《TechTarget：2024人工智能(AI)對醫療保健行業的影響研究報告（英文版）（20頁）.pdf》由會員分享，可在線閱讀，更多相關《TechTarget：2024人工智能(AI)對醫療保健行業的影響研究報告（英文版）（20頁）.pdf（20頁珍藏版）》請在三個皮匠報告上搜索。

1、E-BookHow AI is Impacting Healthcare Page 1 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentE-BookIn th

2、is ebook:President Bidens executive order focuses on the importance of secure and reliable Artificial Intelligence(AI)in various sectors,including healthcare.One specific area that will be affected is healthcare cybersecurity.Furthermore,a recent partnership has highlighted the advantages of AI in m

3、edical coding and auditing,demonstrating its potential in improving the revenue cycle.Additionally,advanced language models have proven their effectiveness by correctly identifying 93.8 percent of patients with adverse Social Determinants of Health(SDOH),suggesting the potential for additional suppo

4、rt in those cases.Page 2 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentHow the Executive Order on AI

5、Will Impact Healthcare Cybersecurity Jill McKeon,Associate Editor President Bidens executive order on safe,secure,and trustworthy AI emphasizes the need to establish rigorous security standards,which will have an impact on healthcare cybersecurity.Artificial intelligence(AI)continues to become ingra

6、ined into our society,and the regulations and guidance that govern it are evolving to match.In October 2023,President Biden issued an Executive Order on the Safe,Secure,and Trustworthy Development and Use of Artificial Intelligence,building upon previous guidance such as the Blueprint for an AI Bill

7、 of Rights and the AI Risk Management Framework.Together,these works aim to guide the nation through the development and deployment of safe,secure,and transparent AI technologies across all sectors.This includes healthcare,an industry in which AI has become an integral part of care coordination and

8、data management.E-BookPage 3 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentBut the Biden administrati

9、on acknowledged that developing and governing reliable AI tools will not be an easy feat.“Artificial intelligence(AI)holds extraordinary potential for both promise and peril.Responsible AI use has the potential to help solve urgent challenges while making our world more prosperous,productive,innovat

10、ive,and secure,”the executive order stated.“At the same time,irresponsible use could exacerbate societal harms such as fraud,discrimination,bias,and disinformation;displace and disempower workers;stifle competition;and pose risks to national security.Harnessing AI for good and realizing its myriad b

11、enefits requires mitigating its substantial risks.This endeavor demands a society-wide effort that includes government,the private sector,academia,and civil society.”With these risks in mind,two major focus areas of the executive order were security and privacy subjects that healthcare practitioners

12、 were already deeply familiar with.For example,the executive order directed various federal agencies to work on establishing standards and best practices for AI security.The National Institute of Standards and Technology(NIST)will be required to set standards for red-team testing to ensure the safet

13、y of AI tools prior to public release.The“AI red-teaming”will involve a structured testing effort to identify E-BookPage 4 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Deter

14、minantsof Health in EHR DataGetting more Xtelligentexclusive contentflaws,vulnerabilities,and unforeseen behaviors in an AI system.The Biden Administration also called on the Cybersecurity and Infrastructure Security Agency(CISA)to conduct AI red-teaming for generative AI tools.Whats more,the execut

15、ive order instructed the Department of Homeland Security to apply NISTs standards to critical infrastructure sectors and establish the AI Safety and Security Board.The Departments of Energy and Homeland Security were also tasked with addressing AI systems threats to critical infrastructure and cyber

16、security risks.“Together,these are the most significant actions ever taken by any government to advance the field of AI safety,”the executive order noted.To Beth Mosier,healthcare and life sciences director at West Monroe,the red-teaming directives are an area that AI developers and users should kee

17、p an eye on,especially in a highly targeted sector like healthcare.“One of the first things the executive order addresses is this concept of the red team test.And so,its escalating it from what do you as individuals need to be thinking of,to what do we as a nation need to be thinking of to make sure

18、 that when it comes to the highest levels,were identifying those things that pose risks to national security,public health,and safety,”Mosier noted.“Its through that broader lens also that theyre going to say,Heres what the best looks like,can you live up to it?And if not,whats your plan to get ther

19、e?E-BookPage 5 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentHealthcare security and privacy experts

20、have already raised concerns about the danger of AI-assisted cyberattacks and the potential for HIPAA violations related to AI chatbots.For example,in July 2023,HHS issued a threat brief about the ways in which threat actors might use AI to exploit vulnerabilities,overwhelm human defenses,and automa

21、te attack processes.HHS directed defenders to the NIST AI Risk Management Framework as a tool to mitigate these threats.The red-teaming activities required by the executive order will also help to reduce risk across AI tools,making them more reliable for end-users across healthcare and other sectors

