MORGAM Manual

MORGAM derived variables Form: 31 Version: 1 Date: 28.06.2005

© National Institute for Health and Welfare and the MORGAM Project investigators Last updated: 15 February 2007 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi)

The purpose of this form is to provide the definitions and format for the derived variables which are available at the MORGAM Date Centre (MDC) for MORGAM Manuscript Groups. Data in the format specified here can be provided for every member of each MORGAM cohort. The form will be appended (without changing the version number) whenever new derived variables become available.

This format should not be used for transferring the cohort data from the MORGAM Participating Centres to the MORGAM Data Centre, because all these variables can be derived from the data provided on the MORGAM Data Transfer Formats.

Contents:

• Format specification
  • Form identification, key and check items
  • Age
  • Blood pressure
  • Blood lipids
  • Body composition
  • Smoking
  • Schooling years
  • Disease history at baseline
  • CHD1: First CHD Event Type 1 during follow-up
  • STROKE1: First Stroke Event Type 1 during follow-up
  • STROKE3: First fatal or non-fatal likely cerebral infarction during follow-up
  • CVD3: First fatal or non-fatal CHD Event Type 1 or likely cerebral infarction during follow-up
  • Death during follow-up
• Columns of the format specification
• Definitions of the variables

Format specification

Columns of the format specification

SEQ.

Sequence number of the item on the form. It is not an identifier of the item, but it may refer to different items when the form is updated. It is for the MDC purposes only.

ITEM NAME

name used for the item in the MDC.

SPECIFICATION AND CODES

specification and values of the variable. More details can be found in the section "Definitions of the variables" below, or by following the hyperlink in column CHARACTERS.

CHARACTERS

indicates the data format in which the variable is available for the Manuscript Groups. In the cases where the value of the variable is fixed, the value is also given in this column.

COMMENTS

date when the variable was added or the latest revision.

Definitions of the variables

The definitions of each derived variable, with hyperlinks to relevant data items of the Data Transfer Formats, are given below.

Form Identification, key and check items:

Items FORM and VERSION

These items identify this data format.

Items CENTRE...EVENT

These are key items used for merging the different records of the same individual. They should be the same for all records of the same individual.

Item EVENT is used in some of the Data Transfer Formats to identify different end-point events on the same individual, and therefore it also is a key item. However, it is irrelevant for this data format.

Items MBIRTH and SEX

These are the same as items MBIRTH and SEX of the "Data transfer format: MONICA survey data", and they should be used for double checking the key items. If you combine data for this individual from from various sources, please check the consistency of items MBIRTH and SEX.

Age

These age group variables were derived from item EAGE of the "Data transfer format: MONICA survey data"

AGE1 is calculated as a difference between the date of birth (MBIRTH of the "Data transfer format: MONICA survey data") and the date of examination (DEXAM of the "Data transfer format: MONICA survey data"). The missing day of birth in MBIRTH was taken as 15. In cases where the month of the birth was missing, it was taken as 06.

Blood pressure

SYSTC1 is the first systolic blood pressure value, derived from items SYST1 and RZ1 of the "Data transfer format: MONICA survey data" as:

SYSTC2 is the second systolic blood pressure value, defined similarly as SYSTC1, but using items SYST2 and RZ2 of the "Data transfer format: MONICA survey data":

SYSTM is systolic blood pressure, derived from items SYSTC1 and SYSTC2 as:

Note that when only one measurement value is available, that is used for SYSTM.

DIASTC1 is the first diastolic blood pressure value, defined similarly as SYSTC1, but using items DIAST1 and RZ1 of the "Data transfer format: MONICA survey data" as:

DIASTC2 is the second diastolic blood pressure value, defined similarly as SYSTC1, but using items DIAST2 and RZ2 of the "Data transfer format: MONICA survey data" as:

DIASTM is diastolic blood pressure, defined similarly as SYSTM, but using items DIASTC1 and DIASTC2:

BP1 is derived from SYSTC1 and DIASTC1 as:

BP1 is derived from SYSTC2 and DIASTC2 as:

BPM is derived from SYSTM and DIASTM as:

HIGHBP1 is derived using self reported blood pressure drug use (DRUGS of form 20) and SYSTM and DIASTM as:

if (DRUGS = 1 or 3) or (140.0 < SYSTM < 999.9) or (90 < DIASTM < 999.9) then HIGHBP1 = 1, if (DRUGS = 2, 8 or 9) and (0 < SYSTM < 140) and (0 < DIASTM < 90) then HIGHBP1 = 2. In all the other cases, HIGHBP1 = 9.

