王飞跃的个人博客分享 http://blog.sciencenet.cn/u/王飞跃

博文

智能系统与加速提升医健水平

已有 14014 次阅读 2014-6-27 23:30 |个人分类:科研记事|系统分类:观点评述| Healthcare

智能系统与加速提升医健水平


 

今年五月, 美国总统执行办公室和总统科技顾问委员会(PCAST)联合向美国总统提交了一份名为<<更好的医疗保健和更低的成本: 利用系统工程加速改善进程(Better Health Care and Lower Costs: Accelerating Improvement Through Systems Engineering)>>的报告, 声称虽然少数采用了系统工程方法的医疗保健组织已经取得了“戏剧性的正面结果”, 但无法在美国的医疗保健体系中广泛推广, 其“罪魁祸首”就是“按服务计费(fee-for-service)”的医保支付系统,认为它是通向更加有效医疗保健的主要障碍。报告要求总统采取有效措施,在美国建立一支用系统工程基本知识“装备”起来的医疗保健队伍,从各个层面落实系统工程方法,加速提升美国的医疗保健水平。

 

对于美国而言,这一报告的发布时刻十分关键。目前,美国的医疗保健成本已接近其整个经济的五分之一,但其中相当的部分是不必要甚至是浪费掉的,根本没有导致更好的健康或医保水平。结果就是尽管花费远高于其它发达国家,但相比之下美国人寿命更短、婴儿死亡率更高,更可悲的是生命最后的几年过得更差。特别是备受争议的奥巴马医改方案今年正式实施,按美国民主党国会领袖的说法,此方案将使具有医疗保健保险的人口比例从之前的85%提高到96%,十年内花费估计近2万亿美元。如此扩展之下,对医疗保健的需求将会激增,迫使美国政府必须在战略上采取措施,在提高医保效率的同时,确保其质量和可承担性。这就是为什么PCAST现在交此报告,要求利用系统工程,加速降低医疗保健成本并提高其水平的背景原因。

 

作为一门二战后才发展起来的多学科交叉方法,系统工程已在军事国防、航空航天、生产制造、教育服务等领域得到广泛的应用并取得了令人瞩目的效果。例如,由于有效的采用了系统工程的方法,美国商业飞行每年的死亡事故已从上世纪60年代的数百起降到目前几乎为零的水平,死亡风险率仅为四千五百万次飞行之一。报告指出,几个在医疗保健组织中应用系统工程系统方法的案例都取得了十分令人满意的结果,例如,小到利用灵捷制造和丰田生产方法重新设计整个流程和运营的丹佛卫生组织,大到从多个方面多种方式全面采用系统工程概念与手段的全美最大医保组织凯撒(Kaiser Permanente), 许多时候效果常常是80%以上的改善与提高。尽管如此,系统工程方法却至今没有能够在美国的医疗保健行业中得到广泛的应用。

 

报告认为,造成这一现象的原因有很多,但首当其冲的是缺乏合适的医疗保健质量与绩效考核手段以及“按服务计费”支付系统的激励结构错位,鼓励了面向批量(Volume)而主要考虑费用的分片孤立式医保服务,有些地方甚至出现应用系统工程效果越佳,经济效益就越差的现象。为了实施系统工程方法,必须尽快转向面向价值(Value)而主要考虑质量的集成合作式医保服务。其它阻碍系统工程方法的因素及实施挑战还包括医疗保健组织的文化与领导力、技术力量、员工能力以及相关的知识水平等等。为此,PCAST提出六项架构性目标:1)加速促进支付系统与希望的医健结果的一致化整合; 2)扩大接触相关健康数据与解析途径;3)提供系统工程方法的技术支持;4)让社团参与改善医保服务;5)分享从成功改善案例中学到的经验;6)培训卫生专业人员掌握新的系统工程技巧和方法。为了实现这些目标,PCAST还在报告中进一步提出了七项建议和十四个具体措施。

 

当然,并不是所有专业人士都认可此份报告。很多专家认为美国的医疗保健需要更加深刻彻底的变革,现有架构体系不可救药,问题之复杂远非采用系统工程方法就能够解决。有人甚至借用爱因斯坦的名言来“评论”这一报告:“所谓疯子就是反复不断地做同样的事,但期望不同的结果!”

