WWW-publications from the WHO MONICA Project
October 1998
Anu Molarius1, Jaakko Tuomilehto2 and Kari Kuulasmaa1 for the WHO MONICA Project3
1 MONICA Data Centre, National Public Health Institute, Helsinki, Finland
2 Department of Epidemiology and Health Promotion, National Public Health
Institute, Helsinki, Finland
3 Annex: Sites and key personnel of the WHO MONICA
Project
This document includes the main findings of unpublished report:
Thanks are due to Alun Evans 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. 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.
Assessment of the awareness and control of hypertension in the community has been one of the aims of the population surveys carried out by the MONICA Collaborating Centres (MCCs). Before the data on awareness of hypertension and on antihypertensive drug treatment can be effectively used to study the situation in individual populations, or to carry out comparisons between populations, certain basic issues of data quality and comparability must be considered.
The main aim of the WHO MONICA Project is to assess the changes occurring in each study population over time. Therefore not only cross-sectional but also longitudinal aspects of data quality need to be evaluated to ensure adequacy, consistency and comparability of data within each population and between the populations.
The MONICA Project has a standard format for data transfer to the common database. To transfer the data in this format it is necessary to collect the information in such a format that the data required for the data transfer can be extracted. The MONICA Manual includes two questions related to the awareness and drug treatment of hypertension. The exact format of these two questions is given in the MONICA Manual (1):
HIBP "Have you ever been told by a doctor or other health worker that you have high blood pressure?"
1 = yes
2 = no (if no, record 8 in item DRUGS)
9 = insufficient dataDRUGS "Are you taking (in the last two weeks) drugs for high blood pressure?"
1 = yes
2 = no
3 = uncertain
8 if HIBP = 2
9 = insufficient data
The quality assessment of data on hypertension control in the initial (first) and middle (second) surveys of the MONICA Project was distributed in June 1996 (3). This report follows the structure of the earlier quality assessment report. It repeats its main results and evaluates the quality of the questionnaires used and the data provided from the final survey.
The survey periods when the data were collected in the MCCs have been reported in the quality assessment report of age and date of examination (2).
The main purpose of this quality assessment report is:
The results of this document are reported by Reporting Unit Aggregates (RUAs) which are foreseen as potential candidates for units of analyses of the MONICA data. The RUAs, their abbreviations and reporting units are listed in Table 1. Some of the RUAs have several versions because different combinations of Reporting Units (RUs) may be used for cross-sectional and trend analyses if all RUs of the population were not included in some of the surveys. Therefore, in AUS-PER, GER-BRE, GER-EGE, GER-KMS, GER-RDM, RUS-MOI and RUS-NOC there is an overlap of RUs included in the RUAs in some surveys. The RUAs are identified by the abbreviation and a version letter. For UNK-GLA, which carried out four surveys, this report assesses the first (initial), third (middle) and fourth (final) survey. In this report, altogether 54 RUAs are considered for the initial MONICA survey, 43 for the middle survey and 41 for the final survey.
The questionnaires used in the initial survey in their local language and/or their English translation were available in the MONICA Data Centre (MDC) for 53 RUAs (Table 2). The questionnaires used in the middle survey were available for all 43 RUAs and the final survey questionnaires for 39 RUAs. In 1991 a questionnaire on the MONICA Population Survey Procedures (Form VI), including a question as to whether the questionnaire was self-administered or completed during an interview, was sent to all MCCs. This was completed for 51 RUAs. In 1995 the questionnaire was further checked by the MCCs regarding the final survey.
Data on the items of hypertension control were transferred to the MDC using the MONICA core data transfer format (Form 04, versions 3,6,7). The items on hypertension control on the data transfer format remained the same throughout the three surveys.
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-NEW, BEL-LUX, FRA-STR, FRA-TOU, LTU-KAU, NEZ-AUC, POL-TAR, POL-WAR, RUS-MOC, RUS-MOIa, RUS-MOIb and SWI-TIC 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 2). The denominator is "all observations" in Tables 4-5. For Table 6 the denominator includes subjects aware of high blood pressure. No age or sex adjustment was applied to the data.