Laboratory results

Total serum cholesterol to two decimal places in mmol/l, derived from items CHOL and CHOLDL of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest two decimals.

The MPCs provided the cholesterol data either in units mmol/l or mg/dl, depending on which units were used in the laboratory. Item CHOLA converts all measurements to units mmol/l.

HDL serum cholesterol to three decimal places in mmol/l, derived from items HDL and HDLDL of the "Data transfer format: MONICA survey data" as:

*The result is rounded to the nearest three decimals.

RCHOL to two decimal places is derived from items CHOLA and HDLA as:

*The result is rounded to the nearest two decimals.

NONHDL to two decimal places is derived from items CHOLA and HDLA as:

*The result is rounded to the nearest two decimals.

• Item was ímplemented on 2006-09-18

Serum triglycerides to two decimal places in mmol/l, derived from items TRIGL and TRIGLDL of the "Data transfer format: Baseline data addition 2"as:

*The result is rounded to the nearest two decimals.

Plasma fibrinogen to two decimal places in g/l, derived from items FIBR and FIBRDL of the "Data transfer format: Baseline data addition 2"as:

Body composition

BMI to two decimal places in kg/m², derived from items HEIGHT and WEIGHT of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest two decimals.

WHR to two decimal places, derived from items WAIST and HIP of the "Data transfer format: MONICA survey data"as:

*The result is rounded to the nearest two decimals.

• Item was ímplemented on 2006-12-08

Smoking

Daily cigarette smoking status of the person at baseline, derived from item CIGS of the "Data transfer format: MONICA survey data"as:

This item is relevant for ex-daily cigarette-smokers and provides an estimate of the age when they stopped smoking daily. It is derived from items MBIRTH and STOP of the "Data transfer format: MONICA survey data" as follows:

YBIRTH is defined as the last four characters of MBIRTH, i.e. the calendar year of birth. STOPAGE is then defined as:

CIGYRS estimates the number of years of daily cigarette smoking prior to the baseline examination. It does not take into account possible non-smoking periods between smoking periods or smoking after the baseline examination. CIGYRS is derived from items CIGS, EVERCIG, EAGE and CIGAGE of the "Data transfer format: MONICA survey data" and the derived variable STOPAGE as follows:

Note: The data for CIGYRS is misleading for the cohorts where the data for items CIGAGE or STOPAGE are missing commonly. For such cohorts, CIGYRS has value "0" for the never-daily-smokers, but it is "99" for all current and/or past daily smokers (see distributions of items CIGAGE and STOPAGE).

Current smoking status of the person at baseline, derived from items CIGS, CIGAR, PIPE, CIGARSM and PIPESM of the "Data transfer format: MONICA survey data"as:

Schooling years

SCHOOL3 is derived from SCHOOL taking into account the birth year of the subject (extracted from MBIRTH), SEX and CENTRE.

SCHOOL3 is coded 9 if schooling years are missing (SCHOOL=99).

The procedure for determining SCHOOL3 can be presented as follows:

1. The calculations are done separately for each CENTRE and SEX.
2. The subjects are sorted by the year of birth. Years with only few subjects are combined.
3. Schooling years are processed in moving windows of seven years. The moving window for year of birth BYEAR contains subjects who were born in year BYEAR-3, BYEAR-2, BYEAR-1, BYEAR, BYEAR+1, BYEAR+2 or BYEAR+3. Tertiles are calculated in each window.
4. Because the tertiles for certain year are integers, there four possible schemes to define the cut points: the lower tertile may belong to either to the lowest category or  to the middle category and the upper tertile may belong to either to the middle category or  to the highest category. For each scheme the percentages of subjects in each category are calculated and the schemes are scored according to these percentages.  The score is the sum of the following penalty terms:
   • Absolute differences between the percentages and the desired percentage 33.33 %.
   • Threshold penalty if the extreme categories have less than 15 % of subjects.
   • Linear penalty if the extreme categories have less than 25 % of subjects.
   • Absolute differences between the percentages and the percentages of the previous window.
   • Small threshold penalty if the resulting cut points are smaller the cut points of the previous window. (We know that the schooling years should an increasing trend)
5. The scheme with lowest score is chosen and the cut points are defined.
6. The cut points are used to determine the value of SCHOOL3 for each subject.

