not told of this and then finds out later? Not only are the optics for this suboptimal (at best), where does the liability fall if the patient does not do well and it comes out that their rheumatologist was paid to reduce the care, particularly in a way that is not supported in the guideline. Clearly, this appears to be a clinical study without an institutional review board and without patient consent.
There are also the data that are collected from this voluntary “opportunity.” A valid question would be, “What kind of data will this produce if rheumatologists are paid to delay, reduce, or stop the use of biologics/JAKi?” Is it possible that physicians may subconsciously delay putting patients on a biologic and taper more rapidly because of the reimbursement? This could lead to faulty, biased, AI-generated data that erroneously show this type of care is working. It would not be unheard of to wonder whether this once-voluntary opportunity might evolve into mandatory policy because now, they have “data to prove it.” … only this time there is no shared savings.
Low disease activity results in long-term savings
This is not meant to be an indictment of AI in health care, value-based care, or shared savings programs. In reality, AI had very little to do with how poorly this program was presented. Hopefully, it will bring about further discussions on how to achieve savings without sacrificing care. In fact, optimal care in RA and PsA is probably one of the best ways to save money in the long run. Nowhere in this program is there any mention of the high cost associated with uncontrolled disease activity in patients with RA or PsA. The downstream costs can be enormous when long- and short-term sequelae are taken into consideration: joint replacements, cardiovascular disease, certain kinds of malignancies, and all the side effects of increased steroid usage are just a few of the consequences we see with uncontrolled disease activity. It is only recently that we have been able to achieve low disease activity and remission in our patients. The rush to get patients off these medications is not the answer to achieving long-term savings. In addition to the very bad optics of paying rheumatologists to delay, taper, or stop using expensive mediations in their patients, the ultimate data achieved will be biased, and the only real winner will be the health insurance company.
Again, AI machine-learning and shared saving programs are not the guilty parties here. In fact, AI may be helpful in coming up with solutions to long-term health care costs, whether in the realm of economics or scientific research. CSRO and our state member organizations continue to educate the health insurance company on the significant drawbacks to this “volunteer opportunity.” Let’s hope a more reasonable program is put forward with AI-generated data that can be trusted. Hopefully not with a platform named “HAL,” for those of you old enough to remember “2001: A Space Odyssey.”
Dr. Feldman is a rheumatologist in private practice with The Rheumatology Group in New Orleans. She is the CSRO’s vice president of advocacy and government affairs and its immediate past president, as well as past chair of the Alliance for Safe Biologic Medicines and a past member of the American College of Rheumatology insurance subcommittee. You can reach her at rhnews@mdedge.com.