WWW-publications from the WHO MONICA Project

Quality Assessment of Acute Coronary Care Data in the WHO MONICA Project

February 1999

Markku Mähönen1, Zygimantas Cepaitis 1and Kari Kuulasmaa1 for the WHO MONICA Project2

1 MONICA Data Centre, National Public Health Institute, Helsinki, Finland
2 Annex: Sites and key personnel of the WHO MONICA Project

© Copyright World Health Organization (WHO) and the WHO MONICA Project investigators 1999. All rights reserved.

This document includes the main findings of the unpublished reports:

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

The second MONICA hypothesis was set up to study the relationships between acute coronary care and 28 day case fatality rate (1). This report assesses the quality of the acute coronary care (ACC) data and explores the possibilities to use the data for testing the MONICA second hypothesis. The key issues considered are:

The ACC data are complementary to the coronary event registration data. Therefore,   good quality coronary event registration data is a prerequisite for the good quality of the ACC data. The quality assessment of the coronary event data has been reported separately (2). The current document assesses only the complementary ACC data.

The terminology used is this report follows that developed for MONICA event registration in the MONICA manual (3), with later refinement in the collaborative publications (4).

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) and Core Data Transfer Format - Acute Coronary Care (3) have been used. 

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 ACC data. The RUAs, their abbreviations and Reporting Units (RU) are listed in Table 1. In ACC data analyses, all RUs within each MONICA Collaborating Centre (MCC) are grouped together except in GER-EGE, where ACC data were collected in RU19 only. This report considers altogether 32 RUAs.

2.2 Time periods considered

ACC data were initially collected intermittently, with the intention of collecting data from 500 consecutive cases near the start and near the end of coronary event data collection. Because of the dramatic changes in acute coronary care in the late '80s, all except 9 RUAs decided to collect ACC data continuously since 1988 or 1989.

If the ACC data collection period differs between the RUs of a RUA, the data for the RUA will not represent individual RUs adequately, but an RU with a longer monitoring period will weight more in the analysis. To avoid the potential bias related to this, it is recommended that equal ACC data collection periods for each RU within a RUA should be considered in the data analysis. Therefore, periods of equal length for each RU within the RUAs were defined for this quality assessment. The time periods considered are shown in Table 1.

The analysis of standard time periods meant that otherwise relevant data were excluded from the following RUs:

2.3 Age and sex

The quality assessment concerns the age group 25-64 years. The age was calculated in full years at the date of onset. When calculating the age, day 99 was interpreted as 15 and day/month 99/99 as 30/06. The date of birth was reported both on the coronary event data (Form 01) and acute coronary care data (Form 02). If the dates of birth between the two data sets were discrepant the date of birth from the coronary event data (Form 01) was used. No age standardization was used. Data for men and women are combined in the analyses.

2.4 Other inclusion criteria

Individual records have been excluded from the analysis if DIACAT=4, because, according to the Manual, ACC data are expected to be sent to the MONICA Data Centre (MDC) only on events with DIACAT 1, 2, 3 or 9. Otherwise, all data available in the MDC were used in the analysis, regardless of their quality.

2.5 Sources of information

The report is based on the data which the MDC has received from the MCCs on acute coronary care (Form 02) and on coronary events (Form 01). In the first years of registration, version 3 of the Form 02 was used, and after 1989, version 6 (3). Data on ACE inhibitors (data items ACEB, ACED and ACEP), lipid lowering drugs (data items HYPOLB, HYPOLD and HYPOLP),  hospital stay (data item HSTAY) and place of death (data item PLOD) were added in version 6; also, data item INOD and NITROD were categorized in more detail.

