字号·· | 护眼
thenextweb

Perficient推出用于联邦医疗保险优势计划星级评分管理的人工智能工具Perficient launches AI tools for Medicare Advantage star ratings management

点「原文对照」整页切到原文,或双击某段只看那段的原文。

图片来源:© SUMALI IBNU CHAMID via Canva.comCredit: © SUMALI IBNU CHAMID via Canva.com

Perficient推出了一套人工智能套件,旨在帮助联邦医疗保险优势计划(Medicare Advantage)提高其CMS星级评分,该评分决定了每年超过130亿美元联邦奖金的获取资格。就在该套件发布的前一天,美国医疗保险和医疗补助服务中心(CMS)公布了2027年的评级结果,显示符合条件的计划数量有所减少。目前,这家咨询公司由瑞典收购集团EQT所有。

Perficient has launched an AI suite to help Medicare Advantage plans raise their CMS star ratings, which decide access to more than $13B a year in federal bonus payments. It arrived a day after CMS published 2027 ratings showing fewer plans qualifying, and the consultancy is owned by Swedish buyout group EQT.

Perficient推出了人工智能工具,以帮助联邦医疗保险优势计划提高其联邦星级评分。这些工具发布于CMS公布评级结果的一天后,该评级显示有资格获得奖金的计划数量有所下降。

Perficient has launched AI tools to help Medicare Advantage plans raise their federal star ratings. They arrived a day after CMS published ratings showing fewer plans qualifying for bonus money.

据该公司公告称,这些工具利用了预测分析、模拟建模以及所谓的“代理驱动参与”技术。目前尚未披露任何客户信息、价格或具体成果。

The consultancy says the tools use predictive analytics, simulation modelling and what it calls agent-driven engagement, according to its announcement. No customer, price or result was disclosed.

据凯撒家庭基金会(KFF)数据显示,2026年联邦在联邦医疗保险优势计划质量奖金项目上的支出至少达到了134亿美元,约占该项目预计支付总额的2.3%。

Federal spending on the Medicare Advantage quality bonus programme reached at least $13.4B in 2026, about 2.3% of projected payments to the programme, according to KFF.

评级在四星及以上的计划可获得更高的支出基准,通常高出五个百分点。评级在4.5星以上的计划可保留基准与投标之间差额的70%。

Plans rated four stars or better out of five get a higher spending benchmark, typically by five percentage points. Those above 4.5 stars keep 70% of the gap between benchmark and bid.

获得资格的难度正在增加。CMS于10月8日公布了2027年的评级结果,在508份受评合同中,有188份达到四星或以上,而去年同期为516份合同中有229份达标。

Qualifying is getting harder. CMS published the 2027 ratings on 8 October, and 188 of 508 rated contracts reached four stars or more, against 229 of 516 a year earlier.

2026年,处于奖金合格计划中的会员比例已从75%降至68%,为2018年以来的最低水平。五星级合同的数量从22份降至15份。

The share of members in bonus-eligible plans had already dropped to 68% in 2026 from 75%, the lowest since 2018. Five-star contracts fell from 22 to 15.

该评级衡量的内容一直存在争议。为国会提供医疗保险建议的医疗保险支付咨询委员会(MedPAC)认为,这些评级使用的衡量指标过多,未能考虑社会风险,且是以合同而非计划层面进行报告的。

What the ratings measure is contested. MedPAC, which advises Congress on Medicare, argues they use too many measures, fail to account for social risk and are reported at contract rather than plan level.

因此,该软件所针对的指标,正是政府自身顾问所称的衡量医疗质量的“薄弱代理指标”。国会预算办公室曾在2018年估计,取消该项目将在十年内节省近1000亿美元。

So the software targets a number the government’s own advisers call a weak proxy for care. The Congressional Budget Office estimated in 2018 that scrapping the programme would save almost $100B over a decade.

所有者是欧洲企业。瑞典集团EQT管理着约1300亿欧元资产,此前曾与Alphabet就其投资组合中的模型授权进行谈判,最终以约30亿美元收购了Perficient,并于2024年10月将其私有化。

The owner is European. EQT, the Swedish group that manages about €130B and was in talks with Alphabet over model licensing for its portfolio, bought Perficient for roughly $3B and delisted it in October 2024.

Perficient拥有约7000名员工,在罗马尼亚、塞尔维亚和英国运营着欧洲交付中心。

Perficient employs about 7,000 people and runs European delivery hubs in Romania, Serbia and the United Kingdom.

华盛顿方面正从另一端将人工智能应用于同一项目。美国卫生与公众服务部(HHS)已于5月开始在支付前对索赔进行欺诈筛查,涵盖Medicare、Medicaid以及保险市场。

Washington is pointing AI at the same programme from the other end. HHS began screening claims for fraud before payment in May, covering Medicare, Medicaid and the insurance marketplace.

评级问题也进入了法庭。Clover Health成功要求重新计算2026年的评分,KFF估计此举将使2027年的支付金额增加约6亿美元,而美国卫生与公众服务部下属的医疗服务中心(CMS)则于7月对该裁决提出了上诉。

The ratings are also in court. Clover Health won a recalculation of the 2026 scores, which KFF estimates adds about $600M to 2027 payments, and CMS appealed that decision in July.

Humana因一份大型合同评分从4.5星降至3.5星而提起诉讼,但法院目前均站在了CMS一边。无论是诉讼还是如今的人工智能软件,最终都指向了同样的半星差距。

Humana sued after one large contract dropped from 4.5 stars to 3.5, and courts have sided with CMS so far. Litigation and now software are routes to the same half-star.