A GUIDE TO STATE IMPLEMENTATION OF ICD-10 FOR MORTALITY National Center for Health Statistics Centers for Disease Control and Prevention July 16, 1998 TABLE OF CONTENTS Introduction Introduction to the International Classification of Diseases (ICD) Implementation Activities Tabulation Lists and Redesigning Publications Developing Mortality Edits Comparability Ratios and Conversion Tables ICD-10 Computer Applications ICD-10 Instruction Manuals ICD-10 Training NCHS Contact Persons Appendices Appendix I. Vital Statistics Instruction Manuals for Mortality Medical Coding Appendix II. ICD-10 Super-MICAR Input Record Format Appendix III. ICD-10 TRANSAX Output Record Format Appendix IV. Summary of Changes in ICD-10/ICD-9 Record Formats A GUIDE TO STATE IMPLEMENTATION OF ICD-10 FOR MORTALITY National Center for Health Statistics Centers for Disease Control and Prevention INTRODUCTION The purpose of this report is to provide a general background to the Tenth Revision of the International Classification of Diseases (ICD-10), oriented to vital statistics staff, statisticians, and epidemiologists at the State level; to describe the ICD-10 material developed by the National Center for Health Statistics (NCHS); and to provide step-by-step assistance on the key tasks. ICD-10 will be implemented in the U.S. effective with deaths occurring in January 1999. In terms of producing and publishing mortality data under ICD-10, the basic activities will be largely the same as under ICD-9. However, the mortality software (ACME, TRANSAX, and MICAR) has been written to run on personal computers rather than on the mainframe. NCHS will no longer provide mainframe support under ICD-10. Data entry under the automated systems will be largely the same under ICD-10 as ICD-9. However, under ICD-10 multiple cause coding (including MICAR reject coding) and underlying cause coding (including ACME reject coding), which some States continue to do on their own, that is, outside the scope of the Vital Statistics Cooperative Program, differs substantially from coding under ICD-9. Further, publications showing mortality data under ICD-10 will differ substantially from those under ICD-9, because of changes in coding rules, changes in category names and ICD numbers, and -- importantly -- because of changes in the tabulation lists used to produce data under ICD-10. To measure these changes, NCHS is doing a comparability study between the Ninth and Tenth Revisions of the ICD. NCHS is making available a number of products that should facilitate the transition by the States to ICD-10. These include revised instruction manuals, revised software, and training classes; some of this material has been put on the NCHS Mortality Home Page at the following address: http:/www.cdc.gov/nchswww/about/major/dvs/mortdata.htm. In addition, NCHS has provided each of the States with a single three volume set of ICD-10. NCHS is also providing state vital statistics programs with an updated and corrected version of ICD-10 Volumes 1 and 3, for state government use only. NCHS Instruction Manual Part 2a, Instructions for Classifying the Underlying Cause of Death, 1999 (See List of NCHS Instruction Manuals, Appendix I ) replaces ICD-10 Volume 2. INTRODUCTION TO THE INTERNATIONAL CLASSIFICATION OF DISEASES (ICD) A basic understanding of the ICD is essential to the production and use of mortality data, because the ICD provides the essential ground rules for the coding and classification of cause-of- death data. The ICD is a classification system developed collaboratively between the World Health Organization (WHO) and ten international centers, one of which is housed at NCHS, and is called the WHO Collaborating Center for the Classification of Diseases in North America (which includes the U.S., Canada, and the English-speaking Caribbean islands). The purpose of the ICD and of WHO sponsorship is to promote international comparability in the collection, classification, processing and presentation of health statistics, including both morbidity and mortality. The U.S. is required to use the ICD for the classification of diseases and injuries under an agreement with WHO that has the force of an international treaty. Generally, the ICD has been revised infrequently, only once every ten years usually; the purpose of the revisions is to stay abreast with medical advances in terms of disease nomenclature and etiology. The introduction of new classifications is costly to the Federal government and the states, and introduces major disruptions in time series of mortality and morbidity statistics. However, it is an essential task to stay abreast with advances in medical science, and to ensure the international comparability of health