Skip to content

Recommendations from the NHLBI Working Group on Epidemiology

October 22, 2014

000010391045-StarTeamX-small

Ongoing discussions on this Digital Forum since 2012 have focused on ways to creatively transform population science. We previously noted that the NHLBI established a working group consisting of selected members of the  NHLBI Advisory Council and Board of External Experts to consider how to best transform population science. In December of 2013, we posted to this digital forum the working group roster and the charge.

Now, after more than a year of deliberations, the working group has released its draft report and recommendations to the NHLBI Advisory Council for input and consideration.  For your information, the presentation to the NHLBI Council on October 22, 2014 can be found here. 

Dr. Michael Lauer, Director, Division of Cardiovascular Sciences, NHLBI, commented on the Working Group’s recommendations at an NHLBI-hosted webinar on October 24, 2014.  The summary of his remarks can be found here.

We would like to know what you think. Please share your comments about the recommendation with us and the epidemiology community through this Digital Forum.

 

 

 

 

One Comment leave one →
  1. Lewis H. Kuller, MD, DrPH permalink
    October 31, 2014 8:36 am

    1. Epidemiology is a basic science of preventive medicine. The major epidemiology advances in the past, either control of infectious disease, identification of risk factors and control of cardiovascular disease (CVD), identification of the determinants and prevention of major cancers, reduction of infant mortality, identification or removal of environmental and pharmacological hazards to the population, have been the gold standards of “quality epidemiology”.

    2. This report does not mention the role of epidemiology in future preventive medicine or public health or reducing morbidity and mortality in the population nor how we can improve the linkage between epidemiology, preventive medicine or public health. This clearly is the highest priority for epidemiological research.

    The Recommendations suggests epidemiology as a data collection science, which it is not. There is no discussion about how epidemiology can improve the study of the interrelationship of the host, e.g. genetics, epi-genetics, etc., lifestyles, agents, and both the social and physical environment. Large data collections as proposed without specific hypotheses will result in almost worthless fishing expeditions with little likelihood of obtaining any useful new information and identification of many red herrings. These large data sets may have one unique opportunity and, that is, to identify rare or unusual populations, epidemics of disease, e.g. clinical coronary artery disease (CAD) at very young ages, individuals who survive into age 90s-100s with no evidence of CAD, which can then be studied more carefully to identify unique risk factors, genetic host susceptibility or environmental interactions. Studying the average or normal with these large samples will provide little or no useful information.

    3. The first of the listed goals should clearly have the highest priority. Future study designs and data collection will depend almost entirely on decisions about the primary hypotheses and questions and the specific study designs will have to be based on whether the distributions of exposures are related to person-to-person or common source transmissions, vector-born (unlikely), etc. For example, nutrition is primarily a commons source epidemic. Nutrition-related disorders, e.g. much of atherosclerosis, obesity, etc., are common source epidemics and, as has been documented and numerous long term studies that have attempted to identify individual risk factors for obesity and interventions have had little success except for bariatric surgery.

    4. Epidemiology is the study of epidemics and identifying unusual distributions of diseases in time, place or person (genetics). The identification of these unique epidemics is the likely winner. Identification, therefore, of the very unusual, and not the mass collection of poorly collected data, is going to be the winner. This is consistent with previous experiences in which identification of individuals with very low cholesterol, e.g. PCSK9 mutations, etc., low HDL and ApoC3 and other genetic variants as well as attempts to identify the determinants of very low rates of CHD in Japan, France, Italy, etc., and the Mediterranean diet, omega-3 fatty acids, soy proteins, Equol, etc. will likely have the biggest impact.

    5. Epidemiology studies presently suffer from very poor quality data collection. The opportunity of improving data collection technology should have the highest priority, e.g. nutrition epidemiological studies based on prior food frequency questionnaires are probably of little value and find that healthy people, especially upper SES, eat a “high quality” diet. We now have better measures of energy intake and expenditures, use of metabolomic and proteomic approaches that can identify specific nutrient intakes within individuals and have better measures now of exercise. These techniques need to be employed in better epidemiological studies and again are not available within these large proposed data sets.

    6. There are few epidemiological studies that evaluate immune function or the association of autoimmune diseases and CVD yet this is an extremely important area. This area of research has suffered from the use of very limited technology, e.g. measuring acute phase proteins and cytokines in the blood is unlikely to be the answer to understanding the interrelationship of immune function, autoimmune disease and even immunological changes with aging, exposure to viruses, etc. and risk of disease.

    7. Another important recommendation should be to develop better methods to link new technology for measurement of environment and lifestyles to epidemiological studies to test specific hypotheses. Again, this cannot be done with these e-epidemiology cohorts. Epidemiological research has evolved over time by applying these new technologies, whether infectious or chronic diseases, to solve important specific hypotheses. The technologies alone without hypotheses add little but the use of these new technologies as they evolve can provide important answers to questions raised in epidemiological studies.

    8. There is an ever-increasing age of onset of first events for both CHD and stroke, i.e. the incidence of events are now occurring in much older individuals, ages 70-75+. These cases are now bumping into the very high rates of dementia, disability and other aging-related diseases. Epidemiological studies of CVD can no longer be done in isolation from studies of cognition in the brain, muscles (sarcopenia), osteoporosis, kidney diseases, frailty and aging, etc. Cardiovascular diseases, such as congestive heart failure (CHF) and atrial fibrillation (AF) in the elderly are also risk factors for brain disease, dementia, etc. Similarly, changes in cognition, such as mild cognitive impairment, may be important determinants of successful treatment of CHF, AF, CHD, etc. The availability of new technologies has dramatically changed the ability to study brain, muscle and bone diseases as well as CVD in older individuals and “aging phenotypes”. Treatment of CHD, stroke, etc. is the largest component of health care costs in the United States and epidemiological which can focus on prevention and reduction of disability, not only of CHD but also the impact on brain (dementia), kidney disease, etc. could have major impact on the health of the older population.

    9. Should epidemiology be separated from clinical research or clinical epidemiology that collect these large data sets and mines them for relationships that have implications for specific treatments of disease? Should epidemiology essentially focus on etiological research and preventive medicine and public health and, therefore, clinical research focus primarily on people with extensive disease, e.g. treatment and clinical care be a separate entity rather than clinical epidemiology, which seems to have taken the primary role at NHLBI but in many ways is not really epidemiology? The term ‘clinical epidemiology’ is probably useful to justify the field but has matured enough now to be spun off as clinical research. Thus, classical epidemiological approaches, population genetics and preventive medicine perhaps should be more closely linked as well as primary prevention, clinical trials, to test primarily etiological and preventive hypotheses be considered a separate entity within NHLBI and even within NIH. Much of what is called epidemiology now is not epidemiology.

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: