Effective Healthcare Data Governance: How One Hospital System is Managing its Data Assets to Improve Outcomes

INTRODUCTION

As healthcare transitions into an analytics-driven industry, data has become its most valuable asset—and one of the most challenging to manage. To support care delivery and other functions, health systems must integrate data from a complex array of internal applications including inpatient and ambulatory EHRs, laboratory information systems, pharmacy systems, and ERP systems. In a parallel need, these organizations must also pull data from affiliated providers’ EHR systems, payers, external laboratories, pharmacies and benchmarking sources. Good data governance practices are clearly essential for hospitals to ensure that all this data is well understood, trusted, accessible, and secure—particularly to validate the accuracy of dashboards, scorecards, and other analytic outputs in a data-driven setting. Further, establishing data governance practices now positions organizations for value-based care, another significant shift in the healthcare landscape in which timely access to trusted data is increasingly important.¹

为了创造这样一个环境，堪萨斯大学医院已经设定了一个全面的目标，成熟成为一个先进的分析组织，有能力管理非常大的数据集，并执行预测分析，以追求最佳的可能结果。为了实现这些目标，堪萨斯大学医院需要一个“单一的真相来源”，用户可以依赖它从不同的来源和系统获取高质量的数据。虽然这部分可以通过技术实现，但他们还需要为数据治理建立一个坚实的框架。简单地说，这是由建立过程和执行计划的管理机构对数据进行管理，以确保数据的安全性、可用性、可用性和可靠性。²

Certain data governance best practices have been identified that touch on a range of issues, from data stewardship to quality and availability.³为了将这些和其他最佳实践落实到位，例如高级领导的积极承诺和参与，以及将数据治理与现实世界的临床或操作努力结合起来，以实现必要的参与水平，堪萨斯大学医院随后成立了一个数据治理委员会。

THE SHORTFALLS OF MANAGING DATA WITHOUT A DATA GOVERNANCE MODEL

对于医疗保健组织来说，堪萨斯大学医院面临着几个常见的数据挑战，首先是手动收集、获取和验证来自如此多来源的数据所需的时间。这就使得员工没有足够的时间去研究数据以获得改进的机会。随着内部和外部对信息需求的增长，这一过程变得更加昂贵和低效。让正确的决策者掌握正确的数据也变得更加困难。然而，医院需要一种更有效的替代方案，而不仅仅是增加更多的全职员工来管理不断增长的数据量。

数据质量问题也经常出现，说明需要广泛清理不同的数据源。除了数据不准确或完全丢失之外，最常见的问题是不标准的数据术语和定义。相互矛盾的数据是另一个挑战，比如董事会成员在两份不同的报告中发现了不同的停留时间值。这些差异和其他差异通常是由产生信息的时间，或者不同源系统中不同的数据类型定义造成的。

The end result was that users didn’t fully trust the data. To overcome this distrust, The University of Kansas Hospital wanted to implement standard terms and definitions, and put in place better data management processes and practices. They also wanted to make data more accessible to clinical and operational end-users. Finally, they wanted to create a single source of truth for information that could take into account the many different people involved in data gathering and analysis and their varying degrees of training and experience.

AUTOMATED ANALYTICS PAIRS WITH PEOPLE-IMPLEMENTED POLICIES

To address the distrust in data, The University of Kansas Hospital implemented a data governance committee, sanctioned by senior leadership and provisioned with executive authority to govern all business intelligence, data and information assets. To keep up with its growing data reporting needs, rather than continuously adding full-time employees, The University of Kansas Hospital opted instead to invest in a modern analytics infrastructure. The chief features of such a solution significantly aid in the transformation to a data-driven culture.

• Automation of the data collection process
• Timely, accurate data for decision making
• The right data accessible to the right users
• A switch in focus from data reporting to identifying and executing improvements

To automate data collecting and reporting, the Hospital deployed a late-binding enterprise data warehouse and analytics applications from Health Catalyst.

THE DATA GOVERNANCE COMMITTEE CHARTER

从一开始，该委员会就明确表示，其目的不是拥有数据，而是促进数据用于有效决策。To that end, the committee pursues four primary goals for data:

Information governance. This involves bringing clinical, administrative, and technology partners together to jointly design and optimize information assets in alignment with the organization’s strategies and goals. It also includes the declaration of sources of truth, systems of record, roles and responsibilities, information delivery standards and certified enterprise reports/dashboards. Data security and access decisions are within the scope of this governance.

Quality. Successful decision making is enabled by data that is complete, timely, accurate and consistent. The Committee assures data quality through standardization, process engineering and the creation and monitoring of data quality metrics.

Usability. Usability includes easy-to-use applications that promote data interaction and informed decision making, as well as creating a common data language for understanding organizational performance. Tools to create such usability include data dictionaries, training, a metadata repository, and access to trusted data sets as close to real-time as possible.

Availability. The Committee strives to implement the appropriate analytical infrastructure that makes data accessible to clinical and operational users when they need it, defined by their respective clinical and operational needs.