Findings
Table 3 shows a comparison of the MONICA standard questions on "HIBP" and "DRUGS" and those used in the RUAs. In the data transfer format, awareness of hypertension was addressed in item HIBP: "Have you ever been told by a doctor or other health worker that you have high blood pressure?" In the initial survey 36 out of 54 RUAs used a question identical to the standard MONICA question. The local questionnaire was not available for one RUA (BEL-LUX). The discrepancies in questionnaires used by the other 17 RUAs were:
In 12 RUAs (AUS-NEW, AUS-PERa, CAN-HAL, DEN-GLO, FIN-KUO, FIN-NKA, FIN-TUL, FRA-STR, FRA-TOU, MLT-MLT, NEZ-AUC, YUG-NOS) the question on HIBP was "Have you ever been told that you have high blood pressure?" Thus, it did not clearly specify that it was a health professional who told about the high blood pressure.
In two RUAs (GER-BREa, ISR-TEL) the question on HIBP was "Did you suffer or ever suffer from following diseases?" (hypertension was one of the conditions listed).
In three RUAs (HUN-BUD, HUN-PEC, ICE-ICE) the question on HIBP was not asked at all.
During the middle survey, the HIBP question was included in the questionnaire in two (HUN-BUD, ICE-ICE) of the three RUAs which did not ask the question in the initial survey. Thus, only one RUA (HUN-PEC) did not ask HIBP in the middle survey. In addition, two RUAs (FRA-TOU and YUG-NOS) revised their HIBP question to follow the MONICA standard data transfer format. All other RUAs used the same question in the middle as in the initial survey.
In the final survey, one RUA (AUS-NEW) revised the HIBP question after the middle survey to follow the MONICA Manual, as well as three RUAs (CAN-HAL, FRA-STR, NEZ-AUC) which did not carry out the middle survey. For one RUA (GER-ERF), the question was changed from the standard one to "Do you suffer from high blood pressure?". As a result, there were 30 RUAs where the question was identical to the MONICA standard, six RUAs (AUS-PERa, AUS-PERb, DEN-GLO, FIN-KUO, FIN-NKA, FIN-TUL) where the professional was not specified, and three RUAs (GER-BREa, GER-BREb, GER-ERF) where the question was of the type "Did you ever suffer from ---?". For two RUAs (BEL-CHA, BEL-GHE) the questionnaire was not available.
Discussion
In two RUAs in the initial survey (GER-BREa, ISR-TEL), two in the middle survey (GER-BREa, GER-BREb) and three in the final survey (GER-BREa, GER-BREb, GER-ERF) the question was of the type "Did you ever suffer from following disease?" instead of "Have you ever been told by a doctor or other health worker that you have high blood pressure?" In this case the response may have depended not only on the diagnosis by a doctor, but also on a subjective feeling. This could bias the result in either direction depending on the interpretation of the question by different persons of different cultures.
In one RUA (ICE-ICE) during the initial survey HIBP was not asked but was coded "yes" for those who stated that they were taking drugs for high blood pressure. Therefore, the proportion of awareness is too low as it excludes those subjects who have been told that they have high blood pressure but who are not under treatment. In the middle and final survey the definition was the standard one in this RUA.
In 12 RUAs in the initial survey, nine in the middle survey and six in the final survey, the question did not have the specification "by doctor or other health worker". Thus, it is possible that in these RUAs the definition for the awareness of high blood pressure is slightly different from that in the other RUAs, but probably this does not cause a significant bias.
In some RUAs where the question was otherwise identical to the standard data transfer format the wording "or other health worker" was omitted i.e. the question referred to a doctor only. This is only a minor deviation.
