Government

Quick access to information based on government's structure



Ministry of Health

Quarterly Digest
Volume 10 - Number 4 July 2001

  • Preface

  • Map: B.C. Local Health Areas

  • British Columbia: Local Health Areas (LHA) within Health Regions

  • Vital Event Statistics - October 1, 2000 - December 31, 2000 and Year-End
    (Population, Livebirth, Death, Marriage, Stillbirth, Infant Deaths)

  • Selected Birth Statistics - October 1, 2000 - December 31, 2000 and Year-End
    (Low Birthweight, Preterm, Teenage Mother, Elderly Gravida, Cesarean Section)

  • External Causes of Death - October 1, 2000 - December 31, 2000 and Year-End
    (Accidents - [Motor Vehicle Accidents, Poisoning, Falls, Burns/Fire, Drowning, Other], Suicide, Homicide, Other External Causes)

  • Neoplasm Deaths - October 1, 2000 - December 31, 2000 and Year-End
    (Lung, Female Breast, Colorectal, Other G.I., Female Reproductive, Prostate, Blood/Lymph, Other Malignancy, Nonmalignant and Unspecified)

  • Heart Disease Deaths - October 1, 2000 - December 31, 2000 and Year-End
    (Rheumatic/Valvular, Hypertension, Ischemic, Conductive & Dysrhythmic, Heart Failure, Congenital, Other)

  • Respiratory Disease Death Statistics - October 1, 2000 - December 31, 2000 and Year-End
    (Emphysema, COPD, Pneumonia, Influenza, Asthma, Lung Disease from External Agents, Pulmonary Fibrosis, Other Respiratory)

  • Other Selected Death Statistics - October 1, 2000 - December 31, 2000 and Year-End
    (Diabetes, Alcohol-Related, AIDS, Other Infectious Disease, Cerebral and Other Vascular, Liver Disease, Amyotrophic Lateral Sclerosis and Multiple Sclerosis, Alzheimer's Disease, Parkinson's Disease)

  • Summary Article:
    The impact of ICD-10 (in year 2000) on identifying, classifying, coding and UCOD selecting neoplasm(s) on death records
    by J. Macdonald & R. Armour

Preface

This "Quarterly's" year 2000 standard tables of vital event data are for the fourth quarter and year-end.

The death data provided in this and previous year 2000 issues are the first in Canada to have been derived from the newly introduced tenth revision of the International Classification of Diseases (ICD-10). ICD code groupings used in this publication for standard cause of death tables have been "translated" from ICD-9 to ICD-10 and the glossary provides complete new code listings and a note of new inclusions/exclusions that could affect comparative counts. The greatest impact on statistical consistency is not due to the additional detail (more than 2000 new codes) afforded by ICD-10. Of greater significance is some major changes in the rules used to select the single underlying cause of death (UCOD) - most notably as relates to cancers and pneumonias. This Agency, concerned about the statistical impact, medical validity, and the new rules' interpretation of certifier intent, has initiated an extensive daily editing of automated systems output that has resulted in production of two data sets - the primary (edited) for more accurate and consistent use in BC. In addition, the Agency has constructed a more precise automated ICD-9 to ICD-10 translation that will enable our provision of historically accurate data. Included in this process has been the recovery of many thousands of death records back to 1986 that contain codes that cannot "translate" and/or would provide additional desired ICD-10 detail. These identified records are being recoded (currently at 95% completion). This entire process will provide current accuracy with historical consistency as well as recovered detail in previous years with the view to making mortality data in BC the best available anywhere.

As previously noted, there are several areas of statistical discontinuities in cause-of-death statistics resulting from the implementation of ICD-10 based classification and selection rule changes. This issue's feature article is an in depth discussion of the impact of these changes on cancer statistics. It provides information regarding "new" cancer codes and examples of new coding application and selection rules. The accuracy of code rules in reflecting the intent of death record certifiers was examined by a survey of more than 100 physicians. This issue paper provides the result of this survey and explains the steps taken by this Agency (including an editing process) to address these data quality concerns. Also provided is a comparison of edited and unedited data.

Due to the fact that Vital Statistics Agency files are continually being updated, totals compiled by addition of the annual quarters will not correspond exactly to year-to-date and year-end figures. For the same reason, depending on the date the data are extracted, there may be differences in numbers presented in the year-end Quarterly Digest and those reported in the 2000 Annual Report. Therefore, the numbers provided in this publication should be considered provisional. Finally, the usual cautions regarding random fluctuations in values, particularly those involving small numbers, must be noted.

As always, requests for changes and suggestions for articles continue to be welcome. Your support and input into this publication is greatly appreciated.

R.J. Danderfer Soo-Hong Uh
CEO/Director Manager
British Columbia Information and Resource
Vital Statistics Agency Management Branch
  Vital Statistics Agency

[Return to Table of Contents]

British Columbia
Local Health Areas


image


[Return to Table of Contents]

British Columbia:
Local Health Areas (LHA)
within Health Regions

01 East Kootenay
LHA

01 Fernie
02 Cranbrook
03 Kimberley
04 Windermere
05 Creston
18 Golden

02 West Kootenay-Boundary
LHA

06/07 Kootenay Lake/Nelson
09 Castlegar
10 Arrow Lakes
11 Trail
12/13 Grand Forks/Kettle Valley

03 North Okanagan
LHA

19 Revelstoke
20 Salmon Arm
21 Armstrong-Spallumcheen
22 Vernon
78 Enderby

04 South Okanagan-Similkameen
LHA

14 Southern Okanagan
15 Penticton
16 Keremeos
17 Princeton
23 Central Okanagan
77 Summerland

05 Thompson
LHA

24 Kamloops
26 North Thompson
29 Lillooet
30 South Cariboo
31 Merritt

06 Fraser Valley
LHA

32 Hope
33 Chilliwack
34 Abbotsford
75 Mission
76 Aggassiz-Harrison

07 South Fraser Valley
LHA

35 Langley
36 Surrey
37 Delta

08 Simon Fraser
LHA

40 New Westminster
42 Maple Ridge
43 Coquitlam

09 Coast Garibaldi
LHA

46 Sunshine Coast
47 Powell River
48 Howe Sound

10 Central Vancouver Island
LHA

65 Cowichan
66 Lake Cowichan
67 Ladysmith
68 Nanaimo
69 Qualicum
70 Alberni

11 Upper Island/Central Coast
LHA

71 Courtenay
72/84 Campbell River/
Vancouver Island West
83 Central Coast
85 Vancouver Island North

12 Cariboo
LHA

25 100 Mile House
27 Cariboo-Chilcotin
28 Quesnel
49 Bella Coola Valley

13 North West
LHA

50 Queen Charlotte
51 Snow Country
52 Prince Rupert
53 Upper Skeena
54 Smithers
80 Kitimat
87/94 Stikine/Telegraph Creek
88 Terrace
92 Nisga'a