THE DATA GOVERNANCE COMMITTEE STRUCTURE

如图1所示，该委员会被划分为不同的组，具有特定的职责。数据治理执行组(Data Governance Executive Group)负责监督改进工作的愿景和战略，而数据咨询组(Data Advisory Group)则在战术层面发挥作用，处理数据质量问题、工作优先级和工作组的创建。工作组由工作组所研究领域的主题问题专家组成。其他关键委员会成员包括数据治理支持团队，由技术、流程改进和临床专家组成。

BEST PRACTICES FOR HEALTHCARE DATA GOVERNANCE

为了避免可能导致不成功的数据治理项目的错误，堪萨斯大学医院设计并实施了一些最佳实践，以构建一个坚实的数据治理结构，在未来几年为医院服务。

• Secure commitment from senior leaders. The implementation of an advanced analytical infrastructure and an effective improvement strategy is a significant change management effort. It is also a shared responsibility across the organization. Accordingly, the visible and active support of senior leaders is paramount. The Hospital’s own executive team made sure the program is staffed with the necessary technology, operational and clinical expertise.
• Make it about improvement, not an IT project. While IT expertise is required, IT is a steward of the data, not the owner. The overall purpose of such projects is about getting good data into the hands of end users so they can make the best decisions about care. Indeed, the responsibility of data ownership falls to clinical and operational decision-makers.
• 确定一个“燃烧的平台”。世界杯厄瓜多尔vs塞内加尔波胆预测When the project is launched, it should be tied to a real world clinical or operational endeavor to realize the necessary level of engagement. For The University of Kansas Hospital, the burning platform was the implementation of a new revenue cycle system for its clinics— which created a data gap and a disruption in reporting key organizational metrics.
• Aim for transparency instead of gate keeping. The data governance committee should not function as a gatekeeper. Its role is to transparently provide end-users with good quality data to use in their daily jobs.
• Choose the right leader. The most effective person to lead the initiative should have both the necessary technical skills and customer service savvy, in order to develop partnerships with clinical and administrative leaders.
• Provide adequate resources. Data governance cannot be a low priority or side job. Dedicated resources are critical to drive progress and maturity.

RESULTS

With a solution comprised of intelligent technology and prudent best practices, The University of Kansas Hospital achieved its primary goals while further transforming into an analytics-driven culture.

70+ standardized enterprise data definition approvals planned for completion in first year. The organization approved seven definitions in the first quarter to more than 66 in the second quarter— averaging four decisions per months, and growing. End-users are assuming ownership of the data and accepting responsibility to manage data quality. A number of data quality issues have been fixed and improved data quality. In turn, trust in the data has grown among end-users.

System-wide executive and clinical engagement in data governance. Due in no small part to an executive team that has been visibly and strongly supportive of the Data Governance Committee, and provided it with the necessary resources, two important events have happened. First, the committee has assumed overall responsibility for managing the availability, usability, integrity, and security of the organization’s data. Second, key clinical and operational leaders are becoming more engaged in the analytics and improvement efforts, with participation rates increasing to 60 percent in bi-monthly governance meetings.

Creation of a multi-year business intelligence/data governance roadmap. The University of Kansas Hospital is moving from a fragmented business intelligence environment to the more cohesive enterprise perspective ideal for managing large data sets and using predictive analytics to create the best outcomes. In support of this journey, they have developed a business intelligence roadmap (illustrated in Figure 2) against which they track their progress. The Data Governance Committee is focused on making progress on several fronts including:

• Building an advanced business intelligence (BI) architecture
• Extracting quality data from multiple source systems into a single source of truth (the EDW)
• Automating data extraction and reporting as much as possible
• Advanced data modeling based on a common vocabulary and definitions
• Providing end-users advanced analytical tools with access to good data in pursuit of good decision-making and outcome improvements.
• 持续改进数据治理，使最终用户能够有效地拥有和管理其信息需求。
• Education of end-users to provide them the advanced analytical and improvement skills required to optimize performance.
• 在堪萨斯大学医院文化中建立数据驱动的心态。

数据对于做出明智的决定至关重要。数据治理结构确保数据是准确的，我们都在讲同样的故事。”

– Chris Harper, Director Business Architecture & Analytics

WHAT’S NEXT

堪萨斯大学医院认为数据治理是成功的数据驱动文化的关键组成部分。因此，重点仍将放在构建最成功的数据治理方案上。例如，医院计划构建一个数据治理仪表板来跟踪和管理治理指标。考虑的未来指标包括确定的数据所有者的数量、记录的流程数量、节省的美元数量(或避免的费用)和企业数据仓库的使用(例如提高的效率、减少使用非企业数据解决方案，以及衡量应用程序的价值以确保客户正在使用已经开发的商业智能解决方案)。2022卡塔尔世界杯赛程表时间

还在持续关注数据质量和可访问性，以促进信任和使用。Elements include:

• Continuing to build a single source of truth for data and key metrics.
• 改进数据捕获(许多指标不是离散地捕获的，需要使用数据样本来估计实际数量)。
• Automating the many metrics that are still tracked manually.
• Learning how to improve the measurement of cost avoidance.

有了这样一个对数据完整性和可用性的承诺，堪萨斯大学医院不仅可以进行分析，以追求更好的结果，而且可以引领道路。

References

1. Barton, D. A. (2015, April). 2015 and beyond: 6 predictions for health and care and population health.HFM Magazine.
2. Stedman, Craig. Vaughan, Jack. (July, 2007).什么是数据治理，它为什么重要?TechTarget.Data governance (DG).
3. 桑德斯,戴尔。(无日期)。医疗保健数据的7个基本实践。Health Catalyst Insights.

ABOUT HEALTH CATALYST

Health Catalyst is a mission-driven data warehousing and analytics company that helps healthcare organizations of all sizes perform the clinical, financial, and operational reporting and analysis needed for population health and accountable care. Our proven enterprise data warehouse (EDW) and analytics platform helps improve quality, add efficiency and lower costs in support of more than 50 million patients for organizations ranging from the largest US health system to forward-thinking physician practices.

For more information, visithealthcatalyst.com, and follow us onTwitter,LinkedIn,andFacebook.