"Insanity: doing the same thing over and over again and expecting different results."

 

实际上,就是单从技术的角度来看,这一报告的力度、深度和广度都明显不足。尤其是在结合智能系统的智能化系统工程展望方面,如智慧和连通化医疗保健系统、利用虚拟世界和软件定义的系统使医保随时随地化、基于知识自动化的医保决策和服务优化推荐等与系统工程直接相关的智能技术方面,都明显地落后于现有理念与当前美国自己的研究和应用现状。 包括中国在内,目前世界上许多国家的无数个小型甚至个人企业正雨后春笋般地以各种独特的创新方式将人工智能技术引入疾病的治疗、康复以及健康的引导、维护等等。从利用智能算法、网络和众包技术为肾移植所开发的肾源匹配优化系统到一般的自动化全球器官共享联合网络,到各种基于iPad的癌症发现、医疗决策、个性配药机器人到一般的云计算机器诊断系统,国内外许多专家一致认为,如果能够通过系统工程的思维和手段把这些以移动装置、机器学习、语言处理、个性优化、渗入环境、虚实互动、可视引导、社会计算和社会媒体等为特色的智能化新健康技术和应用全面地整合到全社会的医疗保健实践之中,必将引发现有医保体系的革命性变化与改善:服务水平和质量的大大提高,服务时间和成本的大大降低。

 

他山之石,可以攻玉。包括美国在内的其它国家的医保实践和设想可以作为我国医疗保健改革的借鉴。其实,近年来国内各地医院实施的“简化就医流程”中的各种各样方便患者的措施, 从多室合一的“一站式服务”到微信监控咨询医护,效果显著,患者获得服务所花费的时间往往从数小时减到数分钟,都是成功利用系统工程和信息技术的例子。我们不妨从更高更广更系统化的角度来考虑整合利用智能技术与系统工程方法,加速提升我国的医疗保健水平与质量,实现医保事业的跨越式发展。特别是利用低成本的虚拟网络社会技术,充分发挥众包方式,鼓励发展网下实际网上虚拟的医护人员、药物供应商、患者和各类服务提供者的智能化平行交互社区,引入开发在线软件构成的医护机器人、患者助理、社区服务员以及软件定义的各类医院、诊所、实验室等,迅速普及医保知识自动化、健康决策自动化、卫生服务自动化等等,尽可能在不增加甚至减少经济负担的前提之下,全面地提高全民的健康水平与生活质量。



http://news.sciencenet.cn/dz/dznews_photo.aspx?id=20673

http://www.qstheory.cn/science/2014-06/27/c_1111344254.htm

http://www.kxfz.cn/a/yiyuan/20140628/134716.html

http://www.wokeji.com/pl/jjms/201406/t20140627_757921.shtml


Some selected comments from my former colleagues:


1.

Thanks for sending this document.

One observation that I might make about this document is lack of realistic tests and environmental assumptions.  Without realistic assumptions of the current system under study and realistic tests of those assumptions, the systems engineering process will likely fail to produce desired results.

One of the hallmarks of systems engineering process is realistic assessments of assumptions, environment and tests of both.  The assumption is that a system engineering approach will help within this domain.  There is an implicit assumption that the system under study is not corrupt. or in other words the services performed were actual services required and performed.  Audits are an accepted form of tests within the business world.

Despite the many deficiencies within the current healthcare system, at least Baucus managed to insert and provide incentives for audit recovery.

http://www.jdsupra.com/legalnews/cms-publishes-annual-update-on-recovery-48507/

http://www.americanbar.org/newsletter/publications/aba_health_esource_home/aba_health_law_esource_1108_pendleton.html

Coupling the outlined systems engineering process with sufficient audit testing may help the methodology espoused within the document.