14 Peace Liard
LHA

59 Peace River South
60 Peace River North
81 Fort Nelson

15 Northern Interior
LHA

55/93 Burns Lake/Eutsuk
56 Nechako
57 Prince George

16 Vancouver
LHA

161 Vancouver City Centre
162 Vancouver Downtown East Side
163 Vancouver North East
164 Vancouver West Side
165 Vancouver Midtown
166 Vancouver South
Unknown Vancouver

17 Burnaby
LHA

41 Burnaby

18 North Shore
LHA

44 North Vancouver
45 West Vancouver-Bowen Island

19 Richmond
LHA

38 Richmond

20 Capital
LHA

61 Greater Victoria
62 Sooke
63 Saanich
64 Gulf Islands


[Return to Table of Contents]

The Impact of ICD-10 (in year 2000) on identifying, classifying, coding and UCOD selecting neoplasm(s) on death records

by J. Macdonald & R. Armour
Since 1979, all disease/circumstance entities on death records have been classified using the 9th revision of the International Classification of Diseases (ICD-9). Beginning in the year 2000, the BC Vital Statistics Agency (BCVSA), like other Canadian provincial jurisdictions dealing with mortality data, began using ICD-10. Implementation of the new classification system included replacement with ICD-10 based software - MICAR (in use since 1995) for automated assignment of codes and ACME (in use since 1990) for application of the rules to select the single underlying cause of death (UCOD).

Because of classification differences in ICD-10 as well as in the rules that govern what and how cancer codes are applied and selected, there are several changes. If implemented without modification, these will have a major impact on consistency and accuracy of cancer mortality data - both UCOD and multiple codes. The various issues discussed here were initially identified by Vital Statistics coding and research staff following ICD-10 Multiple Code Training in Ottawa. These were subsequently confirmed through examination of National Center for Health Statistics (NCHS) literature and manuals, W.H.O. rules and ICD-10 codes, by preliminary testing of software and via a survey of physicians to determine if application of the new rules accurately reflected certifier intent. The BC Cancer Agency was apprised of BCVSA concerns through early issue paper drafts and was involved in discussions about the course of action taken by this Agency.

Differences in ICD-10 neoplasm codes and/or how they are applied:

  1. Mesothelioma: (C450 - C459) now has a specific code group. Previously, if the site was identified (e.g. pleura, peritoneum) this cancer was counted with others of the specified site; if not site-specific this cancer was included with primary site unknown (1991).

  2. Kaposi's sarcoma: (C460 - C469) now has a specific code group. Previously, this went to specified/unspecified site of skin (1730-1739). Multiple code analysis for Kaposi's is adversely affected by the fact that, when due to AIDS, the code for Kaposi's is apparently dropped for a "Kaposi's resulting from HIV disease" specific code (B210).

  3. HIV and other cancers: As with Kaposi's, codes for Burkitt's lymphoma, non-Hodgkins lymphoma, and unspecified cancer, are dropped in favor of equivalent "resulting from HIV disease" codes. What is more problematic, is that other HIV/cancer codes are less specific (e.g. HIV resulting in other malignant neoplasms [B218] or in multiple malignant neoplasms [B217]) resulting in the loss of non-UCOD, cancer specific, information.

  4. Breast cancer: (C500 - C509) ICD-10 codes for breast cancer are not gender specific.

  5. There is no differentiation between cancers with primary site unknown (ICD-9, 1991) and generalized cancer or carcinomatosis (ICD-9, 1990). These are all coded to C80 in ICD-10.

  6. Some conditions/diseases were not previously considered to be neoplastic. Additions to the neoplasms category include - Waldenstrom's macroglobulinemia, alpha heavy chain disease, Franklin's disease (gamma heavy chain), Mediterranean disease, immunoproliferative disease NOS, all go to C88_ (malignant immunoproliferative diseases). Light chain disease (American addendum) is now included with multiple myeloma. Myelofibrosis and panmyelosis, if acute are coded as leukemias (C944 & C955). There are also apparent changes involving conditions being reclassified based on changing views of cancer activity. For example, a cancer previously considered to be of "uncertain behaviour" might now be considered to be malignant, or conditions previously classified in other chapters such as monoclonol gammopathy or chronic myeloproliferative disease are now being classified as neoplasms of uncertain behaviour.

  7. Infrequently, "unknown etiology" or "unknown source", whether referring to cancer or not, may not always be appropriately assigned by the MICAR software resulting in an inappropriate C80 code or other cancer code where there is no cancer.

  8. "Malignant neoplasms of independent (primary) multiple sites" (C97): The rules for applying this ICD-10 code which had no equivalent in ICD-9 are noted below (2). The code C97 will be applied as UCOD on records with two or more rule-determined primary sites.

  9. Melanoma: Morphology is disregarded for certain specified sites of malignant melanoma that have been classified in ICD-10 to cancer of the site rather than within melanoma subcategories. These include malignant melanomas of the eye, anus, and male and female external genitalia.