statistics.Two classification systems. In the United States, but not in most other countries, two related classification systems are used: namely, the ICD, which is used to code and classify medical information from death certificates, and the ICD-CM, where the "CM" refers to the "Clinical Modification." The ICD-CM is used to code and classify medical information from medical records and from many health surveys, and is used in the U.S. for medical reimbursement. There are many differences between the ICD-9 and ICD-9-CM even though the basic structure and design of the two classifications is similar. Thus, the ICD used for mortality in the U.S. adheres very closely to the ICD promulgated by WHO; changes that are made to the ICD in the U.S. are generally sanctioned by WHO and occur infrequently, usually only to accommodate new medical conditions that are widely recognized as public health threats but are not in the existing classification. In the past such changes between official ICD revisions have included, for example, the categories for SIDS (*795.0, *795.1, and *795.2), which were introduced for 1973 (between the Eighth and Ninth Revisions), and HIV infection (*042, *043, and *044) introduced in 1987, between the Ninth and Tenth Revisions. The asterisk preceding the number indicates that the codes are not an official part of the WHO classification, but are unique to the U.S. These changes are documented in the Technical Appendix of Vital Statistics of the United States, Mortality. Generally, the changes made by the U.S. to the ICD for mortality are carried out in close consultation with and are endorsed by the WHO. Compared with the ICD for mortality, the ICD-CM is far more detailed in order to capture the finer grained level of detail on medical records compared to the less detailed diagnostic information reported as causes of death. Further, changes in the CM are made annually, and under the auspices of a consortium that includes NCHS, the Health Care Financing Administration, and representatives of organizations concerned with medical records. The CM includes a 5th digit compared with ICD's 4-digit structure; the 5th digit allows for much greater detail, and collapses back into and 4- and 3-digit structure generally compatible with the ICD. In contrast to the U.S., many other countries used the ICD promulgated by WHO for both morbidity and mortality. ICD Revisions. Since the beginning of the Century the ICD for mortality has been modified about once every ten years, except for the twenty year interval between the last two revisions, ICD-9 and ICD-10, as shown in Table 1. The rationale for the periodic revisions has been to reflect advances in medical science and changes in diagnostic terminology. Historically, the U.S. accepted the WHO versions, except for the Eighth Revision, when the U.S. produced its own "adapted" version, which is symbolized by the "A" in ICDA-8. The U.S.'s rejection of the WHO version included disagreements at the time on the content of the circulatory chapter. That changes in the ICD for mortality have been made only every ten to twenty years ensures comparability over time in our mortality trend data. Uses of the ICD. The ICD is a classification system for mortality, but it is much more than that. Thus, the present ICD-9 contains over 4,000 categories of causes of death, including diseases -- such as heart disease, cancer, stroke, and diabetes; traumas such as fractures and burns; and external causes such as accidents, homicides, and suicides. ICD-10 has much more detail, about 8,000 categories that are valid for cause of death. (ICD-10-CM has almost 50,000 categories!) In addition to being a classification system for causes of death, the ICD serves a number of other purposes for mortality: (1) The ICD includes coding rules for causes of death. These rules allow a coder to identify the single condition on the death certificate that is considered most informative from a public health point of view, called the "underlying cause of death." The coding rules also play the important role of compensating for errors in the cause-of-death statement, where, for example, the causal change of medical events is reported in reverse order. An example would be a physician's statement of a death from lung cancer due to pneumonia. Such an incorrect sequence would, through the coding rules, be reversed to pneumonia due to lung cancer, with lung cancer selected as the underlying cause of death. (2) The ICD includes definitions such as "underlying cause of death," "live birth," "maternal death," and many others. (3) The ICD includes "tabulation" lists, which indicate the cause-of-death groupings that countries should use to present mortality