Introduction

This concept describes the development and use of a crosswalk between diagnostic codes in three International Classification of Diseases (ICD) versions: ICDA-8, ICD-9-CM and ICD-10-CA into 130 mutually exclusive categories of chronic diseases ( Hamad AF et al., (2021) ). The concept also includes a SAS® macro to use the crosswalk and develop chronic disease histories across different ICD versions using Hospital Abstracts and Medical Claims - Medical Services data (outpatient records).

Methods

Background information

Multiple ICD versions are captured in administrative databases in Canada. For example, in Manitoba, three ICD versions are found in hospital and physician administrative data: eighth revision (ICD-8) / eighth revision-adapted (ICDA-8), ninth revision with clinical modifications (ICD-9-CM), and tenth revision with Canadian adaptations (ICD-10-CA). To facilitate chronic disease research, we mapped these three ICD versions into 130 clinically meaningful disease categories by adapting the Clinical Classification Software (CCS). The CCS was developed by the Healthcare Cost and Utilization Project (HCUP) for the US Agency for Healthcare Research and Quality (AHRQ) and provides a validated crosswalk to group diagnosis codes captured in ICD-9-CM and ICD-10-CM into mutually exclusive and broad categories of diseases See Clinical Classification Software (CSS) for ICD-9-CM on the HCUP website for more information. We revised the CCS crosswalk by narrowing the focus to chronic diseases, extending the mappings to include ICDA-8 and ICD-10-CA codes and using truncated 3-digit ICD-9-CM codes, which are used in outpatient records in most provinces in Canada ( Hamad AF et al., (2021) ).

Data sources

The list of ICD-9-CM codes included in the crosswalk was obtained from HCUP. The list of ICDA-8 codes was obtained from the US Department of Health, Education, and Welfare. The list of ICD-10-CA codes was generated from the Manitoba Hospital Abstracts database from 2005 to 2018.

The SAS® macro can generate chronic disease histories using diagnosis codes from the Manitoba Hospital Abstracts and Medical Claims - Medical Services data from 1970 to the present. Hospital Abstracts include hospitalization records and currently report up to 25 diagnosis codes per episode. See the Hospital Abstracts Data Description for more information. Diagnoses are reported in ICDA-8 between 1970 and 1979, ICD-9-CM between 1979 and 2004, and ICD-10-CA since 2004. Medical Services capture outpatient physician visits and include diagnoses in ICDA-8 between 1970 and 1979, ICD-9-CM (3-digit codes) between 1979 and 2015, and ICD-9-CM (5-digit codes) since 2015. See the Medical Claims - Medical Services Data Description for more information. The macro can be revised for use within other jurisdictions with ICD codes in other types of data (e.g., emergency department records, electronic medical records).

Description of the mapping steps

ICD-9-CM, with 3-digit level codes, was the reference coding system for mapping the other ICD versions. The mapping process included four steps and was conducted independently by two reviewers. Discrepancies in mappings were resolved through deliberation until a consensus was reached. More details are available in Hamad AF et al., (2021) .

1. Identifying ICD-9-CM codes for chronic conditions

ICD-9-CM chapters that did not capture specific chronic conditions were excluded: Infectious and Parasitic Diseases (001-139), Symptoms, Signs, and Ill-Defined Conditions (780-799), Injuries and Poisoning (800-999) and Supplementary Codes of Factors Influencing Health Status (V01-V91). Then, the Chronic Condition Indicator (CCI) was applied to exclude ICD-9-CM codes of non-chronic diseases. See the Chronic Condition Indicator (CCI) for ICD-9-CM on the HCUP website for more information. The CCI is a tool developed by HCUP and sponsored by AHRQ that classifies ICD-9-CM codes into chronic and non-chronic conditions.

2. Creating a crosswalk between 3-digit ICD-9-CM codes and disease categories

The existing CCS crosswalk maps 5-digit ICD-9-CM codes to disease categories. Truncating ICD-9-CM codes into 3-digit codes resulted in many instances where a single 3-digit ICD-9-CM code maps to multiple disease categories. ICD-9-CM codes that mapped to multiple disease categories were either assigned to a modified category that combines the overlapping disease categories or assigned to a single CCS disease category. CCS disease categories were combined if there was a significant overlap in the conditions. The choice was made to map the overlapping ICD-9-CM code into a single CCS disease category when the ICD-9-CM code description fit primarily with one CCS category and less with other categories. This resulted in a revised list of CCS disease categories comprising 130 mutually exclusive categories of chronic diseases.