Findings
In the data transfer format, antihypertensive drug treatment was addressed in item DRUGS: "Are you taking (in the last two weeks) drugs for high blood pressure?" In the initial survey, the questions used in 19 out of 54 RUAs were the same as in the standard MONICA data transfer format (Table 3). However, in many of these RUAs the answer category "uncertain" (code=3) was omitted. The local questionnaire was not available for one RUA (BEL-LUX).
The main differences in the other 34 RUAs were:
In 17 RUAs the question had the format: "Are you taking drugs for high blood pressure AT PRESENT?" Although in these RUAs, as well as in those four specifying "in the last week", the criterion did not completely match the format "in the last two weeks", it gives more or less the same result. The definition of taking antihypertensive drugs was, however, significantly different from the standard transfer format in some of these RUAs. In seven RUAs (GER-BER, GER-COT, GER-EGEb, GER-ERF, GER-HAC, GER-KMS, GER-RDMc) the answer of "yes, as prescribed by the doctor" was considered as a positive response and "yes, depending on the condition" was coded "uncertain".
In two RUAs (BEL-CHA, BEL-GHE) the question referred to the present, but had a different wording: "Are you currently on treatment for high blood pressure?"
"Are you taking drugs for high blood pressure IN THE LAST WEEK?" instead of "in the last two weeks" was used in five RUAs (DEN-GLO, RUS-NOCa, RUS-NOCb, RUS-NOI, USA-STA). This may have only a very small effect on the estimated proportion of treated hypertensive subjects.
One RUA (HUN-PEC) did not ask any question about antihypertensive drug treatment. In one RUA (ICE-ICE) the structure of the question is unknown.
In seven RUAs (FIN-KUO, FIN-NKA, FIN-TUL, GER-BREa, LTU-KAU, MLT-MLT, YUG-NOS) the pattern of the question was "When did you last take antihypertensive drugs?" or "How frequently do you use the following medicine?" (including antihypertensive drugs). When they had converted their data into the MONICA data transfer format there was an obvious discrepancy between the standard format and that used in these seven RUAs (for details see Section 7). In one RUA (YUG-NOS) the response to drugs was positive if the subject had been taking drugs for high blood pressure during "today" to "one week ago", and code 3 (uncertain) was used for the period from "one week to one month ago". In one RUA (GER-BREa) the frequency of taking drugs "at least once or more times a month" was coded as "yes", "rarely or never" as "no".
In one RUA (ISR-TEL) the information about drugs in their original questionnaire was ambiguous: "Are you receiving medical treatment at the present time?" It did not specify what kind of medical treatment or whether it was for high blood pressure.
In the middle survey four of the above mentioned RUAs (ICE-ICE, RUS-NOCa, RUS-NOI, YUG-NOS) revised the DRUGS question to follow the standard MONICA data transfer format, and one (AUS-NEW) replaced the word "drugs" with "treatment". All other RUAs used questionnaires similar to those used in the initial survey.
In the final survey, two RUAs (FRA-STR, POL-WAR) revised their questionnaire to follow the MONICA standard. The RUA (AUS-NEW) which changed the question to refer to "treatment" in the middle survey revised the wording to refer to "treatment with medication". Two RUAs (AUS-PERa, AUS-PERb) changed from the standard question to ask about the last month. Excluding the possible addition of the category "uncertain", other RUAs used questionnaires similar to those in the middle survey (or in the initial survey in those RUAs which did not do the middle survey).
Discussion
The main deviations from the recommended definition of drug treatment for hypertension is related to the time when subjects had last taken the antihypertensive drugs. The standard data transfer format states "in the last two weeks".
In 19 RUAs the time was defined using the word "now" or "at present". It is likely that in these RUAs the patients who had stopped treatment during the last two weeks may have been misclassified. However, the proportion of such people is extremely low, and therefore the possible bias is negligible. The same concerns the two RUAs in the final survey which used the period of "last month". In five RUAs the period for drug treatment was "in the last week", and the likelihood of possible bias compared with "in the last two weeks" is similarly small.
