WWW-publications from the WHO MONICA Project
February 1999
Anu Molarius1, Kari Kuulasmaa1, Alun Evans2, Evelyn McCrum2 and Hanna Tolonen1 for the WHO MONICA Project3
1 MONICA Data Centre, National Public Health Institute, Helsinki, Finland
2 Department of Epidemiology and Public Health, The Queen's University of
Belfast, United Kingdom
3 Annex: Sites and key personnel of the WHO MONICA
Project
This document includes the main findings of unpublished reports:
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. 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.
The main hypothesis of the WHO MONICA Project is to assess whether 10-year trends in incidence and mortality from cardiovascular disease are related to changes in known risk factors. Smoking is one of these risk factors. It is therefore of especial importance to assess carefully the quality of the available data on smoking behaviour to be able to give reliable estimates of smoking trends in the different MONICA populations.
The quality assessment of smoking data in the initial (first) survey of the WHO MONICA Project was distributed in January 1992 (5) and the quality assessment of the middle (second) survey in September 1993 (6). Both these quality assessments revealed that the data on smoking behaviour gathered in MONICA are very complex. In general, almost all of the populations can provide comparable data on the number of regular cigarette smokers, but the usefulness of the other data on smoking behaviour is more limited due to differences in the questionnaires used and changes to them between the surveys. Also the standard MONICA smoking questionnaire and the respective data transfer format used for submitting the data for central storage have been revised several times during the MONICA lifespan.
In this report we give an assessment of the quality of the data available and recommendations upon the use of the different smoking items in cross-sectional and longitudinal analyses. This is mainly done by 1) explaining which data were gathered in the different surveys i.e. describing the standard questionnaire, the data transfer format and changes which have been introduced to them between the surveys, 2) assessing the quality of smoking questionnaires used and the smoking data collected in the final MONICA survey, and 3) drawing comparisons with the two earlier surveys by summarizing the main results of the quality assessments of the initial and middle surveys. The focus will be both on the cross-sectional approach and on the 10-year trends in smoking.
The general purposes of the unpublished versions of the quality assessment reports were:
This final version is issued at the time when the MONICA database for the smoking items is closed for the final analyses of the Project and therefore the purpose 2 is no longer valid.
The results of this document are reported by Reporting Unit Aggregates (RUA) which are potential units of analyses of the MONICA data. The RUAs, their abbreviations and reporting units are listed in Table 1.1. Some of the RUAs have several versions because different combinations of Reporting Units (RU) may be used for cross-sectional and trend analyses if all reporting units of the population were not included in all three or two surveys. Therefore, in AUS-PER, GER-BRE, GER-EGE, GER-KMS, GER-RDM, RUS-MOI and RUS-NOC there is an overlap of reporting units included in the RUAs in some surveys. The RUAs are identified by the abbreviation and a version letter. For UNK-GLAa which carried out four surveys, the first (initial), third (middle) and fourth (final) survey are considered.
Compared with other survey quality assessment reports, in the current document GER-EGE, GER-KMS and GER-RDM have been split into smaller RUAs because of the use of different smoking questionnaires in the different Reporting Units. Altogether 57 RUAs are considered for the initial MONICA survey, 45 for the middle survey, and 43 for the final survey.
The information presented in this report was gathered from several different sources. The original questionnaires used by the RUAs and their English translations were available in the MONICA Data Centre for almost all RUAs (Table 1.2). The Questionnaire on MONICA Population Survey Procedures (Form VI) was completed by the MCCs in 1991 and further checked by them regarding the final survey in 1995. Finally, we also looked at the data on smoking behaviour collected by the MCCs.
For the quality analyses all observations within the age group 25-64 were used, except in FRA-LILa in the final survey and in AUS-NEWa, BEL-LUXa, FRA-STRa, FRA-TOUa, LTU-KAUa, NEZ-AUCa, POL-TARa, POL-WARa, RUS-MOCa, RUS-MOIa, RUS-MOIb and SWI-TICa where the age range studied was 35-64. Age was defined as age in full years on the date of examination (see DEF1 in Reference 1). No weighting was applied to the data.
To gather data on smoking behaviour a questionnaire (7) was recommended for the initial and middle surveys (see Appendix 2a). The questionnaire was based on that given in the book by Rose GA et al. (2) Data was collected on present and past cigarette smoking, number of cigarettes smoked, age of starting to smoke, and pipe and cigar smoking.
A new version of the smoking questionnaire was introduced in the November 1990 version of the MONICA Manual (8, see Appendix 2b). In this version two new questions were added: one enquiring the number of days per week the current smoker smoked cigarettes and another about the number of hours per week the subject was subjected to other people's tobacco smoke (passive smoking).
The latest version of the smoking questionnaire was introduced after the middle surveys in March 1992 (3, see Appendix 2c). The smoking questionnaire was revised after an international training seminar for MONICA risk factor measurements was organized in Helsinki in November 1991. In this seminar several illogicalities of the old questionnaire were detected. For example, it was not known how the question "How old were you when you began to smoke cigarettes?" was interpreted by the interviewers between the Centres or between the surveys. Consequently the question was revised to explicitly refer to the age of starting regular smoking. Similarly the question about stopping smoking was revised to refer to regular smoking. Also, it was noticed that the distinction between occasional past smoking and never smoking was unclear and the question on 'ever smoking' was therefore also revised to refer to regular smoking. The question on number of smoking days per week was revised to be asked only from occasional smokers. In addition, it was noted that there was no way of knowing about the possible past regular smoking of the current occasional smokers. Therefore the question of 'ever smoking' was changed to be asked also from current occasional smokers.
The local questionnaires used in the MCCs were, however, not always similar to that recommended in the MONICA Manual. These deviations will be dealt with in Section 4.
Data on smoking behaviour are sent to the MDC using the MONICA Core Data Transfer Format - Survey Data (Form 04, versions 3,6,7). The initial survey used version 3 (9, see Appendix 3a) of the data transfer format. This version had seven smoking items (CIGS, NUMCIGS, EVERCIG, STOP, IFLYEAR, CIGAR, PIPE). Four more items were asked in the initial version of the standard smoking questionnaire (MAXCIGS, CIGAGE, CIGARSM and PIPESM), but unfortunately they were not included in the earliest version of the data transfer format.
The middle survey used version 6 (8, see Appendix 3b) of the data transfer format. This version had thirteen smoking items. Seven of them were exactly the same as in version 3. MAXCIGS, CIGAGE, CIGARSM and PIPESM were included. There were two new items: DAYCIGS and OTHERSM.
The latest version of the smoking questionnaire was incorporated in version 7 (3, see Appendix 3c) of the data transfer format. There were no new items in this version, but the variables and respective answer categories were revised to correspond to the revised questionnaire.
The data transfer format versions used by the MCCs for providing the data for the different MONICA RUAs in the three surveys are given in Table 1.2.
Tables 2.1 and 2.2 give a summary of the data items of the local questionnaires in each survey by RUA when compared with the corresponding data transfer format version (3,6,7). Table 2.1 gives the results for the seven items common to all the three versions and Table 2.2 for those items which were included only in the last two versions of the data transfer format. Each data item was coded as to how well it could be derived from the questionnaire applied. Five codes were used:
Table 3 shows a description of the local questionnaires used in each RUA in the initial MONICA survey and changes introduced to them between the surveys.
Thirty out of the 57 RUAs had identical questions for all seven items. They were CHN-BEIa, CZE-CZEa, FRA-LILa, GER-BERa, GER-COTa, GER-EGEc, GER-HACa, GER-KMSd, GER-RDMf, GER-RHNa, HUN-BUDa, HUN-PECa, ICE-ICEa, ISR-TELa, ITA-BRIa, POL-TARa, POL-WARa, ROM-BUCa, RUS-MOCa, RUS-MOIa, RUS-MOIb, RUS-NOCa, RUS-NOCb, RUS-NOIa, SPA-CATa, SWE-GOTa, SWE-NSWa, SWI-TICa, SWI-VAFa and UNK-GLAa.
In FIN-KUOa, FIN-NKAa, FIN-TULa, MLT-MLTa and USA-STAa the logic of the questionnaire differed significantly from the MONICA standard questionnaire. In GER-EGEd, GER-ERFa, GER-KMSc and GER-RDMe there were only two questions about smoking: category of smoking and number of cigarettes smoked. Also the questionnaires used in DEN-GLOa, GER-BREa and LTU-KAUa had very few items similar to those in the standard questionnaire.
The most serious problems concern the items which are crucial for defining who is a "smoker" (CIGS and EVERCIG). Typical deviations were:
NUMCIGS frequently excluded occasional smokers. In DEN-GLOa, FIN-KUOa, FIN-NKAa, FIN-TULa and MLT-MLTa the question was also asked from past smokers ("How much do you or did you smoke on an average day?"). STOP sometimes enquired the age, not the year, of quitting (BEL-LUXa) or as in LTU-KAUa, had a different wording ("How many years ago did you stop smoking?"). AUS-NEWa, AUS-PERa and NEZ-AUCa combined STOP with IFLYEAR by asking the month besides the year of quitting. GER-BREa used different options for IFLYEAR (-3, 4-6, 7-12 months). LTU-KAUa did not enquire about pipe or cigar smoking. In GER-EGEd, GER-ERFa, GER-KMSc and GER-RDMe pipe and cigar smoking was one of the smoking categories asked, but there was no question about amounts of pipe and cigars smoked. Otherwise PIPE and CIGAR revealed only minor deviations, such as asking daily instead of weekly amounts.
Six out of the 45 RUAs had identical questions for at least ten of the 13 items. They were SPA-CATa (12), HUN-BUDa, HUN-PECa, CHN-BEIa, SWE-GOTa and UNK-GLAa (10). Additional 18 (CZE-CZEa, GER-BERa, GER-COTa, GER-EGEc, GER-EGEd, GER-ERFa, GER-HACa, GER-KMSc, GER-KMSd, ICE-ICEa, ITA-BRIa, POL-TARa, POL-WARa, RUS-MOCa, RUS-MOIa, SWE-NSWa, SWI-TICa and SWI-VAFa) had identical questions for the seven items included in the initial survey data transfer format (Table 2.1).
For items CIGS, NUMCIGS, EVERCIG, STOP, IFLYEAR, CIGAR and PIPE most of the comments presented for the initial survey hold. In the following RUAs there were changes concerning these items:
Table 2.2 gives the summary codes for the six items which were new in the version 6 of the data transfer format. Lack of data is quite common for items MAXCIGS, CIGAGE, CIGARSM and PIPESM. Several MCCs dropped MAXCIGS or CIGAGE or both, and combined CIGARSM with CIGAR and PIPESM with PIPE, even though they asked these questions in the initial survey. Some MCCs may have misunderstood the importance of these items because they were not included in the earlier version of the data transfer format. Some of the MCCs used the question on MAXCIGS as recommended in the initial survey questionnaire i.e. excluded current smokers (CHN-BEIa, GER-AURa, GER-AUUa and SPA-CATa). CZE-CZEa did not specify the one year period ("What was the average number of cigarettes in the time you smoked most?"). Some MCCs excluded occasional past smokers for CIGAGE. In HUN-BUDa, HUN-PECa and SWE-GOTa it addressed only current smokers. In AUS-PERa, AUS-PERb, GER-BREa, GER-BREb, POL-TARa, POL-WARa, SWI-TICa, SWI-VAFa and USA-STAa, CIGAGE referred to regular smoking.
CIGARSM and PIPESM revealed many problems. Some MCCs combined CIGARSM with CIGAR and PIPESM with PIPE (e.g. "If you smoke cigars, about how many do you smoke per week?"). Even if CIGARSM and PIPESM are extracted in a restricted dichotomous version, the variables will exclude past cigar and pipe smokers, and cannot be compared directly to those which applied the recommended wording. USA-STAa used one question looking at only regular present cigar and pipe smoking ("Do you regularly smoke pipe, cigars or cigarillos?"). Some MCCs combined CIGARSM, CIGAR, PIPESM and PIPE with CIGS and EVERCIG. In these cases CIGARSM and PIPESM can only be partly extracted.
DAYCIGS was asked only in SPA-CATa and UNK-GLAa, and OTHERSM in SPA-CATa, UNK-BELa and UNK-GLAa, because at the time the items were introduced most of the Centres had already started their middle survey. DEN-GLOa asked OTHERSM giving answer categories (never, 0.5-1, 2-4, 5+ hours/day). AUS-NEWa converted OTHERSM from another question.
