WWW-publications from the WHO MONICA Project

Quality Assessment of Coronary Event Registration Data in the WHO MONICA Project

January 1999

Markku Mähönen¹, Hanna Tolonen¹, Kari Kuulasmaa¹, Hugh Tunstall-Pedoe² and Philippe Amouyel³ for the WHO MONICA Project⁴

¹ MONICA Data Centre, National Public Health Institute, Helsinki, Finland
² Cardiovascular Epidemiology Unit, University of Dundee, Ninewells Hospital, Dundee, Scotland
³ Faculty of Medicine and Pasteur Institute of Lille, Lille, France
⁴ Annex: Sites and key personnel of the WHO MONICA Project

• Copyright notice
• Document identification:
  • URL:http://www.ktl.fi/publications/monica/coreqa/coreqa.htm
  • URN:NBN:fi-fe19991072

This document includes the main findings of unpublished reports: :

• Pajak A, Kuulasmaa K, Tunstall-Pedoe H, Ingrid Martin I, Rajakangas A-M for the WHO MONICA Project. Coronary event registration data quality report. MONICA Memo 173A, February 1990.
• Mähönen M,  Rajakangas A-M, Kuulasmaa K, Tunstall-Pedoe H for the WHO MONICA Project.  WHO MONICA Project - quality assessment of coronary event data for 1980-1990. MONICA Memo 257A, February 1994.

Acknowledgements

Thanks are due to Alun Evans and Hermann Wolf who commented on the text.

The MONICA Centres are funded predominantly by regional and national governments, research councils, and research charities. Coordination is the responsibility of the World Health Organization (WHO), assisted by local fund raising for congresses and workshops. WHO also supports the MONICA Data Centre (MDC) in Helsinki. Not covered by this general description is the ongoing generous support of the MDC by the National Public Health Institute of Finland, and a contribution to WHO from the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland, USA for support of the MDC and the Quality Control Centre for Event Registration in Dundee. The completion of the MONICA Project is generously assisted through a Concerted Action Grant from the European Community. Likewise appreciated are grants from ASTRA Hässle AB, Sweden, Hoechst AG, Germany, Hoffmann-La Roche AG, Switzerland, the Institut de Recherches Internationales Servier (IRIS), France, and Merck & Co. Inc., New Jersey, USA, to support data analysis and preparation of publications.

Contents

• 1. Introduction
• 2. Material and methods
  • 2.1 Populations
  • 2.2 Periods of events considered
  • 2.3 Age and sex
  • 2.4 Other inclusion criteria for the data
  • 2.5 Sources of information
• 3. Serial number inventory and routine data checking status
• 4. Coverage of registration
  • 4.1 Fatal events
  • 4.2 Non-fatal events
• 5. Availability and quality of data for diagnostic classification
  • 5.1 Fatal events
  • 5.2 Non-fatal events
• 6. Availability of data on previous MI
• 7. Results from test cases and ECG test coding
• 8. Internal quality control
• 9. Summary
• 10. Comments on individual RUAs
• References to publications
• References to internal MONICA documents
• Specific tables
  • Table 1. Reporting Unit Aggregates, MCCs, Reporting Units and calendar years considered
  • Table 2. Number of coronary event records used for the analysis
  • Table 3. Serial number inventory
  • Table 4 Number of unresolved constraint violations by year
  • Table 5. Number of unresolved suspected duplicate registrations
  • Table 6. Comparison of fatal events in the MONICA register with routine statistics
  • Table 7. Proportions of diagnostic categories in non-fatal and fatal events and crude case fatality
  • Table 8. Availability of data in fatal and non-fatal events
  • Table 9. Proportions of categories on data of previous myocardial infarction
  • Table 10. Summary status of coronary event data
• Appendices
  • Appendix 1. Edit specifications
  • Appendix 2. Unresolved constraint violations
  • Appendix 3. External quality control. Test case histories
  • Appendix 4. External quality control. ECG coding
  • Appendix 5. MONICA coronary event trends quality scores
• Annex: Sites and key personnel of contributing MONICA Centres

1. Introduction

To achieve the goals set for the MONICA Project (1), the quality of the acute coronary event and coronary death registers is of paramount importance. The key issues are:

• the coverage of the registration;
• the availability and quality of the data collected on the events, especially those which form the basis for the diagnostic classification of events; and
• for the assessment of trends within populations, which is the main concern of MONICA, the stability of the methods applied for registration and the availability of data in each MONICA Collaborating Centre (MCC).

This report addresses these issues. It follows the approach of the first quality assessment report (5), the second report (6) and the quality assessment in the publication "Myocardial infarction and coronary deaths in the WHO MONICA Project" (2). The current report covers the full MONICA event registration period, including the years covered by the earlier quality assessment reports.

In the specification of the calculations for this quality assessment report the names of the data items in the Core Data Transfer Format - Coronary Events (3) have been used. The terminology used in this report was developed for MONICA event registration in the MONICA Manual (3), with later refinements in the collaborative publications (2). It is important to note that the reliability of event rates depends also on the quality of demographic data.

2. Material and Methods

2.1 Populations

The report considers the Reporting Unit Aggregates (RUAs) which are seen as potential candidates for units of analysis of the MONICA coronary event data. The RUAs, their abbreviations and Reporting Units (RUs) are listed in Table 1. Some of the RUAs have several versions distinguished by suffix ´a´ and ´b´. Different combinations of Reporting Units (RUs) may be used for analyses concerning coronary events only, and others for analyses involving both coronary event and risk factor data. The reason for the distinction is that some RUs of some RUAs were not included in every risk factor survey. Therefore, the different RUAs from AUS-PER, GER-BER, GER-BRE, GER-EGE, ICE-ICE, RUS-MOI and RUS-NOC the RUAs may include the same RUs. Furthermore, the combination (GER-AUG) of GER-AUR and GER-AUU has also been considered. Altogether 49 RUAs are considered.

2.2 Periods of events considered

Only full calendar years of registration are considered. Individual years for which data were received in the MONICA Data Centre (MDC), were excluded if:

• registration in a population was started after January (year 1984 for AUS-NEW, FRA-LIL, FRA-TOU, GER-BRE, GER-AUG, GER-AUR, GER-AUU, ITA-BRI, SWI-TIC and SWI-VAF) or ended before December (year 1994 for AUS-NEW).
• the MCC has indicated that the first year was a pilot year (year 1983 for YUG-NOS, year 1984 for FRA-STR, SPA-CAT and RUS-MOC).

The years considered for each RUA are shown in Table 1.

2.3 Age and sex

The quality assessment concerns the age group 25-64. The age was calculated in full years of the date of onset, except for mortality comparison, where date of death was used (Table 6). When calculating the age, day 99 was interpreted as 15 and day/month 99/99 as 30/06. No age standardization was used. Data for men and women were combined for the analysis.

2.4 Other inclusion criteria for the data

Individual records have not been included in the analysis if:

• DIACAT does not have a valid value,
• DIACAT=4 and SURVIV=2, and none of the CLIND1-3 is between 410-414,
• DIACAT=4 and SURVIV=1 or 9, and none of the CLIND1-3 is between 410-411.

