Without some tough love, code sets can run away from you and wreak havoc with your business functions.
Within a typical data model, code sets take up very little space; but don’t let their size fool you. An inaccurate or misaligned code set can detrimentally produce seismic shifts in business functions by blocking access to what otherwise are accurate customer records.
For instance, the misapplication of a perfectly good zip code might inadvertently deliver a direct mail piece to the other side of town - even when the rest of the mail list information is correct. Admittedly, that’s a rather benign example. But, when it comes to health care codes (ICD-10), the wrong codes appended to a patient’s medical record might impact their diagnosis, prognosis, medication, and even insurance reimbursement.
Code sets are designed to be reference data, providing both the strategic basis for analytics, as well as facilitating daily access for researching operational business data. Ideally, codes sets need to be represented by a “single version of truth” that can be shared across all mission-critical business systems. In this respect, code sets deserve the same attention and tough love as master data. And like master data, codes sets need to be governed through change management, workflow and have their initial creation strictly monitored by subject matter experts.
Code set mismanagement includes:
- Lack of code set synchronization and misapplication to the master record
- Poor mappings between external, industry standard code sets and their internal counterparts.
- Ad hoc and siloed creation of code sets
It is perhaps the siloed, non-standardized creation of new code sets that perpetuates the greatest degree of inaccuracy. Like master data, poor code set data management guarantees unknown volumes of duplication and redundancy.
But like metadata terms and definitions, code set titles and descriptions are subject to subtle, but often times, mind-numbing variances:
Code Set title: Retired Employees | Description: age >65 years
Code Set title: Retired Employees | Description: Age 65-79 years
Or, one precise code set description might be linked with titles:
Code Set title: Hispanic Employees | Description: 30-58 years
Code Set title: Latino Employees | Description: 30-58 years
With the probable exception of data management implementation styles (centralized is most associated with reference data management), code set governance aligns well with standard MDM (Master Data Management) objectives:
- Ensuring same version of reference data is used throughout the company
- Enforcing change management policies
- Increasing data sharing and collaboration throughout the company
- Ensuring data governance and auditability of all reference data
- Easily publishing data to subscribing systems