• Item was drafted on 2006-09-18 and updated 2007-01-05

Disease history at baseline

BASEMI1 is derived from items HISMI1 and HISMI2 of the "Data transfer format: additional baseline data" (Form 21) as:

BASEMI1 =9 for everybody in the Cohorts where specific data for HISMI1 and HISMI2 are not available (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASEMI2 is derived from items HISMI1, HISMI2 and HISUC of the "Data transfer format: additional baseline data" (Form 21) as:

When specific data for HISMI1 and HISMI2 are not available, BASEMI2 uses data on HISUC, which indicates Coronary Heart Disease but the data are unspecific to separate between angina pectoris, MI and possibly cardiac revascularization (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASESTR1 is derived from items HISSTR1 and HISSTR2 of the "Data transfer format: additional baseline data" (Form 21) as:

BASECVD1 is derived from items BASEMI1 and BASESTR1 as:

BASECVD1 = 9 for everybody who did not have a stroke in the Cohorts where specific data on documented or self-reported history of MI are not available (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

BASECVD2 is derived from items BASEMI2 and BASESTR1 as:

When specific data on documented or self-reported MI are not available, BASECVD2 uses data on self-reported data on MI, angina pectoris or possibly cardiac revascularization, where these three conditions cannot be separated from each other (e.g. Cohort 3 of MPC 57 and all cohorts of MPC 36).

CHD 1: First fatal or non-fatal CHD Event Type 1 during follow-up

CHD Event Type 1 is a fatal or non-fatal coronary event (Form 22) with any of the following diagnostic categories or event type:

• a) EVTYPE = 3 and 
  • DGNCAT = 1 (definite AMI)
  • DGNCAT = 2 (possible AMI or coronary death)
  • DGNCAT = 3 (definite or possible, not specifiable)
  • DGNCAT = 4 (unstable angina pectoris)
  • DGNCAT = 5 (unclassifiable) and SURVIV = 2 only
• EVTYPE = 2 or 4 (cardiac revascularization)

The follow-up for CHD1 events ends at EXDATEC of Form 27 and EXDATE of Form 25, whichever is earlier.

CHD1 is coded 1 if the person had a CHD Event Type 1 during the follow-up and the first such event was fatal (SURVIV = 2).

CHD1 is coded 2 if the person had a CHD Event Type 1 during the follow-up and the first such event was non-fatal (SURVIV = 1).

CHD1 is coded 3 if the person did not have a CHD Event Type 1 during the follow-up.

CHD1 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATEC = 77777777 or 88888888).

CHD1 is coded 9 when there are no follow-up data available.

CHDDATE1 is the date of exit from the follow-up for a CHD1 event. It  is derived from EVDATE of Form 22, EXDATEC of Form 27 and EXDATE of Form 25.

If CHD1=1 or 2, then CHDDATE1 = EVDATE of the coronary event.

If CHD1= 3, then CHDDATE is the earlier one of EXDATEC and EXDATE.

Note: In case if the follow-up for non-fatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calendar period by the event register which was used for the follow-up), the follow-up of fatal CHD1 events  is also considered only upto that time.  An upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa, RUS-NOVa and SWE-NSWa was 65 years. For POL-WARa the follow-up of fatal events is ended on 31.12.1994.

If CHD1 is 8, CHDDATE1 is coded as 88888888.

CHDDATE1 is coded as 99999999 if CHD1 is 9 or EVDATE  is not a valid date. 

CHDTIME1 gives the length of the follow-up for a CHD1 event in days. It is the difference between the CHDDATE1 and DEXAM (Form 20).

If CHD1 is 8, CHDTIME1 is coded as 88888.

CHDTIME1 is coded as 99999  if CHD1 is 9 or CHDDATE1 = 99999999.

CHDAGE1 gives the age at the time of exit from the follow-up of a CHD1 event. It is the sum of the age at the baseline examination (AGE1) and CHDTIME1/365.25.

If CHD1 is 8, CHDAGE1 is coded as 888.88.

CHDAGE1 is coded as 999.99   if CHD1 is 9 or CHDTIME1 = 99999.