3. Serial number inventory and routine data checking status

Table 2 shows a summary of the serial number inventory, which is based on a linkage of the coronary event data to the serial number inventory data  (3) received in the MDC. Its purpose is to check that the MDC database has exactly the records which it should have according to the MCC. Ideally, all entries in 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 several discrepancies in the serial number inventory in the RUAs BEL-GCH, CZE-CZEb, GER-RHNa, RUS-NOVa, SWE-GOTa and YUG-NOSa. In  RUS-NOVa and to a lesser extent in YUG-NOSa, there are quite a lot of serial number inventory forms indicating that ACC forms have been sent to the MDC for which,  there are no corresponding ACC forms in the MDC (Form 06 with COROCARE=1 and STATUS=1 but no Form 02).

When the data were received in the MDC they were routinely checked for the constraints specified in Appendix 1. For example, data on enzyme levels and ECG findings should match the coding of ECG and ENZ in the coronary event form. 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.  There are several unresolved constraint violations especially in:

The data with unresolved constraint violations have been included in the analyses for this report.

4. Coverage of registration

The coverage of the ACC data is complete for the defined time periods if there is a corresponding ACC form (form 02) for each coronary event form (form 01) belonging to diagnostic categories 1, 2, 3 and 9. The time periods are shown in Table 1. The coverage of ACC data is shown in Table 3.2. For comparison, Table 3.1 shows the number of coronary events registered and the number of corresponding ACC records without the limitation to the ACC data collection periods specified in Table 1. A summary of Table 3.2 is given in Table 12a. Summaries that restrict to MONICA coronary event Definition 1 and Non-fatal definite MIs respectively, are given in Tables 12b and 12c.

In most RUAs the coverage was good. There was, however, a notable proportion (>10%) of missing ACC forms in:

In CHN-BEI the proportion of missing data for the years 1984 and 1985 was a high 65%; in GER-RHN 30% in 1984; in RUS-MOS 34% in 1986; and in USA-STA 35% in 1986 and 28% in 1988.

5. Completeness of the data

While calculating the availability of data for the individual data items, these data items relevant to different topics were grouped together. Tables 4-5-6-7a-7b-7c-8-9 show the mean and the median of the proportion of missing data in the individual data items in such topics. If a proportion of missing data is the same for all items included in the corresponding column, then mean and median are the same; mean is sometimes higher than median which indicates that the proportion of missing data is clearly higher for one or a few items than for the other items.

5.1 Data on drugs and procedures

5.1.1 Non-fatal events

The mean and median of the proportion of missing data are shown in Table 4. For drugs, the mean and median were identical in almost all RUAs.  For procedures, the mean and median differed in some RUAs, which indicates that there were more missing data for one or some of the procedures, compared with the other procedures. (See the Section 10 for comments on each RUA.)  Summaries of Table 4 that restrict to non-fatal definite MIs are given in Table 13a for drugs before and in Table 13b for drugs during the event.

Overall, there were few missing data on drugs and procedures in non-fatal events; however, the post-event proportion was somewhat higher  in many RUAs. The median proportion of missing data on drugs was very high in GER-RHN, and high in POL-TAR and POL-WAR on drugs before and post-event.

The median proportion of missing data on drugs before the event was over 10% in:

The median proportion of missing data on drugs during the event was over 10% in:

The median proportion of missing data on drugs post-event was over 10% in:

There was a change in coding in CAN-HAL in 1986: before that, there were no insufficient data. In SPA-CAT the median proportion of insufficient data on drugs was higher before and post-event than during the event, when it was very low. There was a notable declining trend in the proportion of insufficient data on drugs before the event in SWE-NSW.

The median proportion of insufficient data on procedures before and during the event was very low in most RUAs. No data on procedures were available in GER-RHN, and in POL-TAR before the event.

The median proportion of insufficient data on procedures before the event was over 10% in:

The median proportion of insufficient data on procedures during the event was over 10% in:

Data on PACEB were not collected in the early study period in AUS-NEW and AUS-PER which explains the higher proportion of missing data (see Section 10, comments for individual RUAs).

