Skip Navigation LinksHome > May 2013 - Volume 24 - Issue 3 > From “Big Epidemiology” to “Colossal Epidemiology”: When All...
Epidemiology:
doi: 10.1097/EDE.0b013e31828c7694
The Changing Face of Epidemiology

From “Big Epidemiology” to “Colossal Epidemiology”: When All Eggs Are in One Basket

Hernán, Miguel A.a; Savitz, David A.b

Free Access
Article Outline
Collapse Box

Author Information

From the aHarvard School of Public Health, Boston, MA; and bDepartment of Epidemiology, Brown University, Providence, RI.

Editors’ note: This series addresses topics that affect epidemiologists across a range of specialties. Commentaries are first invited as talks at symposia organized by the Editors. This paper was originally presented at the 45th Annual Meeting of the Society for the Epidemiologic Research in Minneapolis, MN, 2012.

Correspondence: Miguel A. Hernán, Harvard School of Public Health, 677 Huntington Ave., Boston, MA 02115. E-mail: miguel_hernan@post.harvard.edu.

Epidemiologic data resources are being consolidated into increasingly large clusters. The resulting collections of databases, cohorts, and case populations are often so large that they become unique, never-to-be-replicated resources. Although these colossal epidemiologic projects offer unprecedented research opportunities, they also create new challenges. For example, researchers need to coordinate their work within collaborative teams, standardize data collection and analysis procedures, foster the career development of junior investigators, secure funding for each of the participating teams, and provide epidemiologists outside the consortium with access in order to ensure maximum benefit from the resource.

The editors of EPIDEMIOLOGY invited the leaders of three colossal epidemiologic projects (the Sentinel Initiative, the CHARGE consortium and the IeDEA network) to discuss their experience in a symposium at the annual meeting of the Society for Epidemiologic Research in 2012. These projects, though not yet household names for all epidemiologists, are well known in their respective fields. The Sentinel Initiative is an assemblage of electronic databases mandated by the US Congress to monitor medical product safety. The CHARGE consortium coordinates more than a dozen population-based cohorts for genome-wide meta-analyses of various phenotypes. The IeDEA network is a collection of the main HIV cohorts outside Europe.

Two of these projects are discussed in this issue of EPIDEMIOLOGY.1,2 Toh and Platt1 describe the organization of the Mini-Sentinel project. Do not be fooled by the name. There is nothing “mini” about Mini-Sentinel. This project, funded by the US Food and Drug Administration, comprises over 100 million US residents registered with major healthcare organizations. The authors use a recently published investigation from this project—on the effect of angiotensin-converting enzyme inhibitors and other drugs on the risk of angioedema—to illustrate the design, management, and conduct of pharmacoepidemiologic studies within Mini-Sentinel. The healthcare organizations initially cited legal and proprietary concerns about sharing their data in this megaproject. The Mini-Sentinel circumvented this obstacle by creating a distributed data network in which no individual data are transferred. Toh and Platt discuss the advantages and disadvantages of this model as a way to conduct integrated analyses without actually integrating the datasets.

Psaty and Sitlani2 describe the organization of the CHARGE consortium for genome-wide association studies, established in 2008 and already with over 150 publications. It is interesting to study the similarities and differences of the collaborative models adopted by CHARGE and Mini-Sentinel. Like Mini-Sentinel, CHARGE uses a distributed data model with meta-analysis of study-specific results. Unlike Mini-Sentinel, CHARGE is funded by multiple investigator-initiated grants at each of the participating institutions. This “distributed funding” may provide a more egalitarian structure for the consortium, but at the cost of less financial certainty.

These projects reflect the latest, if perhaps not the final, stage of the scaling-up of epidemiologic resources. Issues that emerged early in this era of consolidation3–5 have only intensified—centralized versus distributed funding and control, the tradeoffs for individual investigators and research teams of being part of a large group, and the tensions of accommodating the intellectual contributions of those who did not participate in building the resource. The most important feature of these most recent projects (and perhaps the only feature that is qualitatively new) is the unprecedented burden to get it right—because the findings may never be replicated.

Back to Top | Article Outline

REFERENCES

1. Toh S, Platt R. Is size the next big thing in epidemiology? Epidemiology. 2013:349–351

2. Psaty BM, Sitlani C. The Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium as a model of collaborative science. Epidemiology. 2013:346–348

3. Kaplan GA. How big is big enough for epidemiology? Epidemiology. 2007;18:18–20

4. Hoover RN. The evolution of epidemiologic research: from cottage industry to “big” science. Epidemiology. 2007;18:13–17

5. Ness RB. “Big” science and the little guy. Epidemiology. 2007;18:9–12

© 2013 by Lippincott Williams & Wilkins, Inc

Twitter  Facebook

Login

Article Tools

Share