Two RUAs in the initial survey and three RUAs in the middle survey used the term "treatment" instead of "drugs". It is probable that this has not had any major effect on the results.
The review of the questionnaires on MONICA population survey procedures revealed that the survey procedures for HIBP and DRUGS applied in the MCCs did not change between the surveys. No information on survey procedures for HIBP and DRUGS was received from five RUAs. In the initial survey, the questionnaire was administered by an interviewer in 37 RUAs and in 11 RUAs a self-administered questionnaire was used (Table 2). One RUA used a combination of an interviewer and self-administered questionnaire.
The MDC checks all population survey core data received from the MCCs at the time they are included in the MONICA database. All possible inconsistencies in the data are reported to the MCC to enable the correction of errors. The following constraints concern items HIBP and DRUGS.
All violations of these constraints were reported to the MCCs for their correction or confirmation. Data values outside the limits were acceptable, but the MCC had to check that the values were not due to data errors. The MCCs were asked to correct values only if they knew that they were incorrect. The currently unresolved constraint violations concerning data on hypertension control are listed in Appendix 1. There were only two unresolved constraint violations for data on hypertension control.
Table 4 presents the distribution of the replies to HIBP in each RUA in the three surveys. In the initial survey, the proportion of missing data was high in three RUAs (BEL-CHA, BEL-GHE, ITA-LAT, range from 17% to 35%). In addition, in three RUAs (HUN-BUD, HUN-PEC, ICE-ICE) HIBP was not asked.
In the middle survey, only one RUA (HUN-PEC) did not ask about HIBP. The proportion of missing data was high in two RUAs (BEL-CHA 41%, BEL-GHE 26%). More than a 5% change in the proportion of missing data between the initial and middle survey was observed in one RUA (FRA-TOU) where the proportion decreased from 6% to zero.
In the final survey, all RUAs asked about HIBP. The proportion of missing data was high in the same two RUAs (BEL-CHA 39%, BEL-GHE 26%) as in the middle survey. These RUAs also had high proportions of missing data in the initial survey. This is explained by the fact that the smoking questionnaire was applied during a home visit, making the subjects respondents, but HIBP and DRUGS were asked during the clinical examination which a large proportion failed to attend.
More than a 5% change in the proportion of missing data between middle and final survey was observed in one RUA (GER-ERF) where the proportion increased from 0% to 7%. In addition, the proportion of missing data in one RUA (GER-BREa) increased from 1% in the initial survey to 6% in the final survey.
The distribution of the proportion of missing data in the RUAs regarding DRUGS (Table 5) was fairly similar to that observed for HIBP. Excluding the RUA (HUN-PEC) which did not ask about DRUGS, the percentage of missing data in the initial survey was high in 6 RUAs (BEL-CHA, BEL-GHE, GER-RHN, ICE-ICE, ITA-LAT, YUG-NOS) ranging from 10% to 91%.
The highest proportion of missing data was 91% in Iceland. They did not ask the HIBP question directly in the initial survey, but converted DRUGS and HIBP items to the MONICA transfer format from their local coding, and they could not unequivocally distinguish between people who did not take drugs and those for whom data were insufficient. It can be assumed that the vast majority of these did not take drugs, despite the fact that they have been coded as insufficient data.
Compared with the initial survey, the proportion of missing data increased in one RUA (GER-BREa) from 0% to 8%, and decreased in one RUA (YUG-NOS) from 10% to 0% in the middle survey. Iceland was able to report data on DRUGS in the middle survey.
Compared with the middle survey, there were very little changes in the proportion of missing data in the final survey.
Specific instructions were given in the MONICA Manual to code "uncertain", i.e. problematic or insufficient information, under DRUGS: "if the use of blood pressure lowering drugs is reported but the person in question is uncertain if these have been used during the last two weeks or he/she is not sure whether the drugs were used for hypertension". We found only 19 RUAs in the initial survey where the code "uncertain" was used in their original questionnaires exactly as proposed in the manual. 10 additional RUAs could derive "uncertain" codes from their original questionnaires. Several RUAs did, however, add the code to their middle or final survey questionnaire.
