Patterns of disease: diabetes mellitus and the rest

BMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6979.545 (Published 04 March 1995) Cite this as: BMJ 1995;310:545
  1. Ronald E LaPorte
  1. Professor Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh PA 15261 USA (rlaporte@vms.cis.pitt.edu)

    We should be investigating the relations between diseases

    As early as 400 BC Indian doctors observed that diabetes was a disease of well fed people. In 1895 Bose wrote “Amongst the Zemindars and Talookdars, who consider it a pride and honor to lead an indolent life, diabetes is a common disorder.”1 In 1962 my 76 year old Aunt Nina attributed her diabetes to being old, fat, and inactive.

    Much of what I have read about epidemiology of diabetes in the past 20 years has merely refined what my Aunt Nina said about her non-insulin dependent diabetes. We have learnt more about its causes and have more precise estimates of the relative risks of developing it as a function of age, obesity, level of physical activity, and, from Rimm and colleagues' paper in this week's journal, cigarette smoking and alcohol use (p 555).2 But substantial advances in our understanding of its epidemiology and that of other non-communicable diseases have been rare.

    The epidemiology of diabetes has recently seen a change of focus—apparent in the paper by Perry and colleagues in this issue with its examination of the interrelations between diabetes and other risk factors for coronary heart disease (p 560).3 Although Lilienfeld and Lilienfeld defined epidemiology's concerns as “the patterns of disease occurrence in human populations and the factors that influence these patterns,”4 this is not how we have investigated diabetes or in fact any non-communicable disease. We do not look at disease within populations; instead we look at a disease, “our” disease. For 40 years we have been diabetes epidemiologists, cancer epidemiologists, cardiovascular epidemiologists, and AIDS epidemiologists. Few researchers have cut across disease boundaries and examined more than one disease. Fewer still have tried to examine the patterns of diseases in a population.

    An important exception was Omran with his development of the concept of the epidemiological transition. His classic paper, which has received relatively little attention since it was published in the early 1970s, showed clearly the existence of both patterns of diseases and processes that produce these patterns.5 Moreover, the changing patterns of disease are very predictable. Omran charted the rapid fall in infectious diseases with rising socioeconomic status. Increasing life expectancy then unmasked chronic diseases, which occur in a distinctive pattern. Initially, deaths from trauma increase. Next comes a peak of non-insulin dependent diabetes, not initially associated with a rise in coronary heart disease. Coronary heart disease emerges after diabetes, and cancer five years after the others have emerged.

    This pattern is almost universally seen and is almost universally ignored. An episystems approach, which investigates the processes and patterns of diseases, could lead to important insights.6 For example, we need to evaluate the complex interrelations not only between diabetes and cardiovascular disease but also between diabetes and cancer, rheumatic fever, and polio.

    Evolutionary biology may provide the best model. This discipline has been important for understanding and predicting the rise and fall in species. Species are constantly in transition, with the numbers of one species rising and those of another falling. This symphony of change is orchestrated by the environment, which interacts with the genetic background of plants and animals. Different species change in response to each other as the result of conditions affecting both of them. They also change as the result of environmental conditions tending to “run together,” yielding strong interrelations in the ecosystem. Darwin described “how plants and animals remote in the scale of nature are bound together by a web of complex relations.”7

    Isn't this exactly how diseases “evolve” on a population basis? Diseases are constantly in transition, with one disease rising and a second falling in a systematic relation with each other. We should not talk about the origin of “disease” but the origin of “diseases”: within a population diseases are intimately bound together by a web of complex relations, which we should be investigating. We need to examine and model the evolution of patterns of diseases.

    We need to break away from our orientation towards single diseases and begin to focus on the big picture. For example, life expectancy has increased enormously almost everywhere during the past 40 years, which has unmasked diabetes as well as coronary heart disease. Yet for any given life expectancy, say 65, the causes of death within populations are almost identical worldwide. As life expectancy is becoming more and more similar worldwide so are the causes of death. In addition the disease patterns in the epidemiological transition are also related to many other features of society, such as socioeconomic status, war, the status of women, and, as Omran has shown, fertility and population growth.

    An episystems approach to diabetes and other noncommunicable diseases is meant not to replace the existing approaches but rather to complement them and provide new insights. An understanding of what drives the process and the patterns of disease will most certainly be important for forecasting the future diseases in a population. Moreover, if we are able to understand the process then prevention will not be disease specific and preventive approaches may affect a wide variety of different, biologically unrelated diseases.

    The next generation of diabetes epidemiologists needs to break away from existing models and begin to think about ways of understanding what drives the evolution of diabetes and other diseases in a population. A prerequisite is much better monitoring of disease as patterns cannot be detected unless the data have been collected.8


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
    Sign up for a free trial