Cancer Incidence and Mortality
The Scientific Basis
Observed Patterns of Cancer Stage at Diagnosis
Interpreting Changes in Relative Survival Over Time
Disease-Specific and All-Cause Mortality Endpoints
Measures of Risk
Cancer Incidence and Mortality
In 2014, an estimated 1,665,540 people in the United States will be diagnosed with cancer, and 585,720 will die of cancer. Estimates of the premature deaths that could have been avoided through screening vary from 3% to 35%, depending on a variety of assumptions. Beyond the potential for avoiding death, screening may reduce cancer morbidity since treatment for earlier-stage cancers is often less aggressive than that for more advanced-stage cancers.
Several potential harms must be considered against any potential benefit of screening for cancer. Although most cancer screening tests are noninvasive or minimally invasive, some involve small risks of serious complications that may be immediate (e.g., perforation with colonoscopy) or delayed (e.g., potential carcinogenesis from radiation). Another harm is the false-positive test result, which may lead to anxiety and unnecessary invasive diagnostic procedures. These invasive diagnostic procedures carry higher risks of serious complications. A less familiar harm is overdiagnosis, i.e., the diagnosis of a condition that would not have become clinically significant had it not been detected by screening. This harm is becoming more common as screening tests become more sensitive at detecting tiny tumors. Finally, a false-negative screening test may falsely reassure an individual with subsequent clinical signs or symptoms of cancer and thereby actually delay diagnosis and effective treatment.
A 2009 publication of preliminary results of a cancer screening project in Japan using multiple whole-body screening technologies illustrates clearly the problems of false-positive screening results and potential overdiagnosis.
The project enrolled 1,217 healthy volunteers, aged 35 years and older between August 2003 and July 2004. The volunteers were employees of Hamamatsu Photonics K.K. and affiliated companies in Japan. Participants participated in a company-wide health insurance program and most received mandatory annual employee’s health check-ups, including chest x-ray, fecal occult blood (FOB) testing, upper gastrointestinal series, and mammography screeningfor many years. Twenty applicants with a prior history of cancer were excluded, leaving 1,197 participants in the evaluation. All enrollees could be considered at low risk for screen-detected cancer.
All participants were offered annual whole body screening by fluorodeoxyglucose positron emission tomography, chest and abdominal computed tomography, brain and pelvic magnetic resonance imaging (MRI), analyses of serum tumor markers including carcinoembryonic antigen, cancer antigen 19-9 (CA19-9), squamous cell carcinoma antigen, prostate specific antigen for men older than aged 50 years, cancer antigen 125 (CA-125) for women, and FOB testing.
Specific clinically relevant classifications were standardized for all test results and imaging tests received three independent readings with consensus resolution (for details refer to ). A screen was classified positive if any test was suggestive of malignancy, and in these cases, the subject was referred to a local hospital for further testing or retesting at the screening center. Interval cancer events were ascertained at subsequent screenings or by interview.
Twenty-two primary cancers were pathologically confirmed, 19 by annual screening, and 18 at the initial screen. The multitest whole body testing overall sensitivity was reported as 81.8% with a specificity of 70.6% (false-positive rate 29.4%) when the results were considered suggestive of malignancy or malignancy could not be definitively ruled out and 68.2% versus 87.4%, respectively, for the more restricted screen classification suggestive of malignancy only.
The incidence of cancers found on the initial screen exceeded the estimated age-matched annual incidence of cancer in Japan by three to four times. Several cancers found on screening were known to be indolent and might be harmless for a long period without affecting cancer mortality.
The possible harms of repeated multimodal cancer screening programs were also assessed using data from the intervention arm of the randomized, controlled Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. A total of 68,436 participants received screening for three cancers, depending on sex, using prostate-specific antigen (PSA), digital rectal exam, transvaginal ultrasound, serum CA-125, chest x-ray, and/or flexible sigmoidoscopy. The study found that after four screening tests (which occurred within a 1-day period in the trial), the cumulative probability of an individual receiving at least one false-positive result was 37% for men and 26% for women. After 14 tests (or 3 years of screening), the probability rose to 60% for men and 49% for women. The cumulative probability of undergoing an invasive procedure as a direct result of a false-positive exam was 29% for men and 22% for women after 14 tests.
The harms of screening are of particular concern in the case of overdiagnosis, as the individual cannot, by definition, reap any potential benefit of screening, but may experience the associated adverse effects, including those of unnecessary treatments. There are two forms of overdiagnosis: 1) the detection of a lesion with essentially no malignant potential (sometimes called “pseudodisease”); and 2) the detection of a lesion that is slow-growing enough that the individual will go on to die of a competing cause of death first. The use of screening tests in individuals with limited life expectancies can therefore be an important cause of overdiagnosis. Studies of current U.S. practice patterns have demonstrated considerable utilization of cancer screening tests in situations where the likelihood of overdiagnosis is high, and benefit very low. For example, investigators linked data from the Surveillance, Epidemiology, and End Results (SEER) cancer registry to Medicare claims in order to evaluate the use of routine cancer screening in persons aged 65 years and older with a known advanced cancer diagnosis (including stage IIIB–IV lung cancer, stage IV colorectal cancer, stage IV gastroesophageal cancer, stage IV breast cancer, and advanced-stage pancreatic cancer). They found 9% of women with established advanced cancer diagnoses received routine screening mammographies and 6% received routine Pap smears; 15% of men with established advanced cancer diagnoses still received routine PSA screening. The strongest individual predictor of routine screening in the setting of advanced cancer was a history of recent screening prior to diagnosis.
There is an association between cancer diagnosis and suicide behavior (either completed or attempted) in the first year after diagnosis. This association has been seen in both adults and young individuals, with relative risk (RR) ranging from 1.6 to 2.0.[6,7]
In developing the cancer screening summaries, the PDQ Screening and Prevention Editorial Board uses the following definitions:
- Screening is a means of detecting disease early in asymptomatic people.