data that can be compared among countries. The U.S. produces not only cause-of-death tabulation lists that meet WHO requirements, but several other lists that are designed to more specifically meet the needs for public health and medical research in the United States. The U.S. lists are described in detail below. (4) The ICD also prescribes the format of the medical certification of death in the "International Form of Medical Certificate of Cause of Death," which is reflected in a two-part medical certification of death that is part of every death certificate in the United States. (5) The ICD includes "regulations" regarding the compilation and publication of statistics on diseases and causes of death. These regulations require, for example, that member states such as the U.S. must use the ICD for compiling mortality and morbidity statistics, including collecting, coding, age-grouping, definitions and other statistics system. ICD-10 Compared with ICD-9. ICD-10 differs from ICD-9 in a number of respects: (1) ICD-10 is far more detailed than ICD-9, about 8,000 categories compared with 4,000 categories; the expansion was mainly to provide more clinical detail for morbidity applications; (2) ICD-10 uses 4-digit alphanumeric codes compared with 4-digit numeric codes in ICD-9. (3) three additional chapters have been added and some chapters rearranged. Cause-of-death titles have been changed, and conditions have been regrouped. (4) Some coding rules have been changed. ICD-10 is also published in three volumes compared with two volumes in ICD-9. Statistical Impact and Comparability. The introduction of a new revision of the ICD can create major discontinuities in trend data, as shown in Figure 1. Figure 1 shows trends in leading causes of death in the United States from 1950 to 1995 in terms of age-adjusted death rates. The lines on the chart are not continuous, but rather are broken by vertical lines that represent the introduction of a new revision of the ICD. Thus, ICD-9 was introduced in 1979. Further, the level of the rates is sometimes discontinuous between revisions. For example, a large discontinuity occurred between 1978 and 1979 in mortality from the 11th leading cause of death, "Nephritis, nephrotic syndrome, and nephrosis;" the rate for this cause in 1979 was over 70 percent higher than in the previous year, because of the introduction of ICD-9. The extent of the discontinuity is measured using a "comparability ratio," which results from double-coding a large sample of the national mortality file, once by the old revision (ICDA-8), and again by the new revision (ICD-9), and expressing the results of the comparison as a ratio of deaths for a cause of death by the later revision divided by the number of that cause of death coded and classified by the earlier revision. The national Comparability Study for ICD-9 was carried out using a sample of 137,000 deaths (and a special sample of 13,000 deaths for infants) occurring in 1976. The ratios for 1976 were considered applicable to deaths occurring in 1979, and represent the break in trend continuity resulting from introducing the new coding and classification system. The ratio for Nephritis is 1.74 indicating that 74 percent more deaths occurred from this cause in 1979 compared with 1978 only because of the introduction of ICD-9. The comparability ratio for Septicemia of 0.85 indicates that about 15 percent fewer deaths occurred in 1979 compared with 1978, because of the introduction of the new revision of the ICD. An NCHS Comparability Study is in preparation for ICD-10 using deaths occurring in 1996. If the sample size is sufficiently large, comparability ratios will be produced not only for the U.S. as a whole, but also for some states, at least the larger states. A provisional set of comparability ratios should be available at the end of 1998; and a final set of national ratios will appear with the first publication by NCHS of mortality data for 1999, which will occur in the fall of the year 2000. Conversion tables should not be used for or confused with comparability studies. Conversion tables produced by WHO compare individual ICD categories in ICD-10 with the most equivalent categories in ICD-9, and vice versa. The tables show in an approximate way how the category titles of ICD-9 and ICD-10 compare with each other. The conversions are approximate since ICD-10 is so much more detailed than ICD-9; thus, many categories in ICD-9 are split in two or more categories in ICD-10. However, this is not always the case. Some categories convert one for one, and yet other categories are less detailed in ICD-10 compared with ICD-9. While the conversion tables were used by NCHS in developing the automated