Exploring and Building Models

In recognition of CHE’s 10th anniversary, colleagues who have been particularly instrumental to shaping CHE this past decade will be invited to write an introduction. This month’s introduction is by Ted Schettler, MD, MPH, CHE Science Director, Science Director of the Science and Environmental Health Network, and Coordinator of CHE’s Science Working Group.

Beginning a decade ago, CHE formed working groups with interests in specific aspects of environmental health science. Most were organized around health outcomes, since individuals and organizations frequently focus on a specific disease or disorder, often for very personal reasons, and it seemed logical to build on that structure.

Periodically, however, we come up against the limits of our taxonomies. For example, naming the diabetes-obesity spectrum working group was challenging from the beginning-it was once known as the metabolic syndrome working group—because of the common co-occurrence of insulin resistance, type 2 diabetes, obesity, cardiovascular disease, and lipid abnormalities, not only in individuals but also in populations. Moreover, midlife diabetes and obesity are themselves risk factors for cognitive decline, dementia, and certain kinds of cancer. But, since these conditions are so commonly mixed together, what is the disease? Does our routine use of the International Classification of Diseases coding system hinder our ability to see patterns and identify common environmental threads that create the conditions giving rise to the diseases of our time?

I have a strong sense that environmental health sciences are moving toward adopting a more integrated, multifactorial, multilevel framework that will eventually profoundly influence study design, data analysis, and ultimately public health interventions—even as scientists become more adept at probing subcellular secrets. For example, recent data show that diet and nutritional status can modify the response to environmental chemicals directly in target tissues or indirectly by altering chemical metabolism and absorption through changes in the intestinal microbiome. Nutrition in the US is strongly influenced by a complex industrialized food system, heavily reliant on fossil fuels, pesticides, and fertilizers, and characterized by large amounts of unhealthy processed food, junk food advertising beginning in early childhood, and inequities in access to nutritious food. Here we see relationships crossing multiple levels of a micro-macro spectrum.

Traffic-related air pollution increases the risk of developing asthma as well as the frequency of attacks in children who already have the disorder. The risk from a given level of exposure, however, is greater in children living in lower socioeconomic circumstances. Similarly, higher lifetime lead exposures result in greater cognitive decline in older people who live in poorer socioeconomic neighborhoods than in those who are better off. In other words, the totality of the chemical, physical, nutritional, and social environments—at multiple levels—create the system conditions out of which health or disease patterns emerge.

This suggests to me that we might learn important lessons from the ecological sciences—the study of nested, hierarchical systems, comprised of interacting and semi-independent parts, that aggregate into higher orders of complex integrated wholes. Seen this way, the boundaries separating CHE’s working groups can be respected for practical purposes (e.g., limiting the number of email messages), but they should be held lightly and even ignored when they interfere with more integrated insights. Periodic scans of the larger landscape looking for patterns and trends may shed additional light and identify new opportunities.

I’ve been thinking a lot about this framework as it pertains to the origins of breast cancer—a complex set of diseases, with different patterns in different countries, influenced by numerous, multilevel risk factors. Recently, the California Breast Cancer Research Program funded development of a complexity model for postmenopausal breast cancer that was publicly released in the recent IOM report “Breast Cancer and the Environment: A Life Course Approach.” This model incorporates physical/chemical, behavioral, societal/cultural, and biologic dimensions. At first glance, these kinds of models can look like spaghetti, where everything is connected to everything else. But they accomplish several important things: they acknowledge and communicate complexity, help to make sense of it, and support the development of collaborative strategies to study and intervene. They can also be modified as new information becomes available. Thus, they move inevitably toward a more valid representation of reality.

There is, of course, the risk that an ecological, complexity framework could serve as an excuse for diverting deserved attention from one set of variables to another, simply by claiming one or another to be more important. We should push back against that. Complex systems are made up of relationships and interactions, feedback loops, and tipping points, and acquire emergent properties not predictable from knowing about the parts. Complex diseases are…well…complex, and when we try to understand them by drilling down to finer detail within this nested, multilevel system, we need to be careful not to ignore or discard pieces and interactions along the way—many are likely to be very important.

Work on individual variables will always add value, but situating that work within a larger framework will make it possible to see how efforts in one area may be related to those in another, thereby creating opportunities for new kinds of collaboration with common purpose. I look forward to CHE’s next ten years as partners continue to share their information and insights, collectively exploring and even building models that better express our understanding of environmental health sciences. With gratitude for all that this community has accomplished over the past ten years.


One thought on “Exploring and Building Models

  1. YES! “Complex systems are made up of relationships and interactions, feedback loops, and tipping points, and acquire emergent properties not predictable from knowing about the parts.” I hope the systems of health and medicine as a whole will move in a more integrated, multi-level direction. All of our lives depend on it. – Gayle Sulik

Leave a Reply

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