    The effect of the application of NCHS/WHO new rules for classification of ICD-10 neoplasm categories:

    In addition to the impact on neoplasm statistics as a result of the move to ICD-10 in the year 2000, many of which are noted above, there are significant differences (multiple code and UCOD) due to classification rule changes. These rules are provided in NCHS Instruction Manual Part 2b, (reference for all rules noted below). The ICD-10 rules for cancer coding disregard "due to" positioning and duration, allow secondary cancers as underlying cause of death (UCOD), make no distinction among multiple possible primary sites and attempt to assume the intent of certifiers based upon a list of "common" secondary cancer sites and/or combinations of anatomical systems. The Instruction Manual has been subject to amendment since its initial use in January 2000 therefore page reference numbers may no longer correspond to those provided. Coding and selection rules for neoplasms have had only minor changes that do not alter the original intent or approach. Those of greatest concern are noted below.

    1. Neoplasms stated to be secondary (p. 128): If secondary cancer is the only cancer noted on the death record and there is no morphology to suggest a primary, no code for "unknown primary" is to be added. This will result, for the first time ever, in deaths with UCOD of secondary cancer.

    2. Independent (primary) sites (p. 132): If two or more sites are mentioned in Part I of the death record without indication that any are primary or secondary, and none are on the list of "common sites of metastases" (see 3 below), each site is to be coded as a primary site. A neoplasm in a "due to" position is not to be used to determine secondary and primary. Likewise, information describing the interval between the onset of the cancers and the death are not to be considered. In the case of two valid primary sites, there are no criteria (medical or positional) for selecting one of the cancers as UCOD.

    EXAMPLE:
    1. (a) cancer of esophagus [3 months duration]
      Due to, or as a consequence of
      (b) cancer of the stomach [8 months duration]

    Result by rule: coded to primary malignant neoplasm of each site mentioned (C159 and C169) since neither is listed (see 3. below) as a common site of metastases and the underlying cause of death is C97 (multiple primary site neoplasms). Previously, based on position and duration, the esophageal cancer would have been coded as secondary to the underlying cause of death of primary stomach cancer.

    1. Metastases and multiple sites (pp. 135-140): WHO/NCHS have identified certain sites as more commonly secondary than others. These "common sites of metastases" include bone, brain, central nervous system, diaphragm, heart, liver, lung not including bronchus [with qualification], lymph nodes, mediastinum, meninges, peritoneum, pleura, retroperitoneum, spinal cord, and ill-defined sites classifiable to C76.

      Because the lung is considered to be a common site for both metastases and primary malignant neoplasms, there are special instructions within the rules. Cancer of the lung (excluding of bronchus or bronchogenic which is always primary) is coded as secondary if accompanied by site(s) not on the list and as primary when noted with sites on the list.

      With the exception of lymph nodes that are always assumed to be secondary, if any one of these "common secondary" sites is the only cancer mentioned, it is coded as primary. Two unqualified sites not on the list are both coded as primary while two on the list (excluding lung) are both coded as secondary and no "primary site unknown" code is generated.

      An exception to coding the "common" secondary sites (other than lung as secondary), is when a malignant neoplasm of lymphatic, hematopoetic, or related tissue (C81-C96) is noted in Part II in which case, the "common" secondary site is coded as primary (p. 139). This means, for example, that cancer of the brain is secondary if accompanied by lymphoma in Part I and primary if the lymphoma is noted in Part II.

    EXAMPLE:
    1. a) cancer of abdomen [3 months duration]
      Due to, or as a consequence of
      b) cancer of the liver [8 months duration]

    Result by rule: Both sites are coded to secondary neoplasms (abdomen is one of the "ill-defined" sites in C76_ category included in "the list") - C798 and C787 respectively. Because both secondary sites are in Part I of the Death Certification, the UCOD is C80 (multiple secondary sites). If the sites had been divided between Part I and Part II, the secondary site in Part I becomes the UCOD. A code for "primary site unknown" of C80 would only appear as UCOD if so noted on the record. Previously, primary liver cancer would have been the UCOD with metastases in the abdomen.

    1. Metastatic ..., metastatic of ... (p.140): As noted in Instruction Manual 2b, the adjective "metastatic" could sometimes mean a secondary neoplasm from a primary elsewhere and sometimes denote a primary that has given rise to metastases. NCHS/WHO have attempted to establish the intent of the physician who uses this adjective on death certifications by applying the rules governing the use of the "list of common sites" and disregarding the term "metastatic. " That is, unless specified as primary or secondary or morphologically site-defined as not C80, all cancers are to be considered in terms of the list of "common" secondary sites, including the exceptions for lung cancer. This means, for example, a single entry of "cancer of brain" would be coded C719 (primary) and "metastatic cancer of brain" would be C793 (secondary). Exception is also to be made for metastatic mesothelioma or Kaposi's sarcoma (p. 147). Where the site is not specifically indexed, it is coded as mesothelioma or Kaposi's of "specified site NEC" if the reported site is not on the "common secondary" list. If the site is not indexed and is a "common" secondary site, it is double coded with unspecified site of mesothelioma/Kaposi's (first entry) and secondary cancer of the site. This assumption about the intent of certifying physicians means that, for example metastatic Kaposi's of brain (on the "list") would be coded C469 (unspecified site of Kaposi's) & C793 (secondary of brain) while Kaposi's sarcoma of brain would be C467.

    2. More than one malignant neoplasm described as metastatic (pp.149-153): Differentiation of which "metastatic" sites are secondary and which are primary when more than one are described on a single death record, becomes even more problematic (and more arbitrary). In this case, not only must the list of "common" secondary sites be considered but the anatomical system sites (list, p.149), and the morphological types of the described "metastatic" cancers must also be considered. It should also be noted that the breast does not apparently belong in any anatomical system. Rules governing this situation are:

    • If two or more "metastatic" sites are not on the "common" secondary list and are in different anatomical systems, each is coded as secondary. If sites are in the same anatomical system, each is coded as primary.

    EXAMPLE:
      A:
    1. a) metastatic carcinoma of stomach
      Due to, or as a consequence of
      b) metastatic carcinoma of breast

      and

      B:
    1. a) metastatic pancreatic carcinoma
      Due to, or as a consequence of
      b) metastatic carcinoma of stomach

    By the ICD-10 rules, cancer of the stomach and of the breast (Example A) would each be coded as secondaries (C788 and C798 respectively) since stomach and breast are in two different anatomical systems. The UCOD would be C80 (carcinomatosis/unknown primary). Formerly, the UCOD would have been gender appropriate breast cancer. In Example B, cancers of pancreas and stomach would each be coded as primary (C259 and C169) since they are of the same organ system, the same morphologically, and neither is on the list of "common sites of metastases". The UCOD would be C97 (multiple primary site cancer) where formerly it would have been stomach cancer.