software for processing mortality data, including the ACME decision tables, TRANSAX, and the MICAR dictionary, and to provide guidance in developing the NCHS mortality tabulation lists, they cannot provide guidance on the quantitative relationship between ICD-9 and ICD-10 categories. That can only be done with comparability ratios, based on a double-coding study, or "Comparability Study," such as is carried out by NCHS for each revision of the ICD. The WHO conversion tables produced by WHO are available on request from NCHS (for NCHS contact persons see Table 3); the conversion tables cannot be placed on the NCHS Home Page because of WHO copyright restrictions. Tabulation lists for ICD-10 have been produced by NCHS to meet both WHO requirements and U.S. national needs. Published as NCHS Instruction Manual Part 9, ICD-10 Cause-of-Death Lists for Tabulating Mortality Statistics, Effective 1999), these lists were distributed to states last fall to be adapted to state uses. The several ICD-10 tabulation lists, in particular the List of 113 Selected Causes of Death, will be used to present U.S. mortality data, effective with deaths occurring in 1999. We hope that the States will find these tables useful, and will adapt them for their own purposes, so that State vital statistics data can be compared with national data. The Instruction Manual Part 9 can be downloaded from the NCHS mortality Home Page. Implementation Tasks. Implementation of ICD is a complex, extended, and costly undertaking, one that began at NCHS about six years ago and will continue through the year 2001. Implementing ICD-10 is an extremely important task for NCHS and the States; it is an obligation to the U.S. and the international health community, and it will affect all public health agencies at the local, state, and Federal levels. Implementing ICD-10 has been carried out by teams at the NCHS Research Triangle Park (RTP) facility as well as in Hyattsville, Maryland; the tasks involve statisticians, medical advisors, mortality medical coders, and computer programmers. At RTP, the major efforts are focused on software development for mortality medical processing; on developing the NCHS instruction manuals for medical coding for ICD-10 that are distributed to all the states and foreign countries, and on training state medical coders and software managers to implement the programs. In Hyattsville, the major efforts have been focused on designing the benchmarking file that will be used to test software for processing cause-of-death data under ICD-10, developing the tabulation lists, developing the instruction manual for editing the mortality medical and demographic data, developing instruction manuals for State querying of death certificates that are not completed correctly or fully; designing and implementing a Comparability Study between ICD-9 and ICD-10, and redesigning all of the products -- published and electronic -- that will be used by NCHS to disseminate mortality data. International Leadership in Mortality Software Development. In implementing ICD- 10 for mortality, NCHS has experienced additional pressure. The software developed by NCHS for processing mortality information from death certificates has become, in effect, an international standard used by a growing number of countries, such as Canada, England, Scotland Australia, Brazil, and Spain. An international consortium of software users was just organized under the auspices of the International Collaborative Effort on Automating Mortality Statistics under the aegis of NCHS, and with the endorsement of WHO. In November 1996, NCHS sponsored an international conference of mainly developed countries to review the development of automated software for processing mortality data. Over 19 countries participated. Web addresses. To ensure that the public health community is well informed about the ICD and the ICD-CM, NCHS has established an ICD presence on the Internet at the following addresses: For mortality: http://www.cdc.gov/nchswww/about/major/dvs/icd10des.htm For morbidity: http://www.cdc.gov/nchswww/about/otheract/icd9/icd9hp2.htm IMPLEMENTATION ACTIVITIES To implement ICD-10 at the national and State levels, a number of activities will have to be undertaken. These include redesigning publications to incorporate ICD-10 tabulation lists, developing mortality edits, installing new software, training medical coders on new data entry and/or coding procedures, analyzing ICD-10 data with the aid of comparability ratios. Below are described these activities, which are spelled out in greater detail in NCHS instruction manuals listed in Appendix I. A number of these instruction manuals are available on the mortality Home