5.1.2 Fatal events

The proportion of missing data on drugs and procedures is shown in Table 5 for fatal events. In the category 'all fatal events', missing data were common. The notable exceptions with median of missing data below 10% over time were CHN-BEI, FIN-FIN, UNK-BEL and UNK-GLA.

In hospitalized fatal events the median proportion of missing data for drugs and procedures before the event was above 30% for one or more years in 13 RUAs, and for drugs and procedures during the event in 11 RUAs. In fatal events admitted to the CCU the median proportion of missing data  was over 30% for one or more years in 9 RUAs for drugs and procedures before the event, and over 30% for one or more years in five RUAs (BEL-GCH, CHN-BEI,  GER-AUG, GER-RHN and RUS-NOV) for drugs and procedures during the event (Table 5).

It should be noted that the numbers of hospitalized fatal events and fatal events admitted to the CCU are quite small in many RUAs and the proportion of missing data may be relatively high, even though the number of missing data is small.

Summaries of Table 5 for all fatal events are given in Table 13c for drugs before and in Table 13d for drugs during the event.

5.2 Data on ECG, enzymes and other items

The proportion of insufficient data on ECG items, enzyme items, hemodynamic items, items on cardiac arrest and resuscitation and other items are shown in Table 6 for non-fatal events and in Tables 7a-7b-7c for fatal events.

5.2.1 Non-fatal events

5.2.1.1 Data on ECG

Data on ECG were very complete in most RUAs (Table 6). The mean and median did not differ suggesting that the data were consistent. The proportion of missing data was over 10% in:

5.2.1.2 Data on enzymes

Data on enzymes were also very complete (Table 6). The median proportion of missing data was over 10% only in:

5.2.1.3 Data on hemodynamic items

Data on hemodynamic items (SYSBP and PULSE; Table 6) were also complete in most RUAs, the exception being GER-RHN with over 50% of missing data . The median proportion of missing data was over 10% in:

5.2.1.4 Data on resuscitation

Data on resuscitation (CAROUT, RESOUT, RESARR, CARIN, RESIN; Table 6) were very complete, the only exception being GER-RHN with no useful data on resuscitation.

A summary of Table 6 for ECG, enzymes and resuscitation is given in Table 13e.

5.2.1.5 Data on other items

The median proportion of missing data on other items (TIME, INITC, CUNIT, CSTAY; Table 6) was over 10% in AUS-NEW (1988) and YUG-NOS (1995). There was variation between the items, however, and the mean was over 10% in AUS-NEW (1988),   BEL-GCH (1987), GER-RHN (1984-1988), SWE-NSW (1986-1987) and YUG-NOS (1995) (see Section 10 for comments on individual RUAs). This indicates that there can be quite a lot of missing data on one or several items.

5.2.2 Fatal events

The proportion of missing data is tabulated separately for all fatal events (Table 7a), hospitalized fatal events (Table 7b), and hospitalized fatal events admitted to the CCU (Table 7c) because such categories of events might be used separately for analyses, and the proportion of missing data may vary between the categories. The number of hospitalized events and especially the number of events admitted to the CCU are quite small. Therefore, the proportion of missing data may be relatively high even though the number of missing data is small. Anyway, if the proportion of missing data is over 30%, analyses will probably be too biased to be useful.

In fatal events, data on ECG and enzymes are relevant only in hospitalized events. For these, however, the proportion of insufficient data is shown also for all fatal events (Table 7a) to see if there are any discrepancies or coding inconsistencies. Comparing tables 7a-7b-7c, the coding seems to have been quite consistent, and the proportion of insufficient data diminishes as expected in hospitalized events, and in events admitted to the CCU. In all fatal events a change in coding pattern was noted in GER-BRE in 1988: after 1988 the proportion of insufficient data was zero. In some Centres the proportion of missing data was higher in hospitalized fatal events, compared with the category all fatal events; the explanation may be that in out-of-hospital deaths it was known that ECG and enzymes were not taken but in hospitalized fatal events this was not always known. This is seen in particular in ICE-ICE, RUS-NOV and YUG-NOS.