22、.On the privacy front,the Biden Administration noted plans to enforce existing consumer protection laws and implement safeguards against fraud,unintended bias,discrimination,and infringements on privacy,all of which have been lasting concerns surrounding AI use in healthcare.“Such protections are es

23、pecially important in critical fields like healthcare,financial services,education,housing,law,and transportation,where mistakes by or misuse of AI could harm patients,cost consumers or small businesses,or jeopardize safety or rights,”the executive order continued.The executive order also emphasized

24、 that the federal government would work to ensure that the collection and retention of data is lawful,taking steps to make E-BookPage 6 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag

25、 Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentre-identification harder and mitigating privacy and confidentiality risks in the process.“People expect privacy when seeking healthcare,and they should not have to give up their privacy in return for receiving care,”Mos

26、ier added.“Those are fundamental rights that we as Americans expect,and so it is up to our government to provide oversight and make sure that happens.”The executive order and other federal guidance shed light on the federal governments vision for the future of AI governance,which signals an increase

27、d focus on decreasing bias and maintaining privacy and security.For AI developers,that means preparing to document the safety and security of their products.Meanwhile,healthcare organizations can expect to see improvements and refinements in AI technologies in the near future.To Mosier,the actions o

28、utlined in the executive order also align with how the White House has handled governance emerging technologies in the past,signifying a unification of siloed agencies under one common goal.“This play is nothing new.In terms of bringing medical devices and drugs to market,we rely on the FDA.When we

29、think about national standards in technology,thats NIST.We have standardized bodies who provide oversight and guidance in many areas of our lives,and I think this is no different,its just an emerging area,a new need,”Mosier said.E-BookPage 7 of 19 In this e-book How the Executive Orderon AI Will Imp

30、actHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive content“But I think what hopefully theyre trying to do is create more of a fabric,if you will,to connect all of those di

31、sparate parties so that theres some sense of,okay,we all have a role in this.NIST has a role,FDA still has a role,lots of other bodies have roles,but the bodies generated out of this,and the guidance that are coming out of this executive order will be the ones that provide the fabric to tie everyone

32、 together.”Next articleE-BookPage 8 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentHow AI is Becoming

33、a Staple in Medical Coding,Auditing Jacqueline LaPointe,Executive Editor A new partnership highlights the benefits of AI in medical coding and auditing and its inevitable use in this area of the revenue cycle.Troves of data flow through the healthcare revenue cycle.Yet,many providers struggle to mak

34、e sense of the codes and clinical documentation to not only submit clean claims but also understand patient encounters that occur within their organizations.Medical coding and clinical documentation are ripe for innovation,as coders and auditors spend countless hours parsing medical records.Computer

35、-assisted coding solutions are among the popular technology implementations to help optimize coding and billing,but artificial intelligence(AI)could bring technology to the next level.“Machine learning and AI have really accelerated over the last four years,and what were seeing in the coding space i

36、s the application of these algorithms to large amounts of clinical data so that we can start to identify conditions and diseases ahead of coders jumping into the chart for the first time,”says Nicola Sahar,MD,president of Semantic Health,which was recently acquired by AAPC,the nations largest medica

37、l coding training and certification association.E-BookPage 9 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive c

38、ontentAAPCs acquisition expands its solutions portfolio,which aims to elevate the quality and efficiency of healthcare through streamlined workflows.But the acquisition also signals a shift in how medical coders and auditors do their jobs and do them quickly and accurately,according to Rae Jimenez,c

39、hief product officer at AAPC.Sahar and Jimenez spoke with RevCycleIntelligence to break down technologys role in medical coding and billing,what it means for professionals,and the future of automation in this space.Tech reduces manual work“Weve got a lot of clinical data in healthcare,but its diffic

40、ult to make sense of it,”Sahar says.“Its difficult to make sure that the data is accurately reflecting whats happening with the patient.”Still,breaking down clinical documentation into codes the language of medical professionals is critical.One misstep and there can be serious consequences,whether t

41、hats missing a diagnosis,submitting false information on a claim,or contributing to data quality issues for analytics.These mistakes can be a headache for revenue cycle teams looking to submit clean claims and reduce denials,but also for patients who expect a seamless experience from the clinicians

42、exam room to the back office.E-BookPage 10 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive content“Its a criti

43、cal process that sees all this clinical data flowing through it,but that relies on experts,so medical coders and CDI specialists,to review this data as fast as they can and as accurately as they can,”Sahar explains.“Technology can improve these manual reviews,both in terms of productivity and in ter

44、ms of identifying error opportunities or audit opportunities.”Providers are seeing the benefits of applying technology to medical coding even though the process requires expertise.“Theres a healthy interest in looking for efficiencies that technology brings,”Jimenez states.“The goal of coding is to

