Effective October 1, 2015, CMS data requires the use of ICD-10 codes for all diagnoses. ICD-10 code sets are not just an update of the ICD-9 code sets but rather fundamentally change the structure and concepts of the codes. ICD-10 codes are more specific, provide the ability to expand the number of codes in each chapter, and reflect advances in medical knowledge and technology.
The CDC and CMS created a tool, “General Equivalence Mappings” (GEMs) as "general purpose translational" tools. However, they are not “true crosswalks” and have some limitations that impact their usefulness in conducting research projects that span across the conversion (2015 data year).
1. These mappings are suggestions and may be approximates, not an identical match.
2. Not all codes in one system have a translation in the other.
3. The GEMs are not bidirectional – content within a given map (forward or backward) differs based on the direction of the mapping.
Code Mapping Categories
1. 1:1 – ICD-9 codes map to ICD-10 codes and are considered clinically equivalent. These include both exact and approximate matches.
2. many:1– multiple ICD-9 codes map to 1 ICD-10 code
3. 1:many – one ICD-9 code maps to multiple ICD-10 codes.
4. many: many – a group of multiple ICD-9 codes covers the same concept as a group of multiple ICD-10 codes; however, they divide up the concept in fundamentally different ways. Individual ICD-10 codes have no equivalent individual ICD-9 codes.
5. No map – no ICD-10 translation for a given ICD-9 code exists or no ICD-9 code for a given ICD-10- code exists
Recommendations for Research Projects
For research projects that include Medicare/Medicaid data spanning the transition date the DAC recommends taking the following steps:
1. Generate independent lists of ICD-9 and ICD-10 codes that are relevant to your clinical concept (the disease that the project hopes to study).
2. Map each of the independent lists with the GEM data both forward and backward
3. Review the results to confirm the completeness and accuracy of your original lists; make sure you are not missing codes you are interested in, and that none of the codes on your original list have meanings that you did not intend. If you add codes during this step, repeat step 2.
4. Use the DAC Code Trends files to understand discontinuities in coding across the transition date based on your ICD-9 list before, and your ICD-10 list after, the transition.
Medicare Spending and Utilization Dashboards
The CMS Chronic Conditions Warehouse develops and publishes the Medicare spending and utilization dashboards.
“Regardless of when a claim was submitted for payment, services that occurred prior to October 1, 2015, use ICD-9 codes. Chronic conditions identified in 2015 are based upon ICD-9 codes for the first ¾ of the year (January-September) and ICD-10 codes for the last quarter of the year (October-December).” - CMS CCW