A quick look w/in the newspapers suggests that the US can not afford expanded healthcare for it's citizens.  To test this assumption please review within the Federal Reserve Audit

http://www.sanders.senate.gov/newsroom/press-releases/the-fed-audit

http://www.sanders.senate.gov/imo/media/doc/GAO%20Fed%20Investigation.pdf

see Table 8 please (around page 144, to review how private interests moved private dept into public debt).  If we cannot afford healthcare then why did we just hand out $16 trillion to private investment bankers.  I have been told by a former Wall Street VP of derivatives (a Bell Labs colleague of mine) that the mortgage based derivatives leveraged / bet $100-to-$1 against default.  The # of bad mortgages was $200 billion x 100 = $20 trillion.  I guess we were lucky or since there has not been a recent federal reserve audit there is another $4 trillion dollars that have not yet been accounted for. This may indeed explain why the interest rates have been held low for quite some time. If the investment / too-big-to-fail banks are at the Fed window w/ nearly 0% interest loans they can continue to sustain holding bad debt on their books w/o incurring derivative penalty / defaults.

Additionally, there are a wide variety of analytic methods that can be applied but many current ones don't scale to the size of a nation.  I know of a couple that would work but perhaps that is another note in the future after the "associates and colleagues of Terry" argue about the statements made within the verbiage above.


2.

A number of issues were raised yesterday that can lead to progress.

One was related to repetition of requests for consultation WITHOUT clear statement of expectation of a better answer which would impact QUALITY OF LIFE. Had there been threat of non-payment for such requests or for the reply, "healthcare" costs would have been decreased over 50%.

It was clear that BIAS is a large factor in the attempt to apply systems engineering methods to human systems - especially in the context of leaving loopholes that encourage personal gain for individuals and groups. This is common in the way laws get written. Congress is notorious for writing laws that encourage "gaming the system." Is this not the process where we see headlines of

K Street Firms  Scramble To Save Millions In Business ?

Even the review which is now under way in the VA health care system is already showing that bonus money was paid for administrators gaming the appointment system.

The flaws now found in the way our government has joined with the insurance industry to set up payment systems has been riddled with a methodology that sought political rather than public health goals.

Just as our nation has wasted lives and trillions of dollars in fighting wars of convenience based on ideology, so also many lives still continue to be lost in the government-industry model of "health care" that encourages manipulative micro-management based on skills of gaming.

Oversight of a system created by government in the VA, Medicare, Medicaid, and based on free-enterprise profit merely watches over a process riddled with flaws and opportunities to place personal gain over the good of a nation.

One area where all those in surgery, medicine, public health, engineering, law, and government

can likely agree is, " Insanity: doing the same thing over and over again and expecting different results."

Albert Einstein


This is a very important document.

The participants are to be respected.

It deserves discussion in DEPTH.


3.

Would love to participate in a serious discussion of this document.

I read the executive summary, and recalled a story that Terry once used in a beginning class, and which I have used in meetings and conferences for years now.  Here’s how it goes:

If you went to a civil engineer and said I want to build a bridge: she would study the site, look at the river and the soil conditions and this and that, and say, here's a recommendation for a specific type of bridge that would cost a bunch of dollars and here’s my bill.


If you went to a systems engineer and said, I want to build a bridge: he would talk with you a bit and maybe say, you know, your real problem is not that you want to build a bridge, your real problem is you want to cross the river.  How about renting a boat.  (in my version of the story, the systems engineer leans back in his chair, puts his feet up on his desk, his hands behind his head, thinks a while…)

I agree totally with Jack - I think this document is saying a lot of very important things.  But I keep seeing the bridge in there.






https://blog.sciencenet.cn/blog-2374-807204.html

上一篇:关于大数据的问答
下一篇:关于PI的回忆: 有理与无理和有限与无限
收藏 IP: 159.226.177.*| 热度|

13 李伟钢 李本先 周健 彭思龙 罗汉江 王涛 武夷山 耿文叶 陈德旺 shenlu yunmu xqhuang Vetaren11

该博文允许注册用户评论 请点击登录 评论 (5 个评论)

数据加载中...
扫一扫,分享此博文

Archiver|手机版|科学网 ( 京ICP备07017567号-12 )

GMT+8, 2024-12-26 02:46

Powered by ScienceNet.cn

Copyright © 2007- 中国科学报社

返回顶部