MORGAM Manual

Data Transfer Format: Follow-up for Atrial Fibrillation Form: 82 Version: 1 Date: 13.10.2012

© National Institute for Health and Welfare and the MORGAM Project investigators Last updated: 13 October 2012 For more information, please contact Kari Kuulasmaa (firstname.lastname@thl.fi)

The purpose of this transfer format is to provide an exact and common format for the MORGAM Participating Centres (MPCs) to transfer follow-up data on atrial fibrillation to the MORGAM Data Centre (MDC). Data in the format specified here should be submitted to the MDC for every member of those MORGAM cohorts that are followed up for atrial fibrillation. The data should be sent through e-mail, not on paper forms.

Atrial fibrillation is of particular interest in the MORGAM Biomarker SubStudy. The cohorts which have these data available at least for the cases where this is in hospital discharge records should transfer them to the MDC.

Contents

• Format specification
• General instructions
• Specific instructions for each item
• Updates

Format specification

General Instructions

The record format is ASCII fixed length and record size is 38 characters. The position of each data item in the record is given in column "COLUMNS" of this data transfer format. There should be one record of data for every member of each MORGAM cohort.

This format corresponds with the layout of the information on the electronic data file which is transferred, and need not correspond with the format of locally used paper forms or electronic files. Instructions for transferring the data are given in section Data communication between the Participating Centres and the MDC. To avoid errors, special attention should be paid in extracting these data items from the local data set.

The ITEM NAME on the document is a computer variable name used for the item by the MDC.

Blank fields are not allowed in the record.

Instructions for making corrections to data that have already been sent to the MDC are given in Section Data communication between the Participating Centres and the MDC.

Please contact the MDC for instructions if you cannot provide information as specified in this document or if you have any problems with the interpretation of the coding for any specific items.

Specific instructions for each item

Follow these instructions carefully when creating a computer file for the data transfer from the MPC to the MDC.

Number 82 indicates the "Data transfer format: Follow-up for atrial fibrillation".

Enter the version number from the heading of the "Data transfer format: Follow-up for atrial fibrillation" which you are using. If the version number of the data transfer format which you are using is not "1", these instructions do not correspond to the format you are using. Check that you are using the valid version of the format.

Items CENTRE...SERIAL

These are key items used for merging the different records of the same individual. Please make sure that the codes are identical with those used for the same person in the "Data transfer format: MONICA survey data".

Code 88 (irrelevant). The purpose of this item is to reserve fixed columns for the key items in all data transfer formats. This is a key item in the event data formats, but irrelevant for the other data transfer formats.

The first two columns are for the month; code 01 - 12, or code 99 if the month is not known. Enter the year of birth, in full four characters, in the last four columns. If the year of birth is not known, which should be very rare, derive the year of birth from an estimate of the age.

This item is the same as item MBIRTH of the "Data transfer format: MONICA survey data", and it will be used for double checking the key items. If you compile the data for this record from various sources, please extract the data for this item from a different source than for the MONICA survey record.

This item is the same as item SEX of the "Data transfer format: MONICA survey data", and it will be used for double checking the key items. If you compile the data for this record from various sources, please extract the data for this item from a different source than for the MONICA survey record.

Code 1 if atrial fibrillation was diagnosed in the person before or during follow-up.

Code 2 if no atrial fibrillation was diagnosed in the person or it is not known whether or not the person had atrial fibrillation.

Each MPC should decide on the exact definition of atrial fibrillation for their cohorts. The definition should be added, prior to data transfer, to the cohort description in the MORGAM internal web site. The description should cover:

• the data sources;
• follow-up procedures;
• diagnostic procedures; and
• any information on the coverage of the follow-up and the quality of the diagnosis, including references to validation studies if available.

Self-reported diagnosis as the only information source is insufficient during follow-up. However, self-reported information can be used for atrial fibrillation diagnosed before the baseline examination (i.e.  before the examination round in which the person entered the study) if other data sources are unavailable.

When routine clinical and death certificate diagnoses are used for the diagnosis of atrial fibrillation, the relevant ICD codes are usually

• ICD-8: 427.4 (Atrial fibrillation and flutter)
• ICD-9: 427.3 (Atrial fibrillation and flutter)
• ICD-10: I48 (Atrial fibrillation and flutter)

These combine atrial fibrillation and flutter, but flutter is far less common and and closely related.

When necessary, the MPC should adjust these to the local diagnostic practices and possible national adaptations of ICD. (For example, some national adaptations of ICD-9 have a fifth number which specifies atrial fibrillation alone.)

The atrial fibrillation record should be submitted also in the cases where the atrial fibrillation was secondary to other MORGAM end-points, such as a coronary or stroke event.

The exit takes place when the earliest one of the following occur:

• Atrial fibrillation is diagnosed first time in the person (i.e. AF_STATUS = 1). See specific instruction for item AF_STATUS for definition of atrial fibrillation. The date of atrial fibrillation should be coded here even if it takes place before the baseline examination. (However, see instructions for code 99999999 below);
• The person dies;
• The person is alive at and has not had atrial fibrillation by the fixed end-date of a follow-up period or the individual end-date of follow-up, such as date of questionnaire return or date of interview;
• The person is lost to follow-up. Most likely reasons are: change of contact information, emigration to another country or refusal to participate in the study any longer.

For persons lost to follow-up right after baseline, or if there is no follow-up data available, code the date of baseline examination (i.e. DEXAM of the first ROUNDS when the person entered the study).

If atrial fibrillation was diagnosed before the baseline examination and nothing else is known about the date of the diagnosis, then AF_EXIT should be coded 99999999. Otherwise missing data are not accepted. If the exact date is not known, provide the best estimate. For example, if only the month of diagnosis is known, code 15 in the first two columns. If the person was lost to follow-up and there is no other evidence about the date when this happened, use the mid point between the day when the person was last known to be under follow-up for non-fatal events and the date when the loss to follow-up was observed.

Note: The principles of coding AF_EXIT differs from that of items EXDATEC and EXDATES for the exit from the follow-up for non-fatal coronary and stroke events. In case of the disease, the date of diagnosis is coded for AF_EXIT whereas in EXDATEC and EXDATES the end of follow-up assuming no disease event is always coded. The more simple coding for AF_EXIT is made possible by the fact that we are interested only of the first diagnosis of atrial fibrillation, but we are interested also in recurrent coronary and stroke events.

These items intend to gather information on likely predisposing factors to the atrial fibrillation.

Updates