Letter to the editor.

To The Editor

I would like to commend ACPE and Dr. Steve Silverstein for doing a great job in raising awareness for the necessity of evidence-based medicine (EBM) in clinical practice in the article that appeared in the Nov./Dec. 2007 issue of The Physician Executive, 33(6), titled Fulfilling the Promise of Evidence-based Medicine.

I have just two comments for further clarification.

First, Dr. Silverstein states that, "Cause and effect can only be determined in randomized, double-blind, placebo-controlled studies--the 'gold standard' of evidence." While it is certainly true that randomized controlled trials (RCTs) are considered the "gold standard" of evidence for studies relating to therapy, it does not follow that cause and effect can only be determined through RCTs.

RCTs are subject to the same sources of error as observational, analytic epidemiological studies when it comes to internal validity--chance (random error), bias, and confounding. While certain aspects of RCTs minimize these effects (e.g. proper randomization eliminates selection bias and helps reduce the likelihood of confounding), RCTs are by no means a guarantee of finding a valid result.

Issues of compliance, losses to follow-up, and a wide range of other issues can potentially invalidate the trial results. Furthermore, the issue of causation is different from that of association. After the role of chance, bias, and confounding have been evaluated and eliminated or found to be unimportant, the observed association can be either causal or not causal.

To ascertain causal inference, researchers most often employ Sir Austin Bradford Hill's Guidelines for Assessing Causation. These include:

* Strength of the association

* Consistency

* Specificity

* Temporality

* Biological gradient

* Plausibility

* Coherence

* Experiment

* Analogy (1)

Note that only one of the above criteria (experiment) deals directly with the method of assessment. For studies such as those involving prognosis, rare diseases, or establishing the accuracy of diagnostic tests, other research designs may need to be utilized.

Hierarchical listing of the levels for different types of evidence are available from sites such as Oxford's Center for Evidence-based Medicine (http://www.cebm.net/index.aspx?o=1025).

Finally, observational, analytic epidemiologic studies have been used to establish a causal association. For example, our knowledge of the causal link between cigarette smoking and lung cancer was not based on even a single RCT. It would simply be unethical to conduct such a randomized trial.

Instead, numerous case-control, cohort, and basic science studies were used to help establish a causal connection. Without epidemiological studies, we would simply not be able to investigate many exposure-based risks because conducting an RCT is out of the question for ethical reasons.

Second, Dr. Silverstein does a good job in mentioning some of the EBM content suppliers, but fails to identify the United States Preventive Services Task Force (USPSTF). The Task Force is an independent panel of experts in primary care and prevention that uses systematic evidence reviews of effectiveness to develop recommendations for clinical preventive services. (2)

The USPSTF is sponsored by the Agency for Healthcare Research and Quality (AHRQ) which publishes USPSTF's recommendations in The Guide to Clinical Preventive Services. Each of the USPSTF recommendations receives a grade using evidence-based guidelines.

Additionally, the author omits the National Guideline Clearinghouse (NGC) (www.guideline.gov), also sponsored by AHRQ. NGC's mission is to "provide physicians, nurses, and other health professionals, health care providers, health plans, integrated delivery systems, purchasers and others an accessible mechanism for obtaining objective, detailed information on clinical practice guidelines and to further their dissemination, implementation and use." (3)

[ILLUSTRATION OMITTED]

Users are able to conveniently compare guidelines side-by-side, including the degree to which they are supported by evidence-based methodologies.

In summary, EBM requires the use of both RCTs and observational, analytic epidemiologic studies. The USPSTF guidelines and the NGC should also be included in any mention of EBM content suppliers.

Sincerely,

Val W. Finnell, MD, CPE 

Uniformed Services University of the Health Sciences 

Bethesda, MD

References

1. Aschengrau A, Seage III, G. Essentials of Epidemiology in Public Health. Boston: Jones and Bartlett Publishers, 2003.

2. 2. Agency for Healthcare Research and Quality. The Guide to Clinical Preventive Services: AHRQ; 2006.

3. National Guideline Clearinghouse. http://www.guideline.gov, Accessed December 2, 2007.