Trends in 21st Century Epidemiology: From Scientific Discoveries to Population Health Impact

Session 4 Panel and Audience Discussion

Slide 1 of 9

  • Questions:
    • How do we fill evidence gaps in care and prevention using epidemiology?
    • What is the role of observational epidemiology in a "data rich" environment where randomized clinical trials may not always be feasible?
  • Moderator: Sheri D. Schully, Ph.D., EGRP, DCCPS, NCI
  • Panelists:
    • Barnett Kramer, M.D., M.P.H.
      Division of Cancer Prevention, NCI
    • Michael Lauer, M.D.
      Division of Cardiovascular Sciences, National Heart Lung and Blood Institute (NHLBI)
    • Jeffrey A. Meyerhardt, M.D., M.P.H.
      Harvard Medical School
    • Olufunmilayo I. Olopade, M.D. F.A.C.P.
      University of Chicago

Slide 2 of 9

Barnett Kramer, M.D., M.P.H.
Division of Cancer Prevention, NCI


Slide 3 of 9: Description of Evidence in PDQ(Physician Desk Query): Five Domains

  1. Study design: ranked by design strength
  2. Internal validity: "quality" of execution within study design (good, fair, poor)
  3. Consistency (coherence)/volume of evidence
    • One vs. multiple studies
    • Small vs. large studies
    • Consistent direction of outcomes
  4. Magnitude of effects: prefer absolute vs. relative effects
    • Change from 1% to 0.5%, or from 4/1000 to 2/1000
    • Not: 50% decrease
  5. External validity (good, fair, poor)
    • Applicability in usual practice with same effect?

Slide 4 of 9

Michael Lauer, M.D.
Division of Cardiovascular Sciences, National Heart Lung and Blood Institute (NHLBI)


Slide 5 of 9: Trends in Epidemiology (M. Lauer, NHLBI)

[Image] A graph showing a low marginal cost ($50 per patient) being emphasized as one of the benefits of using observation studies as platforms for practical trials. Apart from cost, another emphasis of this image is that observational studies and randomized trials should not be treated as different fields. Observational studies can serve as cohorts through which practical trials can take place.

Source: Chris Granger, IOM, November 27, 2012


Slide 6 of 9

Jeffrey A. Meyerhardt, M.D., M.P.H.
Harvard Medical School


Slide 7 of 9: Integrating Epidemiology into Randomized Clinical Trials

[Image] of an observational study tacked onto a randomized clinical trial showing how host factors (complementary meds, adiposity, physical activity, vitamin D, smoking status/history, vitamin intake, diet, and comorbidities/age) can affect the overall outcome of patients (such as overall survival) and other endpoints (such as new cancers, new precursors and other diseases) employed in a randomized trial. These host factors also impact toxicity of drugs used in the trial, as well as molecular biomarkers, quality of life (QOL) and cost effectiveness measures associated with evaluation of drugs used in the trial.


Slide 8 of 9: MyMoffitt Patient Portal

Olufunmilayo I. Olopade, M.D. F.A.C.P.
University of Chicago


Slide 9 of 9

[Image] showing population differences in breast cancer in a survey of women from different geographical regions (i.e., Nigeria (Ilorin), Nigeria & Senegal, United States, Barbados, Poland and Japan). Image also points out that indigenous African women reveal overrepresentation of triple negative breast cancer.

Source: Huo D, Ikpatt OFR et al. JCO 27:4515-21, 2009
Data abstracted from Adeniji et al. ASCO Proceedings 2010, Yang et al 2007, Kurebayashi et al. 2007

Return to Top