Page; others will be placed on the mortality Home Page in the near future at http://www.cdc.gov/nchswww/about/major/dvs/mortdata.htm. Tabulation Lists and Redesigning Publications. Mortality data released for 1999 will have to reflect the new cause-of-death classification system under ICD-10, which includes alphanumeric codes, new titles, and reorganized cause-of-death categories. To address these needs, NCHS has worked with other Federal agencies and with the states in designing a total of eight new tabulation lists, the most detailed of which includes about 8,000 categories that are valid for underlying cause of death. These lists, which were distributed to all the states last winter, are published in NCHS Instruction Manual Part 9, ICD-10 Cause-of-Death Lists for Tabulating Mortality Statistics, Effective 1999. The list that will be most widely used by NCHS and other Federal agencies is the ICD-10 List of 113 Selected Causes of Death, which replaces the ICD-9 List of 72 Selected Causes of Death, which has been in use since 1979. This list is used to identify and rank the leading causes of death in the U.S. The instruction manual with the ICD-10 tabulation lists for mortality can be downloaded from the NCHS Website at the following address: http://www.cdc.gov/nchswww/about/major/dvs/im.htm Also to be available on the Website in the near future, and now available on request, is a list of each ICD-10 cause-of-death code valid for underlying cause of death and the title of the ICD-10 category. This list is called the "Each Cause" list. Developing Mortality Edits. An important task at the State level is converting mortality edits from ICD-9 to ICD-10. These edits can be modeled after the edits being developed by NCHS in Instruction Manual Part 11, Computer Edits for Mortality Data, Effective 1999. The edits include valid and invalid ICD-10 codes, cross edits, consistency edits for age and cause of death, consistency edits for sex and cause of death, rare causes, and edits for injuries (place of occurrence and activity). This manual, which will be available in early August 1998, will be distributed in hardcopy to the states and will be available on the NCHS Home Page, along with the other instruction manuals related to ICD-10 implementation. Comparability Ratios and Conversion Tables. NCHS will be providing a set of national comparability ratios to the States by the end of 1998, with a final set of national comparability ratios by fall of 2000. (See previous discussion of comparability). ICD-10 Computer Applications. All the software for processing mortality data has been converted from ICD-9 to ICD-10. Accordingly, medical coding effective with deaths occurring in 1999 will be according to ICD-10 specifications. The ICD-10 software is pc-based. It is not designed for mainframe applications. The software is also year 2000 compliant. Below are described computer applications for ICD-10, including data entry packages which are applicable to both MICAR and SuperMICAR, MICAR 100/200, ACME/TRANSAX, changes required in the Master Record Format, impact on mortality medical coders, and the plan for converting the ICD-10 software from DOS to WINDOWS95. The use of SuperMICAR or PC-MICAR does note affect the actions taken in implementing ICD-10. 1. Space Requirements for MICAR System PCMICAR and SuperMICAR have almost identical space requirements; therefore, only one set of numbers is reported. These figures are based upon a batch of 4,500 records. Individual PC's used for data entry and processing Software and Tables Data Entry (per system) 24.00 megabytes MICAR100\200 54.75 megabytes ACME\TRANSAX 2.00 megabytes Total for software 80.75 megabytes Data Files Open file during data entry 10.50 megabytes After MICAR100/200 processing additional 0.50 megabytes After ACME\TRANSAX processing additional 2.00 megabytes Total 13.00 megabytes Recommend 100 megabytes free space available before either system is loaded. These estimates to do not include space for creating subset files when the UTLMICAR system is used with a SuperMICAR file LAN Drive for backup file While file is in progress 0.5 megabytes 2.Data Entry Packages (SuperMICAR and PCMICAR) The programs released in October, 1998 will be designed to generate either ICD-9 or ICD-10 input records for MICAR100/200 processing. This means: a. Neither NCHS nor the States will have to maintain two separate data entry programs. States can immediately install the 1999 version received in October, 1998 and continue to enter the remaining 1998 records. The same software can be used to enter 1999 data. b. For States which import the demographic data before SuperMICAR entry is started, the input record format has been modified. This