HEALTHCARE COST & UTILIZATION PROJECT

User Support

Do Your own analysis
Explore Expert Research & Limited Datasets

Transitioning to ICD-10-CM and ICD-10-PCS

 
Slide presentation of Transitioning to ICD-10 and ICD-10-PCS: Challenges in Trend Analysis. PDF slides (330 KB) are also available for download.

Presented at the March 18, 2014 HCUP Partners Monthly Meeting by Donna Pickett, MPH, RHIA, National Center for Health Statistics/CDC (Centers for Disease Control and Prevention), Co-Chair, CDC ICD-10 Transition Workgroup



Slide 1

Transitioning to ICD-10 and ICD-10-PCS: Challenges in Trend Analysis

Transitioning to ICD-10 and ICD-10-PCS: Challenges in Trend Analysis

March 18, 2014

HCUP Partners Monthly Meeting

Donna Pickett, MPH, RHIA

National Center for Health Statistics/CDC (Centers for Disease Control and Prevention)

Co-Chair, CDC ICD-10 Transition Workgroup
 
Slide 2

Overview

Overview
  • Impact of the Transition on Data
  • Analysis Over Time and Across Code Sets
  • Mapping to and Selecting Appropriate ICD-10-CM/PCS Codes
  • Reconciling Data for Calendar Year 2014
  • Impact on Downstream Users of Data
  • Appendix - Resources
 
Slide 3

 Impact of the Transition on Data

Impact of the Transition on Data
 
Slide 4

Reality of Impact

Reality of Impact
  • ICD codes updated regularly so data have always been impacted
  • Transition will change the way public health defines, identifies, analyzes and reports on many health conditions and healthcare services
  • There will be some level of data discontinuity for analyses over time/across code sets
  • No single best approach to conducting trend analyses; each program/project team will need to determine own approach
 
Slide 5

Analysis Over Time and Across Code Sets

Analysis Over Time and Across Code Sets
 
Slide 6

Possible Universes of Data

Possible Universes of Data
  • Individual diseases (asthma)
  • Ranges of codes (chronic lower respiratory grouping that includes asthma)
  • Sub-categories or categories (all respiratory diseases)
 
Slide 7

Type of Data Received

Type of Data Received
  • Already coded in ICD
  • Self-reported diagnosis information
  • Abstracted from narrative/li>
  • Complete narrative*

* Gold standard for use in comparability analysis; comparability ratios derived from dual coded data
 
Slide 8

Decisions, Decisions, Decisions

Decisions, Decisions, Decisions...
  • Many decisions to be made; majority will be unique to individual organizations, programs and/or projects
    • Mapping between code sets (or not) for surveillance, research, statistical analyses, and evaluation of public health and/or healthcare services
    • Public use data sets and education of users
      • Stakeholders that fund supplements to surveys
      • Stakeholders that create discharge/encounter data sets
  • Coordinated (public health) efforts: external cause of injury matrix and injury diagnosis matrix for injury reporting, reportable neoplasms (cancer registries)
 
Slide 9

Macro Level Possible Strategies Already Coded Data

Macro Level Possible Strategies Already Coded Data
  • Map all ICD-9-CM codes into ICD-10-CM/PCS
  • Start over and define conditions from beginning in ICD-10-CM/PCS
  • Map a few key conditions/events into ICD-10-CM/PCS; start over with the rest
  • Map all ICD-10-CM/PCS codes back into ICD-9-CM
  • Map/aggregate ICD-10-CM/ICD-10-PCS codes to category level of detail
 
Slide 10

Mapping to and Selecting Appropriate ICD-10-CM/PCS Codes

Mapping to and Selecting Appropriate ICD-10-CM/PCS Codes
 
Slide 11

Considerations for Mapping

Considerations for Mapping
  • There are "one-to-many" relationships (742.3 Congenital hydrocephalus maps to four ICD-10-CM codes)
    • Will need to decide which of the "many" codes are needed
  • There are "many-to-one" relationships (21 ICD-9-CM codes for Tuberculosis (TB) of lung map to one ICD-10-CM code, A15.0 TB of lung)
    • Loss of detail for stages of TB and how confirmed
 
Slide 12

Considerations (Continued)

Considerations (Continued)
  • There are new concepts in ICD-10-CM that have no predecessor in ICD-9-CM (Glasgow Coma Scale, under dosing, blood type, blood alcohol level)
    • Can track conditions that were not available in ICD-9-CM
  • Some ICD-9-CM codes do not have a translation option in ICD-10-CM (ICD-9-CM code E927.3; V codes largely excluded)
    • Accept the loss and do nothing
    • Use an alternative translation (V03.6 translates to Z23 Encounter for immunization)
    • Propose to add codes into ICD-10-CM/PCS
 
Slide 13

Implications of Mapping between Code Sets

Implications of Mapping between Code Sets
  • The identification or definition of "cases" has changed
  • You may be identifying more or fewer individuals with condition/s of interest depending on in/exclusion of certain codes/code sets [Transient Ischemic Attach (TIA) included in stroke]
  • For some programs arriving at final ICD-10-CM/PCS codes will require extensive collaboration with external partners
 
Slide 14

Implications (Continued)

