Home > Information Needs Assessment: Epidemiologists

© Lisa C. Wallis, Medical Literature and Reference Work, Fall 1999

Introduction
Epidemiology is "the study of the distribution and determinants of diseases and injuries in human populations" (Mausner & Kramer, 1985, p. 1). Currently, no disease, condition, or health-related event is outside the scope of the field, though its origins were strictly rooted in infectious disease (Mausner & Kramer, 1985). The practice of epidemiology is based primarily on two assumptions: 1) disease is not random in any given population, and 2) disease or health conditions are caused by specific factors that vary in populations by place and time (Association of Schools of Public Health, 1999a). Such assumptions allow for systematic scientific study of disease by specialists known as epidemiologists.

Epidemiologists are highly trained professionals, and many of them are in fact medical doctors as well. However, while it may be considered a component of medical science (Hayes, 1998), epidemiology differs from traditional medical practice in that the focus is on the population rather than the individual (Mausner & Kramer, 1985). With both the globalization of health threats and the growing emphasis on preventive measures, epidemiologists are bound to play a more prominent role in society in the future. In fact, between 1995 and 2000 it was expected that the need for epidemiologists was going to triple (Association of Schools of Public Health, 1995), an indicator of the growing regard for the profession.

Educational Requirements
At minimum, a practicing epidemiologist must possess a master's degree, which is typically a master's in public health, or MPH. Most epidemiologists have a Ph.D. in epidemiology along with post-graduate experience in the field (Hayes, 1998). A recent review of the Public Health Employment Connection, a job site created by the Emory University School of Public Health, revealed that of the most recently posted epidemiology jobs, half required a Ph.D. In some instances where a master's degree was acceptable, applicants were required to have a significantly greater amount of experience in the field (Rollins School of Public Health, 1999).

There are 29 accredited schools of public health in the United States (Association of Schools of Public Health, 1999b), all of which offer master's or doctoral degrees in epidemiology. Degrees may also be earned through other non-School of Public Health programs that may or may not be accredited by the Council on Education for Public Health (Association of Schools of Public Health, 1999b). For an example of the types of classes required for a degree in epidemiology, one can look at the degree descriptions from one of the top programs in the country (Hayes, 1998), that of Harvard University's School of Public Health. An epidemiology Ph.D. candidate is required not only to take classes in epidemiological methods and theories, but also must study statistics, biology, physiology, and toxicology. In addition, it is assumed that anyone entering the program had strong undergraduate preparation in biology and mathematics (Harvard School of Public Health, 1999).

Education does not stop with the earning of the doctoral degree. Here in the United States, epidemiologists have access to the world's premier public health research and policy institute, the Centers for Disease Control and Prevention, or CDC. CDC is a leader in lifelong learning for epidemiologists. Through its Epidemic Intelligence Service, CDC offers a two-year, on-the-job training program in applied epidemiology. Eligible candidates for the program include medical doctors with 1 year of clinical training; doctorate holders in epidemiology or related fields; and dentists, physician assistants, veterinarians, or nurses with a master's degree in public health (Centers for Disease Control and Prevention, 1999). The EIS is just one option for epidemiology degree holders who wish to gain practical hands-on experience.

The Practice Environment
The work of the epidemiologist is intended to "determine the cause of disease, its distribution…, how it spreads, and what can be done to control and prevent it" (Hayes, 1998, p. 252). One central component of the profession is the calculation of rates--cases or events in a population in relation to susceptible people (Mausner & Kramer, 1985). The means for calculating these rates include two types of studies, cohort and case-control. In cohort studies, researchers follow a large group of people who are free of disease at the time of selection for the study. However, members of the group differ from one another in terms of their exposure to certain potentially disease-causing events. Over time epidemiologists measure who develops disease. Case-control studies vary from cohort studies in that some participants have already been diagnosed with a specific health condition. These people are then compared retrospectively to a control group of disease-free individuals to determine how their exposures to disease-causing agents have differed (Mausner & Kramer, 1985).

The disease-causing agents are known in epidemiology as risk factors. Their "presence is associated with an increased probability that disease will develop later" (Mausner & Kramer, 1985, p. 6). The measurement of exposure to risk factors allows epidemiologists to determine a basic statistic in the field, the relative risk. This figure is calculated by dividing the incidence of disease in the exposed group by the incidence of disease in the nonexposed group. A relative risk of any number greater than 1.0 indicates that an exposure to a certain risk factor increases one's chances of developing a disease (Hennekens & Buring, 1987). As will be discussed later, this alone does not identify a risk factor, but it is one essential component of identification.

One of the better known roles of an epidemiologist is that of outbreak investigator. Due to Hollywood films like Outbreak or And The Band Played On, people outside the public health field have become aware of epidemiology, though in a limited capacity. Wearing brightly colored protective space suits, epidemiologists in the movies track down killer viruses and other microorganisms and save the world from destruction. The work going on behind the scenes is as systematic as the cohort or case-control studies conducted by researchers in offices. Out in the field, an epidemiologist follows a specific sequence of steps when a health-condition appears that exceeds expected rates of disease. This is known as an outbreak (Mausner & Kramer, 1985). Reingold (1998) defines the process of outbreak investigation as follows:
  1. Establish case definition
  2. Confirm that cases are "real"
  3. Establish the background rate of the disease
  4. Find cases, decide if there is an outbreak, define scope of the outbreak
  5. Examine the descriptive epidemiologic features of the cases
  6. Generate hypotheses
  7. Test hypotheses
  8. Collect and test environmental samples
  9. Implement control measures
  10. Interact with the press, inform the public
The case definition allows researchers to distinguish between true cases and those with similar symptoms or appearances. Medical tests will sometimes be necessary for clarification. Then, as a means of comparison, it is necessary to determine the normal rate of the condition in a population to see if the occurrence is unusually high. If so, people experience the disease must be identified so epidemiologist can clarify exactly who is being affected and where. Descriptive epidemiology refers to the person/place/time data which characterizes each case (Palmer, 1989). Who was affected where and when? Putting data into such a framework allows epidemiologists to develop and test potential explanations for the outbreak. The wrap-up of an outbreak investigation involves further testing, elimination of the disease-causing agent through control measures, and communication about the situation to prevent panic and further illness.

In general, epidemiology attempts to address health promotion, disease prevention, and the quality of health care (Association of Schools of Public Health, 1999a) by "calculating disease trends and probabilities, communicating findings to the public and policymakers, and designing and implementing interventions based on the data" (Koplan, Thacker, & Lezin, 1999, p. 1154). Some argue that the field should be limited to reporting on the causes and prevalence of disease without attempting to solve the problems (Savitz, Poole, & Miller, 1999). Such a debate is just one example of the issues that currently face the practice of epidemiology.

Current Problems/Issues
The inherent limitations of epidemiologic research methods are at the heart of the issues facing epidemiology. There will always be a degree of uncertainly in study results due to factors known as biases and confounders (Savitz, Poole, & Miller, 1999). Biases are flaws in the design of a study, while confounders are external factors that subtlely influence a study's results (Hennekens & Buring, 1987). Because epidemiologic research involves the study of humans and human disease, these problems are more difficult to eliminate than in more controlled research environments. Incorrect interpretation of study results is, unfortunately, not uncommon.

Often these errors translate into great confusion for the public. Almost daily the media print the results of epidemiologic studies that name a new risk factor for an old disease. Often these results contradict earlier results. Who should the public believe? These constant publications and retractions decrease the profession's respectability among the general population (Taubes, 1995). Many established epidemiologists blame members of their own profession, who exaggerate results, identifying risk factors that are actually barely associated with diseases. It can be difficult to distinguish between low-risk and no-risk (Taubes, 1995).

Also causing problems is the emerging trend in epidemiology away from primary research. Newer, inexperienced researchers are conducting secondary analyses of data sets that were originally developed in completely unrelated studies (Holland, 1995). This can lead to a "discrepancy between the question being asked and the methods used to address the question" (Schwartz & Carpenter, 1999, p. 1175). Furthermore, epidemiologists are paying less attention to biologic probability, instead linking risk factors to health conditions through exploratory efforts, not sound scientific principles (Holland, 1995). Yes, alcohol may be related to certain types of cancers, but is there a medical mechanism that can explain why this may be so? Non-medical and medical epidemiologists must collaborate to ensure research is good research, based not only on social sciences and mathematics but on natural sciences (Holland, 1995).

