I-Corps: An Artificial Intelligence Agent for Healthcare Negotiation

这一I-Corps项目的更广泛影响是减少了与健康保险公司谈判报销和索赔范围时所涉及的交易成本。健康保险公司在大多数医疗保健交易中发挥作用，但保险提供商之间和各州之间的管理制度差异很大。这种复杂性导致患者的覆盖范围低于最佳水平，并为消费者和提供者带来了经济负担。从商业角度来看，这项技术可以增加供应商收到的报销，并减少行政管理费用。最终，这意味着提供商将产生更多的收入，消费者将体验到医疗保健相关财务负债的缓解。这个I-Corps项目进一步开发了一个综合自然语言理解，论证和机器学习功能的平台。自然语言的医疗保健政策文件将被翻译成符合人工智能（AI）自动推理的示意性逻辑表示。 该平台将由复合模式提供信息，其中一些模式是通过监督机器学习发现的，数据定义涉及患者/提供者索赔的成功和失败程度，如保险公司决策、保险公司对上诉的回应、患者/提供者特征以及定义政策的模式化合同;因此，与其他机器学习应用程序相比，机器学习环境中的数据将非常丰富。患者特征和合同是情境化信息，并为这种机器学习方法提供了理论约束，这是理论指导机器学习相对较新领域的一个新实例。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact of this I-Corps project is in reducing transaction costs involved when negotiating reimbursements and claims coverage with health insurers. Health insurers play a role in the majority of healthcare transactions, but administrative regimes vary widely between insurance providers and across state lines. This complexity results in less than optimal levels of coverage for patients and creates a financial burden for consumers and providers alike. From a commercial standpoint, this technology can increase the reimbursements received by providers and reduce administrative overhead costs. Ultimately, this means providers will generate increased revenues and consumers will experience relief from healthcare related financial liabilities.This I-Corps project further develops a platform that synthesizes natural language understanding, argumentation, and machine learning functionalities. Health care policy documents in natural language are to be translated into schematic, logical representations that are amenable to artificial intelligence (AI) automated inferencing. The platform will be informed by composite patterns, some found through supervised machine learning, with data definitions involving the degree of success and failure of patient/provider claims as manifest in insurer decisions, insurer responses to appeals, patient/provider characteristics, and schematized contracts that define policy; thus a datum in the machine learning context will be extraordinarily rich compared to other machine learning applications. Patient characteristics and contracts are contextualizing information and provide theoretical constraints with this machine learning methodology a novel instance of the relatively new field of theory-guided machine learning. Collectively, the platform will synthesize a number of AI functionalities.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.