Attachment I

Development of ICD-10-CM and ICD-10-PCS

Overview of ICD-9-CM

The clinical modification of ICD-9 (ICD-9-CM, Volumes 1 and 2) was adopted in the United States in 1979 for morbidity applications, at the same time that ICD-9 (published by WHO) was adopted for mortality data.  In addition to its use in records and surveys, ICD-9-CM is used to classify diseases and health conditions on health care claims and is the basis for prospective payment to hospitals, other health care facilities and health care providers.

The U.S. also developed its own procedure coding system (ICD-9-CM, Volume 3) for inpatient hospital services in the late 1970’s to use with ICD-9-CM, Volumes 1 and 2 for diagnoses; this was necessary because the WHO had not produced a procedure coding system. Since 1979, procedures performed in hospitals have been coded for hospital statistics and on hospital claims, using ICD-9-CM, Vol. 3.   The Current Procedural Terminology (CPT-4), developed and maintained by the American Medical Association, is used in the United States to code professional services on claims of physicians and other non-inpatient providers.  All users code diagnoses with ICD-9-CM, Volumes 1 and 2.  When the inpatient prospective payment system was implemented in 1983, ICD-9-CM, Volumes 1, 2 and 3 was used as the basis for assigning cases to the DRGs. All diagnostic and procedural information was captured using ICD-9-CM. Because there had been radical changes and advances in health care since the implementation of ICD-9-CM, there quickly arose a need to update and revise the system. This was particularly true for the procedure codes of the system, but users also wanted to update the diagnosis portion to obtain greater clinical detail.  Thus, an annual updating process was established through the ICD-9-CM Coordination and Maintenance Committee.  Although this process continues to allow some addition of new conditions and procedures, and expansion for greater detail, it uses as its base a classification system that was developed nearly 30 years ago.

The two separate timelines, which follow, detail the development of the clinical modification of ICD-10 (ICD-10-CM) by the National Center for Health Statistics, CDC, which is responsible for diagnosis classification in the United States and the development of ICD-10 Procedure Coding System (ICD-10-PCS) by the Centers for Medicare and Medicaid Services.

ICD-10-CM Development Timeline


The International Statistical Classification of Diseases and Health Related Problems (ICD), now in its tenth revision, has become the international standard diagnostic classification for all general epidemiological and many health management purposes.  The ICD is the standard used throughout the world for classifying causes of mortality as recorded at the registration of death, and for reporting these data nationally and to the World Health Organization.

Originally designed to classify causes of death, the scope of the ICD was extended at the Sixth Revision in 1948 to include non-fatal diseases.  The application of the classification to morbidity statistics has expanded with each subsequent revision.  Nonetheless, the United States and a number of other countries continue to find it necessary to develop clinical modifications of the ICD to meet the needs of their respective healthcare systems that require more detailed clinical information from hospital, clinic and physician records.


The Tenth Revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) was first released, by the World Health Organization (WHO) in 1993.


As a first step in implementing ICD-10, NCHS awarded a contract to the Center for Health Policy Studies (CHPS) to evaluate ICD-10 for morbidity purposes within the United States.  A prototype of ICD-10-CM was developed following a thorough evaluation of ICD-10 by a Technical Advisory Panel (TAP) consisting of private and public sector stakeholders.  TAP membership included: AHIMA (Rita Finnegan), Beth Israel Hospital (Dr. Lisa Iezzoni), Mayo Clinic (Karel Weigel), HCFA (Patricia Brooks), AHCPR (Kathleen McCormick), physician perspective (Dr. William Felts).  The TAP concluded that there were compelling reasons for recommending an improved clinical modification of ICD-10 that would overcome most limitations of ICD-10 for morbidity applications.  The TAP strongly recommended that NCHS proceed with implementation of a revised version of ICD-10-CM.


Further work on enhancements to ICD-10-CM were undertaken by NCHS, including a thorough review of ICD-9-CM Coordination and Maintenance Committee proposals for modifications that could not be incorporated into ICD-9-CM and through collaboration with many medical/surgical specialty groups.

