Prognostic systematic reviews: challenges and insights: an example.

BACKGROUND AND

OBJECTIVES

Conducting systematic reviews of prognostic factors is challenging, especially in the presence of abundant evidence. In this paper, we describe challenges encountered and lessons learned from our experience in conducting a systematic review and meta-analysis for prognostic biomarkers used in Autosomal Dominant Polycystic Kidney Disease (ADPKD).

METHODS

Throughout the process of conducting the systematic review, we drafted the challenges encountered at each step and tackled potential solutions and insights for the addressed challenges. We then met as a group to discuss whether any further individual issues needed to be added and what aspects would have made the process easier.

RESULTS

Despite a few issues in multiple steps of the process, the authors thought the most challenging step was data analysis and interpretation, mainly due to the variability in how the data were reported. Other challenges were related to screening and extraction of data due to the large number of studies that were not primarily prognostic studies but included valuable data to this review.

CONCLUSION

Using the worked example of a prognostic systematic review and meta-analysis for biomarkers used in ADPKD, we believe that the challenges mentioned and tips provided will make the process easier for researchers conducting future prognostic systematic reviews focusing on prognostic factors.

PLAIN LANGUAGE SUMMARY: Systematic reviews of prognostic biomarkers help showcase which factors may predict disease progression. However, conducting these reviews can be challenging, especially when there is a lot of evidence and data are reported inconsistently. In this paper, we describe the difficulties we encountered while conducting a systematic review and meta-analysis of serum and urine prognostic biomarkers in Autosomal Dominant Polycystic Kidney Disease (ADPKD). Some challenges included analyzing and interpreting data because studies often reported results differently and study screening and data extraction, as many studies were not designed primarily as prognostic studies but still contained useful information. By sharing our experience and potential solutions, we hope to ease conducting future prognostic systematic reviews focusing on prognostic biomarkers.