A summary of Table 7a for ECG, enzymes and resuscitation is given in Table 13f.

5.2.2.1 Data on ECG

The median proportion of missing data on ECG in hospitalized fatal events (Table 7b) was above 30% in

The median proportion of missing data on ECG in hospitalized fatal events admitted to the CCU (Table 7c) was above 30% in

5.2.2.2 Data on enzymes

The median proportion of missing data on enzymes in hospitalized fatal events was above 30% in 12 RUAs for one or more years (Table 7b).

The median proportion of missing data on enzymes in hospitalized fatal events admitted to the CCU (Table 7c) was above 30% in

5.2.2.3 Data on hemodynamic items

The median proportion of missing data on haemodynamic items in all fatal events was above 30% in 26 RUAs for one or more years (Table 7a).

The median proportion of missing data on haemodynamic items in hospitalized fatal events was above 30% in 20 RUAs for one or more years (Table 7b).

The median proportion of missing data on haemodynamic items in fatal events admitted to the CCU (Table 7c) was above 30% in

5.2.2.4 Data on resuscitation

The median proportion of missing data on resuscitation (CAROUT, RESOUT, RESARR, CARIN, RESIN) in all fatal events (Table 7a) was above 30% in

The median proportion of missing data on resuscitation in hospitalized fatal events and fatal events admitted to the CCU was about 20-30% in GER-AUG for several years (Tables 7b and 7c). Data on resuscitation were missing in GER-RHN.

5.2.2.5 Data on other items

The median proportion of missing data on other items (TIME, INITC, CUNIT, CSTAY) in all fatal events (Table 7a) was above 30% in

The median proportion of missing data on other items (TIME, INITC, CUNIT, CSTAY) in hospitalized fatal events (Table 7b) was above 30% in

The median proportion of missing data on other items (TIME, INITC, CUNIT, CSTAY) in fatal events admitted to the CCU (Table 7c) was below 30% in all RUAs.

5.3 Data on items introduced after the beginning of the study

The data items ACEB, ACED, ACEP, HYPOLB, HYPOLD, HYPOLP, HSTAY and PLOD were introduced in MONICA after the ACC data collection had been going on for some years. To get an idea of the possibilities of using these data items, the proportions of missing data are shown in Table 8. In this table, missing data includes all unfilled entries and insufficient data-responses (code 9). A high proportion means that data on this particular item were not collected. Answers to a query about the data collection on these items are documented in Section 10.

5.4 Data on ACCTIME, SMOKE and REHABP

Data on these items showed a varied pattern and were therefore tabulated separately (Table 9).

5.4.1 Non-fatal events

The proportion of insufficient data on ACCTIME was below 10% in all years  in   AUS-NEW (except 1985), GER-EGE, ITA-FRI and RUS-NOV. In addition to these RUAs, the proportion was below 20% in FIN-FIN and UNK-BEL.

The proportion of insufficient data on SMOKE was over 10% in BEL-GCH, CAN-HAL, CHN-BEI (high in 1984-1985), CZE-CZE, DEN-GLO,  FRA-LIL, GER-EGE, GER-RHN, ICE-ICEb, ITA-BRI, POL-TAR, POL-WAR, RUS-MOS, SWE-NSW (1986-1987, 1995) and USA-STA (1981).

Data on REHABP were not collected in AUS-NEW (1988-1993) and GER-RHN (1984-1985, 1987-1988). There were a lot of insufficient data on REHABP in POL-TAR (1989), SWE-NSW (1987) and in USA-STA (1990-1991).

5.4.2 Fatal events

The proportion of insufficient data on ACCTIME was above 30% in one or several years in 20 RUAs (Table 9).

The proportion of insufficient data on SMOKE was above 30% in most RUAs. It was below 30% over all years in FIN-FIN, ITA-FRI, NEZ-AUC and RUS-NOV.

Obviously, data on ACCTIME and SMOKE in fatal events cannot be used in most RUAs.

6. Reliability of the data

The reliability of the data is assessed using three indicators:

6.1 The internal consistency of the data

Logical errors and inconsistencies of the data were checked routinely when the data were received in the MDC. The checking procedure and the currently unresolved constraint violations are described in Section 3.

6.2 Sudden changes in insufficient data

Notable and sudden changes in the proportion of insufficient data may indicate problems in the availability of data or coding of the data, both of which have an impact on the reliability of the data.

6.2.1 Non-fatal events

6.2.1.1 Drugs and procedures

Sudden changes in the median proportion of missing data (Table 4; over 5 % change in the absolute proportion) were noted in these RUAs:

See Section 10 for specific comments and explanations on individual RUAs.

6.2.1.2 Other items

Sudden changes in the median proportion of missing data (Table 6; over 5 % change in the absolute proportion) were noted in:

6.2.2 Fatal events

6.2.2.1 Drugs and procedures

The proportions of missing data on drugs and procedures before the event were quite high and are not commented on here. Sudden changes in the median proportion of missing data on drugs and procedures during the event (Table 5; over 10 % change in the absolute proportion) were noted in all fatal events in RUAs:

The numbers in hospitalized fatal events and especially in fatal events admitted to the CCU were small. This explains the realitvely high variation in the proportion of missing data for such events.

6.2.2.2 Other items

Sudden changes in the median proportion of missing data on other items in all fatal events (Table 7a; over 10 % change in the absolute proportion) were noted in:

6.3 The proportion of insufficient data

A high proportion of insufficient data may result from problems with access to medical records, low quality of medical records or other unknown reasons. Whatever the reason, a high proportion of insufficient data weakens the reliability of the data and invalidates data analyses. RUAs with a high proportion of insufficient data are discussed in the following. The proportions were also commented upon in Section 5.

6.3.1 Non-fatal events

The data on drugs and procedures from GER-RHN cannot be used because almost all data are coded as insufficient. The proportion of insufficient data was also very high in POL-TAR for drugs and procedures before the event, and post-event were relatively high but below 30 %. The proportion was over 30% in POL-WAR before the event and between 20-30 % post-event. The proportion was between 10-15 % post-event in SPA-CAT. The proportion of insufficient data on procedures before the event was high in AUS-PER in 1984-1988.

6.3.2 Fatal events

The median proportions of insufficient data were commented upon in Section 5.1.2. For most RUAs (total 32) the proportion of insufficient data on drugs before the event was so high that the data probably cannot be used in analyses. The proportion of insufficient data on drugs and procedures before the event was below 10 % in 4 RUAs. The median proportion of insufficient data on haemodynamic items was below 10 % in 8 RUAs. The median proportion of insufficient data on resuscitations was below 10 % in 22 RUAs.

7. Comparability of the data over time

The comparability of the accuracy of the data over time is important for the assessment of trends. The proportion of missing data probably reflects the accuracy  quite well. A trends in this proportion may also bias trend analyses. The proportions of missing data were commented on previously in Sections 5 and 6; here they are reviewed from the point of view of trends.

7.1 Non-fatal events

7.1.1 Drugs and procedures

A trend of over 5 % change in the absolute proportion was noted in (see Table 4):

7.1.2 Other items

A trend of over 5 % change in the absolute proportion was noted in (see Table 6):

7.2 Fatal events

7.2.1 Drugs and procedures

In all fatal events a declining trend in the median proportion of insufficient data on drugs and procedures during the event was noted in AUS-NEW, DEN-GLO, FRA-LIL and SPA-CAT (see Table 5).

7.2.2 Other items

In all fatal events a declining trend in the median proportion of insufficient data was noted in FRA-LIL, ITA-BRI (enzymes) and SPA-CAT (1986-1989 higher) (see Table 7a).