45、make it more enabled by technology and less manual.”“For example,looking up a medical record and then populating codes into a different system,”Jimenez continues.“Those types of operational efficiencies where if you have a more streamlined workflow tool that limits the clicks and the redundancy of p

46、opulating information from one system to another,you can consider that manual.”That is where technology can really impact medical coding and billing:workflow.Workflow software that leverages machine learning and AI can elevate coding quality by reducing manual tasks,making it easier for professional

47、s to verify accurate coding and tackle more complex cases.E-BookPage 11 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligent

48、exclusive contentWorkflow software should consider all the tools a coder or auditor needs to review clinical charts,coding data,and AI suggestions.This is likely technologys happy home in medical coding versus complete process automation.“I dont think were at automation with coding and auditing with

49、in the context of AI,and I dont think it will ever be automated fully,”Sahar adds.“The reason thats the case is a lot of these processes are more half science-half art,so you really need experts in the loop with AI systems to be able to review complicated cases.”Synergy between tech and coders AI ca

50、n eliminate the repetitive tasks that are easier to accomplish,bringing more efficiency to coders and auditors,according to Jimenez.AI-enabled technology can summarize large data sets stemming from hundreds of pages of medical records,which a human would have to parse to code an encounter accurately

51、.“The tool can help summarize and give you a snapshot of what happened through the entirety of that admission by quickly capturing the diagnosis and identifying where in the documentation it came from,”Jimenez elaborates.“Staff can then quickly go to those areas without reading hundreds of documents

52、.That is a game changer with time and efficiency.That is going to enable your humans to make better decisions quicker.”E-BookPage 12 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag So

53、cial Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentThat is the key to implementing AI in medical coding and billing;technology must complement the professionals workload and workflow.Coders and technology need to develop a synergy.“And that process then teaches the AI what

54、 to do and what not to do,”Sahar explains,emphasizing how AI and professionals need to work together in order to maximize benefits and ROI.“So over time,this creates a positive feedback loop that frees up the coder to be able to take on more complex work.”AI cannot work in a black box in order to su

55、ccessfully aid medical billing and coding,Jimenez stresses.Coders and auditors need to be able to verify that output from the AI solution reflects the patient encounter fully and accurately.This approach to technology implementation technology supports human work versus eliminating it aligns with he

56、althcare,which relies on people.Technology needs to make care delivery and administration more efficient,whereas other industries can automate more of their tasks completely to achieve efficiency.The future of AI in medical coding,auditing Theres a significant desire and need for tech-enabled coding

57、 processes,which underlined AAPCs recent acquisition of Semantic Health.E-BookPage 13 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting m

58、ore Xtelligentexclusive content“By putting this tool in the hands of our members,its going to make them more efficient,”Jimenez says.“Its also going to help them understand how their roles will evolve,which is also very important to us.”Healthcare organizations simply cannot keep up with the sheer v

59、olume of claims lately as more patients touch the healthcare system after a years-long pandemic and recent coverage changes that reduced the uninsured rate in the US.An aging population also means that providers are going to see more claims since older populations tend to use more healthcare.“Thats

60、hard to keep up with,”Jimenez explains.“So,if there are tools that can help staff gain efficiency and take the more simplistic encounters to code and automate those,people can focus on the harder cases,which,in most cases,generate the most revenue.AI helps them do more work with the staff they alrea

61、dy have versus having to hire more individuals.”Not only are staff expensive to hire,but also hard to find.A 2023 Medical Group Management Association(MGMA)poll found that 34 percent of medical groups find medical coders the most difficult revenue cycle role to hire.Coders also require specialized e

62、ducation and training compared to other revenue cycle roles another aspect AI-enabled coding and auditing technology can help.“Training people on how to use this technology to be more efficient means youll be able to get more production out of each coder,”Jimenez states.“It might also help with thei

63、r proficiency because theyve got a tool.Its not them making the E-BookPage 14 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtel

64、ligentexclusive contentdecisions on their own as one person.Theyve got a tool to give them confidence that theyre making the right code selections.”Technology is also a tool providers can use to reduce burnout.However,organizations need to ensure they are digitally mature meaning they arent on pen a

65、nd paper and have change management in place to leverage AI-enabled coding and auditing technology,Sahar underscores.“People are going to be using AI,so they need to be comfortable with it,they need to trust,they need to know what it can and cannot do,”Sahar explains.“On the change management side,A

66、APC can provide the education,thought leadership,and training to the next generation of codes and continuing coders and auditors.Next ArticleE-BookPage 15 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerat

67、ive AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentGenerative AI May Help Flag Social Determinants of Health in EHR Data Fine-tuned large language models were able to accurately identify 93.8 percent of patients with adverse SDOH who could benefit fro

