Advertisement

European Journal of Epidemiology

, Volume 26, Issue 4, pp 255–259 | Cite as

Strengthening the reporting of genetic risk prediction studies: the GRIPS statement

  • A. Cecile J. W. Janssens
  • John P. A. Ioannidis
  • Cornelia M. van Duijn
  • Julian Little
  • Muin J. Khoury
  • for the GRIPS Group
Open Access
Guidelines

Abstract

The rapid and continuing progress in gene discovery for complex diseases is fueling interest in the potential application of genetic risk models for clinical and public health practice. The number of studies assessing the predictive ability is steadily increasing, but the quality and completeness of reporting varies. A multidisciplinary workshop sponsored by the Human Genome Epidemiology Network developed a checklist of 25 items recommended for strengthening the reporting of Genetic RIsk Prediction Studies (GRIPS), building on the principles established by prior reporting guidelines. These recommendations aim to enhance the transparency of study reporting, and thereby to improve the synthesis and application of information from multiple studies that might differ in design, conduct, or analysis. A detailed Explanation and Elaboration document is published.

Keywords

Genetic Risk prediction Methodology Guidelines Reporting 

Abbreviations

GRIPS

Genetic risk prediction studies

HuGENet

Human genome epidemiology network

REMARK

Guidelines for reporting of tumor marker studies

STARD

Standards for reporting diagnostic accuracy

STREGA

Strenghtening the reporting of genetic association studies

STROBE

Strengthening the reporting of observational studies in epidemiology

Introduction

The recent successes of genome-wide association studies and the promises of whole genome sequencing fuel interest in the translation of this new wave of basic genetic knowledge to health care practice. Knowledge about genetic risk factors may be used to target diagnostic, preventive, and therapeutic interventions for complex disorders based on a person’s genetic risk, or to complement existing risk models based on classical nongenetic factors such as the Framingham risk score for cardiovascular disease. Implementation of genetic risk prediction in health care requires a series of studies that encompass all phases of translational research [1, 2], starting with a comprehensive evaluation of genetic risk prediction.

With increasing numbers of discovered genetic markers that can be used in future genetic risk prediction studies, it is crucial to enhance the quality of the reporting of these studies, since valid interpretation could be compromised by the lack of reporting of key information. Information that is often missing includes details in the description of how the study was designed and conducted (e.g., how genetic variants were selected and coded, how risk models or genetic risk scores were constructed, and how risk categories were chosen), or how the results should be interpreted. An appropriate assessment of the study’s strengths and weaknesses is not possible without this information. There is ample evidence that prediction research often suffers from poor design and bias, and these may also have an impact on the results of the studies and on models of disease outcomes based on these studies [3, 4, 5]. Although most prognostic studies published to date claim significant results [6, 7], very few translate to clinically useful applications. Just as for observational epidemiological studies [8], poor reporting complicates the use of the specific study for research, clinical, or public health purposes and hampers the synthesis of evidence across studies.

Reporting guidelines have been published for various research designs [9], and these contain many items that are also relevant to genetic risk prediction studies. In particular, the guidelines for genetic association studies (STREGA) have relevant items on the assessment of genetic variants, and the guidelines for observational studies (STROBE) have relevant items about the reporting of study design. The guidelines for diagnostic studies (STARD) and those for tumor marker prognostic studies (REMARK) include relevant items about test evaluation; the REMARK guidelines also have relevant items about risk prediction [10, 11, 12, 13]. However, none of these guidelines are fully suited to genetic risk prediction studies, an emerging field of investigation with specific methodological issues that need to be addressed, such as the handling of large numbers of genetic variants (from 10 s to 10,000 s) and flexibility in handling such large numbers in analyses. We organized a 2-day workshop with an international group of risk prediction researchers, epidemiologists, geneticists, methodologists, statisticians, and journal editors to develop recommendations for the reporting of Genetic RIsk Prediction Studies (GRIPS).

Genetic risk prediction studies

Genetic risk prediction studies typically develop or validate models that predict the risk of disease, but they are also being investigated for use in predicting prognostic outcome, treatment response, or treatment-related harms. Risk prediction models are statistical algorithms, which may be simple genetic risk scores (e.g., risk allele counts), may be based on regression analyses (e.g., weighted risk scores or predicted risks), or may be based on more complex analytic approaches such as support vector machine learning or classification trees. The risk models may be based on genetic variants only, or include both genetic and nongenetic risk factors [14].