8. Data on HISIHD and PREMI

Previous CHD is an important predictor of drug use before the event. Previous MI may affect the treatment decisions during the acute event. In many ACC analyses, data will be stratified according to previous CHD and MI. Therefore, data on the coronary event items HISIHD (history of previous CHD) and PREMI (previous MI) are tabulated for the ACC data events in Tables 10 and 11. Data on previous CHD (HISIHD) were only collected on fatal events in the beginning, but later on it was recommended that the MCCs should also collect data on previous CHD in non-fatal events.

8.1 Non-fatal events

Table 10 shows the proportions of different categories of HISIHD in non-fatal events and gives an idea of the possibility to use this data item. The code 8 (not relevant, data on non-fatal events not collected; in the initial years of the study data on HISIHD was collected only on fatal events but this code was later removed and it was recommended that data on previous CHD should also be collected on non-fatal events) is used in:

The other RUAs have data on HISIHD throughout the ACC data collection period. In these RUAs, the coding of the item HISIHD seems to be consistent over the years. Insufficient data were uncommon, the exceptions being AUS-PER (especially in 1984-1986), CAN-HAL, CZE-CZE in 1986, POL-TAR, and to a lesser extent USA-STA (declining trend).

Data on PREMI in non-fatal events were complete in almost all RUAs (see Section 6  of the quality assessment of coronary event registration data in the WHO MONICA Project, 2) .

8.2 Fatal events

Table 11. shows the proportions of insufficient data on HISIHD and PREMI in all fatal events, in hospitalized fatal events (MANAGE=1) and in fatal events admitted to the CCU (CUNIT=1).

In all fatal events the proportion of insufficient data on either PREMI or HISIHD was over 30% in 18 RUAs for one or several years (Table 11).

In hospitalized fatal events (MANAGE=1) the proportion of insufficient data on either PREMI or HISIHD was over 30% for one or several years in:

In fatal events admitted to the CCU (CUNIT=1) the proportion of insufficient data on either PREMI or HISIHD was over 30% in:

9. Summary

The coverage of ACC data over time is summarized in Tables 12a -12b - 12c.The coverage was good in most RUAs. In CHN-BEI the coverage of the years 1984-1985 was very incomplete and perhaps the data for these years cannot be used. More deficient coverage was noted in the first year of registration in BEL-GCH, GER-RHN and RUS-MOS, compared with later years. The other RUAs with deficient coverage were CZE-CZE, GER-EGE, NEZ-AUC, SWE-GOT and USA-STA. Trends from these RUAs may be biased which should be taken into account when analysing the data.

The completeness of data on drugs is summarized in Tables 13a -13b - 13c - 13d, and on some other important items (ECG, enzymes and resuscitation) in Tables 13e and 13f.  In non-fatal events, the ACC data were quite complete. However, in fatal events the proportion of missing data was quite high. For drugs and procedures before the event, the proportion was over 30% in most RUAs which is a serious problem for data analyses.

Overall, the reliability of the data seems to be quite good. However, there were unresolved constraint violations revealing internal inconsistencies in the data in several RUAs. Especially problematic is the situation in RUS-NOV. Also, there were sudden changes in the coding in many RUAs. According to the answers from the MCCs one explanation for the high proportion of missing data in some MCCs is that codes '9' and '2' have to some exten, been used interchangeably. This may also explain the sudden changes in coding. However, the codes are specified in the MONICA Manual and such changes may also indicate other problems in the reliability of the data - over a long time period changes in personnel and in the interpretation of available evidence may occur.

A high proportion of insufficient data, also in non-fatal events, was seen in POL-TAR, POL-WAR and in particular in GER-RHN. The use of data of these RUAs should be based on individual decisions and will depend on which data items are required.

