In the news
Getting a GRiP on genetic risk prediction
|Study:||Strengthening the reporting of Genetic RIsk Prediction Studies: the GRIPS statement|
|By:||Janssens A.C.J.W. et al. (25 authors total)|
To produce a set of recommendations with the aim of enhancing transparency, quality and completeness of genetic risk prediction study reporting.
A two-day workshop with an international group of risk prediction researchers, epidemiologists, geneticists, methodologists, statisticians, and journal editors was organised to develop recommendations for the reporting of genetic risk prediction studies.
A series of reporting recommendations were produced as a 25 point checklist, including the points below.
- Title and abstract should allow easy identification as a genetic risk prediction study
- Introduction should outline scientific background, what is known on the topic, what gaps exist and existing risk prediction models in the area. Study objectives should also be clearly stated and outline the model to be investigated
- Methods should clearly outline key elements of study design and setting, include descriptions of the participant eligibility criteria, their recruitment and selection, and define all characteristics, risk factors and outcomes measured. Detailed descriptions of data collection methods and handling should also be outlined
- Construction of the risk model (including data used) should also be detailed along with how the model was evaluated and validated
- Results should detail the numbers of participants at each stage with explanation for drop-out. Demographic and clinical characteristics of the study population including risk factors used in the modeling should be reported, along with the associations between the variables used in the risk model and the outcome
- Results should also include estimates (and their precision) from the full risk model, an assessment on how well the model fits the data, its predictive ability and validation
- Discussion should include limitations and assumptions made and how they impact on the study, keeping in mind study objectives and any other relevant evidence. The generalisability of the model should also be discussed.
These guidelines aim to increase the quality and completeness of reporting in genetic risk prediction studies, which vary greatly in a similar manner to genetic association studies, for which the STREGA statement was developed (see previous news). The hope is that by increasing both the transparency and completeness of the reporting researchers can more easily pull together the evidence base and therefore allow applications to move into practice. This piece in PLoS Medicine is accompanied by a more detailed explanation and elaboration paper published in the European Journal of Human Genetics.