will require program modification by the State. See ICD-10 Super-MICAR Input Record Format, Appendix II (this option is not available for PC-MICAR; therefore, no record format is needed). c. There will be no retraining for data entry. The major difference in the current system and the ICD-10 system is the information collected on the header screen. For example, coders currently enter the last digit of the data year; with ICD-10, they will enter the complete 4-digit data year. Updated NCHS Instruction Manual, Part 2g will be provided to users. See Summary of Changes in ICD-10/ICD-9 Record Formats, Appendix IV. d. The decision to generate ICD-9 input or ICD-10 input will be made by the user. The system will not be designed to use the data year to determine which format to create. This will allow the States to continue with ICD-9 until they can make all modifications required to implement ICD-10. However, NCHS is not updating or correcting the ICD-9 system. Any errors in the ICD-9 system are being corrected in the ICD-10 system, but we do not have the nosological staff nor programmers to maintain both systems. 3. MICAR100\200: For ICD-10, this system will only be available on the PC. a. For States currently using the PC system: States must have enough space on their PC's to maintain both the ICD-9 and ICD-10 versions of this program. For estimating purposes, the software and tables require approximately 55 megabytes. b. For NCHS and States currently using the mainframe system: (1) Additional disk space: For software and tables: 55 megabytes For data files: 0.5 megabytes for each batch of 4,500 records (2) Procedures for processing the data files will need to be re-written. Depending upon the current job assignments, some staff may need training in the use of PC's and specific training in processing the data through the MICAR system. c. Multiple Cause Coders: Multiple cause coders will need to attend a one week ICD-9 to ICD-10 conversion class. This class entails some pre-classroom training and some post-classroom coding exercises. Each coder will only be required to re-qualify when there are changes in coding instructions based upon the differences from ICD-9 to ICD-10. (Experience with using the PC system's multiple cause entry screen will be included in the conversion training.) 4. ACME\TRANSAX: For ICD-10, this system will only be available on the PC. a. For states currently using the PC system: States will have to have enough space on their PC's to maintain both the ICD-9 and ICD-10 versions of this program. For estimating purposes, the software and tables require approximately 2 megabytes. b. For NCHS and state currently using the mainframe system: (1) Additional disk space: For software and tables: 2 megabytes For data files: 2 megabytes for each batch of 4,500 records (2) Procedures for processing the data files will need to be re-written. Depending upon the current job assignments, some staff may need training in the use of PC's and specific training in processing the data through the MICAR system. c. Underlying Cause Coders: Underlying cause coders will need to attend a 1 week ICD-9 to ICD-10 conversion class. This class entails some pre-classroom training. Neither post-classroom coding exercises nor a qualification process are applicable for this training. (Experience with using the PC system's ACME reject entry screen will be included in the conversion training.) 5.Changes in the Mortality Master Record Format In designing the ICD-10 system, every effort was made to minimize the changes which must be made in both NCHS and the State's mortality master record format. The ICD-10 TRANSAX Output Record Format is shown in Appendix III. a. The increased number of items required for NCHS control information can be ignored by both NCHS and the States when the master record is created. b. Manner of Death code: This is an additional item NCHS will include on the master record. States can chose to ignore this item. The reported manner of death will be reflected in the coded multiple cause data when appropriate. c. Underlying Cause Field: Increased from 4 character to 5 character. However, the 5th position will not be used immediately. It is recommended that states make provisions for capturing a 5-digit underlying cause at the time the mortality master record format is changed; but it is not required for ICD-10 coding. d. For States which include the entity-axis data in the master record, the field has been increased from 140 positions to 160 positions on the TRANSAX record. However, NCHS recommends that only the first 7 positions of each condition be included in a master record. Therefore, the current allocation of 140 positions does not require