Several MCCs made significant positive changes to their final survey questionnaires, mostly by adding MAXCIGS, CIGAGE, DAYCIGS and OTHERSM.
Fourteen out of the 43 RUAs had identical questions for at least ten of the 13 items. They were:
| 13 | GER-EGEc, GER-EGEd, ICE-ICEa, SWE-NSWa |
| 12 | SWE-GOTa, YUG-NOSa |
| 11 | FRA-LILa, FRA-STRa, FRA-TOUa, ITA-FRIa, NEZ-AUCa, POL-WARa |
| 10 | CHN-BEIa and ITA-BRIa. |
Again, the most serious problems concern the items which are crucial for defining who is a "smoker" (CIGS and EVERCIG). Typical deviations were:
NUMCIGS frequently excluded occasional smokers as in the standard questionnaire. In ITA-FRIa, DAYCIGS was multiplied by the number of cigarettes smoked during a smoking day to derive NUMCIGS for occasional smokers. In DEN-GLOa, FIN-KUOa, FIN-NKAa and FIN-TULa the question was also asked from past smokers ("How much do you or did you smoke on an average day?"). STOP sometimes enquired how many years ago the subject stopped smoking (UNK-GLAa). AUS-PERa, AUS-PERb, DEN-GLOa and NEZ-AUCa combined STOP with IFLYEAR by asking the month besides the year of quitting. CAN-HALa gave categories for STOP (<1, 1-5, >5 years) and IFLYEAR (<6, >= 6 months) which deviated from the MONICA standard. It was also the only RUA which did not enquire about pipe or cigar smoking. Otherwise PIPE and CIGAR revealed only minor deviations, such as asking daily instead of weekly amounts.
In the last revision of the MONICA smoking questionnaire EVERCIG, STOP and CIGAGE were revised to explicitly refer to regular smoking. However, only 11 of the 43 RUAs were able to introduce these changes into EVERCIG and 10 to STOP. 19 RUAs changed their questionnaire with regard to CIGAGE or retained the question referring to regular smoking they already had. In FRA-LILa, FRA-STRa and FRA-TOUa the question enquires about the age of starting regular smoking, but the instructions to the coder refer both to regular and occasional smoking. In GER-ERFa, CIGAGE was derived from date of birth and date of beginning to smoke.
MAXCIGS excluded current smokers in GER-AURa and GER-AUUa. CZE-CZEa had a different wording of the question ("How many cigarettes per day did you smoke during the year you smoked most?").
Some RUAs (AUS-PERa, AUS-PERb, DEN-GLOa, FIN-KUOa, FIN-NKAa, FIN-TULa) combined CIGARSM, CIGAR, PIPESM and PIPE with CIGS and EVERCIG (and BEL-CHAa, BEL-GHEa with CIGS). In these cases CIGARSM and PIPESM can only be partly extracted. USA-STAa used one question looking at only regular current or past cigar and pipe smoking ("Do you regularly smoke a pipe, cigars or cigarillos?"). In GER-ERFa, CIGARSM and PIPESM were derived from CIGAR and PIPE.
DAYCIGS sometimes asked the number of smoking days per week (DEN-GLOa, ITA-FRIa, UNK-GLAa). RUAs also differed in whether the question was asked from all smokers (as in format version 6) or from occasional smokers only (as in version 7). In CAN-HALa, DAYCIGS referred to last 30 days. OTHERSM had many different wordings like "how many hours do you spend with other smokers during an average week?" in POL-TARa, or "how many hours per day, on average, are you near people who smoke?" in RUS-NOCa and RUS-NOIa or "how many hours a day are you exposed to other people's tobacco smoke, in such a way that you notice the smoke in the room?" in DEN-GLOa. It is possible that these deviations reflect problems in translating the standard questionnaire into local language and back into English. In AUS-PERa and AUS-PERb the question referred not to present but to the last 10 years. In AUS-PERa, AUS-PERb and POL-TARa the question asked about hours per week as was incorrectly stated in the 1990 version of the smoking questionnaire and not hours per day as specified in the data transfer format versions 6 and 7 and the 1992 version of the questionnaire. In AUS-PERa and AUS-PERb the item was recoded before submitting the data, but not in POL-TARa.
According to the data transfer format versions used in the different surveys the following smoking items should be available for the MONICA RUAs:
10-year trends in smoking can therefore be calculated only for the seven items which are available for all three surveys. In addition, EVERCIG and STOP were revised to refer to regular smoking in the final survey and are thus not fully comparable with the questions in the initial and middle survey. Only two RUAs (GER-BREa and USA-STAa) used exactly the same questionnaire in all three surveys.
In 43 RUAs the questionnaire was completed by interview, in 15 RUAs it was self-administered and in one RUA (USA-STAa) both approaches were used (Table 1.2). The method of administering the questionnaires was not changed in any MCC between the surveys.
When the survey core data were received in MDC, the data were checked routinely for the constraints described in Appendix 1a. All violations of the constraints were reported to the MCC for their correction or elucidation. For some of the constraints, data values outside the limits were acceptable, but the MCC had to check that the values were not unusual owing to data errors. The MCCs were asked to correct values only if they were incorrect.
The current unresolved constraint violations are shown in Appendix 1b. There are more than ten unresolved constraint violations in ITA- LATa in the initial survey UNK-BEL in the middle survey and in RUS-NOCa and RUS-NOIa in the final survey.
Table 4.1 gives the proportions of missing data for the smoking items among the survey respondents. Missing data are defined as code 9, 99, 999 or 9999 for all the items. The denominator used was all subjects surveyed, except for those data items which were preceded by a 'filter question': NUMCIGS, STOP, DAYCIGS, MAXCIGS, CIGAGE and OTHERSM. In these cases, subjects, who were coded "irrelevant" (8, 88, 888 or 8888) for the item under analysis, were excluded.
The high proportion of missing data for STOP in FIN-KUOa, FIN-NKAa and FIN-TULa in all surveys and in GER-EGEd, GER-ERFa, GER-KMSc, GER-RDMe and MLT-MLTa in the initial survey is due to omission of this variable from the questionnaires. The same concerns the high proportion of missing data for CIGAR and PIPE in GER-EGEd, GER-ERFa, GER-KMSc, GER-KMSd, GER-RDMe and LTU-KAUa in the initial survey, LTU-KAUa in the middle survey and CAN-HALa in the final survey.
