James S. Koopman MD MPH

Developing Theory that Serves Public Health

Department of Epidemiology

 Me and my wife, Helen Fox, Christmas 1998

I dedicate myself to research, service, and teaching in epidemiology with an emphasis on methodological issues in infectious disease epidemiology. My ambition is to contribute to establishing a science of infectious disease transmission system analysis. I currently do this by studying HIV vaccine trial designs, STD & HIV surveillance systems, urinary tract infection transmission systems, treatment effects on HIV transmission dynamics, and the consequences of intersubject dependence for the interpretation of standard risk factor assessments.

The science of transmission system analysis provides, I believe, a broader approach to disease causation and control than traditional epidemiological approaches. Traditional methods force epidemiologists to focus too much upon risk factors that act within or upon individuals. Rather than refuting the traditional approach, the science of transmission system analysis provides more powerful and valid assessment of traditional risk factor effects. It does so, however, in a way that expands the realm of issues where epidemiology seeks to make scientific inferences. Issues of interactions between individuals and social organization come to the fore in this science. It's system analysis orientation naturally focuses upon vitally important structural and functional issues in human relationships which the standard methods developed for non-contagious diseases have just assumed away. This promotes better design of infection control programs. It also provides a productive framework for understanding how new infectious disease agents and transmission systems emerge from the complex adaptive systems in which humans and micro-organisms relate.

I believe that an important step in the development of this science is the creation of discrete event models capturing the essence of the complex systems in which humans interact. These might initially follow the traditions of Monte Carlo models where random events are selected by central criteria that tend to create a desired population distribution of events. But eventually they should become models of discrete automatous agents. The locus of action decisions or random events should move into the individual agents. Broad advances are needed in the science of complexity before this becomes possible. In the mean time, I seek to use discrete event models to optimize the design of epidemiological studies.

My students and my colleagues make this pursuit meaningful and enjoyable. A few special colleagues in this regard are Steve Chick, Carl Simon, Betsy Foxman, Rod Little, and Ken Lange. My professional interests and capacities have been greatly influenced by John Fox and John Jacquez. I have greatly benefitted from the intellectual atmosphere generated by John Holland.

Address: Dept. of Epidemiology SPH-1, 109 Observatory, Ann Arbor, MI 48109-2029

e-mail: jkoopman@sph.umich.edu

Phone: (734)763-5629

Fax: (734)764-3192

Publications

Only projects which I manage on the Web are listed here. Someday I will turn this into something which describes all of my work. But for now this is just a gateway to

Research

  1. ASSESSING EFFECTS OF RISK FACTORS FOR TRANSMISSION PROJECT, This project explores the validity of standard risk factor assessment methods used in epidemiology when the risk factors involved affect the transmission of an infectious agent. In that case an assumption intrinsic to the most common uses of risk ratios, risk differences, and other parameters relating exposure to disease such as logistic regression coefficients is violated. That assumption is that the outcome of exposure in one individual is independent of the outcome of exposure in other individuals. Most epidemiologists do not realize that violation of this assumption can cause severe distortion of risk factor assessments. This project will define when, where, and how such distortions arise and what we can do about them.
  2. U of M Transmission Analysis Group, This is a highly interdisciplinary group that works on very diverse aspects of developing a science of transmission system analysis ranging from analysis of compartmental models, development and analysis of discrete event simulation models, design of influenza and HIV vaccine trials, and analysis of the role of sexually transmitted infection systems in maintaining the circulation of specific strains of E. coli that cause urinary tract infections.
  3. Emerging Objectives and Methods in Epidemiology , This is an early draft of a paper that was published in the American Journal of Public Health and deals with some of the basic paradigms that drive our work.
  4. The Role of Early Infection in the Spread of HIV: This is a compartmental model analysis which demonstrates that the patterns of infection which our earlier models could fit only with big differences between early and middle stage infection can also be well fit with models having smaller differences when other realistic phenomenon which help early infection dominate HIV transmission dynamics are taken into account. It serves to demonstrate that one of the biggest determinants of the importance of a risk factor is who that risk factor links through transmission of infection. Early infection is more often transmitted to those who will carry on the chain of transmnission than is later infection. That is a major reason why early infection dominates HIV transmission dynamics.
  5. I have a prepared a more concise and at the same time more complete argument of why early infection determines the rate at which epidemics rise. I will deliver this at the Spanish AIDS meetings in March. This also discusses the relationship of early infection to the extent of transmission control that one can expect to get by treating infected individuals.
  6. HIV and STD Surveillance That Identifies Key Infected Individuals Which Control Programs are Missing: This is a project in its initial stages with funding from the Michigan Department of Community Health. It was stimulated by observations made in the course of our Social Network Analysis and Transmission System Analysis theory development. It is being pursued at a practical field level with the City of Detroit Health Department and the Washtenaw County Health Department. (This is badly in need of updating!)

    7. A summary of our modeling work to date on this project that outlines our new approach to discrete individual models is presented in a document used for presentation on April 28, 1999 at the CDC conference on sexually transmitted infection models in Santa Fe. It is titled "Models for Evaluating Network Statistics and Surveillance Procedures: Preliminary Results for Gonorrhea"

  7. EPIDEMIOLOGY AND HUMAN INTERACTIONS: EMERGING METAPHORS, MODELS, AND METHODS: A talk delivered to the "Asociación Colombiana de Epidemiología" It outlines paths that epidemiology might pursue to firm up its theoretical foundations.
  8. Individual Causal Models and Population System Models in Epidemiology: This outlines two markedly different causal modeling traditions in epidemiology. It argues that to fully analyze epidemiological data, the two traditions need to be merged. This is an APHA commentary to be published in the near future.

Teaching

  1. Infectious Disease Course, Epidemiology 513 , This was an add on to an old infectious disease course which focused on transmission system dynamics. This material is now being put into the department's new introductory infectious disease epidemiology course in a more integrated manner.
  2. Field Work in Epidemiology, Epidemiology 655 , This site has the materials for a course where second semester students respond to an RFA with a proposal, get actual human subjects approval to carry out their proposal, select feasible objectives for their study, design the study, get funded, carry out the study, analyze the results, and present those results both in oral and written form. They do all of this in one semester and usually get a paper out of their work which is published in a reputable journal. The lecture which best reflects the philosophy of the course is one that deals with tradeoffs in study design.
  3. Advanced Infectious Disease Epidemiology, Epid 606 is a course I jointly teach with Steve Meshnick. I have organized a set of simple exercises for this course that demonstrate the value of theoretically analyzing transmission systems. A broad summary of the value and approaches to analyzing transmission systems is also presented.
  4. The Computer Simulation of Epidemiological Processes, Epid 802: This course teaches compartmental model simulation and analysis methods designed to give a more dynamical and systems oriented approach to epidemiological theory and practice.
  5. Epid519 lecture on HIV dynamics: Assessing the determinants of population patterns of infection and strategies for control of transmission
  6. The philosophical basis of epidemiological analysis, Epid 801, is a doctoral level course that gets students to consider their work within a broader context of advancing the science of epidemiology.
  7. Virology 543 lectures: First lecture: Virus Reproduction Dynamics within the host. Second lecture: Virus Transmission Dynamics Between Hosts in the Population
  8. Comments on Population Smoking Models for the Sept. 13, 1999 Public Health School Symposium