    • If two or more "metastatic" morphological types are described, code each to primary.

    • If the record notes a morphology implying site and an independent anatomical system site, both described as "metastatic", code secondary of each site. In the example provided below, the term "metastatic" is assumed to apply to all sites following that are on the same line or joined by the word "and".

    EXAMPLE:
    1. a) metastatic colonic and
      Due to, or as a consequence of
      b) renal cell carcinoma

    By ICD-10 based rule, both sites are coded as secondary (C785 and C790) with a UCOD of C80. Formerly, both colon and kidney would have been coded as independent primaries and UCOD would have been cancer of colon.

    • If there is a combination of "metastatic" site(s) and one site that is unqualified or not on the list of "common" secondaries, code to primary neoplasm of the site that is not described as "metastatic", irrespective of the order of entry or whether it is in Part I or Part II of the record. Code all other "metastatic" cancers as secondary.

    • If all sites are qualified as metastatic and/or on the list of "common" secondary sites including lung, code to secondary malignant neoplasm of all reported sites (pp. 152-153). For example, metastatic stomach cancer (line a) "due to" lung cancer (line b) would result in secondary site codes for both as lung is on the common sites of metastases list and stomach is described as metastatic.

    1. Effect of new pneumonia coding rules on cancer stats: when pneumonia is the last entry in Part 1 of the Death Certification, due to a change in the ICD-10 direct sequel rule for pneumonia (and influenza), many more conditions can be taken from Part II as direct consequences of pneumonia. Part II of a Death Certification is to record "Other significant conditions contributing to the death but not resulting in the underlying cause given in Part I. " ICD-9 based modification rules did require UCOD preference for some conditions noted in Part II under certain circumstances (e.g. AIDS, COPD sequenced to pneumonia). Especially as far as pneumonia is concerned, new rules (even after slight modification since initial release) result in a marked decline in pneumonia deaths. This selection of a UCOD other than pneumonia makes no consideration of qualifiers, or duration or order of Part II conditions (usually a list of chronic conditions) and disregards the opinion of the certifier. Thus, deaths counted as due to pneumonia in ICD-9 are now attributed to many other causes. Some of the most prominent causes include malignant neoplasms, cardiovascular diseases, Alzheimer's, dementia, uncomplicated diabetes, malnutrition, chronic lower respiratory diseases, and septicemia, and could cause an increase in these conditions being counted as UCOD.
    Intent of physician certifiers and BCVSA survey

    Many of the rules noted above were designed to universally assume the intent of certifiers in distinguishing primary versus secondary sites and in the use of the term "metastatic" as an adjective. In order to ascertain how well these rules reflected the intent of BC physicians, Vital Statistics initiated a survey of current certifiers. Except for the criteria that at least one cancer on a record was described as "metastatic" OR more than one unqualified site was noted, medical certifications included single and multiple cancer entries, and a variety of sites including those on the "common secondaries" list. No physician was queried more than once in order to obtain as diverse a sample as possible. In all, between January 2000 and February 2001, 115 physicians who had certified deaths where cancer was involved, were contacted in writing. A letter explaining the purpose of the survey, and a copy of the relevant death certification were either faxed or mailed to the physician. Certifiers were asked to verify the intent of their documentation in order to clarify discrepancies between "system" and "coder" interpretation of cancer-related mortality certification. Of the 115 physicians surveyed, 111 responded.

    The results of this survey indicated that within the survey sample, automated application of the coding rules resulted in a 75% error rate in the automated coding of cancer-related deaths. Of the 111 responses, 27 were coded correctly, and 84 incorrectly.

    The errors in coding fell generally into four categories:

    1. Coding of sites included in "common sites of metastases" list:

    EXAMPLE:
    1. a) cardiac failure [days]
      Due to, or as a consequence of
      b) lymphangitic carcinomatosis [weeks]
      Due to, or as consequence of
      c) adenocarcinoma of lungs [2 months]

    Automated coding result: lymphangitic carcinomatosis coded to C499 (primary connective tissue malignant neoplasm); adenocarcinoma of lungs coded to secondary lung cancer, with UCOD assignment of connective tissue malignant neoplasm, unspecified. It appears that the UCOD selection was based solely on the fact that lung cancer is included in the list of "common sites of metastases", with placement on the certificate and duration not having been taken into consideration. The most often correctly coded "common site of metastases" was metastatic liver cancer. Of the 27 correctly coded records within the survey, 12 involved secondary liver cancers.

    1. Interpretation of the term "metastatic" (i.e. could be interpreted either as metastatic to, and therefore a secondary site, or as metastatic from and therefore a primary site):

    EXAMPLE:
    1. a) respiratory failure
      Due to, or as a consequence of
      b) metastatic cancer of grade III papillary ovarian cancer and inflammatory breast cancer

    Automated coding result: both sites were coded to secondary cancers, with UCOD selection of unknown primary cancer (C80).

    1. Duration inconsistencies:

    EXAMPLE:
    1. a) carcinoma of pancreas [6 month duration]
      Due to, or as a consequence of
      b) carcinoma of lung [1 year duration]

    Automated coding result: pancreatic carcinoma was coded as primary site and lung carcinoma was coded as secondary site, with UCOD selection of carcinoma of pancreas. Despite the fact that the lung cancer was the site of longer duration, it was coded as being secondary spread from the pancreatic cancer, durations having been ignored.

    1. Morphology inconsistencies:

    EXAMPLE:
    1. metastatic ovarian adenocarcinoma

    2. papillary carcinoma of thyroid

    Automated coding result: secondary ovarian adenocarcinoma from primary papillary thyroid cancer, with UCOD assignment of thyroid cancer. Two distinct morphologies were apparently erroneously linked.