Otherwise, all data available in the MDC were used in the analysis, regardless of their quality. The total number of coronary events available in the MDC are shown by calendar year in Table 2.

2.5 Sources of information

The report is based on the data which the MDC has received from the MCCs on coronary events (Form 01, see MONICA Manual (3)), on mortality (Form C (ICD-8) and Form E (ICD-9), see MONICA Manual (4)), and on other information received from the MCCs.

3. Serial number inventory and routine data checking status

Table 3 shows a summary of serial number inventory, which is based on a linkage of the coronary event data and the serial number inventory data (3) received by the MDC. Its purpose is to check that the MDC database has exactly those records which it should have according to the MCC. Ideally, all entries of the last four columns should be zero. Otherwise there is a possibility that some records have been lost or duplicated  from the time of the data's ascertainment. There are a large number of discrepancies in the serial number inventory for AUS-PER, BEL-LUX, HUN-BUD, HUN-PEC, LTU-KAU and NEZ-AUC.

When the coronary event data were received in the MDC they were checked routinely for the constraints specified in Appendix 1. All violations of the constraints were reported to the MCC for their correction or elucidation. Data values outside the constraint limits were acceptable, but the MCC had to check that the values were not unusual due to data errors. The MCCs were only asked to revise data if they were incorrect. The current unresolved constraint violations are shown in Appendix 2 and summarized by calendar year in Table 4. There is only a small number of unresolved constraint violations for any RUA. The number is more than ten for BEL-CHA, BEL-GHE, FIN-NKA, GER-EGE, GER-RHN, HUN-BUD, and SWE-GOT.

The data are also checked periodically in the MDC for events which have been registered more than once. Table 5 shows the number of suspected duplicate registrations which have neither been confirmed as duplicate registrations nor as separate events. There is only a small number of unresolved suspected duplicate registrations for any RUA. The number is more than ten for BEL-CHA, GER-AUG and GER-EGE.

The data with unresolved constraint violations or duplicate suspicions have been included in the analysis for this report.

4. Coverage of registration

4.1 Fatal events

According to the MONICA Manual (3),  suspect coronary events are collected from death certificates, hospital medical records and  community health services. All deaths with the underlying cause of death coded to 410-414 must be registered even if not found to be a coronary case on examination.

Data should have been submitted to the MDC on:

• every event classified in the diagnostic categories 1, 2 or 9; and
• those fatal events with diagnostic category 4 and with the underlying cause of death, or (if not known) one of the three death certificate diagnoses (item CLIND of the Core Data Transfer Format-Coronary Events, see MONICA Manual (3)) coded to ICD codes 410-414.

To be certain that all coronary deaths are registered, in each RUA the number of fatal events registered with a CHD death certificate diagnosis (CLIND1-3 coded 410-414) should be greater than the number of coronary deaths reported in routine (official) mortality statistics (greater because CHD deaths in routine statistics include only those having an ICD code 410-414 as the underlying cause of death). The number of these fatal events registered in MONICA is the numerator and the number of CHD deaths in the routine statistics is the denominator in the CHD/CHD ratio shown in   Table 6 . The ratio is less than one for at least one year in AUS-PER (1988); BEL-CHA (1983-1987); BEL-GHE (1983-1991); BEL-LUX (1985-1987); CHN-BEI (1984-1986); DEN-GLO (1987-1988, 1990-1991); GER-AUR (1985-1986, 1994); GER-AUU (1986); GER-BREa (1992);  GER-RHN (1984-1987); HUN-BUD (1982-1984, 1986, 1988); HUN-PEC (1984-1989); ITA-BRI (1987); NEZ-AUC (1983-1988, 1990-1992); POL-WAR (1988-1989, 1991-1994); RUS-MOC (1990-1991, 1993); RUS-MOIa (1987-1988, 1990, 1992); RUS-MOIb (1988, 1990, 1992); RUS-NOCa (1989, 1991, 1993); RUS-NOIa (1987); UNK-BEL (1986, 1988, 1991-1992).

In some RUAs the access to the death certificates or the final official code of death is restricted, which may be the reason why some deaths coded as CHD are not registered. Therefore, we also compared the number of all registered fatal events to the number of CHD deaths by routine statistics (Total/CHD ratio: the number of events in MONICA register is the numerator and the number of deaths in the routine mortality statistics is the denominator) (Table 6) to see if the restricted availability of routine code of death is compensated by other registrations. In many MCCs, the breadth of registration is large, although the ratio of MONICA CHD deaths/routine statistics CHD deaths is below 1.0. Total/CHD ratio is less than one in at least one year in BEL-LUX (1985); CHN-BEI (1984-1986); GER-AUR (1994); GER-RHN (1987); HUN-BUD (1986); HUN-PEC (1985-1989); RUS-MOC (1990-1991); RUS-MOIa (1987-1990, 1992); RUS-MOIb (1988,1990, 1992); RUS-NOCa (1989,1991, 1993); SWE-GOT (1985); UNK-BEL (1986, 1991-1992).

In many of these populations the discrepancy is small, and could be explained by errors in the routine mortality statistics, or a different way of determining the year of death in the register and in the routine statistics (UK and Australia).

Especially worrying is the situation in MCCs where low CHD/CHD ratio is insufficiently compensated for by broader total fatal registration or otherwise explained ( see Section 10 for specific comments on each RUA):

• BEL-LUX (1985);
• CHN-BEI (1984-1986);
• GER-AUR (1994);
• GER-RHN (1987);
• HUN-BUD (1986);
• HUN-PEC (1985-1989);
• RUS-MOC (1990-1991);
• RUS-MOIa (1987-1988,1990);
• RUS-MOIb (1988, 1992);
• RUS-NOCa (1989, 1991, 1993);

The comparison with routine mortality statistics was not possible because of missing or incomplete routine mortality statistics data in the following RUAs:

• BEL-LUX in 1988-1991;
• RUS-NOCa in 1984;
• UNK-GLAa in 1993;
• RUS-NOCb in 1984 and in 1991-1993.

There is a clear error in routine mortality statistics data in 1993 in RUS-NOCa and RUS-NOIa.

4.2 Non-fatal events

According to the MONICA Manual (3), the suspect coronary events are collected from hospital medical records and from community health services. All clinically diagnosed myocardial infarctions should be registered.

Data should have been submitted to the MDC on:

• every event classified in the diagnostic categories 1, 2 or 3;
• those non-fatal events with diagnostic category 4 and with one of the three clinical diagnoses (item CLIND of the core data transfer format-coronary events) coded to ICD codes 410-411.

The case identification should be wide enough to catch practically all clinically diagnosed myocardial infarctions which would get the diagnostic category DIACAT 1 (definite MI) in MONICA. Unlike for fatal events, there are no independent data sources against which the coverage of the non-fatal registrations could be checked. However, there are two indicators in the data which may reflect problems in the coverage of non-fatal registrations:

• proportion of non-fatal events with diagnostic category 2 or 4 and
• case fatality.

