next   contents

Description and quality of the cohorts Sangita Kulathinal1,2 and Kari Kuulasmaa1 for the MORGAM Project3

¹ Department of Health Promotion and Chronic Disease Prevention, National Public Health Institute, Helsinki, Finland
² Since January 2007 at Indic Society for Education and Development (INSEED), Nashik, India
³ See Annex for the sites and key personnel of contributing MORGAM Centres

© National Institute for Health and Welfare and the MORGAM Project investigators Last updated: 4 July 2007 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi)

Contents

• 1. Introduction
• 2. Target populations of the cohorts
• 3. Selection and recruitment of the cohort members
• 4. Periods of baseline examinations
• References
• Updates to this document
• Selected tables:
  • Table 1. Number of observations in sex and 10-year age groups
  • Table 2. Availability of major CVD risk factors
  • Table 3. Number of observations by month and year of examination

1. Introduction

The MORGAM cohorts are groups of people who were examined at the baseline and followed up for coronary heart disease, stroke and all-cause mortality. The data from the baseline examinations and the follow-up are described and assessed in separate sections of this document. The purpose of the current section is to assess general aspects of the selection of the cohort members, such as the underlying populations, the selection of the members of the cohorts, the relationship between the cohorts and the underlying populations, and the periods of the baseline examinations.

2. Target populations of the cohorts

Being general population cohorts, the MORGAM cohorts were selected in such a way that they would represent a target population defined by a geographic region, age and sex. The target populations, which the cohort were intended to represent have been described in the Full Descriptions of MORGAM Cohorts.

The sexes included in the cohorts, and the full age range of the cohort members at baseline are summarized in Table 2 of the Description of MORGAM Cohorts. Most of the cohorts include both men and women, with the exception of:

• FIN-ATB, FRA-LIL/STR/TOU, Cohort 21 of RUS-NOV and UNK-BEL, which included only men.

The age distribution of the Cohorts is given in Appendix 1, and a summary by sex and 10-year age group in Table 1. The lower age limit is mostly 25 or 35 years. The upper age limit varies mostly between 65 and 75 years. The youngest people are in ITA-ROM, where three of the cohorts included people of age 20 years. The age range was relatively narrow in:

• FIN-ATB: around 54-75 years of age;
• FRA-LIL/STR/TOU, UNK-BEL: around 49-60 years;
• ITA-SHE: 45-65 years;
• UNK-CAE: around 49-66 years;
• UNK-SHE: around 39-61 years.

The age range was contiguous except in DEN-GLO, where the age was around 30, 40, 50 or 60 years at baseline. Two cohorts were based on re-examinations of earlier cohort:

• DEN-GLO: Cohort 21 is based on re-examination of Cohort 01;
• ITA-ROM: Cohort 24 is based on re-examination of Cohorts 21 and 22.

In these cohorts the age ranges at baseline are shifted accordingly. The re-examinations were included in MORGAM because DNA samples were collected only during the re-examination.

3. Selection and recruitment of the cohort members

The procedures used for selecting the cohorts from the target population are described in the Full Descriptions of MORGAM Cohorts. In most cases a random sample of the population was invited to the baseline examinations. Exceptions are:

• FIN-ATB, where all smoking men were invited, after which eligibility criteria based on health were applied.
• FRA-LIL/STR/TOU and UNK-BEL, where occupational and other groups were selected. Nevertheless, the cohort was recruited to broadly match the social class structure of the background population.

The extent to which the cohorts, selected using probability sampling, represent the target populations depends on the quality of the sampling frame, exclusion criteria applied in the survey and survey non-response. These factors were examined thoroughly for the cohorts that took part in the WHO MONICA Project in the relevant MONICA quality assessment report [1]. In summary. there was an attempt to use the best available sampling frames, although some of them were not very up-to-date. Many baseline surveys excluded institutionalized persons or those who were too ill to participate. The survey response rates were mostly between 60% and 80% (see the Full Descriptions of MORGAM Cohorts). Studies comparing mortality among survey respondents and non-respondents suggest that both all-cause mortality and cardiovascular disease mortality among the non-respondents are twice the mortality among respondents [2 and references therein]. This suggests that the non-respondents have an over representation of high-risk people, and perhaps people that are already severely ill at the time of the baseline examination.

The MORGAM inclusion criterion for cohorts of availability of data on smoking, blood pressure and cholesterol could also affect the representativeness of the cohort. However, comparison of the number of MONICA respondents and the MORGAM cohort sizes reveals that this criterion was not followed by most MPCs. Table 2 shows the availability of data on systolic blood pressure, total cholesterol and daily smoking in the MORGAM cohorts. The availability of all these data is always very high, except for:

• Cohort 01 of RUS-NOV, where it is 89% because cholesterol values were excluded for a period when there were problems with laboratory standardization; and
• UNK-CAE, where it is 86%, also because of missing cholesterol values.

Based on these findings, it is unlikely that missing data on the three cardiovascular risk factors have a visible effect on the representativeness of the cohorts.

4. Periods of baseline examinations

Table 2 of the Description of MORGAM Cohorts shows the calendar years during which the baseline examinations were carried out. Table 3 gives the number of observations by month and year of examination. This is based on data item DEXAM of Form 20. The examination periods vary between the RUAs from few months in a year to five years (UNK-CAE). Perhaps the most critical issue concerning the examination period is the potential seasonal variation of the cardiovascular risk factors, such as blood lipids and blood pressure. There are differences in the survey months between the RUAs, some covering evenly the full year, whereas others were carried out within few calendar months at different times of the year. This should be kept in mind when pooling data from different cohorts. However, where there are several Cohorts in the same RUA, the examination periods, or at least the months of examinations, of the Cohorts are similar. Therefore, the survey months are not usually a critical factor for pooling Cohorts within RUAs.

In general, all baseline measurements were taken on one occasion. There is, however, one major exception:

• In FIN-ATB, where the MORGAM baseline (i.e. data item DEXAM of Form 20) was 1992-93, the DNA samples  and smoking data were collected at that time. Blood pressure and weight were measured 0-8 months and serum total cholesterol and HDL cholesterol 1.5-5 years earlier. The other baseline data were collected in 1985-1988, i.e. 4 to 8 years before the MORGAM baseline.

References

1. Wolf HK, Kuulasmaa K, Tolonen H, Ruokokoski E, for the WHO MONICA Project. Participation rates, quality of sampling frames and sampling fractions in the MONICA Surveys. (September 1998). Available from URL:http://www.ktl.fi/publications/monica/nonres/nonres.htm, URN:NBN:fi-fe19991076.
2. Jousilahti P, Salomaa V, Kuulasmaa K, Niemelä M, Vartiainen E. Total and cause specific mortality among participants and non-participants of population based health surveys: a comprehensive follow up of 54 372 Finnish men and women. J Epidemiol Community Health 2005;59:310-315.

Updates to this document

next   contents