    Summary effect on cancer codes/statistics due to ICD-10 code classification and UCOD selection rule changes

    • Slight decline in cancers of pleura, peritoneum and unspecified primary due to reassignment of mesothelioma. This is unavoidable and would require recoding of historical events to reconcile over time.

    • Slight increase in immunoproliferative malignancy, multiple myeloma, and leukemia (also unavoidable - as above).

    • Counts of Kaposi's sarcoma and other cancers resulting from HIV disease used in multiple code analysis will have to be identified through HIV codes and included with the appropriate non-HIV cancer code. In some cases, detail relating to the cancer will be lost.

    • Breast cancer counts derived by code will include a few male cancers.

    • Not all melanoma will be counted within the morphological category.

    • Carcinomatosis and multiple metastases (used in multiple code analyses) will no longer be available.

    • Every UCOD code C97 is one less specified primary site cancer counted for statistical purposes - affects all records with more than one primary listed.

    • There will be an apparent statistical decline in several primary cancers, especially of brain, liver, lung and peritoneum.

    • In BC, physicians have been encouraged both in Instruction Handbooks and directly, to provide multiple site information (primary and secondary). It has also been our experience that use of the term "metastatic" as an adjective modifier of a primary cancer is a common practice here and not generally used in agreement with WHO/NCHS. There also appears to be a discrepancy in how commonly certain sites are secondary. Initial evaluation indicates that of all records containing the adjective "metastatic" or which provide more than one unqualified site, within B.C. the majority will not receive codes that reflect the intent of the certifier and will generally result in undercounting of a cross section of primary site cancers.

    Year 2000 counts by 3 digit code of deaths due to Malignant Neoplasm, C00-C97
    Comparison of edited and unedited results
    Year 2000 Counts by 3 digit code Note: Counts may not correspond exactly to other published year 2000 data due to different run dates.
    In addition, there may be a slight discrepancy in run time between edited and unedited file creation.

    The British Columbia Vital Statistics Agency approach

    BC Vital Statistics has written extensive edits to identify these and other (e.g. pneumonia as non UCOD) concerns. Among edits for more than 90 other "second look" issues there is a set designed to flag for review, many of the cancer deaths (for all codes, not just UCOD) discussed here. The edit lists are produced from ACME output. All unedited loads are stored in our system and all changes are made directly into the BCVSA operating system (VISION). This edit process occurs daily while original records are available for review.

    In addition to a routine edit for gender appropriateness, the "cancer edit" includes all "primary unknown/carcinomatosis" - C80 to confirm and to differentiate and add VS code C801 for carcinomatosis; C97 "multiple primary sites" for UCOD reassignment; any record with any secondary cancer codes are subject to review. Records missing codes for "unknown primary" have that code added. Also, all deaths with any code for cancer due to HIV disease are identified and double coded for improved multiple code use. For possible code reassignment to melanoma of eye (VS code C331), anus (C435), and male or female external genitalia (C439), ICD-10 codes C699, C210, C519, C600, C609, C632 are included in the edit process. Direct contact with certifying physicians continues where clarification is required.

    In order to best meet research needs both nationally, provincially, and regionally, the Vital Statistics Agency maintains two databases - one with stored unedited output and one with corrected/edited codes stored against the registered death event. This means that statistics produced in BC for BC more accurately reflects certifiers' intent and past practices that are more comparable to ICD-9 years but will not correspond to national statistics. In BC, users will have an option. Statistics and research conducted by this Agency and by most of our users will likely be derived from edited data.

    The table above, provides counts of year 2000 deaths due to malignant neoplasms for unedited and edited data. As demonstrated, edited data contains gains in the majority of site specific causes - most notably for cancers of the lung, colon, and liver and for melanoma and other neoplasms of skin. Much of the gain is reflected in the reassignment of codes from secondary cancers and neoplasms of multiple primary sites. In addition, edits for issues other than cancer have resulted in reassignment of UCOD to cancer. For example an edit for aspiration/pneumonia could result in reassignment of the UCOD to the debilitating condition (e.g. cancer) or circumstance that led to the aspiration. For this reason there were 54 more cancer deaths reported in the edited data.

    In addition to a daily edit and review of software output, the BC Vital Statistics Agency has undertaken development of rules and procedures that will allow the production of consistent causes of death statistics that span ICD-9 and ICD-10 (back to 1986). This process involves four strategies. Three of these strategies involve mapping; "dummy down" of ICD-10; mapping of single ICD-9 codes to single ICD-10 codes where a one to one relationship could be established; mapping of combinations of ICD-9 codes (using multiple codes/factors) to a single or preferred ("dummied down") ICD-10. These three strategies resulted in the creation of over 11,000 mapping rules. When the disease(s) in an ICD-9 code could not be sufficiently differentiated to assign to ICD-10 (e.g. mesotheliomas of various sites) or where improved detail is desirable (e.g. identification of hepatitis C) from historical data, original records are being retrieved and recoded. ICD-9 codes are not actually replaced as ICD-10 recodes of historical data (1986 to 1999) are stored separately in the operational system. The separate storage of ICD-10 recodes allows for flexibility, documentation and verification of recode decisions. The recode process involves many thousands of records and is at approximately 95% completion at this time.

    Some of the issues around ICD-10 based cancer mortality identified here have also been noted in several other provinces and countries and there is general concern about the integrity and accuracy of cancer statistics. Use of provincial Vital Statistics data is more extensive in BC than in any other jurisdiction and a need to find a compromise among the needs for consistency, comparability and accuracy while still meeting national requirements is the rationale behind the approach we have taken.

    Glossary

    [Return to Table of Contents]

    Alcohol-Related:
    This category includes all deaths considered as being directly or indirectly related to alcohol as indicated by inclusion by the certifier of selected alcohol identifying conditions anywhere on the death record (including "lifestyle" field). It should be noted that where alcohol is an indirect cause of death (i.e. not UCOD) and the direct underlying cause of death falls within one of our selected causes (e.g. motor vehicle accidents), then this death may be counted in both columns. That is, not all of "alcohol-related" are exclusive. This category includes ICD-10 codes: F100-F109, K700-K709, O993, P043, O354, Q860, G312, G621, G721, I426, K292, K860, X45, X65, Y14, T510-T512, T519. Note: now excludes acute pancreatitis, and cirrhosis not specifically identified as alcohol induced.