Especially important for the assessment of trends is the comparability of data over time. The data should be equal in accuracy (or inaccuracy) during the study period. The comparability was assessed using the proportion of missing data as an indicator of the accuracy of the data. The results show that the proportion of missing data declined in many MCCs over time (Tables  13a -13b - 13c - 13d). The quality of medical records has improved notably in many MCCs over the ten year period. Also, the involvement of the register team in the documentation of treatments has resulted in better quality of medical records. Although the decline in the proportion of missing data improves the accuracy of cross-sectional estimates, it does not remove the possible bias in the estimates of trends.

Considering the second MONICA hypothesis, since the proportion of missing data in fatal events is quite high in most RUAs, approaches using only the non-fatal events to investigate the impact of the changes in the acute coronary care and in the treatment of CHD on CHD mortality should also be explored.

Even though the ACC data are not ideal for the testing of the MONICA second hypothesis, at least directly, and even though the proportions of insufficient data vary and are quite high in fatal events, the ACC data are valuable and useful for many other analyses. In particular, data on non-fatal events are very complete. Coronary event data on previous CHD and previous MI are also quite complete for non-fatal events and allow the stratification of data by previous CHD. The absolute proportions of insufficient data as well as changes in coding and possible trends should be taken into account while analysing the data,  before deciding wh RUAs could be included in the planned analyses. The persons analysing the data should read this report carefully and explore the possible biases in the results caused by missing data.

10. Comments on individual RUAs

(ACE* and HYPOL* refer to the items ACEB, ACED, ACEP, HYPOLB, HYPOLD and HYPOLP.)

AUS-NEWa

Data collection method: Hot pursuit
Comments:

AUS-PERb

Data collection method: Cold pursuit
Comments:

BEL-GCHa

Data collection method: Hot pursuit
Comments:

CAN-HALa

Data collection method: Hot pursuit
Comments:

CHN-BEIa

Data collection method: Cold pursuit
Comments:

CZE-CZEb

Data collection method: Mixed
Comments:

DEN-GLOa

Data collection method: Cold pursuit
Comments:

FIN-FINa

Data collection method: Hot pursuit
Comments:

FRA-LILa

Data collection method: Mixed
Comments:

FRA-STRa

Data collection method: Cold pursuit
Comments:

FRA-TOUa

Data collection method: Cold pursuit
Comments:

GER-AUGa

Data collection method: Hot pursuit
Comments:

GER-BREb

Data collection method: Cold pursuit
Comments:

GER-EGEd

Data collection method: Mixed

GER-RHNa

Data collection method: Hot pursuit
Comments:

ICE-ICEb

Data collection method: Cold pursuit
Comments:

ITA-BRIa

Data collection method: Hot pursuit
Comments:

ITA-FRIa

Data collection method: Cold pursuit
Comments:

LTU-KAUa

Data collection method: Cold pursuit
Comments:

NEZ-AUCa

Data collection method: Hot pursuit
Comments:

POL-TARa

Data collection method: Cold pursuit
Comments:

POL-WARa

Data collection method: Cold pursuit
Comments:

RUS-MOSa

Data collection method: Cold pursuit
Comments:

RUS-NOVa

Data collection method: Hot pursuit
Comments:

SPA-CATa

Data collection method: Cold pursuit
Comments:

SWE-GOTa

Data collection method: Hot pursuit
Comments:

SWE-NSWa

Data collection method: Cold pursuit
Comments:

SWI-SWIa

Data collection method: Cold pursuit
Comments:

UNK-BELa

Data collection method: Mixed
Comments:

UNK-GLAa

Data collection method: Cold pursuit
Comments:

USA-STAa

Data collection method: Cold pursuit
Comments:

YUG-NOSa

Data collection method: Hot pursuit
Comments:

References

  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. Mähönen M, Tolonen H, Kuulasmaa K, Tunstall-Pedoe H, Amouyel P for the WHO MONICA Project. Quality assessment of coronary event registration data in the WHO MONICA Project. (January 1999). Available from: URL: http://www.ktl.fi/publications/monica/coreqa/coreqa.htm, URN:NBN:fi-fe19991072
  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. 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 rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583-612.