Article type
Year
Abstract
Objective: The ACE-inhibitor Myocardial Infarction Collaborative Group conducted a systematic overview of large-scale trials evaluating ACE-inhibitor therapy. The Gruppo ltaliano per lo Studio della Sopravvivenza nell'Infarto Miocardico Acuto (GISSI) was responsible for data collection, and data analysis of "early" trials, in which ACE-inhibitor treatment started in acute phase of myocardial infarction and continued for a short period of time (CCS-1, CONSENSUS-2, GISSI-3, ISIS-4 trials).
Methods: Between November 1995 and February 1996, a common protocol including 88 items was adopted to collect individual data from 98,496 enrolled in the four trials and representing about 98% of all the patients randomized in the trials retrieved for the overview. An automatic procedure was implemented to check data for completeness and internal consistency: all the discrepancies detected through the data checking process were notified to investigators and corrections included in the main database.
Results: The table shows the number of anomalies issued and solved and the type of checking:
Type #checking #issued #solved
Missing data 5 24 24
Numeric range 17 714 703
Data range 32 38 29
Logical 30 2902 2902
Total 84 3678 3658
Data were also checked for consistency with published results, treatment allocation and patient lost to follow-up, but no relevant discrepancies were found.
Discussion: An ad hoc protocol and checking procedure allowed to quickly collect and clean the 88 requested items. Protocol, data request and checking procedure will be discussed.
Methods: Between November 1995 and February 1996, a common protocol including 88 items was adopted to collect individual data from 98,496 enrolled in the four trials and representing about 98% of all the patients randomized in the trials retrieved for the overview. An automatic procedure was implemented to check data for completeness and internal consistency: all the discrepancies detected through the data checking process were notified to investigators and corrections included in the main database.
Results: The table shows the number of anomalies issued and solved and the type of checking:
Type #checking #issued #solved
Missing data 5 24 24
Numeric range 17 714 703
Data range 32 38 29
Logical 30 2902 2902
Total 84 3678 3658
Data were also checked for consistency with published results, treatment allocation and patient lost to follow-up, but no relevant discrepancies were found.
Discussion: An ad hoc protocol and checking procedure allowed to quickly collect and clean the 88 requested items. Protocol, data request and checking procedure will be discussed.