Reproducibility and impact factors on revision rate published concerning the cemented Lubinus SP II stem

Background and purpose Clinical studies and register data collections have different organizational prerequisites and base data that might have a relevant influence on the result. The objectives of the present paper are to identify potential bias factors inherent to scientific data, evaluate the quality of datasets, and examine the outcome of the implant with respect to its revision rate by comparative literature analysis. Methods By using a standardized methodology, a metaanalysis of clinical literature and register data concerning the cemented Lubinus SPII stem was carried out with the main criterion being the revision rate, which was calculated by means of the indicator “Revisions per 100 observed component years.” Results The results of the Lubinus SPII stem are to be rated as good. The datasets do not exhibit relevant bias factors. However, clinical studies showed a considerably greater variance in outcome than register data. The numbers of cases recorded through register data are 115 times higher than those of all the clinical studies together. The differences in outcome among the various countries vary by a factor of 2.3, with Sweden showing the best results. In individual cases also, other implants show deviations from the worldwide mean by a factor of 2–3. In Finland, it was possible to reduce the revision rates by one third within a period of 10 years. The learning curve is thus shown to have a significant impact on the results. Interpretation The results published on the Lubinus SPII stem are good and reliable. Owing to better basic data and lesser influences through individual circumstances, register data are superior to clinical literature. A standardization of evaluation and reporting procedures or combined evaluations of national register data can provide essential contributions to the scientific discussion.