In the initial survey GER-BREa, GER-KMSd, ISR-TELa and SWE-NSWa, in the middle survey AUS-NEWa, and in the final survey AUS-PERa and AUS-PERb had increased proportions of missing data for CIGAR and PIPE. RUS-MOIb had high proportion of missing data for STOP, CIGAR and PIPE in the initial survey. Concerning the other smoking items in Table 4.1 there were increased proportion of missing data for STOP in the initial survey in CHN-BEI, GER-RHNa, HUN-BUDa, HUN-PECa, ITA-FRIa, SWE-NSWa and YUG-NOSa, in the middle survey in SWE-GOTa, SWI-VAFa and UNK-GLAa, and in the final survey in CAN-HALa, CHN-BEIa, DEN-GLOa, GER-AURa, GER-AUUa, POL-TARa, SWE-GOTa and UNK-GLAa. For NUMCIGS increased proportion of missing data was observed in SWE-NSWa in the middle survey and in BEL-GHEa and SWE-GOTa in the final survey. In FIN-KUOa and FIN-NKAa there were increased proportions of missing data on EVERCIG in the final survey and on IFLYEAR in the middle survey. Other RUAs which had high proportion for IFLYEAR were GER-EGEd, GER-ERFa, GER-KMSc in the initial survey and FIN-TULa in the middle survey.
As to the items in Table 4.2, all RUAs except SPA-CATa and UNK-GLAa had missing data for DAYCIGS and all except AUS-NEWa, SPA-CATa, DEN-GLOa, UNK-BELa and UNK-GLAa for OTHERSM in the middle survey. The proportion of missing data was however high in AUS-NEWa and DEN-GLOa, because their questions differed significantly from the standard question. In addition, many RUAs had high proportions of missing data on CIGAGE, MAXCIGS, CIGARSM and PIPESM in the middle survey, and on DAYCIGS and OTHERSM in the final survey due to the omission of these variables in their questionnaire. HUN-BUDa (Mid), POL-TARa (Mid), RUS-NOCa, RUS-NOIa and SWE-GOTa (Mid) had increased proportions of missing data on CIGAGE. In HUN-BUDa (Mid) and SWE-GOTa (Mid) this was because the question was asked from current smokers only. Similarly, several RUAs had increased proportion of missing data for MAXCIGS in the middle survey because the question was asked from past smokers only. AUS-PERa and AUS-PERb had increased proportions of missing data on CIGARSM and PIPESM in the final survey. CAN-HALa did not ask CIGARSM or PIPESM in their final survey questionnaire. UNK-GLAa had increased proportion of missing data for DAYCIGS in the initial survey, and FIN-KUOa, FIN-NKAa, FIN-TULa, ITA-BRIa, RUS-NOCa, RUS-NOIa, SWI-VAFa and UNK-GLAa in the final survey. AUS-NEWa, FRA-STRa, GER-EGEc, LTU-KAUa, RUS-NOCa, RUS-NOIa and SWE-NSWa had high proportion of missing data for OTHERSM in the final survey.
Tables 4.1 and 4.2 reveal some discrepancies between the questionnaires and the data provided. Some data were not submitted to the MDC although the questionnaires show that they should be available. In these cases 100% of insufficient data should not be acceptable. This concerns MAXCIGS, CIGARSM and PIPESM in CHN-BEIa and CZE-CZEa in the middle survey.
Another inconsistency has been found. According to the questionnaires, some of the items were not included in the middle or final survey. The database does not reflect this as some data are available for these "missing" items. This applies to:
In initial survey:
In final survey:
There are following explanations to these inconsistencies:
Next, the distributions of the smoking items in the three surveys will be briefly reviewed concentrating on the most obvious inconsistencies in the data. Breakdown by sex are presented in Tables 6 and 8-18. Tables 8, 11, 12, 13, 15 and 17 exclude the initial survey because they concern items which were not included in the initial survey data transfer format (version 3). Consequently, only the RUAs which carried out the middle or final survey or both are included in these tables.
1. "Do you smoke cigarettes now?" (see Table 5).
This question has two aims: firstly, to separate current cigarette smokers from non-smokers and secondly, to separate current regular cigarette smokers from current occasional cigarette smokers. In some RUAs, the data provided for the item has difficulties in satisfying these aims as discussed in Section 4. In AUS-NEWa (Mid), BEL-CHAa (Ini&Mid), BEL-GHEa (Ini&Mid), GER-BREa and GER-BREb occasional smokers cannot be separated from regular smokers, because CIGS allows only yes/no options. In AUS-NEWa (Ini), AUS-PERa, AUS-PERb, CAN-HALa (Ini) and USA-STAa, CIGS addressed only regular smokers so that occasional smokers cannot be separated from non-smokers. In AUS-NEWa (Ini), AUS-PERa, AUS-PERb, BEL-CHAa, BEL-GHEa, DEN-GLOa, FIN-KUOa, FIN-NKAa and FIN-TULa the question addressed also pipe and cigar smokers. Those who only smoked pipe or cigar were, however, excluded from CIGS by the MCCs.
The following observations were made from the data:
2. "On average how many cigarettes do you smoke a day?" The main function of NUMCIGS is to give information about how much the regular smokers smoke. In this respect no inconsistencies were found in the data (Tables 6.1 and 6.2).
Another aspect of this item is its efficiency in separating regular and occasional smokers. In the quality assessments of the initial (5) and the middle (6) survey it became apparent that NUMCIGS will not be useful for this purpose because different MCCs applied different coding practices for NUMCIGS among occasional smokers and there were changes within RUAs in coding practices between the surveys. Therefore the item should not be used in analyses for occasional smokers.
The following observations were made from the data (Tables 7.1 and 7.2):
3. "On how many days a week do you smoke cigarettes?" (see Tables 8.1 and 8.2).
Only SPA-CATa and UNK-GLAa asked DAYCIGS in the middle survey. In DEN-GLOa the question was not asked and the item has been erroneously coded 8. In the final survey, those RUAs where the question was asked differed in relation to whether the question was asked from all smokers (as in transfer format version 6) or from occasional smokers only (as in version 7) i.e. whether the category 'every day' was included. Therefore the data on this item is of limited use for comparison between RUAs.