STROKE 1: First fatal or non-fatal Stroke Event Type 1 during follow-up

Stroke Event Type 1 is a fatal or non-fatal stroke event (Form 23) with any of the following diagnostic categories:

• DGNCAT = 1 (definite)
• DGNCAT = 9 (unclassifiable)

The follow-up for STROKE1 events ends at EXDATES of Form 28 and EXDATE of Form 25, whichever is earlier.

STROKE1 is coded 1 if the person had a Stroke Event Type 1 during the follow-up and the first such event was fatal (SURV28 = 2).

STROKE1 is coded 2 if the person had a Stroke Event Type 1 during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE1 is coded 3 if the person did not have a Stroke Event Type 1 during the follow-up.

STROKE1 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE1 is coded 9 when there are no follow-up data available.

STRDATE1 is the date of exit from the follow-up for a STROKE1 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE1=1 or 2, then STRDATE1 = EVDATE of the stroke event.

If STROKE1= 3, then STRDATE1 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only upto that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa and SWE-NSWa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE1 is 8, STRDATE1 is coded as 88888888.

STRDATE1 is coded as 99999999 if STROKE1 is 9 or EVDATE is not a valid date .

STRTIME1 gives the length of the follow-up for a STROKE1 event in days. It is the difference between the STRDATE1 and DEXAM (Form 20).

If STROKE1 is 8, STRTIME1 is coded as 88888.

STRTIME1 is coded as 99999 if STROKE1 is 9 or STRDATE1=99999999.

STRAGE1 gives the age at the time of exit from the follow-up of a STROKE1 event. It is the sum of age at the baseline examination (AGE1) and STRTIME1/365.25.

If STROKE1 is 8, STRAGE1 is coded as 888.88.

STRAGE1 is coded as 999.99 if  STROKE1 is 9 or STRTIME1 is 99999.

STROKE3: First fatal or non-fatal likely cerebral infarction during follow-up

Likely cerebral infarction is a fatal or non-fatal event based on the specific MORGAM diagnosis of cerebral infarction (item CI)  when that is available, and on the clinical or death certificate diagnosis when CI is undetermined. The definition is less reliable than an event definition based on validated diagnosis, but unlike the validated diagnosis, it is available for all cohorts.

More specifically, the diagnosis is based on the data of Data Transfer Format: Stroke Events (Form 23) and Data Transfer Format: Follow-up Data ( Form 25). Potential events are those whose diagnostic category is "definite stroke" or "unclassifiable" (i.e. DGNCAT of Form 23 = 1 or 9):

• If the diagnosis was cerebral infarction (i.e. CI = 1), then the event is a likely cerebral infarction. 
• If the diagnosis was not cerebral infarction (i.e. CI = 2), then the event is not a likely cerebral infarction. 
• If the data was insufficient for the diagnosis (i.e. CI = 9) but the clinical or death diagnoses are available, then the event is a likely cerebral infarction if:
  • any of the clinical diagnoses available indicates cerebral infarction (i.e. CLIND1 or CLIND2 or CLIND3 has value:
    • 432, 433 or 434 in ICD-8;
    • 433 or 434 in ICD-9; or
    • I63 in ICD-10; or
  • the event is fatal and the underlying, direct or antecedent cause of death indicates cerebral infarction (i.e. SURV28 = 2 and DEATHDU, DEATHDA, DEATHDB or DEATHDC has one of the above ICD-values).

The follow-up for STROKE3 events ends at EXDATES of Form 28 and EXDATE of Form 25, whichever is earlier.

STROKE3 is coded 1 if the person had a likely cerebral infarction during the follow-up and the first such event was fatal (SURV288 = 2).

STROKE3 is coded 2 if the person had a likely cerebral infarction during the follow-up and the first such event was non-fatal (SURV28 = 1).

STROKE3 is coded 3 if the person did not have a likely cerebral infarction during the follow-up.

STROKE3 is coded 8 (irrelevant) if there was no follow-up for non-fatal events for the person (EXDATES = 77777777 or 88888888).

STROKE3 is coded 9 when there are no follow-up data available..

• Item was implemented on 2006-12-20

STRDATE3 is the date of exit from the follow-up for a STROKE3 event. It is derived from EVDATE of Form 23, EXDATES of Form 28 and EXDATE of Form 25.

If STROKE3 = 1 or 2, then STRDATE3 = EVDATE of the stroke event..

If STROKE3 = 3, then STRDATE3 is the earlier one of EXDATES and EXDATE.