change. e. For States which include the record-axis data in the master record, no change in size is require. In both ICD-9 and ICD-10, the record-axis field is 100 characters. f. The program used to move the medical data items from the TRANSAX output record to the mortality master file must be modified to accept the new ICD-10 TRANSAX record format. To minimize the amount of programming changes required by ICD-10, the PC-ACME/TRANSAX system will include a provision for generating the basic medical items with ICD-10 codes in the ICD-9 TRANSAX record format. This record would include the data year, state code, certificate number, underlying cause (4 characters), place of injury, entity-axis codes (140 characters), and record-axis codes. g. Affect of ICD-10 alphanumeric codes: The underlying cause field must be defined as alphanumeric. However, the ICD-9 underlying cause field should have been defined as alphanumeric as well. In ICD-9, the underlying cause was either three or four digits with the fourth position being blank for valid 3-digit codes. Some states may have added a - (dash) in the 4th position when the code was only 3-digits which then allowed the field to be defined as numeric. The definition of the underlying cause must be changed from numeric to alphanumeric. The entity- axis and record-axis fields must also be defined as alphanumeric. If State procedures for ICD-9 codes included adding a - (dash) to all 3-digit codes for the purpose of declaring a numeric field, this practice must be discontinued for ICD- 10. 6.General Effect on Mortality Medical Coders a. For a period of time, medical coders will need to use both ICD-9 and ICD-10 coding material. In addition to the current ICD-9 material each coder is using, NCHS will provide the following ICD-10 documentation: In a 1" looseleaf binder: Instruction Manual Part 2a, NCHS Instructions for Classifying the Underlying Cause of Death, 1999 In a 2" looseleaf binders: Instruction Manual Part 2b, NCHS Instructions for Classifying the Multiple Causes of Death, 1999 Instruction Manual Part 2c, ICD-10 ACME Decision Tables for Classifying the Underlying Causes of Death, 1999 International Statistical Classification of Disease and Related Health Problems, Tenth Revision, Volume 1, Corrected and Updated by NCHS. International Statistical Classification of Disease and Related Health Problems, Tenth Revision, Volume 3, Corrected and Updated by NCHS. b. This use of alphanumeric codes will affect the productivity of coders who simultaneously code and data enter the codes in electronic form. With ICD-9, a coder is able to keep one hand on the keyboard and use the other hand to look through reference material. With alphanumeric codes, it will be much harder to use only one hand on the keyboard. Production standards will need to be adjusted. 7.Conversion of Mortality Medical Data System software from DOS to WINDOWS95 a. PC requirements PC: Pentium 120 MHZ OS: WINDOWS95 (including WINDOWS95 NT) RAM: 32 megabytes CD: CD-ROM (used during installation only) Hard Drive: 100 megabytes free space (minimum) b. The Windows version of the software will not be available until later in 1999. The system delivered in October, 1998 will be a DOS based system. c. Users should be familiar with WINDOWS applications before the updated system is implemented. Once the WINDOWS version is available, the DOS system will not be supported. ICD-10 Instruction Manuals. Instruction manuals for mortality medical data under ICD-10 have been revised. Some of these have already been published; others will be available by the end of 1998. The mortality medical manuals manuals are listed in Appendix I, with their expected distribution dates. In addition, the states are being provided with corrected and updated versions of ICD-10 volumes 1 and 3 in both hardcopy and on disk. These volumes have been specially prepared by NCHS for use in State mortality medical coding. The manuals may not be reproduced and distributed in any way, because of copyright considerations of the World Health Organization. ICD-10 Training. ICD-10 training will be carried out in two phases, one for experienced coders; these are caused "conversion" courses to be taught in late 1998 and early 1999. A subsequent set of courses will be offer in 1999 for new coders. In addition, it is hoped that a new course will be introduced for PC Managers of mortality data entry and coding systems. A new course is also being designed for statisticians that will cover key concepts in ICD-10 mortality medical coding and mortality analysis. The conversion courses are shown in Table 2. NCHS Contact Persons on ICD-10 Implementation are shown in Table 3. [figure 1 age-adjusted death rates for the 15 leading causes of death: US 1950-95 not available in ASCII copy] TABLE 1. ICD IMPLEMENTATION DATES IN U.S. ____________________________________________________________________________ Revision Years in Effect ____________________________________________________________________________ First (ICD-1) 1900-1909 Second ICD-2) 1910-20 Third (ICD-3) 1921-29 Fourth (ICD-4) 1930-38 Fifth (ICD-5) 1939-48 Sixth (ICD-6) 1949-57 Seventh (ICD-7) 1958-67 Eighth, Adapted (ICDA-8) 1968-78 Ninth (ICD-9) 1979-1998 Tenth(ICD-10) 1999- ___________________________________________________________________________ Table 2. Conversion Courses, ICD-10 Mortality Medical Coding Training 1. Multiple Cause Coding. 