    Assignment of Health Region:
    Cases are assigned to Health Regions by the aggregation of appropriate LHAs.

    Assignment of Local Health Area (LHA):
    Allocation of LHA, in the case of births and deaths is based upon the usual residence (by postal code) of the mother and deceased respectively. Marriages are assigned to LHAs according to the place of event. Community name, is used in the absence of postal code.

    Elderly Gravida:
    Any mother who was 35 years of age or older at the time of delivery of a live born infant.

    External Causes of Death:
    Deaths due to environmental events, circumstances and conditions as the cause of injury, poisoning, and other adverse effects. Broad categories include accidents, suicide, medical or abnormal reactions (considered accidents), homicide, legal intervention, misadventures (counted as accident) and injury from war operations. Standard "Quarterly" tables under this heading include deaths due to accidents, suicide, homicide, and other. Accidents are subdivided by the following categories; motor vehicle accidents (MVA) (ICD-10 V020-V049, V090-V092, V093, V120-V149, V190-V196, V200-V249, V260-V349, V360-V449, V460-V549, V560-V649, V660-V749, V760-V799, V803-V805, V820-V821, V823-V839, V840-V875, V877-V8999, Y850), poisoning (X40-X49), falls (W00-W19), burns/fire (X00-X19), drowning (V900-V909, V920-V929, W65-W74), other accidents (V010-V019, V050-V069, V091, V099, V100-V119, V150-V189, V198-V199, V250-V259, V350-V359, V450-V459, V550-V559, V650-V659, V750-V759, V800-V802, V806-V819, V822, V876, V910-V919, V930-V949, V950-V978, V98-V99, W20-W64, W75-W99, X20-X39, X50-X59, Y40-Y849, Y859, Y86, Y880-Y883). Suicide ICD-10 codes are X60-X84, Y870; homicide (X85-Y09, Y871); "other [external]" consists of events of undetermined intent, legal interventions, and operations of war (Y10-Y369, Y890-Y899).

    Note: the late effects of accidental poisoning, falls, and burns/fire are no longer identified separately for inclusion in these categories and are now part of "other accidents"). Trains are now considered motor vehicles in ICD-10 but for consistency, have been excluded from MVA counts to still be considered as "other transport".

    Heart Disease:
    Tables under this heading include deaths due to:

    • rheumatic/valvular: (I050-I099, I340-I38)
    • hypertension/hypertensive: (I10-I159)
    • ischemic: (I200-I259) (Note: now includes cardiomyopathy specified as ischemic)
    • conductive & dysrythmic: (I440-I499)
    • heart failure: (I500-I509)
    • congenital: (Q200-Q249)
    • other: pulmonary (I260-I289), inflammatory (I300-I339, I400-I409), cardiomyopathy (I420-I429)(Note: now excludes ischemic),
      other ill-defined or unspecified heart disease (I510-I519)(includes myocardial degeneration)

    ICD-9:
    The ninth revision of International Classification of Diseases, World Health Organization, Geneva, 1977. An internationally used system of approximately 12,000 four (and some three) digit numbers representing a system of categories to which morbid entities are assigned according to an established criteria. ICD provides a common basis of disease and injury classification that facilitates storage, retrieval, and tabulation of statistical data.

    ICD-10:
    The tenth revision of International Classification of Diseases and Related Health Problems, World Health Organization, 1992. In use beginning with year 2000, update of ICD-9 revised with alpha-numeric system and increased code detail (approximately 18,000). The BC Vital Statistics Agency and all their provincial counterparts utilize an ICD-10 that has been modified by the National Center for Health Statistics (NCHS) for use in the classification and analysis of medical mortality data in the United States (October, 1998).

    Infant Deaths:
    Deaths of children under one year of age.

    Live birth:
    The complete expulsion or extraction from its mother, irrespective of the duration of the pregnancy, of a product of conception in which, after the expulsion or extraction, there is:

    • breathing;
    • beating of the heart;
    • pulsation of the umbilical cord; or
    • unmistakable movement of voluntary muscle, whether or not the umbilical cord has been cut or the placenta attached.

    Low Birth Weight:
    Any liveborn infant weighing less than 2500 grams.

    Neoplasms (ICD-10 C000-D489):
    Although the vast majority of deaths in this category are due to malignant cancer, also included are benign, in-situ, and unspecified "tumours". Detailed ICD-10 breakdown used in "Neoplasm Deaths" tables are;

    • lung: includes trachea, bronchus, lung and pleura (C33, C340-C349, C384, C450). Note: now excludes mesothelioma of lung and trachea.
    • female breast: (C500-C509)
    • colorectal: (C180-C218)
    • other G.I.: includes esophagus, stomach, small intestine and duodenum, liver & intrahepatic bile ducts, gallbladder and extrahepatic ducts, pancreas, peritoneum, other and ill-defined within digestive organs (C150-C179, C220-C269)
    • female reproductive: includes uterus, cervix, placenta, ovary and adnexa, vagina & external genitalia (C510-C58)
    • prostate: C61
    • blood lymph: includes lymphatic and haematopoietic tissue (C810-C969, C463).
    • other malignancy: includes malignant neoplasms of other (e.g. lip, oral cavity, pharynx, nose, ear, larynx, heart, bone and connective tissue, urinary tract, eye, brain, endocrine glands)and ill-defined or unspecified sites (C000-C148, C300-C449, C451-C462, C467-C499, C600-C609, C620-C768, C5099*, C80*). Note: * codes used exclusively by BC Vital Statistics Agency for male breast cancer and for unknown primary site cancer.
    • non-malignant & unspecified: includes benign, in-situ, and neoplasms of uncertain or unknown behaviour (D000-D489).
      Note: This neoplasm group now includes myeloproliferative disease, thrombocythemia, monoclonal gammopathy, and lymphoproliferative disease which were not previously considered neoplastic in ICD-9 and were counted in other ICD chapters.