The specialty groups included: American Association of Dermatology, American Academy of Neurology, American Association of Oral and Maxillofacial Surgeons, American Academy of

Orthopedic Surgeons, American Academy of Pediatrics, American College of Obstetricians and Gynecologists, American Burn Association, American Diabetes Association, American Nurses Association, American Psychiatric Association, American Urology Association, ANSI Z16.2 Workgroup (Workers Comp), National Association of Childrens Hospitals and Related Institutions.


December 1997

The draft of the Tabular List of ICD-10-CM, and the preliminary crosswalk between ICD-9-CM and ICD-10-CM were made available on the NCHS website for public comment. Comments received during the three-month open comment period, which began December 1997 and ended February 1998, were requested to be in writing. At the conclusion of the open comment period the draft version of the ICD-10-CM (and the draft crosswalk) were removed from the NCHS homepage.  More than 1,200 comments were received from 22 individuals and organizations representing a variety of groups, including one governmental agency, two research institutions, three information system developers, four professional organizations, and several health care providers.  Comments ranged from general observations to very specific and detailed analyses.  The comments received were analyzed and categorized by a contractor into the following draft categories: 480 should be rejected; 238 required no response or necessitated no further action; 180 merited direct incorporation into ICD-10-CM; 268 were recommended as stated; and 77 had merit but required further review to determine possibility of inclusion into ICD-10-CM.


ICD-10 is implemented in the United States for mortality reporting.  Final death statistics and leading causes of death for data years 1999 and 2000, using ICD-10, have been published and are available on the NCHS website.  Preliminary data for year 2001 was published March 2003.

An overview of the comments received during the ICD-10-CM comment period were posted on the NCHS website in 1999.  A summary of the comments was also presented at the November 1999 ICD-9-CM Coordination and Maintenance Committee meeting and posted on NCHS website.


Development of ICD-10-CM continued with changes being made in response to the open comment period, as well as, input from physician specialty groups.


May 29, 2002

NCHS posted, on its website, a pre-release version of ICD-10-CM.  Consistent with ICD-10, the classification is alphanumeric.  Codes are either six or seven digits.  (ICD-9-CM, Volume 1 is five digits numeric.)  The descriptions are consistent with current medical terminology and findings.  Some of the modifications featured in ICD-10-CM have already been incorporated into clinical modifications of ICD-10 implemented in Australia (ICD-10-AM) and Canada (ICD-10-CA).  NCHS will make available the following tools that are essential to a transition to ICD-10-CM:  database version, crosswalk, educational materials, official coding guidelines, and comparability study/ratios for trend analyses.  NCHS has already conducted the comparability study for mortality and will also conduct similar study for morbidity.  The ratios assist users of coded data to discriminate between real changes in utilization by diagnosis and those resulting from artifacts of the coding system


June 2003

An updated pre-release draft of ICD-10-CM is posted on the NCHS web site at:  The files, in PDF (Adobe) format include the tabular list, alphabetic index, the external cause of injury index and the Table of Neoplasms, as well as draft coding guidelines.

August 2003

The American Health Information Management Association (AHIMA) and the American Hospital Association (AHA) jointly conduct a pilot test of ICD-10-CM during June/July 2003.  The study involved dual coding records in ICD-9-CM and ICD-10-CM.  More than 6100 records from a broad cross section of health care community were dual coded by 180+ participants.  The initial results indicate that: there is general support for adoption of ICD-10-CM; ICD-10-CM is seen as an improvement over ICD-9-CM; and ICD-10-CM is more applicable to non-hospital settings than ICD-9-CM.