15 students, November 2-6, 1998. At Homewood Suites, Crabtree Valley, Raleigh, North Carolina 2. Underlying Cause Coding. 3 concurrent classes, 60 students. December 7-11, 1998. At Sheraton Four Points, Crabtree Valley, Raleigh, North Carolina 3. Multiple Cause Coding. 3 concurrent classes, 45 students, December 14-18, 1998. At Sheraton Four Points, Crabtree Valley, Raleigh, North Carolina 4. Underlying Cause, 2 concurrent classes, 40 students, January 11-15, 1999, Homewood Suites, Crabtree Valley, Raleigh, North Carolina 5. Underlying Cause, 20 students, February 8-12, 1999. Homewood Suites, Crabtree Valley, Raleigh, North Carolina TABLE 3. NCHS CONTACT PERSONS FOR ICD-10 IMPLEMENTATION Telephone E-mail Mortality Medical Software - Mary M. (Margie) Trotter 919-541-5566 mmt0@cdc.gov Nosological Inquiries - Julia Raynor 919-541-4408 jer3@cdc.gov Tanya Pitts 919-541-7872 twp1@cdc.gov Tabulation Lists and Data Presentation - Ken Kochanek 301-436-8884 kdk2@cdc.gov ext. 172 Conversion Tables - Ken Kochanek 301-436-8884 kdk2@cdc.gov ext. 172 Margie Trotter 919-541-5566 mmt0@cdc.gov Edits - Jeff Maurer 301-436-8884 jdm5@cdc.gov ext. 174 Querying - Donna Hoyert 301-436-8884 dlh7@cdc.gov ext. 168 Comparability Study - Bob Anderson 301-436-8884 rca7@cdc.gov ext. 179 Training - Tanya Pitts 919-541-7872 twp1@cdc.gov Julia Raynor 919-541-4408 jer3@cdc.gov Analysis of Mortality Medical Data - Mortality Statistics Branch staff 301-436-8884 ` APPENDIX I VITAL STATISTICS INSTRUCTION MANUALS FOR MORTALITY MEDICAL DATA (Expected Availability) Part 2a Instructions for Classifying the Underlying Cause of Death, 1999 (October 1998) Part 2b Instructions for Classifying the Multiple Causes of Death, 1999 (October 1998) Part 2c ICD-10 ACME Decision Tables for Classifying Underlying Causes of Death (October 1998) Part 2d NCHS Procedures for Mortality Medical Data System File Preparation and Maintenance (October 1998) Part 2f ICD-10 TRANSAX Disease Reference Tables for Classifying Multiple Causes of Death (October 1999) Part 2g Data Entry Instructions for the Mortality Medical Indexing Classification and Retrieval System (MICAR), 1999 (October 1998) Part 9 ICD-10 Underlying Cause-of-Death Lists for Tabulating Mortality Statistics, Effective 1999 (October 1997) Part 11 Computer Edits for Mortality Data, Effective 1999 (August 1998) Part 20 Cause-of-Death Querying, 1999 (October 1998) Appendix II ICD-10 Super-MICAR Input Record Format [see Wordperfect version or pdf at http://www.cdc.gov/nchswww/about/major/dvs/medsof.htm] Appendix III ICD-10 TRANSAX Output Record Format [see Wordperfect version or pdf at http://www.cdc.gov/nchswww/about/major/dvs/medsof.htm] APPENDIX IV March 19, 1998 Summary of Changes in ICD-10/ICD-9 Record Formats A. All record formats: 1. Data year Increased from single digit to 4 digit field 2. Lot Number Increased from 3 digit field to 4 digit field 3. Section Number Additional NCHS control information; 1 digit 4. Shipment Number Increased from 2 digit field to 3 digit field Changed to accept alpha-numeric values 5. Receipt Date Additional NCHS control information; 6 digits 6. Version control Additional NCHS control information SuperMICAR\PC-MICAR: 4 positions ACME input: 8 positions ACME output\TRANSAX: 12 positions 7. State Specific Data Increased from 17 characters to 30 characters B. Super-MICAR 1. Dates of Injury/Surgery: Removed requirement for using "/" between month and day and day and year 2. Place of Injury Added field to permit importing coded data C. ACME\TRANSAX record formats: 1. Manner of Death Additional data item for ACME and TRANSAX; 1 position 2. Underlying Cause Increased from 4 characters to 5 characters. 5th position will be blank. This position is reserved for use if 5-digit codes are implemented. 3. ACME Input Data Increased from 79 characters to 120 characters 4. Entity-Axis Data: Increased length of each condition from 7 positions to 8 positions. NOTE: The value coded in the 8th position should not be moved to the final mortality data record.