SBIR Phase II: Reducing Claims Denials in Healthcare Through Blockchain and Machine Learning

这个小型企业创新研究（SBIR）项目的重点是改善医疗保健报销和收入周期管理。两个关键问题是，医疗保健索赔中的错误会导致频繁的付款拒绝，而未能事先获得授权可能会使这些服务无法报销。这些问题可能导致不适当的直接向患者计费和/或医疗保健提供者注销和转移成本以弥补损失。此外，支付和事先授权规则的复杂性要求在索赔提交、核对和返工方面产生相当大的管理开销，从而增加了管理成本。拟议的项目解决了医疗保健报销系统中的这些低效率问题。这个SBIR二期项目建议使用先进的分析、机器学习和区块链技术来解决以下研究目标：(1)分析和预测医疗保健索赔风险的拒绝付款；(2)预测在进行临床干预之前需要事先授权的可能性；(3)激励索赔提交过程的准确性，减少相关的管理负担和成本。该研究将进行高级索赔解析、数据提取、建模和机器学习，以定义特定的风险模式，并构建可重复、高效和准确的预测算法。这些方法将应用于索赔拒绝和预测事先授权的临床数据。区块链技术将用于激励人口统计和临床数据收集和索赔处理工作流程，以提高数据准确性、收集效率和预测质量。技术成果将是一个准确的、可预测的、持续学习的、高效的机器学习工具集，与生产力引擎集成在一起。该奖项反映了美国国家科学基金会的法定使命，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。

This Small Business Innovation Research (SBIR) project is focused on improving healthcare reimbursement and revenue cycle management. Two key concerns are that errors in healthcare claims result in frequent payment denials, and failure to obtain prior authorization can make those services non-reimbursable. These issues can lead to inappropriate direct billing to patients and/or a write-off by the healthcare provider and cost-shifting to cover losses. In addition, complexity within the rules for payment and prior authorization require considerable administrative overhead for claim submission, reconciliation, and rework, inflating administrative costs. The proposed project addresses these inefficiencies in the health care reimbursement system. This SBIR Phase II project proposes to use advanced analytics, machine learning, and blockchain technologies to address the following research objectives: (1) analyze and predict healthcare claim risk for denial of payment; (2) predict the likelihood of a prior authorization requirement before a clinical intervention is undertaken; and (3) incentivize accuracy in the claim submission process and decrease associated administrative burden and cost. The research will conduct advanced claims parsing, data extraction, modeling, and machine learning to define specific patterns of risk, and build reproducible, efficient, and accurate predictive algorithms. These approaches will be applied to both claim denials and to clinical data predictive of prior authorization. Blockchain technology will be utilized to incentivize demographic and clinical data collection and claims processing workflows for improvements in data accuracy, efficiency of collection, and predictive quality. The technical result will be an accurate, predictive, continually learning, highly efficient machine learning toolset integrated with a productivity engine.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.