Data governance guiding principles for health systems

Some form of data governance is necessary for all organizations, regardless of industry or size. Data governance accomplishes many goals, from meeting regulatory requirements to fueling strategic initiatives that depend on data. While most organizations recognize the importance of data governance, they have trouble keeping momentum after starting. One issue is that data governance models are often boilerplate adoptions, but in reality only their core concepts are consistent, and processes should be right-sized and adapted to fit each organization’s specific needs. For example, traditional data governance includes defining roles for people and functions, such as data producers, stewards, custodians, etc. The result is a very centralized and layered model that may be too heavy for many organizations who then face adoption and maintainability challenges:

• Many organizations do not have ample resources for each data domain (e.g. customer, product, vendor), increasing the risks of tribal knowledge and gaps when individuals leave

• Many layers and delineated business domains hinder collaboration and communication of top-down initiatives

• Data questions and requests may take many rounds of escalations and referrals to reach the right person who can help, and data issues may take a long time to bubble up

• Large modifications to organizational structures and individual responsibilities are often met with cultural resistance and spur even more unmonitored workarounds to formal processes

In order to right-size a data governance model for your organization, data governance should be built iteratively while taking advantage of the processes that happen to work well today. Start by uncovering each business group’s pain points and then understand how they execute and manage their data projects in the current state. This approach will identify the areas of the business requiring standardized processes and organizational oversight, as well as the areas that business owners can manage themselves. Establishing a vision and guiding principles will help you answer questions when design choices have polar pros and cons. No two organizations will have the same governance needs, but core concepts typically stay the same:

• Governance should prioritize solving the actual pain points felt by business groups today around executing data projects and trusting data

• A long-term vision can provide the proper strategic direction, but it must be flexible as data needs and process will change over time with new system implementations and evolving priorities

• Building iteratively means only instituting governance in areas that need it and understanding that your people will have to learn over time, so a centralized PMO and change management are critical for successful adoption

• Find the right balance between a large, centralized model and one that is comprised of several decentralized, disparate parts – the model must be able to align business groups and influence them to take action

• Governance should seek to maximize business value for the whole organization and treat data as an asset, but fruition of these philosophies requires support from executive stakeholders

Implementing a data governance program can be a daunting undertaking, but starting small, iteratively tackling real pain points, and finding an organizational structure that fits your culture are principles that will get the quick success that you can build upon. Read this case study to learn how we helped a healthcare provider do just that.

West Monroe designed an innovative data governance model for a large healthcare provider that balanced change with practical cultural adoption

West Monroe was challenged to design a data governance model for a health system struggling with widespread data issues for many years. During our discovery sessions, we investigated these data-related aspects of the organization:

• Processes and allocated resources that handle data requests today

• Landscape of data ownership and expertise

• How data is validated and how exceptions are handled and remediated

• Transparency and communication of data-related activities across the organization

• Data accessibility, appropriate use, and security

We heard many of the challenges healthcare organizations nationwide face today:

• Acquisitions have resulted in additional data and metric definitions and an overall lack of standardization and agreement on how to interpret data

• Silos of knowledge resulting from factors like geographical dispersion, uneven team distributions of data analysts or analysis skillsets, and lack of direct communication channels between teams

• Large ERP implementations or upgrades with varied levels of adoption across hospitals, and development of training and institutional knowledge resources that fail to meet day-to-day business needs or keep up with demand

• Chaotic reporting and analytics environment with many disjointed source systems, causing difficulties in finding and accessing the right information

• An overwhelming queue of data requests that IT is unable to fulfill in time to meet the business need

After distilling the common pain points among business groups, West Monroe designed a federated (non-centralized) governance model that would leverage and scale pockets of existing data expertise. We also helped define formal process flows, decision frameworks, and data governance supporting tools that would focus on enabling the following capabilities over time:

• Through peer-to-peer mentorship and training, grow more multi-functional analysts from beginner roles like data entry and report development to true data analysts

• Processes to reduce duplicate work, enforce standardization, document solutions to data requests, and efficiently escalate requests that need to be met immediately or need top-down support from business leadership

• “Translators” who can decipher business and technical language barriers between requesters and IT, to develop clearer requirements and build trust in existing data assets

• Tools to monitor changes to reports, data definitions, etc., making the side effects of ERP implementations/upgrades and business rule changes more transparent and easier to anticipate

The resulting governance model is now able to support an organization of thousands of care providers while relying on less than 1 percent of people resources. We were able to accomplish this by limiting governance only to core areas of the business, supporting development of institutional knowledge resources to help business users self-service, and establishing a roadmap that would simplify and make the organization’s current reporting and analytics environment easier to navigate.