4. "Did you ever smoke cigarettes in the past?" (see Tables 9.1 and 9.2).
In the final survey, in the RUAs which referred to regular smoking (FRA-STRa, FRA-TOUa, FRA-LILa, GER-EGEc, GER-EGEd, GER-ERFa, ICE-ICEa, ITA-FRIa, NEZ-AUCa, SWE-GOTa, SWE-NSWa, USA-STAa, YUG-NOSa) by EVERCIG, past occasional smokers are included in the "No, never" category. The same concerns AUS-PERa, AUS-PERb and USA-STAa where the question addressed only regular smokers in all surveys. In AUS-NEWa (Mid), BEL-CHAa (Ini&Mid), BEL-GHEa (Ini&Mid), GER-BREa and GER-BREb which allowed only yes/no options, occasional past smokers may have been included in the "Yes, regularly" category.
5."When did you stop smoking cigarettes?" There is no specific table referring to this item. The questions asked in the local questionnaires are presented in Section 4 and the proportions of missing data are given in Table 4.1.
6. The proportion of past smokers who stopped smoking in the last 12 months (see Tables 10.1 and 10.2).
IFLYEAR estimates the proportion of recent quitters. The instructions were unambiguous so relatively few problems occurred.
7. "What is the maximum number of cigarettes you have ever smoked per day as long as a year?" (see Tables 11.1 and 11.2).
Ten RUAs asked the question and have provided data on MAXCIGS in the middle survey. CHN-BEIa asked the question but has not sent the data to the MDC. AUS-PERa and AUS-PERb converted data from another question. Most of the data are insufficient in GER-AURa, GER-AUUa and SPA-CATa because the question was asked from past smokers only. The number of subjects with insufficient data is missing for RUS-MOCa and RUS-MOIa, because they submitted data using version 3 of the data transfer format which did not include this item.
In the final survey, MAXCIGS was asked in all RUAs except in FIN-KUOa, FIN-NKAa, FIN-TULa, GER-ERFa, GER-BREa, GER-BREb and USA-STAa. For FIN-KUOa, FIN-NKAa and FIN-TULa the data are however available for almost half of the respondents. The MCC should provide an explanation for this. In GER-AURa and GER-AUUa the question was asked from past smokers only.
8. "How old were you when you began to smoke cigarettes?" (see Tables 12.1 and 12.2).
In the middle survey, CIGAGE was asked and data has been provided for 27 RUAs. In HUN-BUDa the proportion of insufficient data is high because past smokers were skipped. The reason for high proportion of insufficient data in POL-TARa, RUS-NOCa and RUS-NOIa is unknown. The number of subjects with insufficient data is missing for RUS-MOCa and RUS-MOIa, because they submitted data using version 3 of the data transfer format which did not include this item.
In the final survey, CIGAGE was asked in all RUAs. It referred to regular smoking in 19 RUAs. In RUS-NOIa the proportion of missing data was 44%.
9. Smoking cigars/cigarillos (see Tables 13.1, 13.2, 14.1 and 14.2).
CIGAR and CIGARSM both refer to cigar smoking. CIGAR was included in the initial survey data transfer format but not CIGARSM. Regarding CIGAR very few problems occurred. In the initial survey, CIGAR was not asked in GER-EGEd, GER-ERFa, GER-KMSc, GER-RDMe and LTU-KAUa. In the initial and middle survey, CIGAR addressed only regular smokers in AUS-NEWa, AUS-PERa, AUS-PERb and USA-STAa. DEN-GLOa, FIN-KUOa, FIN-NKAa, FIN-TULa, GER-BREa, GER-BREb and YUG-NOSa asked the daily amount smoked but recoded the item before submitting the data to the MDC.
Due to a high level of missing data, CIGARSM is not a useful indicator of smoking behaviour in the middle survey. Only 25 RUAs provided data concerning this item, four of them (AUS-NEWa, AUS-PERa, AUS-PERb and USA-STAa) skipping occasional cigar smokers. CHN-BEIa has not provided data they have collected. AUS-NEWa derived the values for CIGARSM from CIGAR, and in FIN-KUOa, FIN-NKAa and FIN-TULa the item was generated by a computer algorithm.
In the final survey, most of the RUAs could provide information both on CIGARSM and CIGAR. CAN-HALa did not ask about cigar smoking at all and AUS-PERa, AUS-PERb and USA-STAa addressed only regular cigar smokers. In FIN-KUOa, FIN-NKAa and FIN-TULa, CIGARSM was generated by a computer algorithm. DEN-GLOa, FIN-KUOa, FIN-NKAa, FIN-TULa, GER-BREa and GER-BREb asked the daily amount smoked but CIGAR could be fully extracted using this information.
10. Pipe smoking (see Tables 15.1, 15.2, 16.1 and 16.2).
Cigar and pipe smoking were surveyed with the same methods in MCCs so the same remarks largely concern PIPESM and PIPE as CIGARSM and CIGAR. In the initial survey, PIPE was not asked in GER-EGEd, GER-ERFa, GER-KMSc, GER-RDMe and LTU-KAUa. In the middle survey, the same 25 RUAs provided data on PIPESM as for CIGARSM. AUS-NEWa, AUS-PERa, AUS-PERb and USA-STAa skipped occasional pipe smokers. AUS-NEWa derived the values from PIPE and in FIN-KUOa, FIN-NKAa and FIN-TULa this item was generated by computer. Data are not available for CHN-BEIa although their questionnaire indicates that they have collected data on this item. The only RUA which did not ask PIPE in the middle survey was LTU-KAUa. In the initial and middle survey, FIN-KUOa, FIN-NKAa, FIN-TULa, GER-BREa and GER-BREb recorded the daily amounts smoked but recoded the item before submitting the data to the MDC.
In the final survey, most of the RUAs could provide information both on PIPESM and PIPE. CAN-HALa did not ask about pipe smoking at all and AUS-PERa, AUS-PERb and USA-STAa addressed only regular pipe smokers. In FIN-KUOa, FIN-NKAa and FIN-TULa PIPESM was generated by computer. FIN-KUOa, FIN-NKAa, FIN-TULa, GER-BREa and GER-BREb recorded the daily amount smoked but PIPE could be fully extracted using this information.
11. "For how many hours, on average each day, are you subjected to other people's tobacco smoke?" (see Tables 17.1 and 17.2).
