THE EFFECTS OF MAGIC IN MEDICAL EPIDEMIOLOGY

Magic does, indeed, occur in medical epidemiology for the simple reason that it is an imprecise science. All too often critical errors are made, some systematic across studies. Philip Cole [cited in 1] referred to systematic errors of bias and confounding in epidemiological studies as ‘the plague upon the house of epidemiology’. These problems arise in part from inattention to a priori hypotheses and lack of strict and consistent operational definitions. Dr Poikolainen [2] is correct in his assessment of these studies. It takes but a flick of a wrist to alter the findings of medical epidemiological investigations. Any scientist performing cohort studies is well aware of this. Alter one operational definition, ever so slightly, and the results change. Go fishing for results and you are likely to land at least a minnow. It is unknown the degree to which these processes are due to outside influences pressing on the objectivity of science, a desire to confirm one's favorite hypothesis or sheer ignorance and lack of thought about the meaning of the subject matter at hand. Our own work [3] was based on a firm well-articulated a priori hypothesis, generated from the classic work of Shaper, Wannamethee and Walker [4]. It explicated the operational definitions of abstinence as strict ones because we were well aware that (a) respondents' drinking can change over the life course and (b) respondents can misinterpret the meaning of questions asked if those questions are vague. It is in this sense that our work sought to improve on past efforts in this domain of science. We eliminated studies from the error-free category with wordings such as ‘do you rarely/never drink?’ or ‘never or almost never drink?’ because the questions were vague and could even include binge drinkers (infrequent but heavy drinkers) in that abstainer group. Dr Poikolainen perhaps misunderstood our definition of a long-term abstainer as it applied to an error-free study: (a) The study had to separate former drinkers from abstainers; (b) it had to separate occasional drinkers from abstainers and (c) it had to specifically identify the time period of abstinence as more than one year. Dr Poikolainen questions a study that we classified as error-free. That particular study eliminated from the lifetime abstainer category those who consumed at least 12 drinks in the past year (current drinkers) and included as lifetime abstainers those who had consumed less than 12 drinks in their entire life. Former drinkers were defined as those drinking 12 or more drinks in one year but not in the past year. We are in full agreement that one year is probably inadequate to identify long–term abstinence but we are also aware that recall bias is a major problem in such research. It is quite clear (and we feel that Dr Poikolainen would agree) that the solution to carefully delineating drinking status over long periods in the life course resides in performing multiple measurement mortality/morbidity studies. Unfortunately few of these are in existence. Our conclusion is that alcohol most probably benefits the heart as evidenced from laboratory findings. It is by no means magic to report what appears to be a weak or non-existent protective effect for at least three reasons. First, if the protective effect is, in fact, real, it should also be remembered that other substances, lifestyles and behaviors also ‘protect’ and the majority of the latter are much less likely to cause harm. Such knowledge should most probably be conveyed as the strongest public health message in this domain. Second, it suggests that the size of the protective effect for populations has been overestimated in the past. Third, it should be viewed within a broader context of the tendency in recent years in medical epidemiology to report significant results that do not take into account the systematic errors and biases from confounders that can obliterate statistical significance. We are pleased that greater recent attention is devoted to carefully assessing the problems in studies contributing to the protection thesis on the population level (e.g. Flávio et al.[5]; Jackson et al.[6]) and that efforts are being made to explicitly test the Shaper et al.[4] hypothesis in individual studies (e.g. Harriss et al.[7]). These thoughtful efforts should lay out a better map for evaluating the associations between alcohol use and disease incidence. The research discussed in this letter was funded by the Alcohol Education and Rehabilitation Foundation (Australia). Seed money for hypothesis development and initial coding of the studies to the first author came from Robert Newcomer (Chair, Department of Social and Behavior Science, University of California, San Francisco, USA), Office of the Dean (School of Nursing, University of California, San Francisco, USA) and NordAN (Stockholm, Sweden). The majority of funding for research performed by Kaye Fillmore has been derived from the US National Institutes of Health (NIAAA). She has received a minor amount of seed money from NordAN, a collection of Scandinavian groups interested in the control of the accessibility of alcohol, and a minor amount of money from the International Center for Alcohol Policies to support an in-house paper on the nature of contemporary alcohol-related research and has received travel expenses from the same group at an earlier time. She has consulted for NIAAA and for WHO (Geneva and Europe). Tim Stockwell periodically conducts consulting work for WHO and Health Canada on alcohol and other drug research issues. He is in receipt of funding from the Centre for Addictions Research of BC, the BC Ministry of Health, WHO and the Canadian Institutes for Health Research. He has previously received travel expenses from the International Center for Alcohol Policies but has not received personal fees or research funds from alcohol or tobacco manufacturers or from pharmaceutical companies. Tanya Chikritzhs has received all of her research funding from the National Drug Strategy Commonwealth, Department of Health and Ageing, Australia, through competitive grants with no links to the alcohol beverage industry. She has performed minor consulting to the health department in Western Australia.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.