Proportion of non-fatal events with DIACAT 2 or 4:

In order to capture all events with MONICA DIACAT 1, the selection of clinical diagnoses used for registration will in practice be such that it will also produce a significant proportion of DIACAT 2 (possible MI) and 4 (no MI). Therefore, the case ascertainment for definite coronary events may be incomplete, if the proportion of events belonging to DIACAT 2 or 4 is small.

Table 7 shows the proportion of different diagnostic categories of non-fatal events. The proportion of possible and not AMIs is particularly small (below 20%) in the following RUAs: BEL-CHA (1983-1986); BEL- GHE (1983-1987, 1992); CHN-BEI (1985); DEN-GLO (1982); GER-RHN (1987); ICE-ICEa (1989-1990, 1992); ICE-ICEb (1990-1992); NEZ-AUC (1992); SWE-GOT (1992-1993); YUG-NOSa (1984-1987, 1990-1992). The reason for the low proportion in BEL-CHA, BEL-GHE and YUG-NOS is that DIACAT 4 events are not sent to the MDC. In ICE-ICEa and ICE-ICEb the reason may be related to low numbers and clinical practice patterns. In SWE-GOT data for the MI register are collected only from clinically suspected MIs with the clinical diagnoses of 410-411. (See Section 10 for specific comments for each RUA.)

Case fatality:

If there is under-ascertainment of non-fatal events, it may lead to an unusually high case fatality. Case fatality, calculated using MONICA definition I: (F1+F2+F9)/(F1+F2+F9+NF1), for men and women combined, is shown in Table 7.

Case-fatality was over 65% in CHN-BEI (1990-1991, 1993); HUN-BUD (1982-1987, 1989); POL-TAR (1984-1993); RUS-MOC (1991, 1993); RUS-MOIa (1988-1989, 1992-1993); RUS-MOIb (1992-1993); RUS-NOCa (1990, 1992); RUS-NOCb (1989, 1990); RUS-NOI (1987-1988, 1993).

The particularly high case fatalities in HUN-BUD and POL-TAR are explained in Section 10.

5. Availability and quality of data for diagnostic classification

5.1 Fatal events

The results of autopsy give valid data for the diagnostic classification of fatal events. The frequency of autopsies done is shown in Table 8.

There was a large variation in the frequency of autopsies between the MCCs. However, within each RUA the frequency of autopsies was fairly stable except in DEN-GLO, GER-EGEa, GER-EGEb and SWE-GOT where there was a drop in the frequency of autopsies, and YUG-NOS where there was first a rise in the frequency of autopsies and later a marked decline.

In events where autopsy was not carried out or the results of autopsy are not available (NECSUM= 8 or 9), other diagnostic information is very important for reliable classification of fatal events. For such fatal events, the availability of data on symptoms, the number of ECGs, the availability of data on enzymes, on previous CHD and previous MI are shown in Table 8.

For events with no autopsy, the RUAs in which the range in the yearly proportion of the availability of data was over 40% are shown in Table A.

Changes in the proportion of DIACAT=9 events (Table 7) were notable in GER-EGEa and GER-EGEb (rising trend), GER-RHN (varies from 15% to 63%), HUN-PEC (varies from 0% to 38%) and SWE-GOT (a rise in 1990-1994).

5.2 Non-fatal events

The availability of the data has been very good in non-fatal events, as shown in Table 8. However, the availability of data from year to year varies in some RUAs. In DEN-GLO, 1982 deviates from other years (lower availability of data on ECGs); in FRA-TOU there is a drop in the availability of data on ECGs; in GER-EGEa, GER-EGEb, GER-RHN, HUN-BUD and HUN-PEC the availability of data on ECG varies from year to year; in RUS-NOCa, RUS-NOCb and RUS-NOI the availability of data on more than three ECGs drops to zero in 1988. (See also Section 10 for specific comments and clarification for each RUA.)

6. Availability of data on previous MI

The data item on previous MI is important when the first ever coronary events need to be identified. This concerns for example the calculation of incidence rates. The proportions of different categories of data on previous MI are given in the Table 9.

The proportion of indeterminate codes (4 (yes, from ECG) and 9 (insufficient data)) was over 30 % in:

• BEL-CHA (1986)
• CAN-HAL (1984-1993)
• GER-EGEb (1988)
• GER-RHN (1984)
• HUN-BUD (1982-1984)
• HUN-PEC (1984)
• POL-TAR (1984-1993)
• RUS-MOIa (1985)

In addition to these the proportion of indeterminate codes was over 10 % in:

• AUS-NEW (1985)
• AUS-PERa (1984-1987)
• AUS-PERa (1984-1987)
• BEL-CHA (1983-1992)
• BEL-GHE (1983-1992)
• BEL-LUX (1985, 1988-1989)
• CZE-CZE (1984-1993)
• DEN-GLO (1982-1984, 1987, 1989)
• FRA-LIL (1985-1994)
• FRA-STR (1985)
• GER-AUU (1985-1987, 1989-1994)
• GER-BREa (1985-1992)
• GER-BREb (1985-1992)
• GER-EGEa (1986, 1988, 1990-1993)
• GER-EGEb (1990-1993)
• GER-RHN (1985)
• LTU-KAU (1985)
• POL-WAR (1984-1994)
• RUS-MOC (1986-1991)
• RUS-MOIa (1986-1990)
• RUS-MOIb (1985-1990)
• SPA-CAT (1985-1987, 1989)
• SWI-VAF (1985-1989, 1992-1993)
• USA-STA (1981-1982, 1986)

It is important to note that these proportions are calculated for all registered events. If we consider only non-fatal definite MIs, the proportions of indeterminate events are very low in nearly all RUAs. On the other hand, for fatal events the proportions are higher than those shown in Table 9.

7. Results from test cases and ECG test coding

The MONICA Quality Control Centre for event registration in Dundee, Scotland, circulated test case histories to the MCCs. The test case histories firstly focused on training of coders and later on assessing the comparability between the MCCs.  The results of test case history exercises are summarized in Appendix 3. Likewise, the MONICA Quality Control Centre for ECG Coding in Budapest, Hungary, circulated test sets of ECGs for coding by the MCCs. The results of the ECG test coding are summarized in Appendix 4.

8. Internal quality control

It was suggested in the MONICA Manual (3) that the identification of events should start with a wider range of death certificate and hospital record diagnosis ICD codes. Event identification could proceed with a narrower range, if it had been established (and documented) in a pilot study (and reviews) that only a negligible number of definite coronary events were excluded by adopting a narrower range of ICD codes.

The results of the initial pilot studies and later repetitions are known for AUS-PER, ICE-ICE and UNK-GLA only.

9. Summary

Table 10 gives an overall summary of the data quality for each RUA by year. Data for the RUAs and years coded "2" have sufficient quality for calculation of attack rates and case fatality. Code "2*" indicates that   the data do not fulfil all the criteria for the good score "2" but the difference is minor, or there is an acceptable explanation for the deviant pattern. Code "1" indicates that a decision needs to be made separately for each publication. Score "0" is used for cases where the data are insufficiently complete for use in analyses.

After the previous quality assessment (6), many MCCs have done a lot of work in completing the data collection and clarifying the unresolved issues. There are data for ten or more years from 26 MCCs out of the 32 current MCCs. In 17 MCCs, ten or more years have score "2". There are several years in three MCCs having ten years' data for which the assessment cannot be completed because of missing data on routine mortality statistics. Four MCCs have data for nine years and will not be able to complete the tenth year: one of these with nine years data (AUS-NEW) score "2" for all nine years; there have been problems in the availability of  ECGs in FRA-TOU; and the coding of enzymes is deviant from the other MCCs in  SWI-TIC and SWI-VAF. Bremen (GER-BRE) has data for 8 years, all coded "2".   Luxembourg (BEL-LUX)  has data for seven years but routine mortality statistics are missing for four years and the quality assessment cannot be completed for this MCC.

There are still unresolved data queries in some MCCs. Clarifications which have been received from the MCCs are listed in Section 10. Several MCCs have notified that they are unable to make any data amendments.

The explanations for scores "2*",  "0" and "1" are included in Section 10. It is important that those who use the coronary event data for analyses read this section carefully so as to be able to understand the data better and make decisions about their use.

Appendix 5 defines a risk score for trends in coronary events, and shows the values of the score and its components for the RUAs and years to be considered in the the forthcoming publication of the 10-year trends.

10. Comments on individual RUAs

AUS-NEW

ICD-version used: 9/ICD9-CM (ICD9-CM used since 1st Jan 1988)
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data from Aug 1984 to August 1994, but only full calendar years are considered in this report.
• ICD9-CM has been used in most hospitals since Jan 1988.

AUS-PER

ICD-version used: ICD9/ICD9-CM
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• ICD9-CM is used from 1988
• The MCC applied hot pursuit in 1989 for four months to check the completeness of registration; 86% of definite MI cases were registered using cold pursuit only. The 'extra' cases identified through hot pursuit were not included in the MDC database. 
• There is a major concern in the consistency of the registration over the 10-year period:   The MCC applied cold pursuit, registering events from the ICD codes 410-411. When   ICD9-CM was introduced in 1988 with a new ICD code 411.1 - intermediate coronary syndrome -  the number of registered non-fatal events increased.  Before 1988, the number of registered non-fatal events was lower (about 700/year) than after 1988 (about 1000/year). It is difficult to evaluate the effect of this bias on the trends, and therefore no corrections are recommended for the data analyses. It is possible that also DIACAT 1 events were detected from the ICD code 411.1 (which would most probably have been coded as 413 during ICD-9 and thus escaped the registration which was done using ICD codes 410-411).
• The ratio of non-fatal  DIACAT=1 events to all registered non-fatal events is somewhat higher during the first years of registration (77% in 1985, compared with 51% in 1991).
• Availability of data on previous MI in fatal events was lower during the first years of registration.
• A few fatals are missing in 1988 - the explanation is errors in the routine mortality statistics.

BEL-CHA

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Restricted

Comments:

• Fatal events: low CHD/CHD ratio, but total fatal/routine CHD ratio high. The MCC has explained this by errors in routine statistics; also, the death certificate codes often change for the routine statistics after the MCC has reviewed them; the coverage of registration is secured since all nominal death certificates are reviewed by the MONICA team before they reach the Health Statistics Department.
• The availability of data on >=3 ECGs was below 80% in 1991; however, the availability of data on >=1 ECG and on enzymes was nearly 90%.
• CLIND missing in 1983-1985 and 1990-93 in non-fatal events.
• There are no DIACAT 4 events since these are not sent to the MDC which explains why the proportion of non-fatal DIACAT=1 events from all non-fatal events is so high; however, this proportion was higher during the first years of registration (up to 91% in 1984, declining to 63% in 1992).
• Crude case fatality is  62% in 1986 and 60% in 1987, declining to 42% in 1991.
• The case ascertainment of non-fatal events is secured by checking all sickness leaves with a suspicion of cardiovascular disease regularly against the MI register. Also, letters to GPs are sent once a year.
• There are 24 unresolved constraint violations and 20 suspect duplicates.

BEL-GHE

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Restricted

Comments:

• Fatal events: low CHD/CHD ratio, but total fatal/routine CHD ratio high. The MCC has explained this by errors in routine statistics; also, the death certificate code often change for the routine statistics after the MCC has reviewed them; the coverage of registration is secured since all nominal death certificates are reviewed by the MONICA team before reaching the Health Statistics Department. Less than 5% of death certificates are unavailable because the death occurred outside the registration area. However, lists of all inhabitants of Ghent who have died were received from the population department on a weekly basis and the certifying physician was known and data for these cases were reviewed.
• The availability of data on >=3 ECGs was below 80% in 1983; also the availability of data on >=1 ECG was somewhat lower than later; however, the availability of data on enzymes was over 90%.
• The proportion of non-fatal DIACAT=1 events of all non-fatal events is high because no DIACAT=4 events are sent to the MDC.
• GPs were contacted regularly, and sickness leaves for cardiovascular reasons were checked regularly.
• In 1983 the availability of data on ECG was lower than during later years in non-fatal events.
• In non-fatal events the codes for the item CLIND were given by the MONICA team on the basis of medical records and do not represent the diagnosis made by the clinicians. (Their diagnoses have not been available before 1993). In 1983-1985 CLIND was not coded at all in non-fatal events.
• The case ascertainment of non-fatal events was secured by checking all sickness leaves with suspicion of cardiovascular disease regularly against the MI register. Also, letters were sent to GPs once a year.
• There are 10 unresolved constraint violations and 32 Forms 01 with no Form 06 (coronary event records for which there is no serial inventory record).

BEL-LUX

ICD-version used: 9
Non-fatal case finding method: Mixed
Access to death certificates: Restricted

Comments:

• The MCC has not been able to complete the full ten years' of registration, and data for the last three years are missing.
• Fatal events: The coverage of fatal events is incomplete in 1985: CHD/CHD ratio low; it is presumed, that the high total/CHD registration compensates for this in 1986 and in 1987. The death certificate codes often change for the routine statistics after the MCC has reviewed them. According to the MCC, the Principal Investigator gets all the death certificates from persons dying in the area and identifies the interesting ones. Then he calls the certifying doctor and collects relevant information. The MCC has been validating the routine mortality statistics which seemed to have a lot of errors. The results are not available.
• The number of fatal events is lower in 1990-1991 than in the earlier years, but because of missing routine mortality statistics, the coverage of fatal registration cannot be assessed.
• There are about 15 hospitals treating the patients; the principal investigator visits the hospitals regularly and reviews the admission lists and picks the relevant cases. Until 1991 there were no official hospital discharge lists for the area.
• Routine mortality statistics are missing for the years 1988-1991.
• There are a lot of discrepancies in the serial number inventory.

CAN-HAL

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Restricted

Comments:

• Availability of data on previous MI is low, especially in fatal events.

CHN-BEI

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• Comparison with routine mortality statistics showed a lower number of fatal events in MONICA register in 1984-1986 and also in 1987 and 1989 (low CHD/CHD ratio in 1984-86; the number of registered fatal events is higher after 1990). However, there are three explanations:
  1. The routine mortality statistics have been very unreliable before MONICA which has been a key factor in developing them. Death certificate registration was not a legal requirement in China until 1990.
  2. In the early study period the MONICA team did not send data on false positive events to the MDC. The death certificates were written by administrators and were unreliable. The MONICA team was unable to register these missing events during the later years of the study.
  3. The size of the population is growing.
• The registration of non-fatal events is lower in 1984-1985 than during the later years. The size of the population is growing fast, however.
• All non-fatal events were registered from ICD codes 410-411 in 1984-1985; later, also other ICD codes were used. The exact cause of the change is unknown; however, before 1985 diagnostic tests were seldom done in suspected CHD, and therefore ICD codes 412-414 were not used in the absence of diagnostic evidence of CHD.
• The availability of data on enzymes and ECG was lower in 1993 than during previous years. There has been a trend in some hospitals for enzyme tests and ECGs to be taken less often. The reason is unknown to the MCC.
• The availability of data on  ECG and on enzymes was a little bit lower than 90% in 1990.
• Crude case fatality is relatively high during the whole data collection period.
• The proportion of non-fatal DIACAT=1 events of all registered non-fatal events is relatively high.
• The coverage of registration was checked every year using a door-to-door data collection procedure for a sample of 5% to 10% of the population in the RUA. Over 90% of events had been registered; however, there may be more missing events for the last year of registration (1993) and possibly for the first years of registration (1984-1985).
• There are some discrepancies in the serial number inventory.

CZE-CZE

ICD-version used: 9
Non-fatal case finding method: Mixed
Access to death certificates: Full

Comments:

• Almost all non-fatal DIACAT 1 and DIACAT 2 events are registered from CLIND 410-411; according to the MCC the explanation is that clinicians use only these diagnoses in acute hospitalizations.
• The proportion of non-fatal DIACAT=1 events of all registered non-fatal events is relatively high.

DEN-GLO

ICD-version used: 8
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• Fatal events: CHD/CHD ratio below one indicates that there are missing fatal events in 1987, 1988 and 1990-1991. Total fatal events registration compensates for this, however. The MCC has answered that they cannot register these missing fatals any more.
• Non-fatal events: availability of data on ECG is clearly lower in 1982 than during later years. The item NUMECG was not recorded in 1982, because the item not was introduced in the MONICA forms until later.
• The proportion of non-fatal DIACAT=1 events of all registered non-fatal events is relatively high in the first years of registration.
• The availability of data on >=1 ECG and on enzymes was a little bit less than 90% in 1989.
• Fatal events with no necropsy: in 1983-1984 the availability of data on previous CHD is much lower than during other years; the availability of data on previous MI is about the same, however, and the number of fatal DIACAT 1+2+9 events is about the same than during the previous and later years which indicates that this problem does not affect the diagnostic classification of fatal events.
• In fatal events the coding of the item MANAGE is probably not reliable in 1982-1983.

FIN-KUO

ICD-version used: 8/9 (During 1983-1986 ICD 8, from 1987 ICD 9)
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• None.

FIN-NKA

ICD-version used: 8/9 (During 1983-1986 ICD 8, from 1987 ICD 9)
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• None.

FIN-TUL

ICD-version used: 8/9 (During 1983-1986 ICD 8, from 1987 ICD 9)
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• None.

FRA-LIL

ICD-version used: 9
Non-fatal case finding method: Mixed
Access to death certificates: Restricted

Comments:

• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• The availability of data on >=3 ECGs was below 80% over all years, but well compensated with the high availability of data on enzymes and >=1 ECG; in 1987 the availability of data on >=1 ECG was a little bit less than 90%.
• Crude case fatality is relatively high, about 60%.
• Fatal events: there are several events coded MANAGE 9: about 90% of these are deaths at home but there was no information available to allocate a specific code for MANAGE.

FRA-STR

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Restricted

Comments:

• The MCC has provided data for 1984 but only data from 1985 are considered in this report, since the data from 1984 are not comparable with the later years, and the MCC has suggested that this year should be omitted. In 1984 the number of non-fatal DIACAT 2 events was clearly lower than during later years. The explanation is that the access to various documents was not so good as during the following years. The number of DIACAT 9 events was somewhat lower in 1984 than later, especially in fatal events with no necropsy. There were 110 death certificates with very incomplete information: the name of the patient was missing, as well as the name of the physician who signed the death certificate. Therefore it was not possible to validate these events and they were not sent to the MDC. According to the Principal Investigator many of these events probably would have ended up as DIACAT 9 if there had been more information available. Fatal events with no necropsy: the availability of data on previous CHD or MI and symptoms lower in 1984 than during later years. The explanation is that access to information concerning out-of-hospital deaths was not so good in 1984 and partly in 1985 than during later years.
• In 1985 there where quite a lot of fatal MANAGE=9 events. This is explained by the fact that access to information was not as good in 1985 as later.

FRA-TOU

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Restricted

Comments:

• The MCC has provided data since July 1984 but only full calendar years are considered in this report.
• The MCC has registered full nine years; it cannot complete the full ten years of registration because of constraints of finance and of obtaining data from private hospitals (especially, ECGs are given to the patients at discharge at these hospitals, and cold pursuit-approach is not feasible).
• In non-fatal events, the availability of data on ECG is lower from 1987 onwards than previously. This is explained by a change in the middle of 1987; key hospital staff members moved from the public hospital to a private hospital, and more patients later on were treated in the private hospital. In that hospital the practice was that ECGs were given to patients at discharge and were thus unavailable to the MCC personnel using cold pursuit-approach.
• There is a drop in fatal events in 1988; the same drop is seen in routine mortality statistics; there is no clear-cut explanation for this. The number of DIACAT F1+F2+F9 events declines more smoothly, however.

GER-AUG

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Restricted

Comments:

• The comments apply to both GER-AUR and GER-AUU.
• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• Fatal events are missing for 1994 in GER-AUR.
• Restricted access to death certificates explains why not all routine mortality statistics CHD deaths are registered. Total/CHD ratio compensates this except for 1994.
• Only events getting the diagnosis 410-411 are registered. After reviewing a broad range of admission diagnoses, the MCC team asks the attending physician for the diagnosis; if the physician thinks it is an MI the event is registered. Once a year all discharge diagnoses are reviewed and missing MIs are registered. Registration methods are described in detail in the publication 'Case finding, data quality aspects and comparability of myocardial infarction registers: results of a south German register study. J Clin Epidemiol 1991: 44: 249-260'.
• Crude case fatality is relatively high, about 60%.
• There are 12 suspected duplicates and  a few unresolved constraint violations.

GER-BRE

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• The registration of the last two years is missing.
• A few fatals may be missing in 1992 in RU 1 (CHD/CHD ratio below one); however, total fatal registration compensates for this.
• The proportion of non-fatal DIACAT=1 events from all registered non-fatal events is relatively high.
• Almost all non-fatal DIACAT 2 events are registered from CLIND 410-411.
• The coding of the item MANAGE differs from that used in other MCCs. Code 5 (medical consultation) is used quite often, code 4 (medically unattended death) seldom. In addition, code 9 is used too (according to the MCC, these events may have been hospitalized in a hospital outside the study area).