The proportion of uncertain information was more than 5% in two RUAs (GER-ERF, LTU-KAU) in the initial survey (Table 5). In the middle and final surveys the proportion of uncertain data was less than 1% in all RUAs. Consequently, there was a more than 5% change in the proportion of "uncertain" in GER-ERF and LTU-KAU between surveys. In GER-ERF those who responded "yes, depending on the condition" were coded as "yes" in the final survey, while they were coded as "uncertain" in the initial and middle surveys. For LTU-KAU the MCC informed that there are two possible explanations for the decrease in the proportion of uncertain data on DRUGS: 1) the question was interpreted by physician in the middle and final survey and 2) the awareness of the patients about the drugs has increased during the last years.
For those who answer negatively to the HIBP question, the response to the DRUGS question should be negative. The response to DRUGS question can, however, be positive even if the response is "no" or "insufficient data" in HIBP, if the MCC has confirmed this as a reply to unresolved constraint violations. Except in HUN-BUD in the initial survey where the HIBP question was not asked, this happened only in occasional cases.
Table 6 shows the responses to the question on antihypertensive drug treatment for persons who are aware of their hypertension (HIBP = 1). There was more than 10% of "uncertain" on antihypertensive drug treatment among those aware of high blood pressure in GER-ERF (Ini), GER-RDMc (Ini), LTU-KAU (Ini), and a more than 5% change in the proportion of uncertain data between surveys was observed in GER-COT, GER-EGEb, GER-ERF and LTU-KAU.
Large fluctuations in the prevalence of awareness and treatment of hypertension between surveys could indicate quality problems. Therefore, we analyzed the stability of the proportions of "yes"-answers for HIBP and DRUGS questions within RUAs between surveys (Tables 4 and 5). Excluding the RUAs which did not ask these questions in some surveys, a more than 10% change in the proportion of "yes"-answers to awareness of hypertension (HIBP=1) among all respondents was observed in five RUAs (FRA-TOU, ICE-ICE, RUS-MOC, RUS-MOIa, RUS-NOI). In ICE-ICE and at least partly in FRA-TOU this was due to a change in the questionnaire. RUS-NOI has informed that the decrease in awareness of hypertension was probably due to the deterioration of the health care system and changes in individuals' attitudes towards health during an economically and socially difficult period. The reason in the other RUAs is unknown. A more than 10% change in the proportion of "yes"-answers to antihypertensive drug treatment (DRUGS=1) among all respondents was observed in one RUA (RUS-NOCa). The MCC has explained that this may be due to a dramatic change in strategy of hypertension treatment during the last years, progress in medical knowledge, promotion efforts of pharmacological industry and appearance of lot of new drugs in Russia.
The MCCs in question were asked to try to identify the reasons for these large changes. In particular, it is important to exclude the possibility that they could be due to errors or changes in coding practices between the surveys.
Tables 7 and 8 give the overall scores for the quality of data for the items HIBP and DRUGS respectively in each RUA. The scores were defined as follows:
a) Questionnaire compliance
| Score= | 0 | if no questionnaire available at the MDC, no HIBP/DRUGS question in the questionnaire, or the formulation of the question is significantly different from the standard format; |
| 1 | if an alternative formulation of the question was used, but it is considered of having only minor or no effect on prevalence estimates; | |
| 2 | for all other RUAs. |
b) Insufficient data
| Score= | 0 | if proportion of "insufficient" > 10%; |
| 1 | if proportion of "insufficient" is 5-10%; | |
| 2 | for all other RUAs. |
c) Proportion of uncertain data (DRUGS only, Table 8)
| Score= | 0 | if proportion of "uncertain" among respondents aware of high blood pressure is > 10%; |
| 1 | if proportion of "uncertain" among respondents aware of high blood pressure is 5-10%; | |
| 2 | for all other RUAs. |
d) Trends
| Score= | 0 | if the question was not asked in one or several surveys, OR there is a more than 10% change in proportion of "yes"-answers in the RUA between surveys and no confirmation received from the MCC that the changes are real (i.e. not due to error or changes in coding practices), OR there is a more than 5% change in proportion of "uncertain" among respondents aware of high blood pressure (for DRUGS only); |
| 1 | if there is a 5-10% change in proportion of "yes"-answers in the RUA between surveys and no confirmation received from the MCC; | |
| 2 | for all other RUAs. |
Note: No summary score was assigned, because the adequacy of the data for the various cross-sectional and longitudinal analyses will vary depending on the circumstances and will therefore have to be judged on a case by case basis. Consult Tables 7 and 8 to judge the quality limitations of the data in each RUA.