    Other Selected Death Statistics:
    Tables under this heading include deaths due to:

    • diabetes (E100-E149).
    • alcohol related - see above.
    • AIDS/HIV: (B200-B24).
    • other infectious and parasitic disease: (A000-B199, B250-B999) Note: Now includes obstetrical and neonatal tetanus.
    • cerebro and other vascular disease: (I600-I698, I700-I879, I950-I959, I880-I899, I970-I979, I99). Includes cerebrovascular disease, disease of arteries and veins, hypotension, and other circulatory system disease. Note: "Other circulatory system disease" now includes post procedural disorders of the circulatory system (I970-I979) which are never selected as the UCOD. However they are confirmed by editing and either recoded to the more specific disease (embolism, stroke, M.I.) or double coded if the complication is confirmed.
    • liver disease: (K700-K7699). Note: Now includes toxic liver disease with cholestasis.
    • ALS/MS: Amyotrophic lateral sclerosis and multiple sclerosis: (G122, G1221, G35). Note: In order to maintain continuity with ICD-9, unspecified motor neuron disease (G122) is included in this category as it was previously not distinguishable from ALS.

    Premature/Pre-term:
    Any live born infant less than 37 weeks gestation at delivery.

    Respiratory Disease Death Statistics:
    Tables under this heading include deaths due to the following:

    • emphysema: (J430-J439) Note: Now excludes when described as or resulting in obstructive disease -see note at COPD.
    • COPD: (440-J449). Note: Now includes specific code within the group for COPD when accompanied with acute lower respiratory infection, or with acute exacerbation. This inclusion has no statistical impact on UCOD. Also, this category now includes asthma and emphysema described as obstructive not previously included in ICD-9.
    • pneumonia: (J120-J181, J188-J189) Note: ICD-10 has a new code for chlamydial pneumonia (J160). It is uncertain if this condition would have previously been coded to "pneumonia due to other specified bacteria" (ICD-9 4828) or to "other diseases due to viruses and chlamydia" (0788), or to both. This disease is very rare on death records so if not coded to 4828, the impact to comparison of historical data would still be minimal. Daily VS edits have been implemented to unsure consistent selection of pneumonia.
    • influenza: (J100-J118)
    • asthma: (J450-J459, J46) Note: Now excludes when described as obstructive - see note at COPD.
    • lung disease due to external agents: (J60-J709)
    • pulmonary fibrosis: (J841)
    • other respiratory diseases: (J00-J069, J182, J200-J42, J47, J80-J840, J848-J9899) Note: Now includes post procedural respiratory disorders (J950-J959) which formerly used to be injury codes. These codes are never selected as the UCOD so their impact would only effect multiple code analyses. The Vital Statistics Agency includes these in daily data edits to confirm them as post procedural and to double code for the specific respiratory condition (e.g. pneumonia). As a result, respiratory disease statistics in BC are more consistent with historical data.

    Stillbirth:
    The complete expulsion or extraction from its mother after at least twenty weeks of pregnancy or after attaining a weight of at least 500 grams, of a product of conception in which, after expulsion or extraction, there is no breathing, beating of the heart, pulsation of the umbilical cord or unmistakable movement of voluntary muscle.

    Teenage Mother:
    Any mother who was age 19 or less at the time of delivery.

    UCOD:
    Underlying cause of death - based upon application of standard international coding rules for determining sequential relationships of conditions and diseases from immediate cause backwards to underlying cause.

    [Return to Table of Contents]

    Index of Quarterly Digest Articles to date:

    Volume 1 - Number 1 & 2 - October 1991

    1. Alcohol Related Deaths.
    2. Status Indians in British Columbia, 1989 - A Vital Statistics Overview.

    Volume 1 - Number 3 - January 1992

    1. Fatal Poisonings in British Columbia - 1989 - by J.M. Macdonald.
    2. Fatal Head Injuries in British Columbia - 1990 - by J.M. Macdonald.

    Volume 1 - Number 4 - April 1992

    1. Mortality Mapping in British Columbia - by M.C.R. Edgell & L.T. Foster.
    2. Suicide Deaths in British Columbia, - by T.A. Tuk & J.M. Macdonald.
    3. Selected Health Status Indicators in British Columbia, 1985-1990 by K.F. Burr, L.T. Foster, & J.H. Mohamed.

    Volume 2 - Number 1 - July 1992

    1. Charting Birth Weight of British Columbia Newborns: How do we compare? - by W.J. Kierans.
    2. Cardiovascular Disease Death in British Columbia: Still number one - by J.M. Macdonald, T.A. Tuk & J.H. Mohamed.

    Volume 2 - Numbers 2 & 3 - November 1992

    1. A Deadly Affair: Cigarette Smoking - Attributable Deaths, British Columbia, 1985 and 1989 - by R.M. Strohmaier & W. Hu.
    2. Health Status Registry: A Health Planning Tool Revisited and Revitalized by W.J. Kierans & A.K. McBride.

    Volume 2 - Number 4 - February 1993

    1. Recent Advances in Community Health Related Information: A Vital Statistics Perspective - by T.A. Tuk, M.A. Collison, L.T. Foster.
    2. Cesarean Section Rates: A British Columbia Overview Prepared by Division of Vital Statistics.

    Volume 3 - Number 1 - May 1993

    1. Cancer Mortality in British Columbia Part I: Cancer as an Underlying Cause of Death - by C. Cranfield, T.A. Tuk & J.M. Macdonald.
    2. Estimates for Health Effects Attributable to Second-Hand Smoke in British Columbia - by M.E. Thomson.

    Volume 3 - Number 2 - August 1993

    1. Cancer Mortality in British Columbia Part II: Cancer Mortality Multiple Conditions - by J.M. Macdonald, T.A. Tuk & C. Cranfield.
    2. Technical Notes - An Alternative Approach to Mapping Mortality: A Bayesian Procedure - by Ying C. MacNab.

    Volume 3 - Number 3 - November 1993

    1. Injury Facts and Prevention Strategies for Children and Youth in British Columbia - by Office for Injury Prevention.
    2. Ethnicity and Health Status, Part One: The IndoCanadian Community - by W.J. Kierans.
    3. Technical Notes - Measurement of Mortality Part I: Crude Rate - by Ying C. MacNab & T.A. Tuk.

