As the health of the public improves, many more people are living into their ninth and tenth decades. Should these people undergo screening tests? If not, at what age do the costs and harms of screening outweigh its benefits? Few studies of preventive medicine interventions have enrolled patients past the age of 75. Closing this gap in our understanding of disease prevention should be high on the list of national health research priorities. For the present, however, each patient will require an individualized decision. What rules of thumb should we use to help older patients decide when we should stop screening them for high cholesterol levels, cancer, and other diseases?

From the individual patient’s perspective, the most useful heuristic is to compare the harms and benefits of screening for that patient. Most of the benefit in screening for a disease is reduced disease-specific mortality, and this benefit happens to people who would have died of the disease except for the interventions triggered by a positive screening test. Harm can happen to anyone, including people who never develop the disease. Because benefits increase as the risk of disease increases, the balance of harms and benefits will shift toward increasing benefit as the risk of disease increases. For most diseases, and certainly for cancer and coronary heart disease, the risk increases with advancing age. So, all other things being equal, screening older people should be more likely to lead to net benefit than screening younger people.

Of course, all other things are not equal as one ages. We know very little about the efficacy of screening in older persons, but advancing age brings changes that reduce the chance that the benefits of screening will exceed its harms.

A risk factor in younger people may not be a risk factor in older persons. Cholesterol screening is a case in point. Cohort studies suggest that an increased serum cholesterol level is a weaker risk factor for coronary heart disease in older people and may not increase the risk in older men.1 Although treatment may reduce serum cholesterol levels in older people, the coronary death rate won’t change if it is independent of the serum cholesterol level.

The risk of disease may fall in older people. After age 65, the risk of cervical cancer is very low in monogamous women who have had consistently negative cervical cytology.

Screening fails in older people in part because they die of other diseases before they can benefit from screening. A person’s total mortality rate is the sum of the mortality rates of the patient’s diseases and the patient’s age-specific mortality rate. The age-specific mortality rate increases with age, reducing the overall impact of any reduction in disease-specific mortality rate, and the improvement in life expectancy from treating a disease shrinks steadily as a person ages from the seventh into the ninth decade.2 With benefit declining during advancing age, the balance of harms and benefits will usually shift in the direction of net harm, as the harms of follow-up testing and treatment tend to increase with advancing age.

Many years of periodic screening may be required to yield a substantial difference in the mortality rate between screened and unscreened persons. In women aged 50 to 69 years, the breast cancer death rate does not change during the first four years of regular screening. After 4 years, the breast cancer death rate of screened women steadily diverges from the death rate of unscreened women, albeit slowly.

The psychological benefits of screening may change as a person ages. Reassurance that one does not have cancer may mean less as a person ages and grows to accept the closeness of death.

Screening may be increasingly burdensome as people age. Sensory and cognitive problems, physical disability, and difficulty getting transportation all increase the hardship of getting to the place where screening occurs.

Are physicians alert to the possibility that the harms of screening older people with many comorbid illnesses might outweigh the benefits? Two articles in this issue indicate that physicians do fewer screening tests in older people and in sick people. The two articles complement one another. Kiefe and her colleagues measured the frequency of cervical and breast cancer screening as a function of the comorbidity score, an indicator of the number and severity of comorbid illnesses.3 The age range of the patients, 50 to 74 years, covers the period during which breast cancer screening is most effective and avoids the years after age 75, when advancing age alone might motivate less aggressive screening. Dr. Kiefe found that the screening frequency went down steadily with increasing comorbidity score, taking factors such as age into account.

Burack and his colleagues focused on the effect of advancing age and tried to separate the effect of aging from that of worsening health.4 This study included women who were more than 50 years old and participated in a national, household-based health survey. The interviews provided information on their health status (good, fair, or poor) and limitations in activity (one or more limitations vs none). Mammography rates fell with advancing age and with worsening health status in univariate analyses, and age predicted mammography rates in multivariate analyses in which age and health status were predictor variables. In a subgroup analysis of women who reported good or better health, women aged years 75 or older had 40% lower mammography rates than women aged 50 to 64 years.

These two studies strongly suggest that advancing age and comorbid illnesses are associated with less cancer screening in women. What is the reason? Physicians appear to take comorbidity into account when deciding about screening, as the decline in screening rates with increasing comorbidity occurs at ages when mammography is strongly recommended.3 The dearth of evidence-based recommendations for older women, however, could account for the reduction in screening in women aged 75 years or older. Perhaps physicians do a mental calculation of expected length of life and withhold screening from those with a short expected life span. This hypothesis could account for the effects of comorbidity and advancing age on screening rates. Alternatively, older women and sicker women may seek screening less aggressively. Perhaps older women are more likely than younger women to defer decisions about screening to their physicians.

Are physicians right to be less aggressive in screening older people and sick people? Until there is evidence from good clinical trials, we will have to rely on indirect evidence to make patient-specific decisions. Here is a rule of thumb for deciding whether to start screening or to continue screening. First, determine the life expectancy for the patient’s physiologic age (see Table 1 in reference 2), which is the period during which screening could be helpful. Second, determine how many years of screening are required before screened people have lower disease-specific mortality rate than unscreened people. This period is approximately 5 years when screening for breast cancer  5 or for colorectal cancer.6 Third, ask how the patient feels about screening, diagnosis, and treatment. If the patient will live long enough to experience the potential benefit from screening, would accept treatment, and would value the years that could be gained, screening may be appropriate. With this rule of thumb, we will occasionally find ourselves offering screening to a healthy, vigorous, 90-year-old person.

Even for healthy older patients, most screening decisions are likely to be “close calls,” which makes talking with the patient about screening even more important. With enough discussion, the right decision about screening will become obvious.—Harold C. Sox, MD,Dartmouth Medical School, Hanover, NH.


  1. Top of page
  2. References
  • 1
    Kronmal RA, Cain KC, Ye Z. Total serum cholesterol levels and mortality risk as a function of agea report based on the Framingham data. Arch Intern Med. 1993 153:1065 73.
  • 2
    Welch HG, Albertsen PC, Nease RF, Bubolz TA, Wasson JH. Estimating treatment benefits for the elderly: the effect of competing risks. Ann Intern Med. 1996 124:577 84.
  • 3
    Kiefe CI, Funkouser E, Fouad MN, May DS. Chronic disease as a barrier to breast and cervical cancer screening. J Gen Intern Med. 1998 13:357 65.
  • 4
    Burack RC, Gurney JB, McDaniel AM. Health status and mammography use among older women. J Gen Intern Med. 1998 13:366 72.
  • 5
    Nystrom L, Rutqvist LE, Wall S, et al. Breast cancer screening with mammography: overview of Swedish trials. Lancet. 1993 341:973 8.
  • 6
    Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by screening for fecal occult blood. N Engl J Med. 1993 328:1365 71.