Aims and use of the GRIPS statement

The 25 items of the GRIPS statement are intended to maximize the transparency, quality, and completeness of reporting on research methodology and findings in a particular study. It is important to emphasize that these recommendations are guidelines only for how to report research and do not prescribe how to perform genetic risk prediction studies. The guidelines do not support or oppose the choice of any particular study design or method, e.g., the guidelines recommend that the study population should be described, but do not specify which population is preferred in a particular study.

The intended audience for the reporting guidelines is broad and includes epidemiologists, geneticists, statisticians, clinician scientists, and laboratory-based investigators who undertake genetic risk prediction studies, as well as journal editors and reviewers who have to appraise the design, conduct and analysis of such studies. In addition, it includes “users” of such studies who wish to understand the basic premise, design, and limitations of genetic prediction studies in order to interpret the results for their potential application in health care. These guidelines are also intended to ensure that essential data from future genetic risk prediction studies are presented in standardized form, which will facilitate information synthesis as part of systematic reviews and meta-analyses.

Items presented in the checklist are relevant for a wide array of risk prediction studies, because GRIPS focuses on the main aspects of the design and analysis of risk prediction studies. GRIPS does not address randomized trials that may be performed to test risk models, nor does it specifically address decision analyses, cost-effectiveness analyses, assessment of health care needs, or assessment of barriers to health care implementation [15]. Once the performance of a risk model has been established, these next steps toward implementation require further evaluation [10, 16]. For the reporting of these studies, which go beyond the assessment of genetic risk models as such, additional requirements apply. However, proper documentation of genetic predictive research according to GRIPS might facilitate the translation of research findings into clinical and public health practice.

Development of the GRIPS statement

The GRIPS statement was developed by a multidisciplinary panel of 25 risk prediction researchers, epidemiologists, geneticists, methodologists, statisticians, and journal editors, seven of whom were also part of the STREGA initiative [11]. They attended a 2-day meeting in Atlanta, Georgia (US) in December 2009 that was sponsored by the US Centers for Disease Control and Prevention on behalf of the Human Genome Epidemiology Network (HuGENet) [17]. Participants discussed a draft version of the guidelines that was prepared and distributed before the meeting. This draft version was developed on the basis of existing reporting guidelines, namely STREGA [11], REMARK [13], and STARD [12]. These were selected out of all available guidelines (see http://www.equator-network.org) because of their focus on observational study designs and genetic factors (STREGA), prediction models (REMARK), and test evaluation (REMARK and STARD). During the meeting, methodological issues pertinent to risk prediction studies were addressed in presentations. Workshop participants were asked to change, combine, or delete proposed items and add additional items if necessary. Participants had extensive post-meeting electronic correspondence. To harmonize our recommendations for genetic risk prediction studies with previous guidelines, we chose the same wording for the items wherever possible. Finally, we tried to create consistency with previous guidelines for the evaluation of risk prediction studies of cardiovascular diseases and cancer [2, 18]. The final version of the checklist is presented in Table 1.
Table 1

Reporting recommendations for evaluations of risk prediction models that include genetic variants

Title and abstract

 

1

(a) Identify the article as a study of risk prediction using genetic factors. (b) Use recommended keywords in the abstract: genetic or genomic, risk, prediction

Introduction

 Background and rationale

2

Explain the scientific background and rationale for the prediction study

 Objectives

3

Specify the study objectives and state the specific model(s) that is/are investigated. State if the study concerns the development of the model(s), a validation effort, or both

Methods

 Study design and setting

4a

Specify the key elements of the study design and describe the setting, locations, and relevant dates, including periods of recruitment, follow-up, and data collection

 Participants

5a

Describe eligibility criteria for participants, and sources and methods of selection of participants

 Variables: definition

6a

Clearly define all participant characteristics, risk factors and outcomes. Clearly define genetic variants using a widely-used nomenclature system

 Variables: assessment

7a

(a) Describe sources of data and details of methods of assessment (measurement) for each variable. (b) Give a detailed description of genotyping and other laboratory methods

 Variables: coding

8

(a) Describe how genetic variants were handled in the analyses. (b) Explain how other quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen, and why