Past Information Needs Studies
Few, if any, studies have been conducted to assess the information needs of practicing epidemiologists. However, some articles have been written about the development of information products for the practice of epidemiology. Computer technology in particular has greatly expanded the options available to epidemiologists in their research. Sparks (1996) and others (Fernandez, Sobreques, and Schiaffino, 1999; Joffres & LaPorte, 1998) have discussed at length the availability of electronic epidemiology journals and full texts via the Internet. Reagan (1997) provided practical advice for choosing computer systems in hospital infection control. Other articles also have discussed information systems and networks in clinical settings (Thomson, 1997; Verdier & Flory, 1994). Clarke, McLaffery, and Tempalski (1996) tackled the interesting issue of global mapping of disease in a review of new technology in geographic imaging systems. These systems are invaluable to epidemiologists, providing the "place" data in the person/place/time information need. Unfortunately, one can only identify information needs of epidemiologists by making inferences from the problems facing the profession currently.

Creating an Information Product
In identifying problems that are facing the practice of epidemiology, Nasca (1997) names access to data as a primary concern. Indeed, others (Hierholzer, 1991) have written of the high variation of data from source to source and noted that few standards have been incorporated into database development. The National Center for Health Statistics was cited as one model collection of epidemiological data (Hierholzer, 1991). As the value of epidemiology is in its contributions to public health decision making (Association of Schools of Public Health, 1999a), it is essential that epidemiologists have access to valid data when they are developing their research projects.

As mentioned earlier, the relative risk is one way of identifying risk factors for causation of disease. However, that alone is not enough, and in fact, most epidemiologists require "a very strong association between disease and risk factor [i.e. RR > 3.0]" (Taubes, 1995, p. 168) as well as "a highly plausible biological mechanism" (Taubes, 1995, p. 168) before they accept a given risk factor as the cause of disease. Plus, causation cannot be determined from a single study (Taubes, 1995). Therefore, it would be valuable for epidemiologists to have access to a single standardized database of epidemiological studies, their numerical results, and subsequently identified risk factors.

A project of this magnitude is not easily accomplished. It will require the effort and collaboration of numerous agencies, organizations, and individuals. Fortunately, it will be possible to model this project on an already existing database of genetic information, known as HuGE Net (Human Genome Epidemiology Network, 1999). Associated with the human genome project, HuGE Net is a collection of information on basic and applied population based research on genes. It acts as a crossroads between genetic and molecular epidemiology. HuGE Net has the following goals:
A project of the Centers for Disease Control and Prevention, universities, research institutes, health departments, and journals, HuGE Net provides an exemplary framework for developing an epidemiology risk factor network--RiskNet.

The goals of RiskNet will be similar to those of HuGE Net. However, the information contained in the collection will be peer-reviewed studies on health conditions and their risk factors. It will of necessity be a retrospective project, perhaps going as far back to the earliest conclusive studies that occurred in the beginning of this century. As such, the project will never be completed. RiskNet will continually be a work in progress. Housed perhaps at the CDC, RiskNet will be staffed by information professionals who have an interest or advanced degree in public health. Their task will be to collect, organize, and disseminate information from epidemiological studies past and present.

Contributions will be accepted from authors, publishers, and journals. The one criterion for acceptance will be peer review. Each report will have had to be published in a recognized epidemiological journal or monograph. The role of the information professionals involved in the project will be the classification of terminology using standardized MeSH terms, development of a searchable Internet-based data system, and ongoing entry and maintenance of the database. Ideally, the database will be a fairly simple construct, with fields for bibliographical information, the disease(s) studies, identified risk factors, and relative risk figures. What will distinguish this database from MEDLINE will be its attempt to summarize epidemiological data in an easy-to-read tabular format with which researchers can visually compare results. In addition, future efforts may include summary statements of findings, making the database valuable to healthcare consumers and other non-medical users as well.

Evaluation of the Product
Evaluation will be conducted regularly via the Internet. Though users will not have to pay for the service, they will be required to register to use RiskNet. Periodically, an online evaluation will appear, and users will be required to complete it before they can access more information. The evaluation will consist of open- and closed-ended questions regarding ease of use, perceived value of the data to their work, and suggestions for improvement. It will be necessary for RiskNet to internally monitor itself as well. A board of evaluators will be formed at the outset, consisting of representatives from each participating agency, publishing house, and journal, as well as the entire information professional team. This board will concern themselves more with the issues of database management and upkeep of RiskNet. They will review user comments and attempt to incorporate suggestions into the design of the collection.