The data on HIBP and DRUGS in the MONICA RUAs will be used to estimate the prevalence of hypertension control and changes in these estimates between the surveys. Also, one of the definitions of high blood pressure commonly used in MONICA (i.e. SBP>=160 mmHg or DBP>=95 mmHg or DRUGS=1) uses data on DRUGS as it considers all those on drug treatment as having high blood pressure. Before the data on hypertension control in the MONICA RUAs can be reported, bias arising from these variables must be excluded.
Half of the MCCs did not follow the period definition of "in the last two weeks" for the DRUGS question. However, in many RUAs the difference from the standard one was small and may have not introduced a bias. In the initial survey, in three RUAs (ICE-ICE, ISR-TEL, GER-BREa) the deviation of definition from the standard data transfer format was too large to be acceptable, and in one RUA (HUN-PEC) the question was not asked. In addition, in three RUAs (ICE-ICE, HUN-BUD, HUN-PEC) the awareness of hypertension could not be determined.
The availability and comparability of data improved by the middle survey: only one RUA (HUN-PEC) did not ask about hypertension control and several RUAs revised their questions to be in concordance with the MONICA standard. The same concerns the final survey where all RUAs asked these questions.
In two situations there would be difficulties in using data for trend analyses: 1) if there is a large change in the proportion of insufficient or uncertain data between the surveys within the RUA; and 2) if the main items for generating variables HIBP and/or DRUGS are equivocally described on the original questionnaire used in one or more surveys. These problems were uncommon.
In conclusion, data for most RUAs can be used for trend analyses of awareness and treatment of hypertension as summarized in Tables 7 and 8. Nevertheless, there are some minor problems that can cause some variation, although not a major bias, in the results of the comparative analyses either within or between RUAs. Moreover, even though the questions would look identical and strictly follow the MONICA standard data transfer format, it is very difficult to prove that the responses to the questions are fully comparable across many different countries and areas. A diagnosis of hypertension by a doctor may have been based on differing criteria for hypertension in the different RUAs and the criteria may have changed over time. The subject may also have misunderstood or forgotten what the doctor said. Similarly, there may be drugs for hypertension which have different effects.
The following list includes only those RUAs with specific findings or exceptional background information relevant for the use of data.
AUS-NEW
AUS-PERa
AUS-PERb
BEL-CHA and BEL-GHE
BEL-LUX
CAN-HAL
CHN-BEI
DEN-GLO
FIN-KUO, FIN-NKA and FIN-TUL
FRA-LIL
FRA-STR
FRA-TOU
GER-BREa
GER-BREb
GER-BER
GER-COT
GER-EGEb
GER-ERF
GER-HAC
GER-KMS
GER-RDMc
GER-RHN
HUN-BUD
HUN-PEC
ICE-ICE
ISR-TEL
ITA-LAT
LTU-KAU
MLT-MLT
NEZ-AUC
POL-TAR
POL-WAR
RUS-MOC and RUS-MOIa
RUS-MOIb
RUS-NOCa and RUS-NOI
RUS-NOCb
SWE-GOT
UNK-GLA
USA-STA
YUG-NOS