GER-EGE

ICD-version used: 9
Non-fatal case finding method: Mixed
Access to death certificates: Full

Comments:

• After the reunification of Germany, the frequency of necropsies fell in East Germany. Therefore the proportion of necropsies was lower during 1991-1993 than previously,
• 1984 differs from other years: the number of non-fatal DIACAT 2 events was lower, also the number of non-fatal DIACAT 1 events was somewhat lower than later on; the proportion of non-fatal DIACAT 4 events was higher.
• The number of ECGs was not recorded in 1984 and 1985, and data on these items as well as on HISIHD has only been collected later, which explains the low proportions.
• The availability of data on >=3 ECGs was below 80% over most of the years; however, the availability of data on >=1 ECG and on enzymes was 90% but somewhat less for enzymes in 1985-1988.
• There are 92 unresolved constraint violations, among them 12 DIACAT_CHECK_1 violations. There are 27 suspected duplicates.

GER-RHN

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Restricted

Comments:

• The MCC stopped event registration in 1988, and there has been no contact with the MCC since 1990.
• The coverage of fatal events is incomplete in 1987; CHD/CHD ratio is low in 1984-1985 but total fatal registration compensates for this.
• The proportion of non-fatal DIACAT=1 events from all registered non-fatal events is relatively high, nearly 80%.
• The number of DIACAT 9 events is much higher in 1984-1985 than later.
• The availability of data on ECGs (NUMECG) is lower in 1984-1985 than during the later years.
• In fatal events, the availability of data on previous CHD and MI is lower in 1984-1985 than subsequently.
• There are several discrepancies in the serial number inventory, 10 suspected duplicates, and  several unresolved constraint violations.  

HUN-BUD

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The coverage of fatal registration was possibly incomplete (low CHD/CHD ratio); total/CHD ratio compensates for this. In 1986 the number of total fatal registrations was lower than the number of CHD deaths in routine mortality statistics.
• The registration of non-fatal events is incomplete, which is also shown by the high case fatality. The MCC has noticed that about 30% of non-fatal events were treated in hospitals outside the monitoring area.
• Data on previous MI are available only from 1985 onwards.
• The data on NUMECG is probably missing before 1985.
• In fatal events without necropsy, the availability of data on previous CHD is low before 1985.
• In fatal events without necropsy, the availability of data on symptoms varies from 30% to 72%.
• There are 24 unresolved constraint violations and several discrepancies in the serial number inventory.

HUN-PEC

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The registration of fatal events is incomplete.
• Data on previous MI not available in 1984.
• There are unresolved constraint violations and discrepancies in serial number inventory. 

ICE-ICE

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• A few more deaths in routine statistics than in the MONICA register explained by errors in routine statistics.
• The proportion of non-fatal DIACAT=1 events of all registered non-fatal events is quite high and over 85% in 1992, which can occur by chance because the total number of events was low.
• Almost all non-fatal DIACAT 2 events are registered from CLIND 410-411; however, all discharges with CLIND 410-414 are screened; studies done in 1980, in 1984 and 1992 showed that all MIs by MONICA criteria have been coded 410-411.

ITA-BRI

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• In 1986 and 1987 the number of fatal events with CHD was a little less than the number in routine statistics. Total fatal registration compensates for this.

ITA-FRI

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• No comments.

LTU-KAU

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• The number of fatal events is lower in 1983 than afterwards, both in the MONICA register and in routine mortality statistics. The MCC has looked at the data very carefully and is convinced that this is the real situation; before 1983, the number of CHD deaths was even lower than in 1983.
• Fatal events: CHD/CHD ratio is below 1 in 1982, 1983 and 1991; however, total/CHD ratio compensates for this.
• The availability of data on enzymes in non-fatal events is between 62%-74% which is lower than in most other MCCs.
• The availability of data on >=3 ECGs was below 80% in 1988 and 1990-1990; however, the high availability of data on >=1 ECG compensates for this.
• Almost all non-fatal DIACAT 1 and DIACAT 2 events are registered from ICD codes 410-411.
• Crude case fatality is high which may indicate an incomplete case ascertainment of non-fatal events; however, the proportion of non-fatal DIACAT 4 events is quite notable which suggests complete registration.
• There are some discrepancies in the serial number inventory form: 266 forms 01 with no form 06 (coronary event records for which there is no serial inventory record)

NEZ-AUC

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• Fatal events: CHD/CHD ratio is below 1; explained by errors in routine statistics. Erroneous date of birth or code of residence was found in at least 32 records during the ten years. This represents the minimum of erroneous records in the routine mortality statistics and explains why the CHD/CHD ratio is below one. An additional problem is that the routine mortality statistics records the death in the year when the registry is notified but MONICA registers death in the year of event. On the other hand, total fatal registration compensates the low CHD/CHD ratio.
• Taking the above explanations into account, the registration of fatal events is probably complete over the whole time period; however, there are more fatal events missing in 1992 than earlier (between 4-11 missing fatal events/year; in 1992 there were 25 missing fatal events), which is the explanation for code '1' in the summary table.
• According to the MCC, in 1992 there were 17 fatal cases in the routine mortality statistics coded as CHD who refused to participate and could therefore not be registered. The number refusing to participate is clearly bigger in 1992 than previously (about 4/year).
• Non-fatal DIACAT 4 events have not been sent to the MDC. This explains why the proportion of non-fatal DIACAT=1 events from all registered non-fatal events is is relatively high.
• NUMECG was assigned '3' in the first years of registration in non-fatal events with ECG taken.
• The number of registered events fluctuates during the last years 753 (in 1990) - 1073 (in 1991) - 767 (in 1992). The reason is that in 1991 acute coronary care data were collected and therefore more non-fatal DIACAT 2 events were registered.
• The proportion of non-fatal DIACAT 2 events is higher in 1992; see the explanation above for the fluctuation.
• CLIND has not been recorded for non-fatal events.
• There are over 200 discrepancies in the serial number inventory. The MCC no longer has resources to locate these records and to correct this problem.

POL-TAR

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Restricted

Comments:

• The event rates and case fatality are not comparable with other populations because the availability of data for diagnostic classification of events, that is, on >=3 ECGs and on enzymes, is lower than in other MCCs; therefore most of the non-fatal events are DIACAT 2 events.
• The proportion of non-fatal DIACAT 4 events fluctuates too - see the explanation above.
• The number of registered events is lower in 1984-1985 than later on, both fatal and non-fatal. However, the number of CHD deaths in routine mortality statistics is also clearly lower in 1984-1985 than later on, which suggests that the numbers are true, especially for fatal events.
• The number of non-fatal DIACAT 1 events declined in 1992-1993. Also, CHD mortality in the routine statistics declined. According to the MCC the registration procedures did not change.
• The number of non-fatal DIACAT 2 events is much lower in 1984 than later. The reason is that there was a change in the coding of SYMPT. In 1984 SYMPT=1 were coded only when exact duration of pain was given in the medical record. Such detailed data was infrequent, and after 1985 wording in the medical record that indicated a long duration of pain was accepted as SYMPT=1 if there was no mention of pain duration in minutes. This mainly involved a shift from DIACAT 4 to DIACAT 2.

