With AI technologies taking the digital revolution to the next level so quickly, it can be easy to bypass common sense best practices regarding sensitive data. Agile companies are in a race to take advantage of the AI part of the digital movement, which is why core principles such as data governance and data management might take a back seat. There’s nothing sexy about:
1. putting together a data governance framework
2. hosting a few sessions to hammer out a set of data management practices and operations based on that strategic framework, then
3. socializing it to the various IT cloud teams that handle data
Even though data governance and management may not seem exciting, it’s crucial to have policies and agreed-upon practices in place to handle complicated situations involving data integrity. One such scenario arises when there are conflicting versions of entity data after recovering it from a damaged backup database. When data is considered the master, it is assigned a special identifier called a gold ID. If this master record is replicated to other databases, it’s imperative to ensure that all copies remain consistent, even in backup databases. In the event that one of these replicated databases becomes corrupted, it becomes necessary to have a defined process for accurately and safely restoring that database.
So what does data governance look like? There are some templates online to get started. I put together this governance guiding document based on previous consulting work I did as part of a strike force of architects with financial and healthcare corporate clients. I also found an outline on the AHIMA website a good point of reference for structuring a governance document, which I used as an outline below. In subsequent articles, I’ll dive into more details and provide the basics for cloud data management principles and operations, especially when it comes to data I/O ingestion in a multitenant architecture.
Here’s a healthcare industry example of a core data governance document to lay the foundation for a framework in any industry:
The characteristics of data quality can be summarized as follows:
· Accuracy: Data should be error-free and correct.
· Accessibility: Measures should be in place to ensure data availability when needed.
· Comprehensiveness: Data should contain all the necessary elements.
· Consistency: Data should be reliable and consistent throughout the entire patient encounter.
· Currency: Data should be up to date and current.
· Definition: All data elements should have clear definitions.
· Granularity: Data should be at an appropriate level of detail.
· Precision: Data should be collected in its exact form with precision.
· Relevancy: Data should be relevant to its intended purpose of collection.
· Timeliness: Documentation should be entered promptly, updated regularly, and available within specified time frames.
Many healthcare organizations understand the need for data governance, but may not understand where to begin or how to establish a robust data governance program. One potential obstacle to implementing healthcare data governance at an organizational level is a lack of understanding among key stakeholders regarding data as an asset. This can lead to data silos and delays in developing a comprehensive program.
Healthcare data governance should encompass the entire organization, involving interdisciplinary teams comprising subject matter experts. The primary aim of healthcare data governance is to foster an organizational culture that prioritizes data security, reliability, and accessibility for authorized individuals. When the entire organization is engaged, a data governance culture is established, resulting in a strong program. To cultivate a healthcare data governance culture, it’s good to start with baby steps that demonstrate the value and benefits of data governance.
In order to implement a healthcare data governance plan or program, the first step is to define the concept of data governance and determining its scope. This entails establishing a foundational framework for collecting, retaining, utilizing, accessing, and sharing healthcare data. This framework encompasses various aspects such as policies, procedures, standards, ownership, decision rights, roles and responsibilities, and accountability associated with the data.
To effectively establish healthcare data governance plans or programs, organizations should form a dedicated team called the Data Governance Management Team (or a similar name). This team should include key individuals such as the Chief Data Officer (or an equivalent position) working in collaboration with the Chief Medical Information Officer. Together, they will spearhead the efforts to develop and implement robust healthcare data governance initiatives.
If interested in more healthcare data governance, check out the PDF file on the AHIMA website here.
DATA RETENTION POLICY
A compliant data retention policy defines the duration for which data must be retained to meet regulatory and/or organizational requirements. On top of that, it specifies the actions to be taken once the retention period has expired. Organizations may choose to delete (or destroy) or archive data based on their retention guidelines.
These policies and procedures, or standard operating procedures (SOPs), provide a structured framework for efficient Data Governance, guaranteeing data integrity, appropriate access, privacy compliance, secure data sharing, and quality data retention.
To restate, in the fast-paced digital landscape IT finds itself today, driven by AI technologies like ChatGPT, it is crucial for technical organizations to not overlook fundamental best practices concerning sensitive data. While agile companies strive to harness the power of AI within the digital revolution, core principles such as data governance and data management can sometimes be neglected. Establishing a robust data governance framework and defining data management practices and operations based on a strategic framework may not seem glamorous. However, it is a necessary step for organizations to take. This involves conducting sessions to develop a comprehensive data governance framework and subsequently sharing and integrating it with the various IT cloud teams responsible for data handling.