Advertisement
Mayo Clinic Proceedings Home

Clinician Attitudes Toward Biostatistics

      To the Editor: The article by West and Ficalora
      • West CP
      • Ficalora RD
      Clinician attitudes toward biostatistics.
      on clinician attitudes toward biostatistics raises important issues regarding training and continuing education of physicians in statistics. The authors found that more than two-thirds of their respondents at Mayo Clinic Rochester disagreed or strongly disagreed with the statement that “the current level of medical training in biostatistics in medicine is adequate.”
      Appropriate training is particularly important for physicians who do not pursue additional training through a Master of Public Health degree or a research fellowship. For these physicians, the only required training in statistics could be a short unit in medical school reinforced by some additional exposure during their residency (using, it is hoped, the integrated approach of teaching in the context of clinically relevant medical discussions proposed by West and Ficalora). Matthews and McPherson
      • Matthews DR
      • McPherson K
      Doctors' ignorance of statistics [editorial].
      caution that “Innumerate doctors…are doomed to have to accept without reservation the statements made in summaries, discussions, or conclusions, and their clinical practice may thus be altered on the basis of flimsy or inconclusive evidence.”
      An additional complication to consider is the increasing sophistication of statistical methods used in the medical literature. Our review of original articles published in the New England Journal of Medicine found a marked increase in the complexity of statistical methods.
      • Horton NJ
      • Switzer SS
      Statistical methods in the journal [letter].
      For example, the use of multiple regression increased from 14% in 1989 to 51% in 2004 to 2005.
      • Emerson JD
      • Colditz GA
      Use of statistical analysis in the New England Journal of Medicine.
      Multiple regression is an important concept to communicate to all physicians because it is key to understanding confounding variables, effect modifications, and interactions that arise in many articles that communicate clinically important research findings. The current level of statistical education in medical schools and residency programs might not provide students with a working knowledge of these and other intermediate-level statistical topics.
      Medical educators, journal editors, and statisticians should be encouraged to implement and disseminate more of the innovative educational approaches that West and Ficalora describe. It might also be time to revisit proposals to require statistical training as a prerequisite to medical school.

      REFERENCES

        • West CP
        • Ficalora RD
        Clinician attitudes toward biostatistics.
        Mayo Clin Proc. 2007; 82: 939-943
        • Matthews DR
        • McPherson K
        Doctors' ignorance of statistics [editorial].
        Br Med J. 1987; 294: 856-857
        • Horton NJ
        • Switzer SS
        Statistical methods in the journal [letter].
        N Engl J Med. 2005; 353: 1977-1979
        • Emerson JD
        • Colditz GA
        Use of statistical analysis in the New England Journal of Medicine.
        in: Bailar III, JC Mosteller F Medical Uses of Statistics. 2nd ed. NEJM Books, Boston, MA1992: 45-57

      Linked Article

      • Clinician Attitudes Toward Biostatistics–Reply–I
        Mayo Clinic ProceedingsVol. 82Issue 12
        • Preview
          I fully agree with Dr Horton's comments concerning biostatistics in the current medical literature. As he notes, the statistical methods used in modern research reports are becoming increasingly complex, compounding the problem of poor clinician understanding of basic statistical concepts. Furthermore, there is compelling evidence that researchers often apply statistical techniques inappropriately, perhaps as a result of their own incomplete comprehension of the methodology.1 As neatly stated by Altman,2 “the main reason for the plethora of statistical errors is that the majority of statistical analyses are performed by people with an inadequate understanding of statistical methods.
        • Full-Text
        • PDF