Implications (Continued)
  • Mapping from ICD-9-CM to ICD-10-CM differently for the same health conditions/procedures
    • Judgment is needed to determine which ICD-10-CM/PCS codes to utilize going forward
    • Others who study same/similar condition/s may select a different set of ICD-10-CM/PCS codes; without collaboration among the respective entities studying the same conditions, multiple condition definitions by those who analyze the same condition could develop
  • Collaborate on/share mappings or lists of ICD-10-CM/PCS codes with partners
    • Leverage "ICD-10-CM/PCS Transition" community on phConnect

Link to the ICD-10-CM/PCS Transition Community located on phConnect, which is an online collaboration tool built to support geographically dispersed professionals working in the field of public health: http://www.phconnect.org/group/icd-10-cm-pcs-transition-for-public-healthExit Disclaimer
 
Slide 15

Reconciling Data for Calendar Year 2014

Reconciling Data for Calendar Year 2014
 
Slide 16

Possible Strategies

Possible Strategies
  • Backwards map the 3 months of ICD-10-CM/PCS coded data into ICD-9-CM and report together
  • Let 9 months of ICD-9-CM coded data represent entire year
  • Report on first 9 months and last 3 months of data separately
  • Forward map the 9 months of ICD-9-CM coded data to ICD-10-CM and report together
  • Convert reporting from Calendar Year to Fiscal Year, if feasible
 
Slide 17

Impact on Downstream Users of Data

Impact on Downstream Users of Data
 
Slide 18

For Data Users: Selection and Interpretation of Data

For Data Users: Selection and Interpretation of Data
  • Users of data derived from ICD codes need to understand how the data have changed
    • Understand origin of the data
    • Selection/extraction of data for independent analysis
    • Understand published data analysis/reports
  • Interpretation of data
    • May be considered new baseline year
    • May not be able to draw conclusions for first year or two
 
Slide 19

Possible Trending Impact

Possible Trending Impact

Image: Screen shot of a graph depicting the age-adjusted death rates for Nephritis, nephritic syndrome, and nephrosis in the United States from 1968 to 2005. Three lines exist representing ICD-8, ICD-9 and ICD-10 coding for a given time period. From 1968 to 1980, the death rate for Nephritis, nephritic syndrome, and nephrosis slightly decreased, remained steady from 1980 to 2000, and then slightly increased from 2000 to 2005. This graph reveals that when transitioning from one ICD coding classification system to the next, examining trends over time may be misleading since there is not always a one-to-one correlation between code sets.

Reference: Robert Anderson (2011). Coding and Classifying Causes of Death: Trends and International Differences. In Rogers, Richard G., and Crimmins, Eileen M (Eds), International Handbook of Adult Mortality (467-489). Springer Science.
 
Slide 20

Possible Trending Impact (Continued)

Possible Trending Impact (Continued)
  • Behavioral Risk Factor Surveillance System
    • Changed sampling methodology in 2011 to conduct landline and cellular telephone-based surveys, as well as changed weighting methodology
    • 2011 considered baseline year for data and 2011/2012 data not directly comparable to previous years
  • National Immunization Survey
    • Also changed sampling methodology in 2011 to include both landline and cellular telephone-based surveys
    • Sampling change noted as limitation in publications but no break in trends over time
 
Slide 21

Possible Strategies for Deciding How to Trend

Possible Strategies for Deciding How to Trend
  • Comparability analysis for already coded data
    • Take existing year/s of ICD-9-CM-based reporting and map them forwards into ICD-10-CM
    • Evaluate how original estimates (based on ICD-9-CM) compare with new estimates (ICD-10-CM)
    • Make determination from there (e.g., refine what ICD-10-CM codes to use and redo comparison, move forward with selected ICD-10-CM codes, etc.)
  • Comparability ratios for narrative information
    • Dual code data
    • Divide estimate for one coding scheme by estimate for other
    • Use to estimate values if coded in other coding scheme
 
Slide 22

Summary

Summary
  • Many challenges to come with trending data over time and across code sets
  • Pathways will depend on unique circumstances of each condition, range of codes used, reason for analysis, level and type of data collected, and ability to analyze and understand comparability between code sets
 
Slide 23

Questions?

Questions?

Email: ICD-10Transition@cdc.gov

For more information please contact Centers for Disease Control and Prevention

1600 Clifton Road NE, Atlanta, GA 30333

Telephone, 1-800-CDC-INFO (232-4636)/TTY: 1-888-232-6348

E-mail: cdcinfo@cdc.gov

Website for Centers for Disease Control: www.cdc.gov
 
Slide 24

APPENDIX - RESOURCES

APPENDIX - RESOURCES
 
Slide 25

Helpful Links

Helpful Links
 

Internet Citation: Transitioning to ICD-10-CM and ICD-10-PCS. Healthcare Cost and Utilization Project (HCUP). July 2016. Agency for Healthcare Research and Quality, Rockville, MD. www.hcup-us.ahrq.gov/datainnovations/icd10_transitioning_pres.jsp.
Are you having problems viewing or printing pages on this website?
If you have comments, suggestions, and/or questions, please contact hcup@ahrq.gov.
Privacy Notice, Viewers & Players
Last modified 7/28/16