Detecting Causal Nonlinear Exposure-Response Relations in Epidemiological Data

Cox Jr., Louis Anthony
April 2006
Dose-Response;2006, Vol. 4 Issue 2, p119
Academic Journal
The possibility of hormesis in individual dose-response relations undermines traditional epidemiological criteria and tests for causal relations between exposure and response variables. Non-monotonic exposure-response relations in a large population may lack aggregate consistency, strength, biological gradient, and other hallmarks of traditional causal relations. For example, a u-shaped or n-shaped curve may exhibit zero correlation between dose and response. Thus, possible hormesis requires new ways to detect potentially causal exposure-response relations. This paper introduces information-theoretic criteria for identifying potential causality in epidemiological data that may contain nonmonotonic or threshold dose-response nonlinearities. Roughly, exposure variable X is a potential cause of response variable Y if and only if: (a) X is INFORMATIVE about Y (i.e., the mutual information between X and Y, I(X; Y), measured in bits, is positive. This provides the required generalization of statistical association measures for monotonic relations); (b) UNCONFOUNDED: X provides information about Y that cannot be removed by conditioning on other variables. (c)PREDICTIVE: Past values of X are informative about future values of Y, even after conditioning on past values of Y; (d) CAUSAL ORDERING: Y is conditionally independent of the parents of X, given X. These criteria yield practical algorithms for detecting potential causation in cohort, case-control, and time series data sets. We illustrate them by identifying potential causes of campylobacteriosis, a food-borne bacterial infectious diarrheal illness, in a recent case-control data set. In contrast to previous analyses, our information-theoretic approach identifies a hitherto unnoticed, highly statistically significant, hormetic (U-shaped) relation between recent fast food consumption and women's risk of campylobacteriosis. We also discuss the application of the new information-theoretic criteria in resolving ambiguities and apparent contradictions due to confounding and information redundancy or overlap among variables in epidemiological data sets.


Related Articles

  • The marginalization of hormesis. Calabrese, E.J.; Baldwin, L.A. // Human & Experimental Toxicology;Jan2000, Vol. 19 Issue 1, p32 

    Despite the substantial development and publication of highly reproducible toxicological data, the concept of hormetic dose-response relationships was never integrated into the mainstream of toxicological thought. Review of the historical foundations of the interpretation of the bioassay and...

  • HORMESIS: IMPLICATIONS FOR CANCER RISK ASSESSMENT. Borak, Jonathan; Sirianni, Greg // Dose-Response;2005, Vol. 3 Issue 3, p443 

    Current guidelines for cancer risk assessment emphasize a toxicant's "mode of action", rather than its empirically derived dose-response relationship, for determining whether linear low-dose extrapolation is appropriate. Thus, for reasons of policy, demonstration of hormesis is generally...

  • Some implications for quantitative risk assessment if hormesis exists. Sielken Jr, R.L.; Stevenson, D.E. // Human & Experimental Toxicology;1998, Vol. 17 Issue 5, p259 

    The existence of hormesis should impact quantitative risk assessment in at least seven fundamental ways. (1) The dose-reponse models for bioassay and epidemiological data should have greater flexibility to fit the observed shape of the dose-response data and no longer be forced to always be...

  • Implications of hormesis on the bioassay and hazard assessment of chemical carcinogens. Teeguarden, J.G.; Dragan, Y.P.; Pitot, H.C. // Human & Experimental Toxicology;1998, Vol. 17 Issue 5, p254 

    Hormesis has been defined as a dose-response relationship which depicts improvement in some endpoint (increased metabolic rates, reduction in tumor incidence, etc.) at low doses of a toxic compound followed by a decline in the endpoint at higher doses. The existence of hormetic responses to...

  • A general classification of U-shaped dose-response relationships in toxicology and their mechanistic foundations. Calabrese, E.J.; Baldwin, L.A. // Human & Experimental Toxicology;1998, Vol. 17 Issue 7, p353 

    The development of a comprehensive database of chemical hormetic responses (i.e., U- or inverted U-shaped dose-response relationships) using objective a priori study design, statistical and study replication criteria has recently been reported.[sup ] An assessment of this database reveals the...

  • Detecting hormesis using a non-parametric rank test. Deng, C.; Zhao, Q.; Shukla, R. // Human & Experimental Toxicology;Dec2000, Vol. 19 Issue 12, p703 

    When a dose–response experiment is conducted, the enhanced responses can be observed at low doses. This phenomenon is often called hormesis. The enhanced responses at low doses does not necessarily mean the existence of hormesis. It is important to conduct statistical analyses to...

  • Chemical hormesis: its historical foundations as a biological hypothesis. Calabrese, E.J.; Baldwin, L.A. // Human & Experimental Toxicology;Jan2000, Vol. 19 Issue 1, p2 

    Despite the long history of hormesis-related experimental research no systematic effort to describe its early history has been undertaken. The present paper attempts to reconstruct and assess the early history of such research and to evaluate how advances in related scientific fields affected...

  • Risk communication and the challenge of hormesis. Calabrese, Edward // Human & Experimental Toxicology;Jan2003, Vol. 22 Issue 1, p1 

    Analyzes the approaches for estimating risk communication implications of Hormesis. Nature of the dose-response relationship; Reduction in the incident of adverse effects by carcinogenic agents; Impact of concept of hormesis on their communication strategies.

  • Testing the Dose-Response Specification in Epidemiology: Public Health and Policy Consequences for Lead. Rothenberg, Stephen J.; Rothenberg, Jesse C. // Environmental Health Perspectives;Sep2005, Vol. 113 Issue 9, p1190 

    Statistical evaluation of the dose-response function in lead epidemiology is rarely attempted. Economic evaluation of health benefits of lead reduction usually assumes a linear dose-response function, regardless of the outcome measure used. We reanalyzed a previously published study, an...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics