Introduction Data and Methods

2. Background

2.1 Prevalence as an indicator of population health

The prevalence of diseases in the population is an indicator of the current stock of health. It is the result of past rates of disease incidence, disease progression, and survival. Over time, the prevalence of disease may change because of increases or decreases in risk factors for various diseases and because of increasing ability to treat diseases in order to delay their progression to disability and death. The relative size of the change in incidence rates and survival rates will determine change in disease prevalence. Simulation analysis relating population disability levels to change in incidence of disability and mortality has clarified the primary role of incidence change in affecting population health as well as the potential negative role of increasing survival without changing incidence [Crimmins, Hayward, and Saito 1994].

The major diseases and conditions of old age can be divided into fatal and non-fatal conditions. For the lethal diseases of old age, the potential end point of an incident case may be death. Many of the causes of disability in old age, however, are not lethal and one survives with them until death from some other cause. Therefore, the prevalence of non-mortal disease is determined by the incidence of those conditions, the length of life after incidence, and the death rate from lethal conditions. Trends in the prevalence of fatal and non-fatal condition may differ.

Trends in disease presence do not necessarily represent trends in disease of a specified severity. Even with a constant prevalence of disease over time, the severity of diseases could change. It is possible that in more recent years, people are learning of the presence of less severe disease at an earlier stage because of the growing ability to diagnose noninvasively. Prevalence of disease could also increase because of increased use of medical services.

2.2 Empirical Evidence of Changes in Disease Prevalence

There have been a number of studies of trends in the prevalence of diseases and conditions among the elderly population. Manton and associates, using the U.S. National Long-Term Care Survey for the 1982 to 1989 period, examined changes in the prevalence of 16 diseases and conditions [Manton, Stallard, and Corder 1995]. After controlling for age, sex, and disability status, they found reductions in the prevalence of arthritis, arteriosclerosis, hypertension, stroke, other circulatory diseases, emphysema and dementia. They found increases in the percent of people with Parkinson's, heart disease other than heart attack, bronchitis, pneumonia, and broken hip. In this study there was no change in the prevalence of diabetes, heart attack, cancer, and asthma. Crimmins and associates [Crimmins, Reynolds, and Saito 1999] found that during the 1980s, there were decreases in cerebro and cardiovascular diseases and arthritis among persons in their 60s. Robine and his associates found increases in the prevalence of almost all diseases between 1981 and 1991 among the old in France [Robine, Mormiche, Sermet 1998]. They found large increases in the prevalence of the most important disabling diseases - cardiovascular conditions and arthritis. In France prevalence change differed for older and younger age groups; for a number of diseases where there were prevalence increases among the old, no increases were found for younger age groups.

Other researchers have concentrated on trends in specific diseases. The reported prevalence of heart disease in the U.S. has increased by about 50% between 1972 and 1989 [Feinleib 1995]. Several countries have reported increases in the prevalence of diabetes [Bloomgarden 1994]. Hann and associates found an increase in the prevalence of stroke and angina examining two cohorts from a large HMO in the U.S. from the beginning of the 1970s and 1980s [Hann et al 1996]. Stroke prevalence also increased in Minnesota during this period [Brown et al 1996, Shahar et al 1997].

Some of these studies have investigated change in both the prevalence of diseases and the processes by which the prevalence changes have come about. Generally, they attribute increases in prevalence to lengthening survival after disease diagnosis with varying pattern of change in incidence [Brown et al 1996, Feinleib 1995, Hann et al 1996, Liebson 1997, Shahar et al 1997]. Other researchers have modeled trends in incidence, mortality, and prevalence of specific diseases to see the implications for disease prevalence of assumed changes in incidence and mortality. Bonneux and associates [Bonneux et al 1994] predict that the falling mortality rates from heart disease should result in a rising prevalence of persons with heart disease. Their model, in fact, predicts that even with steady decreases in incidence of heart disease at younger old ages, in the older old ages the trend is likely to be toward both increasing incidence and prevalence among the older old.

Thus the literature is somewhat mixed in the picture that is provided of disease trends in the older population. Although the bulk of the empirical evidence is that disease prevalence is increasing, there is some indication that trends for the old may differ from those for the total population. The theoretical evidence makes it clear that if mortality from diseases is declining, there may be an effect on prevalence. If there are no accompanying declines in disease incidence but only increases in survival, prevalence of disease should increase. If both incidence and case fatality rates are declining, the effect on prevalence depends on the relative change [Crimmins, Hayward, and Saito 1994].


Introduction Data and Methods

Change in the Prevalence of Diseases among Older Americans: 1984-1994
Eileen M. Crimmins, Yasuhiko Saito
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871