68、m additional support.Researchers from Mass General Brigham have found that specialized large language models(LLMs)can identify under-documented social determinants of health(SDOH)in electronic health records(EHRs),according to a study published this week in npj Digital Medicine.SDOH have a significa

69、nt impact on patient outcomes,but the research team underscored that SDOH documentation in clinical notes is often incomplete or missing.As a result,extracting this information to provide additional support to patients can be a challenge,leading the researchers to investigate generative AIs potentia

70、l in this area.“Our goal is to identify patients who could benefit from resource and social work support,and draw attention to the under-documented impact of social factors in health outcomes,”said corresponding author Danielle Bitterman,MD,a faculty member in the Artificial Intelligence in Medicine

71、(AIM)Program at Mass General Brigham and a physician in the Department of Radiation Oncology at Brigham and Womens Hospital,in a news release detailing the study.E-BookPage 16 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical C

72、oding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive content“Algorithms that can pass major medical exams have received a lot of attention,but this is not what doctors need in the clinic to help take better care of patients each day.Algori

73、thms that can notice things that doctors may miss in the ever-increasing volume of medical records will be more clinically relevant and therefore more powerful for improving health,”Bitterman continued.The research team noted that clinicians often summarize SDOH information in their visit notes,but

74、these data are rarely systematically organized in EHRs.To address this,the researchers sought to fine-tune LLMs for SDOH data extraction.They began by manually reviewing 800 clinical notes from 770 cancer patients who received radiotherapy at the Department of Radiation Oncology at Brigham and Women

75、s Hospital.During this process,the research team flagged sentences that referred to one or more of six SDOH categories:employment status,housing,transportation,parental status,relationships,and the presence or absence of social support.After annotating these data,the information was then used to tra

76、in existing LLMs to identify clinician references to SDOH.Each model was then tested using an additional cohort of 400 patients who received immunotherapy at Dana-Farber Cancer Institute and patients admitted to critical care units at Beth Israel Deaconess Medical Center.E-BookPage 17 of 19 In this

77、e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentThis testing revealed that fine-tuned LLMs can accurately and consist

78、ently identify SDOH references in EHRs.Official diagnostic codes included these data in approximately two percent of cases,whereas the generative AI models flagged 93.8 percent of patients with adverse SDOH.However,the“learning capacity”of the LLMs was limited by the relative rarity of SDOH document

79、ation in the training data,as only three percent of sentences in clinician notes referenced social determinants.The researchers addressed this by leveraging ChatGPT to generate 900 additional synthetic samples of SDOH references that could be utilized for extra model training.The specialized LLMs we

80、re also less prone to bias than other generalist models,such as GPT-4.The research team found that their fine-tuned LLMs were significantly less likely to alter their determinations based on a patients race,ethnicity,and gender.Despite this,the research team underscored that it is difficult to under

81、stand the origins of algorithmic bias,so more research is needed.“If we dont monitor algorithmic bias when we develop and implement large language models,we could make existing health disparities much worse than they currently are,”Bitterman said.“This study demonstrated that fine-tuning LMs may be

82、a strategy to reduce algorithmic bias,but more research is needed in this area.”E-BookPage 18 of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataG

83、etting more Xtelligentexclusive contentThis is just one potential use case for generative AI in healthcare that researchers are exploring.In June,a research team from Beth Israel Deaconess Medical Center(BIDMC)found that generative AI tools like ChatGPT have significant potential to assist clinician

84、s with complex diagnostic cases.To evaluate the accuracy of these tools,the researchers tested ChatGPT-4s performance on 70 complex diagnostic reasoning challenges.The model was then tasked with providing potential diagnoses based on each case.The final diagnosis was included in the tools differenti

85、al in a majority of cases,and the models top diagnoses matched the final diagnosis in just under 40 percent of cases.The study had multiple limitations,but the research team noted that the study plays an important role in expanding the literature exploring the promise of healthcare AI.E-BookPage 19

86、of 19 In this e-book How the Executive Orderon AI Will ImpactHealthcare CybersecurityHow AI is Becoming aStaple in Medical Coding,AuditingGenerative AI May HelpFlag Social Determinantsof Health in EHR DataGetting more Xtelligentexclusive contentGetting more Xtelligent exclusive content As an Xtellig

87、ent member,you have access to TechTargets entire portfolio of 80+websites.Xtelligent access directs you to previously unavailable“platinum members-only resources”that are guaranteed to save you the time and effort of having to track such premium content down on your own,ultimately helping you to sol

88、ve your toughest IT challenges more effectivelyand fasterthan ever before.Take full advantage of your membership by visiting https:/ Stock 2024 Xtelligent Media,TechTarget.No part of this publication may be transmitted or reproduced in any form or by any means without written permission from the publisher.E-Book