In the middle survey, OTHERSM was asked only in SPA-CATa, UNK-BELa and UNK-GLAa, and in a restricted form in DEN-GLOa. In AUS-NEWa it was converted from another question. Proportion of missing data was high in AUS-NEWa (22%), DEN-GLOa (34%) and UNK-GLAa (18%). DEN-GLOa, where the question was asked giving answer categories, recoded the answers as follows.
| Answer category | Recoded as (hours/day) |
|---|---|
| Never | 0 |
| 0.5-1 | 1 |
| 2-4 | 3 |
| 5+ | 99 |
In the final survey, OTHERSM was asked in all but 9 RUAs (BEL-GHEa, BEL-CHAa, GER-AURa, GER-AUUa, GER-ERFa, GER-BREa, GER-BREb, SWE-GOTa and USA-STAa).
An individual can start to smoke but cannot then ever return to be a 'never smoker'. Therefore the proportion of never smokers can decrease by time but cannot, by definition, increase by time within a birth cohort in a population. If the questionnaires and the data collection measures did not change between the surveys, there was no change in the sampling frame and no large differences in the participation rate, there should be no essential increases in the proportion of never smokers within birth cohorts. Large increases in the proportion of never smokers indicate a probable change in the subpopulation which the responders represent.
We investigated the stability of the proportions of never smokers (never smoker defined as CIGS=2 and EVERCIG=2 or 3, see Section 6.1) by calculating the difference in proportion of never smokers by 10-year birth cohorts between the surveys. 10-year birth cohorts were defined by the years of birth corresponding closest to the 10-year age groups 25-34, 35-44, 45-54 and 45-54, in the middle of the initial survey in each RUA. Table 18 gives the differences in proportion of never smokers between the three surveys by these birth cohorts. DEN-GLO was excluded from this analysis because they surveyed men and women of ages 30, 40, 50 and 60 years and did not therefore examine the same birth cohorts in the three surveys. Also, AUS-PERb and GER-BREb, which did not carry out the initial survey, and RUAs which only did the initial survey were excluded from this analysis.
To identify the RUAs where there is possibly a bias in the cohort trend, either through a measurement bias or a change in the population which the sample represents, a Cohort Trend Score (CTS) was defined. The score is based in the estimated changes and their standard errors for men and women in the common age groups 35-44 and 45-54 in three steps:
| CTS = | 2 | if all four changes are within limits; |
| 1 | if one of the four changes is out of limits; | |
| 0 | if at least two of the four changes are out of limits. |
If the score is 2, there is no evidence of bias between the surveys. If the score is 1, there may be a bias, at least concerning the representativeness of the sample in some sex/birth cohort. A score 0 is a sign of concern about a more general bias.
When comparing the estimates to the respective standard errors, in the following RUAs outstanding increases in the proportion of never smokers were observed: AUS-PERa (oldest birth cohort in women), FRA-STRa (birth cohort 35-44 in men), GER-HACa (birth cohort 35-44 in men) and RUS-MOIa (birth cohort 35-44 in women). The MCCs concerned were asked to carefully review the possible reasons for these increases within the birth cohorts.
In the initial survey, the smoking questionnaire should have been validated by measurement of serum thiocyanate. According to the Manual thiocyanate should have been measured in a 10% sub-sample if it could not be done for every subject. In the 1990 version of the Manual, the smoking questionnaire was asked to be validated in 100% of subjects by expired carbon monoxide, serum cotinine or serum thiocyanate. Thiocyanate remained thus an option for validating the questionnaire but it was no more part of the core study. Measurement only in sub-samples was possible, but not recommended, and should have contained a minimum of 10% of the subjects.
Table 19 gives the availability of data on serum thiocyanate, serum cotinine and expired air carbon monoxide in the RUAs by survey. Serum cotinine and carbon monoxide were included in the data transfer format versions 6 and 7 only. Twenty RUAs in the initial survey, 10 in the middle survey and 29 RUAs in the final survey did not collect data on serum thiocyanate at all. Serum cotinine and carbon monoxide were measured only rarely. Two RUAs (SWE-NSWa and UNK-GLAa) collected data on cotinine in a sub-sample of the subjects in the middle survey and also two RUAs (UNK-GLAa and YUG-NOSa) in the final survey. Two RUAs (UNK-GLAa and USA-STAa) in the middle survey and eight RUAs (FRA-LILa, FRA-STRa, FRA-TOUa, ICE-ICEa, UNK-BELa, SWI-VAFa, UNK-GLAa and USA-STAa) in the final survey collected data on carbon monoxide at least in a sub-sample of the subjects.
In the quality assessment of smoking data in the initial survey (5) the definition of "a smoker" was extensively discussed giving various optional definitions and assessing the advantages and disadvantages of their use based on the data available in the MONICA Project. Tables 20.1 and 20.2 give the proportion of smokers in each RUA by survey using these optional definitions. The middle survey quality assessment provided further information on the changes in the questionnaires and their implications for the definitions and the RUAs which could provide reliable estimates of trends in regular smoking. Over the years there has also been continuing discussion on the terms used for the different smoking categories.
WHO has also been working on standardizing the terminology and definitions of smoking (4). The terminology is quoted below:
Any population can be divided into two categories, smokers and non-smokers.
- A smoker is someone who, at the time of the survey, smokes any tobacco product either daily or occasionally.
Smokers may be further divided into two categories:
- A daily smoker is someone who smokes any tobacco product at least once a day, (with the exception that people who smoke every day, but not on days of religious fasting, are still classified as daily smokers).
- An occasional smoker is someone who smokes, but not every day.
Occasional smokers include:
- Reducers - people who used to smoke daily but now do not smoke every day.
- Continuing occasional - people who have never smoked daily, but who have smoked 100 or more cigarettes (or the equivalent amount of tobacco) and now smoke occasionally.
- Experimenters - people who have smoked less than 100 cigarettes (or the equivalent amount of tobacco) and now smoke occasionally.
- A non-smoker is someone, who, at the time of the survey, does not smoke at all.
Non-smokers can be divided into three categories:
- Ex smokers are people who were formerly daily smokers but currently do not smoke at all.
- Never smokers are those who either have never smoked at all or have never been daily smokers and have smoked less than 100 cigarettes (or the equivalent amount of tobacco) in their lifetime.