ICD-10-PCS Development Timeline


The Centers for Medicare and Medicaid Services (CMS), formerly the Health Care Financing Administration (HCFA), began developing a replacement to ICD-9-CM, Volume 3 in 1990 in order to address a number of limitations with the current procedure classification.  ICD-9-CM procedure codes are made up of four digits. There are only 10 codes available within each category (fourth digits 0-9). Once a category is full, one must either combine types of procedures under one code, or find room in another section of the code book to place a new code. The benefit of such a system is that one can easily collapse the codes into categories when doing research to capture a wide range of similar procedures. When one puts a similar code in a separate part of the book, coders and researchers may not easily find it. There can be gross errors when trying to identify particular types of cases or make projections of numbers of procedures performed when codes are not carefully placed within a system such as the current ICD-9-CM.

A specific example of a generic problem with ICD-9-CM is that there is not a consistent identification of the procedure approach. While the approach may not have been an important issue during the 1970s, there have been considerable advances in surgery since that time. With a coding system that is 23 years old, not only is the technology bound to be outdated, the premises on which the coding system was established are outdated. There are a number of approaches and techniques used for procedures such as lasers and the use of scopes that were not anticipated when the structure of ICD-9-CM was developed. Therefore the basic categories were established on technology that is currently outdated. Updating this system has been quite difficult and involves making compromises.

The development of ICD-10-PCS adhered to the criteria established by NCVHS for a procedure classification system (Appendix I).  ICD-10-PCS is made up of seven alphanumeric characters. It should be noted that ICD-10-PCS is not based on an international classification because currently none exists.  The system provides exponentially greater code capacity, where all substantially different procedures can have a unique code. Whereas ICD-9-CM procedure classification contains fewer than 4,000 codes, the current draft of ICD-10-PCS contains 197,769 codes (1). There is also significant room for easy expansion as new technologies are developed.  The substantial increase in codes means that a coder can quite readily find the appropriate code that describes a particular procedure, rather than having to use a less specific or clinically detailed ICD-9-CM, Volume 3 code.

ICD-10-PCS was developed using an open process.  A Technical Advisory Panel (TAP) provided review and comments throughout development.  The TAP included representatives of AHA, CPRI/ANSI-HISB, AHIMA, AMA, American College of Surgeons, managed care industry, medical informatics, American Association of Medical Transcriptionists, NIH/NLM, AHCPR, state health system, hospital information, NCHS, ProPAC, and the National Association of Childrens Hospitals and Related Institutions.

ICD-10-PCS Timeline


May 1986, a meeting was convened by HCFA to open discussion among people interested in procedure coding.  As a result of that meeting, AHA and AMA agreed to co-chair a task force that would outline and set priorities for the objectives of a common procedure coding system.  This task force was charged to undertake a thorough evaluation of the purpose and scope of Volume 3 of ICD-9-CM and CPT to evaluate the feasibility of developing a new procedure coding system to achieve the objectives of a common system.


Earlier review efforts uncovered structural problems in both Volume 3 of ICD-9-CM and CPT. Concern for data quality issues and the cost of submitting data in more than one classification is significant.  The feasibility of creating a single procedure coding system that will satisfy all users is as yet unknown.  The AMA sponsored a study to investigate the costs and benefit of a single system for physician payment.  The study conducted by Coopers and Lybrand, compared two alternatives: 1) a major restructuring of CPT to serve uses beyond physician offices; and 2) a replacement of both Volume 3 of ICD-9-CM and CPT. The results of the AMA study showed that the costs of a replacement system were significant and that the identification of benefits was difficult, thus the consultants concluded that a replacement system, for measuring physician services, was not justified.

HCFA awarded a contract for a pilot project that would review the cardiovascular procedures in all of the procedure coding systems (CPT, HCPCS, local HCPCS and Volume 3 of ICD-9-CM), standardize the nomenclature with definitions, recommend a standard format amenable to updating and expansion, and develop a cardiovascular chapter for ICD-9-CM, Volume 3.


HCFA funded a continuation of the pilot project for the development of a revision of the respiratory system chapter consistent with the approach, design and format of the cardiovascular chapter.  The project focuses on the respiratory chapter.


HCFA announced plans to initiate a solicitation for a contract to develop a new procedure coding system for use with hospital inpatients to replace Volume 3 of ICD-9-CM.  The new system is referred to as ICD-10-PCS.