 Analysis: risk model construction

9

Specify the procedure and data used for the derivation of the risk model. Specify which candidate variables were initially examined or considered for inclusion in models. Include details of any variable selection procedures and other model-building issues. Specify the horizon of risk prediction (e.g., 5-year risk)

 Analysis: validation

10

Specify the procedure and data used for the validation of the risk model

 Analysis: missing data

11

Specify how missing data were handled

 Analysis: statistical methods

12

Specify all measures used for the evaluation of the risk model including, but not limited to, measures of model fit and predictive ability

 Analysis: other

13

Describe all subgroups, interactions, and exploratory analyses that were examined

Results

 Participants

14a

Report the numbers of individuals at each stage of the study. Give reasons for nonparticipation at each stage. Report the number of participants not genotyped, and reasons why they were not genotyped

 Descriptives: population

15a

Report demographic and clinical characteristics of the study population, including risk factors used in the risk modeling

 Descriptives: model estimates

16

Report unadjusted associations between the variables in the risk model(s) and the outcome. Report adjusted estimates and their precision from the full risk model(s) for each variable

 Risk distributions

17a

Report distributions of predicted risks and/or risk scores

 Assessment

18

Report measures of model fit and predictive ability, and any other performance measures, if pertinent

 Validation

19

Report any validation of the risk model(s)

 Other analyses

20

Present results of any subgroup, interaction, or exploratory analyses, whenever pertinent

Discussion

 Limitations

21

Discuss limitations and assumptions of the study, particularly those concerning study design, selection of participants, and measurements and analyses, and discuss their impact on the results of the study

 Interpretation

22

Give an overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from similar studies, and other relevant evidence

 Generalizability

23

Discuss the generalizability and, if pertinent, the health care relevance of the study results

Other

 Supplementary information

24

State whether databases for the analyzed data, risk models, and/or protocols are or will become publicly available and if so, how they can be accessed

 Funding

25

Give the source of funding and the role of the funders for the present study. State whether there are any conflicts of interest

aMarked items should be reported for every population in the study

The GRIPS explanation and elaboration article

Accompanying this GRIPS statement, an Explanation and Elaboration document has been written, modeled after those developed for other reporting guidelines [19, 20, 21, 22]. The Explanation and Elaboration document illustrates each item with at least one published example that we consider transparent in reporting, explains the rationale for its inclusion in the checklist, and presents details of the items that need to be addressed to ensure transparent reporting. The Explanation and Elaboration document was produced after the meeting. The document was prepared by a small subgroup and shared with all workshop participants for additional revisions and final approval.

Concluding remarks and future directions

High-quality reporting reveals the strengths and weaknesses of empirical studies, facilitates the interpretation of the scientific and health care relevance of the results—especially within the framework of systematic reviews and meta-analyses—and helps build a solid evidence base for moving genomic discoveries into applications in health care practice. The GRIPS guidelines were developed to improve the transparency, quality and completeness of the reporting of genetic risk prediction studies. As outlined in the introduction, GRIPS does not prescribe how studies should be designed, conducted, or analyzed, and therefore the guidelines should not be used to assess the quality of empirical studies [23]. The guidelines should be used only to check whether all essential items are adequately reported.

Finally, the methodology for designing and assessing genetic risk prediction models is still developing. For example, newer measures of reclassification were first introduced in 2007 [24], and several alternative reclassification measures have been proposed [25]. Which measures to apply and when to use measures of reclassification are still subject to ongoing evaluation and discussion [26]. Furthermore, alternative strategies for constructing risk models other than simple regression analyses are being explored, and these may add increased complexity to the reporting. In formulating the items of the GRIPS statement, these methodological advances were anticipated. It is for this reason that the GRIPS statement recommends how a study should be reported and not how a study should be conducted or analyzed. Therefore, methodological and analytical developments will not immediately impact the validity and relevance of the items, but the GRIPS statement will be updated when this is warranted by essential new developments in the construction and evaluation of genetic risk models.