References

Association of Schools of Public Health. (1995). You can make a difference: Pursue a career in public health [brochure]. Washington, DC: Author.

Association of Schools of Public Health. (n.d./1999a). Epidemiology [WWW document]. Available: http://www.asph.org/whatepi.htm

Association of Schools of Public Health (1999b). List of accredited schools of public health [WWW document]. Available: http://www.asph.org/listsph.htm

Centers for Disease Control and Prevention. (1999). Epidemic Intelligence Service (EIS), CDC [WWW document]. Available: http://www.cdc.gov/epo/dapht/eis/index.htm

Clarke, K. C., McLafferty, S. L., & Tempalski, B. J. (1996). On epidemiology and geographic information systems: A review and discussion of future directions. Emerging Infectious Diseases, 2(2) [WWW document]. Available http://www.cdc.gov/ncidod/eid/vol2no2/clarke.htm

Fernandez, E., Sobreques, J., & Schiaffino, A. (1999). Epidemiology and public health journals on the Internet. Journal of Epidemiology and Community Health, 53, 510-512.

Harvard School of Public Health. (1999). Register: Department of Epidemiology [WWW document]. Available: http://www.hsph.harvard.edu/Register/epi-int.html

Hayes, D. (Ed.). (1998). Exploring health care careers (vol. 1). Chicago: Ferguson Publishing Company.

Hennekens, C. H. & Buring, J. E. (1987). Epidemiology in medicine. Boston: Little, Brown and Company.

Hierholzer, W. J. (1991). Health care data, the epidemiologist's sand: Comments on the quantity and quality of data. American Journal of Medicine, 91, 21S-26S.

Holland, W. W. (1995). The hazards of epidemiology [editorial]. American Journal of Public Health, 85, 616-617.

Human Genome Epidemiology Network. (1999). About HuGE Net [WWW document]. Available: http://www.cdc.gov/genetics/hugenet/about.htm

Joffres, M. R. & LaPorte, R. E. (1998). Bringing epidemiology manuals and books onto the Internet through the Epilink. American Journal of Epidemiology, 147, 325-329.

Koplan, J. P., Thacker, S. B., & Lezin, N. A. (1999). Epidemiology in the 21st century: Calculation, communication, and intervention. American Journal of Public Health, 89, 1153-1155.

Mausner, J. S. & Kramer, S. (1985). Epidemiology: An introductory text (2nd ed.). Philadelphia: W. B. Saunders Company.

Nasca, P. C. (1997). Current problems that are likely to affect the future of epidemiology. American Journal of Public Health, 86, 907-911.

Palmer, S. R. (1989). Epidemiology in search of infectious diseases: Methods in outbreak investigation. Journal of Epidemiology and Community Health, 43, 311-314.

Reagan, D. R. (1997). Microcomputers in hospital epidemiology. Infection Control and Hospital Epidemiology, 18, 440-448.

Reingold, A. L. (1998). Outbreak investigations: A perspective. Emerging Infectious Diseases, 4(1) [WWW document]. Available: http://www.cdc.gov/ncidod/EID/vol4no1/reingold.htm

Rollins School of Public Health. (1999/n.d.). By classification [WWW document]. Available: http://web.sph.emory.edu/phemploy.nsf/By+Classification?OpenView&Start=1&Count=30&Expand=5#5

Savitz, D. A., Poole, C., & Miller, W. C. (1999). Reassessing the role of epidemiology in public health. American Journal of Public Health, 89, 1158-1161.

Schwartz, S. & Carpenter, K. M. (1999). The right answer for the wrong question: Consequences of type III error for public health research. American Journal of Public Health, 89, 1175-1180.

Sparks, S. M. (1996). Use of the Internet for infection control and epidemiology. American Journal of Infection Control, 24, 435-439.

Taubes, G. (1995). Epidemiology faces its limits. Science, 269, 164-169.

Thomson, W. M. (1997). A medication capture and analysis system for use in epidemiology. Drugs and Aging, 10, 290-298.

Verdier, C. & Flory, A. (1994). An information system for epidemiology based on a computer-based medical record. Methods of Information in Medicine, 33, 496-501.