March 1995

HCFA awarded a contract to 3M HIS to develop the procedure classification to replace Volume 3 of ICD-9-CM (hospital inpatient procedures).  The contract is based on the prototype 7-digit alphanumeric procedure classification system developed by 3M HIS in previous contracts.  Primary objective is developing a new procedure coding system to replace the current Volume 3 of ICD-9-CM; the project’s additional objectives are to improve the accuracy and efficiency of coding, to reduce training efforts, to improve communications with physicians, and to be compatible with the current billing infrastructure.

ICD-10-PCS was developed using an open process.  A Technical Advisory Panel provided review and comments throughout development.  The TAP included American Health Information Management Association (Sue Prophet), American Hospital Association (Nelly Leon-Chisen), American Medical Association (Barry Eisenberg), CPRI/ANSI-HISSP (Dr. Simon Cohn), American Association of Medical Transcription (Claudia Tessier), NIH/NLM (Betsy Humphreys), AHCPR (Michael Fitzmaurice), state health system representative (Kevin Ray), hospital information (Laura Green), NCHS (Donna Pickett), ProPAC (Julian Pettingil), Dr. Clement  McDonald, American College of Surgeons (Dr. George Spaulding), United HealthCare (James Cross and Philip Bryson), and the National Association of  Children’s Hospitals and Related Institutions (John Muldoon).  ICD-10-PCS was sent to approximately 30 specialty groups for their review and comments.  Additionally, HCFA also provided an onsite presentation at the AMA to more than 20 specialty groups.


A training program was developed, and informal testing and training were conducted.  AHIMA national conventions trained 65 coders in two half-day sessions.  Another 70 volunteered to test the system, but weren’t at the training.  The coders received 400 records that were coded with ICD-10-PCS during next 3 months (received by January 1997).


Clinical Data Abstraction Centers (CDACs) conducted formal testing of ICD-10-PCS.  HCFA trained CDACs (5/14-15/97) with follow-up training after the CDACs informally coded 30 records (6/3/97).  Final training session 6/18/97 where CDACs tested system on 5,000 medical records (2500 per CDAC) identifying cases with a wide distribution of ICD-9-CM procedure codes.


Additional formal testing of ICD-10-PCS using ambulatory records conducted 10/98 – 2/99.

582 ambulatory records obtained by CDACs.  369 records of the 582 had procedures that were tested using ICD-10-CM and reported as part of the 3/99 on findings.  CMS was not able to obtain obstetrics records.   CMS also tested ICD-10-PCS on list of problem cases from Editorial Advisory Board for Coding Clinic for ICD-9-CM submitted by AHA.

Spring 1998

Final version of ICD-10-PCS, training material and crosswalk to ICD-9-CM procedure codes posted on CMS website released spring 1998.  The CMS written testimony on April 9, 2002 states that CMS has already awarded a contract to 3M to undertake any DRG conversion activities that would be required should ICD-10-PCS be named as a national standard.  Information on a conversion and ICD-10-PCS’s incorporation into the DRG system would be made available as part of the annual inpatient prospective payment system proposed and final rule process.


ICD-10-PCS is updated every October 1 to accommodate changes made to ICD-9-CM, Volume 3.

Appendix I

NCVHS Recommendations for a Single Procedure Classification System, November 1993

An Outline of the Characteristics of a Procedure Classification System

Hierarchical structure

Ability to aggregate data from individual codes into larger categories
Each code has a unique definition forever – not reused


Flexibility to new procedures and technologies (Aempty@ code numbers)
Mechanism for periodic updating
Code expansion must not disrupt systematic code structure


Provides NOS and NEC categories so that all possible procedures can be classified somewhere


Each procedure (or component of a procedure) is assigned to only one code

Ease of Use

Standardization of definitions and terminology
Adequate indexing and annotation for all users

Setting and Provider Neutrality

Same code regardless of who or where procedure is performed


Body system(s) affected
Technology used

Limited to classification of procedures

Should not include diagnostic information
Other data elements (such as age) should be elsewhere in the record