Notes

Acknowledgments

The members of the GRIPS group are: A. Cecile J. W. Janssens, John P. A. Ioannidis, Sara Bedrosian, Paolo Boffetta, Siobhan M. Dolan, Nicole Dowling, Isabel Fortier, Andrew N. Freedman, Jeremy M. Grimshaw, Jeffrey Gulcher, Marta Gwinn, Mark A. Hlatky, Holly Janes, Peter Kraft, Stephanie Melillo, Christopher J. O’Donnell, Michael J. Pencina, David Ransohoff, Sheri D. Schully, Daniela Seminara, Deborah M. Winn, Caroline F. Wright, Cornelia M. van Duijn, Julian Little, and Muin J. Khoury. Workshop was sponsored by the Centers for Disease Control and Prevention on behalf of the Human Genome Epidemiology Network (HuGENet). The findings and conclusions in this report are those of the authors and do not necessarily reflect the views of the Department of Health and Human Services. A. Cecile J. W. Janssens is financially supported by grants from the Erasmus University Medical Center Rotterdam, the Center for Medical Systems Biology in the framework of the Netherlands Genomics Initiative (NGI) and the VIDI grant of the Netherlands Organisation for Scientific Research (NWO). John P.A. Ioannidis: Tufts CTSI is supported by the National Institutes of Health/National Center for Research Resources (UL1 RR025752). Opinions in this paper are those of the authors and do not necessarily represent the official position or policies of the Tufts CTSI. Julian Little holds a Canada Research Chair in Human Genome Epidemiology. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of interest

John Ioannidis is a member of the PLoS Medicine Editorial Board.