3. Creating a crosswalk between 3-digit ICDA-8 codes and CCS categories

3-digit ICDA-8 codes were manually mapped to the 130 disease categories produced in step 2 using existing crosswalks between ICDA-8 and ICD-9-CM from the Swedish National Board of Health and Welfare, the US Department of Health and Human Services, the Manitoba Centre for Health Policy, and additional published literature listing diseases-specific equivalent ICD codes.

4. Creating a crosswalk between ICD-10-CA and CCS categories

ICD-10-CA codes were first mapped to ICD-9-CM codes based on a crosswalk developed by the Canadian Institute for Health Information (CIHI) and then mapped to the 130 disease categories using the crosswalk produced in step 2. ICD-10-CA codes with fair or poor mapping grades based on the CIHI grading system were reviewed independently by two reviewers; the final disease category assignment was based on the existing CCS crosswalk for ICD-10-CM.

How to use the crosswalk files

The crosswalk is provided in two Microsoft Excel csv (Comma Separated Values) files.

The file CCS_diag_crosswalk.csv (CSV) contains the mappings between ICDA-8, ICD-9-CM and ICD-10-CA codes and the 130 chronic disease categories. The file has three variables:


1. CCS_category_id : an ID assigned to each of the 130 chronic disease categories based on the source CCS crosswalk.
   
2. diagnosis_coding_system : an indicator of the coding system; ICDA-8, ICD-9-CM or ICD-10-CA.
   
3. diagnosis_code_match_pattern : lists the ICD codes that map to the chronic disease categories, formatted into uppercase, without whitespace.

The file CCS_categories.csv (CSV) contains descriptions of the 130 chronic disease categories. The file has 2 variables:


1. CCS_category_id : an ID assigned to each of the 130 chronic disease categories based on the source CCS crosswalk.
   
2. CCS_category_descr : a short description of the chronic disease categories, based on the descriptions assigned to the original CCS categories.

Research using the crosswalk

The crosswalk was used in research predicting the risk of asthma in children using complete health histories of the children and their parents (Hamad AF et al., (2023)). The health histories were constructed using diagnosis codes in Hospital Abstracts and Medical Claims - Medical Services data from 1974 to 2019. The crosswalk was used to map all children and parental ICD codes in ICDA-8, ICD-9-CM or ICD-10-CA to 130 chronic disease categories that were used for asthma risk prediction using machine learning approaches.

Limitations / Cautions

• The ICD-10-CA crosswalk includes the codes recorded in the Manitoba Hospital Abstracts database over a defined period (2005-2018); therefore, it may not include the entire list of ICD-10-CA codes.
  
• The mappings were developed using ICD versions relevant to Manitoba and other provinces in Canada. Some barriers could be associated with applying the crosswalk to ICD versions used in other countries.
  
• The crosswalk has not been independently validated.

SAS Macro Code and Input Files

There are several files required to facilitate applying the crosswalk to map codes in different ICD versions to chronic disease categories:


• SAS macro: identify_CCS_health_status-SAS-macro.txt (TXT) - this macro uses the crosswalk to construct disease status of 130 chronic disease categories given a set of ICD codes in any of ICDA-8, ICD-9-CM or ICD-10-CA versions. It also produces the status of 30 high-level categories “clusters” of chronic disease categories that are based on the source CCS crosswalk.
  
• SAS Macro documentation: identify_CCS_health_status-SAS-macro-documentation.pdf (PDF) - this file describes in detail how the SAS macro works.
  
• Four csv input files required for the macro to function:
  
  • CCS_diag_crosswalk.csv (CSV) - provides a crosswalk between three ICD versions (ICDA-8, ICD-9-CM, ICD-10-CA) and the CCS categories,
    
  • CCS_cluster_mappings.csv (CSV) - contains the mappings between CCS categories and high-level clusters,
    
  • CCS_clusters.csv (CSV) - lists the 30 high-level CCS categories of chronic diseases,
    
  • CCS_categories.csv (CSV) - lists the 130 CCS categories