POL-WAR

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Restricted

Comments:

• CHD/CHD ratio below one in 1988-1989 and 1991-1992 indicates that some fatal events may be missing. Total fatal registration compensates.
• The availability of data on >=3 ECGs was below 80% during the registration period; however, the high availability of data on >=1 ECG and on enzymes compensates for this; for 1991-1994 the availability of data on >=1 ECG and on enzymes is somewhat less than 90%.
• The coding of the item MANAGE varies in fatal events.
• In fatal events the availability of data on previous CHD or MI was much higher in 1984-1985 than later.
• In non-fatal events availability of data on enzymes (76%-94%) and NUMECG>=3 (64%-80%) is somewhat lower than in most other MCCs.

RUS-MOC

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data for 1984 but has indicated that this year was a pilot year.
• The number of non-fatal events decreased while the number of fatal events was stable or increasing, which arises suspicion of incomplete case ascertainment. The same methods, however, were applied over the whole time period: 53 hospitals, 19 out-patient clinics and 12 morgues are covered. All hospitals where visited every year to control the quality of registration; also, the Bureau of Medical Statistics was visited every year to ascertain the completeness of registration. There are some 'closed medical institutions' which also treat MI patients; these are, however, inaccessible for ordinary people and very expensive. The hospitalization for MI decreased in Moscow because of the reorganization of primary care; after 1988 in 6 of the 8 biggest hospitals in Moscow the total number and the proportion surviving and hospitalized with diagnoses of ICD codes 410-411 decreased constantly. At the same time pre-hospital sudden (surviving less than 24 hours) deaths (definite or possible MI) increased in men from 85/100 000 (in 1985) to 197/100 000 (in 1993) and similarly in women from 11 to 32. The increase in sudden coronary deaths is in contradiction with the decrease in non-fatal events. Either the increase is real (in which case there is probably incomplete coverage of non-fatal events: there should be some increase in non-fatal events as well if the CHD rates were increasing) or there has been a change in the coding of death certificates from other diagnoses to CHD diagnoses. It seems probable that over time more non-fatal MIs went unrecognized because people did not go to hospital.
• Because of administrative changes in Moscow (population data cannot be calculated for the last year) the MCC will not be able to collect data for the last year of registration, and the coronary event registration covers years 1985-1993. Some data have been received for the year 1984; the coverage of registration is, however, clearly incomplete for 1984 (CHD/CHD and total/CHD ratio below one, CHD/CHD 0.79) and data for this year cannot be used in the analyses.
• In fatal events the proportion of MANAGE 9 (unknown) is quite high in 1985 compared to later years.
• Comparison with routine mortality statistics indicates that a few fatal events are missing, especially in 1990. The MCC has no possibility of completing the registration because some patients are treated in medical institutions to which the MCC has no access.
• In non-fatal events the availability of data on enzymes (59%-74%) is lower than in most other MCCs.
• Crude case fatality is quite high which may indicate an incomplete case ascertainment of non-fatal events. Therefore, the coverage of non-fatal events may be incomplete. The number of patients treated in hospitals has decreased and it may be possible that patients with suspected MI do not go to hospital as often as previously.
• The number of non-fatal DIACAT 1 events is clearly lower in 1991, in 1992 and especially in 1993 than during previous years; however, the proportion of DIACAT 4 events is notable; for hospitalized events the coverage of non-fatal events seems to be complete but non-hospitalized non-fatal MIs may be increasingly missed.
• According to the MCC, the drop in non-fatal events in 1991-1993 is real because the number of non-fatal events and hospital admissions with diagnoses 410-411 decreased in most of the hospitals. The reason may be, however, an increase in unrecognized non-hospitalized MIs.
• In fatal events the proportion of MANAGE 9 is quite high in 1985.
• In fatal events without necropsy the availability of data on symptoms, previous CHD or previous MI is clearly lower in 1985 than later.
• Almost all non-fatal events are registered from CLIND 410-411.
• There are a few discrepancies in the serial number inventory.

RUS-MOI

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments (see the comments for RUS-MOC, too):

• The MCC has provided data for 1984 but has indicated that this year was a pilot year.
• Comparison with routine mortality statistics indicates that a few fatal events are missing, especially in 1988 and 1992. The MCC has no possibility of completing the registration because some patients are treated in medical institutions in which the MCC has no access.
• Crude case fatality is quite high which may indicate an incomplete non-fatal case ascertainment especially in 1992 and 1993 in RU 2. See the comments for RUS-MOC.
• The number of non-fatal events declines and is quite low in 1992 in RU 2 which may indicate an incomplete case ascertainment. See the comments for RUS-MOC.
• In fatal events with no necropsy the proportion of DIACAT 9 events is much higher in 1985-1986 than later on.
• In fatal events with no necropsy the availability of data on previous CHD or MI was lower in 1985-1986 than later on.
• Almost all non-fatal DIACAT 2 events are registered from CLIND 410-411.
• In non-fatal events the availability of data on enzymes (60%-71%; somewhat less in RU 2) is lower than in other MCCs.
• There are a few discrepancies with the serial number inventory.

RUS-NOC

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data for 1984 but has indicated that this year is a pilot year.
• Because of a difficult financial situation, routine mortality statistics have been only obtained for RU 3 (i.e. RUS-NOCa) for the years 1991-1993. Therefore, the quality assessment could not be completed for RUS-NOCb (RUs 3 and 4) for the years 1991-1993.
• In non-fatal events NUMECG>=3 is low and drops to zero in 1988. The reason is that ECGs were not taken routinely more than once after 1988. In non-fatal events the availability of data on enzymes is lower in 1984-1987 than later on.
• The availability of data changed because of the rapid social changes in Russia.
• Fatals events are probably missing in the RUS-NOCa (RU 3), especially in 1993 but also in 1989 and 1991. The assessment of the completeness of the registration of fatal events was not possible for the years 1984 (both RUs) and for the years 1991-1993 for RU 4 because of incomplete routine mortality statistics.
• Crude case fatality is relatively high, above 60% over several years which might indicate an incomplete case ascertainment of non-fatal events. On the other hand, the proportion of non-fatal DIACAT=1 events of all non-fatal events is quite low which argues against incomplete case ascertainment.
• There are some discrepancies with the serial number inventory and 2 suspect duplicates.
• Routine mortality data are incorrect for RUS-NOCa (i.e. RU 3) in 1993.

RUS-NOI

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments (see the comments for RUS-NOC, too):

• In non-fatal events NUMECG>=3 drops to zero in 1988.
• Most of the registered fatal events belong to DIACAT 2.
• In fatal events the proportion of DIACAT 4 events varies by year. The proportion of necropsies done was high (62%-83%).  In non-fatal events the availability of data on enzymes was somewhat lower during the first years of registration, compared with later years. The availability of data changed because of the rapid social changes in Russia.
• In 1993 routine mortality data are incorrect. High case fatality in 1993; probably missing non-fatals.

