LIFETIME EVENTS AND SCREENING TERMINOLOGY

When mass-screening programs for chronic diseases are developed, it is common practice to describe episodes that occur during the patient's lifetime by an accepted terminology that describes and defines the evolving events up to the point of death. These lifetime periods may be illustrated as in Figure 1.

Assuming initial health, the first event is the initiation of a biological disease process. This continues until, in the absence of a screening program, symptoms appear, leading to presentation and eventual clinical diagnosis.

The time from disease initiation to clinical diagnosis is defined as the total preclinical phase (TPCP) (8). Clearly, the TPCP is a theoretical concept, as the time at which the disease is initiated can never be known with certainty; indeed, with reference to neoplasia, it is likely that multiple events are required to establish the cancerous process, making it almost impossible to define precisely the start time of the disease.

If a mass-screening test has been developed for a disease, it becomes possible to detect its presence during the TPCP. The period between the earliest time at which the disease is detectable by the test and the time at which clinical symptoms become apparent is defined as the detectable preclinical phase (DPCP), also known as the ''sojourn time'' (9). Clearly, the length of the DPCP depends on the sensitivity of the screening test, and this is likely to be different for each disease under consideration. It is also apparent that different forms of screening tests for any one disease will be associated with different DPCP times; thus, breast self-examination will have a shorter DPCP than mammography, and digital rectal examination will have a shorter DPCP than a prostate-specific antigen (PSA) screening test. Naturally, the effectiveness of any proposed mass-screening program will be determined by the ratio of DPCP to TPCP—the longer the detectable phase, the more effective the program. Occasionally, DPCP is mistakenly understood to stand for precancerous phase. This is misleading, as a variable portion of the correctly

Figure 1 Lifetime events and terminology of screening programs. DPCP: detectable preclinical phase; TPCP: total preclinical phase. Depending on the sensitivity of the test, screening may be undertaken at any time t during the DPCP. Lead-time bias is defined as the time between the initiation of the test and the time of symptomatic clinical presentation. The death of the patient may occur earlier or later depending on whether the screening process is harmful or beneficial to the individual.


Figure 2 Length bias. Solid lines indicate duration of detectable preclinical phase. Biologically active tumors develop and present more quickly than indolent tumors that progress over long periods. Some tumors may never present clinically. Screening events (vertical dotted lines) automatically select a preponderance of the more indolent tumors. Source: Refs. 12 and 13.


Defined DPCP may be related to invasive (i. e., not precancerous) yet asymptomatic cancerous growth. Hence, screening for diseases that have a significant invasive asymptomatic period during the DPCP (perhaps breast) is likely to be less effective than screening for diseases with extensive noninvasive periods during the DPCP (perhaps prostate).

The detection of disease during the DPCP by the screening test itself results in difficulties when attempts are made to judge the effectiveness of any single program. The time between the positive result of the screening test and the time at which clinical symptoms would have appeared is defined as the lead-time bias. It will be evident that if treatment for the disease is ineffective, the patient will die at the same point as if it had originally been detected because of clinical symptoms. Survival, however, will apparently have been increased by the amount of time between the positive screening test and the clinical presentation; hence, it is necessary to account for lead-time bias when assessing the efficacy of screening for any particular disease (10). For severe chronic disease such as cancer, the elimination of leadtime bias is best achieved by randomized controlled trials of the screening program, as described below.

The death of the patient is an event that is determined by the severity of the disease and the efficacy of treatment. As noted above, it is necessary to demonstrate that a screening program results in prolonged survival after lead-time bias and other distortions have been taken into account. It should not be forgotten that screening programs may also lead to a shortening of the patient's life—if, for instance, death occurs because of septicemia following transrectal biopsy for presumed prostate cancer suspected on the basis of a screening test, the program (depending on the frequency of the complication) will be judged as being of questionable value with regard to public health.

A further distortion that may be observed when mass screening for chronic disease is described as length bias (11). In the case of neoplasia, different tumors grow at different rates, and, thus, fast-growing cancers are characterized by a short DPCP. A screening test undertaken during the DPCP will, of necessity, detect more slow-growing cancers than fast-growing cancers (Fig. 2); thus, the disease so detected will have a more favorable outcome than the disease detected by standard clinical tests. This length-bias effect may be reduced by repeat screening, but once again it can be seen that improved prognosis in the screening group may not necessarily be related to a real improvement in survival for the disease in question.

It is often difficult to persuade the general public— and occasionally the medical profession—that earlier detection of disease is not automatically associated with a better prognosis. Figures 1 and 2 well illustrate the complexity of the issues raised when mass screening is undertaken for chronic disease.

  • Contact
  • Category: Men's diseases