Essay Essays are opinion pieces on a topic of broad interest to a general medical audience.

See all article types »

Observational Research, Randomised Trials, and Two Views of Medical Science

  • Jan P Vandenbroucke
  • Published: March 11, 2008
  • DOI: 10.1371/journal.pmed.0050067

Reader Comments (5)

Post a new comment on this article

Two views of science and the real world

Posted by plosmedicine on 31 Mar 2009 at 00:23 GMT

Author: Philippe Autier
Position: Coordinator, Biostatistics & Epidemiology Cluster Group Head, Epidemiology Methods and Support Group
Institution: International Agency for Research on Cancer
Submitted Date: March 07, 2008
Published Date: March 10, 2008
This comment was originally posted as a “Reader Response” on the publication date indicated above. All Reader Responses are now available as comments.

The inverse hierarchy proposed by JP Vandenbroucke for discovery (and explanation) on the one hand, and for verification (and evaluation) on the other is appealing as it corresponds quite well with scientific activity in medical sciences. Conceptual developments by JP Vandenbroucke are largely rooted in an idealistic perspective of scientific activity, in which pure uncorrupted research is limited only by the capacities of the human brain. In this ideal world, the more verification that can be done, the better, as it secures the certainty of the benefits to be gained thanks to the new medical discovery, or the certainty that controlling a newly discovered risk factor will lead to decreasing the burden of the disease caused by this risk factor. This contrasted view may be helpful for the clarification of the actual objectives pursued by a specific type of research. A scientific discovery (even after apparent verification) is always susceptible to contradiction by some arguments or by a new discovery. If the argument or new discovery is verified as being true, then it might be that the initial discovery is false. This natural property of scientific discoveries (or of scientific theories) has been called “falsifiability” (or refutability, or testability) by the philosopher Karl Popper [1].

Numerous discoveries are made every day, but in the public health and medical world, those that successfully went through a verification process are more likely to be recognized as evidence-based and thus valid for being used in public health actions or medical practice. Positive results of verification studies have a greater chance of being published by high impact factor medical journals than results of discovery studies. Hence, the key issue in medical scientific activity is not so much discovery but rather verification.

The "verification" process proposed by JP Vandenbroucke could be considered as the need to convince (the medical community, the lay public and the regulatory authorities) about the validity of findings made during the discovery process. In the real world, the amount of effort needed to convince about the validity of a discovery may vary considerably according to the sensitivity of the topic at stake. This sensitivity, and thus the amount of effort needed to convince, is further complicated by the sad reality that scientific research may be another way to achieve a non-scientific agenda. Ideological or religious beliefs, corporate agendas and economic interests often underlie scientific research. Lobbies may be active in keeping verifications to a minimum, or in trying to avoid having independent groups performing a verification study. Many discoveries originating in public health and medical research do not have to, or cannot, be tested via compulsory phase III types of trials before becoming available to doctors, patients or the public (e.g., preventive actions). Also, many scientists will consider with good faith that their discovery does not need verification.

The story of post-menopausal hormone treatment (MHT) is probably one of the most spectacular examples of misplaced conviction: that MHT formulated a combination of estrogens and progestogens improved the health status of post menopausal women, including reducing their chance of cardiovascular events. This conviction was built on a succession of discovery studies (mainly case-control and prospective cohort designs), without proper verification of the alleged benefits and harmful effects. For twenty years, this conviction was actively supported by the many companies marketing MHT [2] and shared by a large part of the medical community. Formal verification of the actual health impact of MHT was begun only after a relatively small randomized trial, the HERS study indicated that this medication could increase the occurrence of cardiovascular diseases and breast cancer without reduction in hip fracture incidence [3]. Questions raised by this trial and systematic review of observational studies [4] prompted the launch of one of the largest randomized trials ever conducted was mounted, the Women’s Health Initiative, that demonstrated that long-term use of post-menopausal hormone treatment (MHT) increased, among other things, the risk of cardiovascular disease and breast cancer [5]. All of the apparently beneficial health effects found in observational studies were due to the damaging influence of bias by indication, i.e., women who took MHT had healthier lifestyle than non takers, which in turn explained the apparent health benefits of MHT. In 2006, the International Agency for Research on Cancer classified MHT as carcinogenic to humans [6].

Those involved in science for the pursuit of a non-scientific agenda will tend to concentrate their efforts on the verification side rather than on the discovery side. As a consequence, research institutions, scientific publishers and regulatory authorities impose a number of rules on verification studies to protect them against distortion or bias from ideologies, beliefs, hidden agendas and economic interests. This explains the considerable importance of fighting against publication bias (usually non publication of negative studies), disclosure of conflict of interests and increasingly the registration of study designs before the collection of data.

While this may seem like a pessimistic interpretation, it is probably a realistic way to assess how these sophisticated views outlined by JP Vandenbroucke may help medical scientists to cope with the harsh daily routine of scientific activity.


1. Lipton P (2004) The Medawar Lecture 2004: The truth about science. Phil Trans R Soc B 360: 1259-1269.

2. Fugh-Berman A, Pearson C (2002) The Overselling of Hormone Replacement Therapy. Pharmacotherapy 22:1205–1208)

3. Hulley S, Furberg C, Barrett-Connor E, et al (2002) Noncardiovascular disease outcomes during 6.8 years of hormone replacement therapy: Heart and Estrogen/progestogen Replacement Study follow-up (HERS II). JAMA 288:58–66

4. Clinical Synthesis Panel on HRT (1999). Hormone replacement therapy (HRT). Lancet 354: 152-155.

5. Chlebowski RT, Hendrix SL, Langer RD, et al (2003) Influence of estrogen plus progestogen on breast cancer and mammography in healthy postmenopausal women: The Women’s Health Initiative Randomized Trial. JAMA 289:3243–3253

6. Cogliano V, Grosse Y, Baan R, et al for the WHO International Agency for Research on Cancer (2005) Carcinogenicity of combined oestrogen-progestagen contraceptives and menopausal treatment. Lancet Oncology 2005;6:552–553.

No competing interests declared.