    Volume 3 - Number 4 - May 1994

    1. Ethnicity and Health Status, Part Two: The Chinese Immigrant Community - by W.J. Kierans.
    2. Potential Years of life Lost: British Columbia, 1985-1992 - by R.M. Strohmaier.
    3. Technical Notes - Measurement of Mortality Part II: Age Standardized Mortality Rate - by Ying Cai MacNab & T.A. Tuk.

    Volume 4 - Number 1 - June 1994

    1. Vital Statistics in British Columbia: An Historical Overview of 100 Years, 1891 to 1990 - by J.M. Macdonald.
    2. Technical Notes: Measurement of Mortality Part III: Standardized Mortality Ratio - by Ying Cai MacNab & T.A. Tuk.

    Volume 4 - Number 2 - August 1994

    1. Status Indians in British Columbia: A statistical Overview [1987-1992] - prepared by Medical Services Branch of Health Canada and BC Division of Vital Statistics.

    Volume 4 - Number 3 - November 1994

    1. Drug-Related Deaths in British Columbia: 1981 to 1993 - by T.A. Tuk and J.M. Macdonald.

    Volume 4 - Number 4 - February 1995

    1. Sudden Infant Death Syndrome in British Columbia: 1981 to 1993 - by H. Amershi.

    Volume 5 - Number 1 - May 1995

    1. AIDS/HIV Related Mortality in British Columbia: 1985 to 1994 - by N. Fast.

    Volume 5 - Number 2 - August 1995

    1. Suicide, Homicide, and Gun Deaths, British Columbia, 1985 to 1993 - by T.A. Tuk & J. Macdonald.

    Volume 5 - Number 3 - November 1995

    1. Respiratory Disease Mortality in British Columbia, 1985 to 1994 by Y.C. MacNab, J. Macdonald, T.A. Tuk.

    Volume 5 - Number 4 - March 1996

    1. Diabetes in Birth and Death: British Columbia, 1987 to 1994 - by K. Stenning.

    Volume 6 - Number 1 - July 1996

    1. Increased Maternal Age and the Outcome of Pregnancy: An eight year population based study, British Columbia, 1987 - 1994 by Y.C. MacNab, J. Macdonald, T.A. Tuk.

    Volume 6 - Number 2 - October 1996

    1. Marriage and Family in British Columbia: 1931 - 1994 - by Z. Kashaninia.

    Volume 6 - Number 3 - January 1997

    1. Accident Fatality in British Columbia, 1987 - 1995 [Introductory chapter of in progress longer report] - by E. Demaere.

    Volume 6 - Number 4 - April 1997

    1. A Review of Delivery Mode in British Columbia, 1987 - 1995 - by Y.C. MacNab.

    Volume 7 - Number 1 - July 1997

    1. Pregnancy Outcomes in British Columbia, - by Cathy Hull.

    Volume 7 - Number 2 - October 1997

    1. Women and Cancer: Lung and breast cancer among women in BC, 1974-1996 - by Z. Kashaninia.

    Volume 7 - Number 3 - January 1998

    1. Women and Cancer (Part II): Ovarian, uterine and cervical cancer among women in B.C., 1980 to 1996 - by Z. Kashaninia.

    Volume 7 - Number 4 - May 1998

    1. The Declining Trend of Sudden Infant Death Syndrome: Comparison with other major causes of infant mortality and deaths due to unknown causes, BC, 1985 to 1996 - by R. Fisk, J Macdonald, W. Vander Kuyl with editorial comments by S. Peck.

    Volume 8 - Numbers 1 & 2 - December 1998

    1. Animal Caused Fatalities, British Columbia, 1969 to 1997 - by R. Armour and J. Macdonald.

    Volume 8 - Number 3 - March 1999

    1. The Impact of Infectious Diseases on Mortality in BC, 1990-1997 - by Z. Kahsaninia.

    Volume 8 - Number 4 - August 1999

    1. Hepatitis Deaths in British Columbia, 1990-1998 - by Z. Kashaninia.

    Volume 9 - Numbers 1 & 2 - October 1999

    1. Drowning and Other Water-Related Accidental Fatalities, British Columbia, 1990 to 1998 -
      by Z. Kashaninia, J. Macdonald and R. Armour
      .

    Volume 9 - Number 3 - March 2000

    1. Fire Deaths in British Columbia, 1986 to 1998 - by Z. Kashaninia.

    Volume 9 - Number 4 - June 2000

    1. Beautiful and Deadly British Columbia: Natural Environmental Deaths, 1985 to 1998 -
      by T. Stubbings and J. Macdonald
      .

    Volume 10 - Number 1 - October 2000

    1. Tuberculosis and Mycobacterium: Impact on Mortality in British Columbia, 1990 to 1999 -
      by Z. Kashaninia
      .

    Volume 10 - Numbers 2& 3 - December 2000

    1. The Health Status Registry - adapted from original report by M. Silver.

    Volume 10 - Number 4 - July 2001

    1. The Impact of ICD-10 (in year 2000) on identifying, classifying, coding and UCOD selecting neoplasm(s) on death records - by J. Macdonald and R. Armour.

    Editor's Note:

    Due to a new Government cost constraint directive, the Quarterly Digest, beginning with this issue, will only be published on the BC Vital Statistics web-site. Regular subscribers will be receiving notification of availability and information regarding distribution options.

    Contributors' Note:

    The editorial staff would like to invite any researchers of health-related topics who wish to contribute an article or paper summary for publication in this Quarterly Digest to contact the Information and Resource Management Branch of the British Columbia Vital Statistics Agency. Articles should focus on health status issues in British Columbia. It is preferable that submissions be in "electronic media" format (e.g. Word, Word Perfect, Excel, Power Point, Corel, Pagemaker, etc.). Article presentation will be subject to space allowances and publishing deadlines.

    Readers' Note:

    Re: "Letters to the Editor", or mailing and distribution.

    Comments, suggestions, questions, or criticisms regarding this publication?

    Use the Feedback Form