Open Access

This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

References

  1. 1.
    Khoury MJ, Gwinn M, Yoon PW, Dowling N, Moore CA, et al. The continuum of translation research in genomic medicine: how can we accelerate the appropriate integration of human genome discoveries into health care and disease prevention? Genet Med. 2007;9:665–74.PubMedCrossRefGoogle Scholar
  2. 2.
    Hlatky MA, Greenland P, Arnett DK, Ballantyne CM, Criqui MH, et al. Criteria for evaluation of novel markers of cardiovascular risk: a scientific statement from the American Heart Association. Circulation. 2009;119:2408–16.PubMedCrossRefGoogle Scholar
  3. 3.
    Kyzas PA, Denaxa-Kyza D, Ioannidis JP. Quality of reporting of cancer prognostic marker studies: association with reported prognostic effect. J Natl Cancer Inst. 2007;99:236–43.PubMedCrossRefGoogle Scholar
  4. 4.
    Kyzas PA, Loizou KT, Ioannidis JP. Selective reporting biases in cancer prognostic factor studies. J Natl Cancer Inst. 2005;97:1043–55.PubMedCrossRefGoogle Scholar
  5. 5.
    McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, et al. REporting recommendations for tumor MARKer prognostic studies (REMARK). Nat Clin Pract Urol. 2005;2:416–22.PubMedCrossRefGoogle Scholar
  6. 6.
    Kyzas PA, Denaxa-Kyza D, Ioannidis JP. Almost all articles on cancer prognostic markers report statistically significant results. Eur J Cancer. 2007;43:2559–79.PubMedGoogle Scholar
  7. 7.
    Tzoulaki I, Liberopoulos G, Ioannidis JP. Assessment of claims of improved prediction beyond the Framingham risk score. JAMA. 2009;302:2345–52.PubMedCrossRefGoogle Scholar
  8. 8.
    von Elm E, Egger M. The scandal of poor epidemiological research. BMJ. 2004;329:868–9.CrossRefGoogle Scholar
  9. 9.
    Simera I, Moher D, Hoey J, Schulz KF, Altman DG. A catalogue of reporting guidelines for health research. Eur J Clin Invest. 2010;40:35–53.PubMedCrossRefGoogle Scholar
  10. 10.
    von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med. 2007;4:e296. doi: 10.1371/journal.pmed.0040296.CrossRefGoogle Scholar
  11. 11.
    Little J, Higgins JP, Ioannidis JP, Moher D, Gagnon F, et al. STrengthening the REporting of genetic association studies (STREGA): an extension of the STROBE statement. PLoS Med. 2009;6:e22. doi: 10.1371/journal.pmed.1000022.PubMedCrossRefGoogle Scholar
  12. 12.
    Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. BMJ. 2003;326:41–4.PubMedCrossRefGoogle Scholar
  13. 13.
    McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, et al. Reporting recommendations for tumor marker prognostic studies. J Clin Oncol. 2005;23:9067–72.PubMedCrossRefGoogle Scholar
  14. 14.
    Janssens ACJW, Van Duijn CM. Genome-based prediction of common diseases: methodological considerations for future research. Genome Med. 2009;1:20.PubMedCrossRefGoogle Scholar
  15. 15.
    Khoury MJ, Gwinn M, Ioannidis JP. The emergence of translational epidemiology: from scientific discovery to population health impact. Am J Epidemiol. 2010;172:517–24.PubMedCrossRefGoogle Scholar
  16. 16.
    Moons KG, Altman DG, Vergouwe Y, Royston P. Prognosis and prognostic research: application and impact of prognostic models in clinical practice. BMJ. 2009;338:b606.PubMedCrossRefGoogle Scholar
  17. 17.
    Khoury MJ, Dorman JS. The human genome epidemiology network. Am J Epidemiol. 1998;148:1–3.PubMedGoogle Scholar
  18. 18.
    Freedman AN, Seminara D, Gail MH, Hartge P, Colditz GA, et al. Cancer risk prediction models: a workshop on development, evaluation, and application. J Natl Cancer Inst. 2005;97:715–23.PubMedCrossRefGoogle Scholar
  19. 19.
    Altman DG, Schulz KF, Moher D, Egger M, Davidoff F, et al. The revised CONSORT statement for reporting randomized trials: explanation and elaboration. Ann Intern Med. 2001;134:663–94.PubMedGoogle Scholar
  20. 20.
    Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, et al. The STARD statement for reporting studies of diagnostic accuracy: explanation and elaboration. Ann Intern Med. 2003;138:W1–12.PubMedGoogle Scholar
  21. 21.
    Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6:e1000100. doi: 10.1371/journal.pmed.1000100.PubMedCrossRefGoogle Scholar
  22. 22.
    Vandenbroucke JP, von Elm E, Altman DG, Gotzsche PC, Mulrow CD, et al. Strengthening the reporting of observational studies in epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4:e297. doi: 10.1371/journal.pmed.0040297.PubMedCrossRefGoogle Scholar
  23. 23.
    Vandenbroucke JP. STREGA, STROBE, STARD, SQUIRE, MOOSE, PRISMA, GNOSIS, TREND, ORION, COREQ, QUOROM, REMARK and CONSORT: for whom does the guideline toll? J Clin Epidemiol. 2009;62:594–6.PubMedCrossRefGoogle Scholar
  24. 24.
    Cook NR. Use and misuse of the receiver operating characteristic curve in risk prediction. Circulation. 2007;115:928–35.PubMedCrossRefGoogle Scholar
  25. 25.
    Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med. 2008;27:157–72.PubMedCrossRefGoogle Scholar
  26. 26.
    Janssens ACJW, Khoury MJ. Assessment of improved prediction beyond traditional risk factors: when does a difference make a difference? Circ Cardiovasc Genet. 2010;3:3–5.PubMedCrossRefGoogle Scholar

Copyright information

© The Author(s) 2011

Authors and Affiliations

  • A. Cecile J. W. Janssens
    • 1
  • John P. A. Ioannidis
    • 2
    • 3
    • 4
    • 5
    • 6
  • Cornelia M. van Duijn
    • 1
  • Julian Little
    • 7
  • Muin J. Khoury
    • 8
  • for the GRIPS Group
  1. 1.Department of EpidemiologyErasmus University Medical CenterRotterdamThe Netherlands
  2. 2.Department of Hygiene and EpidemiologyUniversity of Ioannina School of MedicineIoanninaGreece
  3. 3.Biomedical Research InstituteFoundation for Research and TechnologyIoanninaGreece
  4. 4.Department of MedicineTufts University School of MedicineBostonUSA
  5. 5.Center for Genetic Epidemiology and Modeling and Tufts CTSI, Institute for Clinical Research and Health Policy StudiesTufts Medical CenterBostonUSA
  6. 6.Stanford Prevention Research CenterStanford University School of MedicineStanfordUSA
  7. 7.Department of Epidemiology and Community MedicineUniversity of OttawaOttawaCanada
  8. 8.Office of Public Health GenomicsCenters for Disease Control and PreventionAtlantaUSA

Personalised recommendations