- Ex-occasional smokers are those who were formerly occasional, but never daily, smokers and who smoked 100 or more cigarettes (or the equivalent amount of tobacco) in their lifetime.
These definitions can be used to classify the population according to their lifetime smoking status. In particular:
- Ever smokers are defined as those who have ever smoked at least 100 cigarettes (or the equivalent amount of tobacco) in their lifetime.
A specific sub-category of interest are those who have smoked, or now smoke, every day.
- Ever daily smokers are defined as persons who are currently daily smokers, reducers or are ex-smokers.
The WHO classification is a general one where all tobacco products are taken into account in the definition of a smoker. In the MONICA Project only cigarette, pipe and cigar smoking are considered. In the MONICA countries other types of tobacco consumption are unknown or at least rare. In addition, in most of the RUAs the proportion of pipe or cigar smokers is small. In MONICA the smoking questionnaire has been mainly directed towards cigarette smoking and therefore gives the best information about it.
The table below gives the main categories of cigarette smoking and their definitions in the MONICA data, using the WHO terminology where applicable:
| Current smoking | CIGS | Past smoking | EVERCIG |
|---|---|---|---|
| A. Cigarette smoker | 1,3 | ||
| A.1 Daily cigarette smoker | 1 | ||
| A.2 Occasional cigarette smoker | 3 | A.2.1 Cigarette reducer1 | 1 |
| A.2.2 Continuing occasional cigarette smoker or experimenter1,2 |
2 | ||
| A.2.3 Insufficient data1 | 9 | ||
| B. Non-(cigarette) smoker | 2 | B.1 Ex-cigarette smoker3 | 1 |
| B.2 Never or ex-occasional cigarette smoker4 | 2,3 | ||
| Insufficient data | 9 | ||
| Insufficient data | 9 |
1: EVERCIG is available for occasional cigarette smokers only in version 7
of the data transfer format (Form 04).
2: Continuing occasional cigarette smokers cannot be separated from
experimenters in MONICA data.
3: Subcategories can be defined according to the time since smoking (IFLYEAR
and/or STOP).
4: Code 3 of EVERCIG ("Yes, occasionally in the past, but not now")
is available only in versions 3 and 6 of the data transfer format (Form 04).
The MONICA data do not have specific questions which separate different types of former cigar and pipe smoking. Therefore, the total smoking, also including cigar and pipe, can only be used to address the current smoking status satisfactorily. The following table defines the main categories of current smoking using the MONICA data.
| Smoking category | CIGS | CIGARSM | PIPESM | |
|---|---|---|---|---|
| Smoker | Daily smoker | 1 | 1,2,3,4 | 1,2,3,4 |
| 1,2,3 | 1 | 1,2,3,4 | ||
| 1,2,3 | 1,2,3,4 | 1 | ||
| Occasional smoker | 3 | 2,3,4 | 2,3,4 | |
| 2,3 | 3 | 2,3,4 | ||
| 2,3 | 2,3,4 | 3 | ||
| Non-smoker | 2 | 2,4 | 2,4 | |
| Insufficient data | 9 | any | any | |
| any | 9 | any | ||
| any | any | 9 | ||
This definition, called definition A, uses the data items CIGARSM and PIPESM, which unfortunately were collected only in the latest versions (version 6 and 7) of the core data transfer format, which were used in the middle and final surveys. Only about half of the MCCs, however, collected these data in the middle survey.
Alternatively, the smoking categories can be defined (definition B) using data items CIGAR and PIPE instead of CIGARSM and PIPESM:
| Smoking category | CIGS | CIGAR | PIPE |
|---|---|---|---|
| Smoker | 1,3 | 0-887 | 0-887 |
| 2 | 1-887 | 0-887 | |
| 2 | 0-887 | 1-887 | |
| Non-smoker | 2 | 0,888 | 0,888 |
| Insufficient data | 9 | any | any |
| any | 999 | any | |
| any | any | 999 |
Definition A addresses cigar and pipe smoking more explicitly than definition B, but the latter definition has the advantage that it is available in all three MONICA surveys. Therefore, where definition A can be used for cross-sectional analysis in the final survey only, definition B can also be used for trend analysis.
The computation of the proportions of (definition A and B) smokers is straightforward. Note, however, that the proportions should not be calculated when the proportion of missing data on CIGARSM and/or PIPESM (definition A) or CIGAR and/or PIPE (definition B) is high because it will lower the accuracy of the estimates. The computation of the proportions of the categories of cigarette smokers which use both data items CIGS and EVERCIG is not straightforward, because the latter is defined only in a subcategory of the former. The problem can be illustrated using an example:
Let us assume a sample of 100 subjects, where the item CIGS is distributed as:
| CIGS | 1 | 2 | 3 | 9 | Total |
|---|---|---|---|---|---|
| n | 40 | 60 | 0 | 0 | 100 |
Let us assume also that among the non-cigarette smokers (i.e. CIGS=2), the item EVERCIG is distributed as:
| EVERCIG | 1 | 2 | 9 | Total |
|---|---|---|---|---|
| n | 20 | 30 | 10 | 60 |
We can now estimate the proportions of the different categories of cigarette smoking. As we do not know the smoking status of those with insufficient data, we exclude them from the analysis and get the following percentages:
| % | |
|---|---|
| A.1 Daily | 44 |
| A.2 Occasional | 0 |
| B.1 Ex | 22 |
| B.2 Never-daily | 33 |
| Total | 100 |
This, however, conflicts with the fact that we know that 40% of the sample are cigarette smokers and 60% of them are non-cigarette smokers (i.e. ex-cigarette smokers or never-daily cigarette smokers).
The obvious bias in the example is caused by the fact that we have insufficient data for data item (EVERCIG) which is defined in a subcategory of another variable (CIGS). To avoid this bias we must also follow the hierarchy of the data items in the estimation of the percentages. In the case of the example above, we do it in three steps:
| % | |
|---|---|
| A.1 Daily | 40 |
| A.2 Occasional | 0 |
| B. Non-smokers | 60 |
| Total | 100 |
| % | |
|---|---|
| B.1 Ex | 40 |
| B.2 Never-daily | 60 |
| Total | 100 |
| % | |
|---|---|
| A.1 Daily | 40 |
| A.2 Occasional | 0 |
| B.1 Ex | 24 |
| B.2 Never-daily | 36 |
| Total | 100 |
The mathematical basis for the estimation of the proportion of ex-cigarette smokers is presented in Appendix 4. Note that the proportion of ex-smokers can be calculated in this way in each survey only. In the trend analysis, if the trend is modelled using the year of survey examination as the explanatory variable, the calculation of the proportion of ex-smokers using the three steps is not possible.
