Background estimations in neutrinoless double beta decay experiments (0nubetabeta) require reliable statistical limits on gamma-spectrometric low-level material screening measurements. For this purpose a custom method based on Bayesian statistics with reference to the international standard ISO 11929-7 is presented. The analysis combines the data from sample- and background spectra and comprises the physical knowledge of non-negative counting rates. It allows to incorporate multiple gamma lines of radionuclides. The confidence intervals pass continuously from two-sided intervals into single-sided upper limits.