SPA-CAT

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data for 1984 but considers the year 1984 as a pilot year since there may be incomplete case ascertainment; the number of non-fatal events is lower in 1984 than later.
• The number of registered non-fatal events rises steadily but there is no evidence of bias; the population is growing fast.
• The availability of data on three or more ECGs was below 80% in 1985; however, the availability of data on >=1 ECG and on enzymes was nearly 90% and compensates for this.

SWE-GOT

ICD-version used: 8/9 (ICD 8 until 1986, from 1987 ICD 9 )
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The MCC is registering events simultaneously for the MONICA MI register and for another MI register with its own criteria; this has resulted in some confusion. After a lot of work done in the MCC, the problems have been mainly resolved.
• There were problems with the classification of fatal events; however, all fatal events without necropsy have now been reclassified by one investigator using the same criteria.
• non-fatal DIACAT 4 events were not sent to the MDC which explains why the proportion of non-fatal DIACAT=1 events of all registered non-fatal events is relatively high.
• A few fatal events are missing in 1985. Otherwise, the registration of fatal events has been completed after a lot of work done in the MCC.
• The proportion of necropsies declined from 97% to 66%. A new law in Sweden lead to a drop in frequency of necropsies performed. Often, the GP writes the death certificate without any adequate information. This explains the increase in the proportion of DIACAT 9 events in the 1990s, and the lower availability of data on ECGs and enzymes (they were not taken) in fatal events with no autopsy.
• All non-fatal events have got the clinical diagnoses 410-411.
• In non-fatal events the year 1991 is different from the other years; the number and proportion of DIACAT 1 events is lower and DIACAT 2 higher. There is no explanation for that, however, and no indication for coding errors.
• The proportion of fatal events dying in hospital (MANAGE=1) decreased notably in the 1990s; the MCC could not find any coding errors.
• There are 30 unresolved constraint violations.
• There are discrepancies in the serial number inventory and 2 suspect duplicates.

SWE-NSW

ICD-version used: 8/9 (ICD 8 until 1986, from 1987 ICD 9 )
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• None.

SWI-TIC

ICD-version used: 8/9 (ICD 8 is used for non-hospitalized fatal events and ICD 9 for hospitalized events)
Non-fatal case finding method: Cold pursuit
Access to death certificates: Very restricted

Comments:

• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• The MCC has used its own derived codes for enzymes in events treated with thrombolytics using clinical diagnoses.
• Data on women were not collected. Data for men were collected for nine full years.
• In fatal events the coding of the item MANAGE differs in 1985-86 from later years. The reason for this problem is that the MCC had no possibility of obtaining more information on fatal events not found in their hospital register other than that included in the official computerized mortality records. For the years 1985-86 some information was available but the format of the mortality records changed in 1987 and this information was no longer available.

SWI-VAF

ICD-version used: 8/9 (ICD 8 is used for non-hospitalized fatal events and ICD 9 for hospitalized events)
Non-fatal case finding method: Cold pursuit
Access to death certificates: Very restricted

Comments:

• The MCC has provided data since Oct 1984 but only full calendar years are considered in this report.
• The MCC has used its own derived codes for enzymes in events treated with thrombolytics using clinical diagnoses.
• Data on women were not collected. Data for men were collected for nine full years.
• In fatal events the coding of the item MANAGE differs in 1985-86 from later years. The reason for this problem is that the MCC had no possibility of obtaining more information on fatal events not found in their hospital register other than that included in the official computerized mortality records. For the years 1985-86 some information was available but the format of the mortality records changed in 1987 and this information was no longer available.

UNK-BEL

ICD-version used: 9
Non-fatal case finding method: Mixed
Access to death certificates: Full

Comments:

• The number of fatal events is lower in MONICA register in 1986 compared with the number in routine statistics; the 'excess' deaths were registered in the routine mortality statistics from death which had occurred in 1985.
• The proportion of non-fatal DIACAT=1 events from all registered non-fatal events is relatively high.

UNK-GLA

ICD-version used: 9
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• CHD/CHD ratio below one in 1986; this is explained by errors in routine mortality statistics.
• Routine mortality statistics are missing in 1993.

USA-STA

ICD-version used: 9-CM
Non-fatal case finding method: Cold pursuit
Access to death certificates: Full

Comments:

• The number of fatal events drops in 1987, both in routine mortality statistics and in the MONICA register. The MCC has not found any explanation for this change.
• Both the number and the proportion of non-fatal DIACAT 1 events declines steadily over time. There is no indication of bias, however; the proportion of DIACAT 4 events is notable and about the same over time, crude case fatality is about the same level over time. Overall, more non-fatal events were registered later.

YUG-NOS

ICD-version used: 9
Non-fatal case finding method: Hot pursuit
Access to death certificates: Full

Comments:

• The MCC has provided data for 1983 but considers this first year of registration as a pilot year since the coverage of fatal events may be incomplete (CHD/CHD ratio below one). Also, the proportion of necropsies done is clearly lower in 1983 than later. The proportion of non-fatal DIACAT 1 events is higher in 1983 than later on.
• All non-fatal events have the clinical diagnoses 410-411.
• The proportion of non-fatal DIACAT=1 events from all registered non-fatal events is relatively high.
• In fatal events with no necropsy, the availability of data on previous CHD is lower during 1983-86 than later; however, the availability of data on previous MI is high and on symptoms about the same level as during the other years.
• The proportion of necropsies is lower in 1994-95 than in the previous years.

References to publications

1. Tunstall-Pedoe H for the WHO MONICA Project. The World Health Organization MONICA Project (Monitoring Trends and Determinants in Cardiovascular Disease): A major international collaboration. J Clin Epidemiol 1988;41:105-14.
2. Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas A-M, Pajak A for the WHO MONICA Project. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project. Registration procedures, event rates and case fatality in 38 populations from 21 countries in 4 continents. Circulation 1994;90:583-612.
3. WHO MONICA Project. MONICA Manual. Part IV: Event registration. Section 1: Coronary event registration data component. (March 1999). Available from: URL: http://www.ktl.fi/publications/monica/manual/part4/iv-1.htm, URN:NBN:fi-fe19981154.
4. WHO MONICA Project. MONICA Manual. Part II: Annual Statistics. Section 1: Population demographics and mortality data component. (December 1998). Available from: URL: http://www.ktl.fi/publications/monica/manual/part2/ii-1.htm, URN:NBN:fi-fe19981149.

References to internal MONICA documents

5. Pajak A, Kuulasmaa K, Tunstall-Pedoe H, Martin I, Rajakangas A-M for the WHO MONICA Project. Coronary event registration data quality report. MONICA Memo 173A, February 1990.
6. Mähönen M, Rajakangas A-M,  Kuulasmaa K, Tunstall-Pedoe H for the WHO MONICA Project. WHO MONICA Project - quality assessment of coronary event data for 1980-1990.  MONICA Memo 257A, February 1994.