If we want to categorise the ex-cigarette smokers further using data items IFLYEAR and/or STOP, we can get unbiased estimates of the proportions of such categories in the entire sample by multiplying three proportions: the proportion of non-cigarette smokers in the sample, the proportion of ex-cigarette smokers among the non-cigarette smokers and the proportion of the subcategory of ex-cigarette smokers among the ex-cigarette smokers.
A summary concerning the smoking items actually available for the different RUAs
according to the questionnaires used by the MCCs and the data provided for the different
surveys is given in Appendix 5.
RUAs which have used a question not corresponding to the standard question should either
be excluded from the analyses or a note should be made in the publication to clearly
indicate the deviation of the question from the standard. Exclusion of a particular RUA
from a cross-sectional analysis concerning a particular item even though data is available
depends on the nature of the analysis, the data item in question and the nature of the
deviation. Therefore, it is not feasible to give quality scores which could be referred to
in all types of analyses. The person carrying out cross-sectional analyses should consult
Tables 2.1, 2.2 and 3 to judge the quality limitations of the data. Particularly,
exclusion of RUAs which have converted the item from another question and RUAs with more
than 10% of missing data for the particular item (Tables 4.1
and 4.2) should be considered.
In most situations the item NUMCIGS should not be used for occasional smokers, since different MCCs have applied different coding practices which vary between RUAs within surveys and within RUAs between surveys.
The definition for a regular cigarette smoker used in MONICA is CIGS=1 as presented in Section 6.1. There are seven RUAs with a particular concern about the quality of the trend estimates in the proportion of regular smokers using this definition. The problems are the following:
In order to estimate the bias caused by these problems we looked at the proportion of smokers who smoke very little in the three surveys (Table 7.1), the proportion of missing data for items CIGS and NUMCIGS (Table 4.1) and the change in the proportion of occasional smokers (Table 5) in the RUAs concerned. The following conclusions were made:
In addition to the RUAs listed above, AUS-PERa and AUS-PERb had a minor change in the specification of CIGS to the final survey. No other RUAs changed their question concerning regular smoking between the surveys.
As mentioned in Section 4.4, 10-year trends can be calculated only for the seven items (CIGS, NUMCIGS, EVERCIG, STOP, IFLYEAR, PIPE and CIGAR) included in all three versions of the data transfer format. Also, it has to be taken into account that EVERCIG and STOP were revised to refer to regular smoking in the final survey, although only 11 RUAs introduced this change to EVERCIG and 10 to STOP. Exclusion of RUAs with more than 10% of missing data for a particular item or with a more than 5% change in the proportion of missing data between surveys should be considered (Table 4.1). Item NUMCIGS should not be used for occasional smokers, as mentioned in Section 6.2.
Constructing a solid smoking questionnaire has been a development process in MONICA: the questionnaire has been revised several times during the project. This has had both advantages and drawbacks. The drawbacks are not difficult to point out. The standard questionnaires are not comparable between the different surveys for many of the smoking items. Also the MCCs have not always been able to introduce the changes in the standard questionnaire to their local questionnaire which has created even more deviations between the standard questionnaire and the data provided. Additional complications emerge from the changes in the data transfer format and the different formats used by the MCCs for submitting the data to the MDC in each survey. There have also been technical problems, for example the fact that not all the items of the initial survey questionnaire were included in the data transfer format version used, so that not all the data collected by the MCCs have been available for centralized storage and analysis. But there have also been clear advantages in the process. When the comparability problems, differences in coding practices between MCCs and between surveys and inconsistencies in the standard questionnaire have been highlighted, it has been possible to create a more consistent and better standardized smoking questionnaire, of which the questionnaire used in the final survey is an example.
Relatively few problems have been discovered concerning data on the smoking items which are crucial for the classification of smokers. Therefore the 10-year trend analysis can be carried out in almost all RUAs without concerns for the quality of data. Due to the reasons mentioned above, the availability and quality of data for the other smoking items was not very good in the initial and middle surveys, but has improved remarkably in the final survey. Thus better cross-sectional data are available for analysis for the final survey enabling broader knowledge about the smoking behaviour in the MONICA populations. The only major disadvantage is that only a part of the MCCs were able to introduce the changes in the latest version of the standard questionnaire, such as specifying the age of starting to smoke to refer to regular smoking and asking about past smoking from current occasional smokers, to their local questionnaire. Therefore it is of especial importance that those carrying out longitudinal and cross-sectional analyses take into account both the changes in the standard questionnaire between the surveys and the deviations of the local questionnaires from the standard ones.
Below is a summary of the main findings for the individual RUAs concerning the quality of data on smoking behaviour in the three surveys. These findings should be borne in mind when using the data for analysis.
AUS-NEWa
AUS-PERa
AUS-PERb
BEL-CHAa
BEL-GHEa
BEL-LUXa
CAN-HALa
CHN-BEIa
CZE-CZEa
DEN-GLOa
FIN-KUOa, FIN-NKAa and FIN-TULa
FRA-LILa
FRA-STRa
FRA-TOUa
GER-AURa and GER-AUUa
GER-BREa
GER-BREb
GER-BERa and GER-COTa
GER-HACa
GER-EGEc
GER-EGEd
GER-ERFa
GER-KMSc
GER-KMSd
GER-RDMe
GER-RDMf
GER-RHNa
HUN-BUDa and HUN-PECa
ICE-ICEa
ISR-TELa
ITA-BRIa
ITA-FRIa
ITA-LATa
LTU-KAUa
MLT-MLTa
NEZ-AUCa
POL-TARa
POL-WARa
ROM-BUCa
RUS-MOCa and RUS-MOIa
RUS-MOIb
RUS-NOCa and RUS-NOIa
RUS-NOCb
SPA-CATa
SWE-GOTa
SWE-NSWa
SWI-TICa and SWI-VAFa
UNK-BELa
UNK-GLAa
USA-STAa
YUG-NOSa