Low number suppression is a process in which if results are below a certain threshold that a result of zero is returned instead. As an example, if we found out there was only one person who was treated in Oxford on a set day, who had a certain religion and of a certain ethnicity, it may be possible to establish who that person was. Therefore, even in datasets that have been pseudonymised, there may be a series of data points that result in only a small number of individuals having those characteristics. In this case, there is a risk that individuals could be identified. A threshold must be set for each Data Source at which the software will state there are no results, instead of the real number. Most Data Partners have set this threshold to 10, meaning any result that should return a result below 10, will instead always state 0 results.