Note: In case if the follow-up for nonfatal events was ended prior to the follow-up for fatal events (either due to upper age-limit or due to coverage of calender period by the event register which was used for the follow-up), fatal follow-up is also considered only upto that time.  The upper age limit for the follow-up of nonfatal events for LTU-KAUa, POL-WARa was 65 years. For  RUS-NOVa and SWE-NSWa it was 75 years. For POL-WARa, follow-up of fatal events ended on 31.12.1994.

If STROKE3 is 8, STRDATE3 is coded as 88888888.

STRDATE3 is coded as 99999999 if STROKE3 is 9 or EVDATE is not a valid date.

• Item was implemented on 2006-12-20

STRTIME3 gives the length of the follow-up for a STROKE3 event in days. It is the difference between the STRDATE3 and DEXAM (Form 20).

If STROKE3 is 8, STRTIME3 is coded as 88888.

STRTIME3 is coded as 99999 if STROKE3 is 9 or STRDATE3 = 99999999.

• Item was implemented on 2006-12-20

STRAGE3 gives the age at the time of exit from the follow-up of a STROKE3 event. It is the sum of age at the baseline examination (AGE1) and STRTIME3/365.25.

If STROKE3 is 8, STRAGE3 is coded as 888.88.

STRAGE3 is coded as 999.99 if  STROKE3 is 9 or STRTIME3 is 99999.

• Item was implemented on 2006-12-20

CVD3: First fatal or non-fatal CHD Event Type 1 or likely cerebral infarction during follow-up

CVD3 is derived from CHDAGE1, STRAGE3, CHD1 and STROKE3 as follows:
 

CVDDATE3 is the date of exit from the follow-up for a CVD3 event. It is derived from CHD1, STROKE3, CHDAGE1, STRAGE3, CHDDATE1, STRDATE3 as follows:

CVDTIME3 gives the length of the follow-up for a CVD3 event in days. It is derived from CHD1, STROKE3, CHDAGE1, STRAGE3, CHDTIME1 and STRTIME3 as follows:

CVDAGE3 gives the age at the time of exit from the follow-up of a CVD3 event. It is derived from CHD1, STROKE3, CHDAGE1 and STRAGE3 as follows:

Death: Death during follow-up

DEATH is coded 1 if the person died during the follow-up (EXREAS of Form 25 = 1).

DEATH is coded 2 if the person did not die during the follow-up (EXREAS of Form 25 = 2, 3, 4, 5).

DEATH is coded 9 when there are no follow-up data available due to any reason.

For the analysis of DEATH as an end-point, the date, follow-up time and age at the time of exit from the follow-up of DEATH are the same as MORTDATE, MORTTIME and MORTAGE.

CHD2 is coded 1 if the person died due to coronary heart disease during the follow-up (SURVIV of Form 22 = 2 and DGNCAT of Form 22 <>7).

CHD2 is coded 2 if the person did not die due to coronary heart disease during the follow-up.

CHD2 is coded 9 when there are no follow-up data available due to any reason.

Note that CHD2 = 2 includes those who survived during the follow-up or who died due to cause other than coronary heart disease. For the analysis of CHD2 as an end-point, the date, follow-up time and age at the time of exit from the follow-up of CHD2 are the same as MORTDATE, MORTTIME and MORTAGE.

STROKE2 is coded 1 if the person died due to stroke during the follow-up (SURV28 of Form 23 = 2 and DGNCAT of Form 23 <> 4).

STROKE2 is coded 2 if the person did not die due to stroke during the follow-up.

STROKE2 is coded 9 when there are no follow-up data available due to any reason.

Note that STROKE2 = 2 includes those who survived during the follow-up or who died due to cause other than stroke. For the analysis of STROKE2 as an end-point, the date, follow-up time and age at the time of exit from the follow-up of STROKE2 are the same as MORTDATE, MORTTIME and MORTAGE.

MORTDATE is the date of exit from the mortality follow-up and is equal to EXDATE of Form 25.

If DEATH=9, then MORTDATE = 99999999.

MORTTIME gives the length of the mortality follow-up in days. It is the difference between the MORTDATE and DEXAM (Form 20). When DEATH is 9, MORTIME is coded as 99999.

MORTAGE gives the age at the time of exit from the mortality follow-up. It is the sum of the age at the baseline examination (AGE1) and (MORTTIME/365.25). When DEATH is 9, MORTAGE is coded as 999,99.