Extracting data in file-based forensic psychiatric research: Some methodological considerations

This article explores possibilities and problems of the use of file-based data in forensic psychiatric research. Attention is paid to methodological problems with regard to data quality, operationalization of information into variables, and procedures for reliable data extraction, including interrater reliability assessments. We report on experiences from an ongoing Swedish forensic psychiatric research project based on complete forensic psychiatric investigation records, the National Police Register and the National Hospital Discharge Register. We conclude that when forensic psychiatric investigation records and population-based registers are combined, various factors for research can be extracted reliably. However, contrary to expectations, several categories of variables could not be consistently extracted. Examples of both types of variables are provided.