The Stockholm Birth Cohort Study has two parts, the first part being the original data from the so called Metropolitan project, covering the period 1953-1983. The Metropolitan project data was de-identified in 1985, following a debate in the media and a public concern for personal integrity.
It was however established that the personal integrity of the persons included in the project was never breached during this time. As the data were de-identified, researchers continued to analyse the data that had been collected up to that point, especially studying how childhood social circumstances influenced children’s life chances. More than 100 scientific publications have resulted from the Metropolitan project.
The second part of the Stockholm Birth Cohort Study consists of follow-up data from the Swedish Work and Mortality Data Base, which is held at CHESS. This is a completely anonymous data base covering the whole Swedish population, with information about occupation, income, health and causes of death among other things. Logically, this data base must include most of the people that had once been members of the metropolitan study. By a “probability matching” we managed to link follow-up data to the original data, thereby providing a more than 50-year long follow-up period. This linkage was based on information that was common to both data sets, for instance a person’s district of residence, occupation and housing conditions. If the information was the same in both data sets we assumed that the person could be the same. We estimate that the matching is correct for over 90% of the cohort.
We have considered the ethics of matching two anonymous data sets to each other, including the uncertainty in linking. At no point in this procedure have we breached the anonymity of those individuals that are part of the cohort. Even attempting to identify a certain individual behind a specific anonymous observation would be a criminal offence. Researchers have access only to small segments of the data base. The data are stored in encrypted form, and the master copy is locked up in a safe. The risk that a person’s integrity is breached is extremely small. We find that such a risk is by far outweighed the positive value of being able to do research into determinants of children’s life chances.
The Swedish Research Council has supported the project financially and its ethical committee has approved of the project. The regional ethics committee has also approved the project (dnr 03-629). Please contact CHESS if you would like to see our ethics approval for the Stockholm Birth Cohort Study and for the Work and Mortality Data Base.
December 9, 2011
Page editor: Ylva Brännström Almquist