- Positive results of examinations, tests, or procedures used in screening are usually not diagnostic but identify persons at increased risk for the presence of cancer who warrant further evaluation.
- Diagnosis is confirmation of disease by biopsy or tissue examination in the work-up following positive screening tests. (Following a positive screening result, cancer can often be ruled out by procedures other than biopsy or tissue examination.)
The purpose of this summary is to present an explicit evidence-based approach used in the development of the screening summaries. In reaching conclusions, evidence on the balance of risks and benefits is weighed. Cost and cost-effectiveness, however, is not taken into account. Assignment of levels of evidence associated with such screening tests is also discussed.
By design, the PDQ does not issue any clinical practice guidelines. Although many public health organizations present guidelines for health care and screening activities, the quality of these guidelines vary widely, partly because they are opinion based and rely on systematic reviews of evidence that are of varying quality. The best guidelines as assessed by the Appraisal of Guidelines for Research and Evaluation (AGREE) are those based on the best systematic reviews as evaluated by the Assessment of Multiple Systematic Reviews (AMSTAR).Summary Development
The cancer screening summaries are based on various levels of published scientific evidence and collective clinical experience. The highest level of evidence is taken as mortality reduction in controlled, randomized clinical trials. The results of clinical studies, case-control studies, cohort studies, and other information are also considered in formulating the summaries. In addition, the incidence of cancer, stage distribution, treatment, and mortality rates are considered. The summaries are subject to modification as new evidence becomes available.The Scientific Basis
At least two requirements must be met for screening to be efficacious:
- A test or procedure must be available to detect cancers earlier than if the cancer were detected as a result of the development of symptoms.
- Evidence must be available that treatment initiated earlier as a consequence of screening results in an improved outcome.
These requirements are necessary but not sufficient to prove the efficacy of screening, which requires a decrease in cause-specific mortality. For example, these two criteria are met in the case of screening for childhood neuroblastoma by assessment of urinary catecholamine metabolites. On the basis of these criteria, a mass screening program was conducted in Saitama Prefecture, Japan, between 1981 and 1992 for 6-month-old infants. During that 12-year period, the annual incidence of neuroblastoma in children younger than 1 year increased from about 28 per million to 260 per million but without a significant reduction in incidence in children older than 1 year. Because there also was no reduction in mortality for the disease, this experience provided strong evidence of overdiagnosis—diagnosis of some neuroblastomas detectable by screening, which would not have been clinically diagnosed later. Similar experiences have been reported elsewhere in Japan  and in the Quebec Neuroblastoma Screening Project (QNSP) in Canada. The history of screening for neuroblastoma also provides a useful illustration of the benefit of undertaking well-designed evaluations of emerging screening technologies before implementing screening programs. Although such studies are very costly, it has been shown that the QNSP itself averted unnecessary morbidity for thousands of children and did so while returning a yield plausibly estimated at a cost savings 64.5 times the investment in the study.Detection
Direct or assisted visual observation is the most widely available examination for the detection of cancer. It is useful in identifying suspicious lesions in the skin, retina, lip, mouth, larynx, external genitalia, and cervix.
The second most available detection procedure is palpation to detect lumps, nodules, or tumors in the breast, mouth, salivary glands, thyroid, subcutaneous tissues, anus, rectum, prostate, testes, ovaries, and uterus and enlarged lymph nodes in the neck, axilla, or groin.
Internal cancers require procedures and tests such as endoscopy, x-rays, MRI, or ultrasound. Laboratory tests, such as the Pap smear or the FOB test have been employed for detection of specific cancers.
The performance of screening tests is usually measured in terms of sensitivity, specificity, and positive-predictive values (PPV) and negative-predictive values (NPV). Sensitivity is the chance that a person with cancer has a positive test. Specificity is the chance a person without cancer has a negative test. PPV is the chance that a person with a positive test has cancer. NPV is the chance that a person with a negative test does not have cancer. PPV and to a lesser degree, NPV are affected by the prevalence of disease in the screened population. For a given sensitivity and specificity, the higher the prevalence, the higher the PPV.High-Risk Populations
Some individuals are known to be at high risk for cancer, such as those with a personal history of cancer or a strong family history of cancer (in two or more first-degree relatives); increasingly, as genetic mutations and polymorphisms are found to be associated with specific cancers, high-risk individuals will be identified through genetic testing. The type, periodicity, and commencement of screening in high-risk populations for most cancers reflect the judgment of practitioners rather than evidence from scientifically conducted studies. Physician judgment is needed in such circumstances to determine the most appropriate application of available screening methods. Prudence suggests increased vigilance in the higher-risk populations. At a minimum, this means that the high-risk person is identified, is counseled appropriately, and regularly undergoes those screening procedures that have been shown to be of benefit to the general population.Cancer Recurrence
Please refer to the PDQ Adult Treatment summaries for information on cancer recurrence.Improved Outcomes
For nearly all cancers, treatment options and survival are related to stage, which is generally characterized by the anatomic extent of disease. On this basis, it is assumed that early detection of cancer at an earlier stage may yield better outcomes. In the 1940s, a generalized staging classification of localized, regional, and distant disease was developed to show long-term trends, and it is still useful. In the more detailed TNM system, which has been periodically modified, the (T)umor size, the status of the lymph (N)odes, and the status of distant (M)etastases are also categorized. These elements are grouped into stages 0, I, II, III, and IV according to their association with survival. In general, larger primary malignant tumors have a higher incidence of metastasis to regional lymph nodes and to distant sites. Stage has such a profound effect on outcome that all randomized treatment trials require the comparison of similar stages in evaluating differences in outcome. Shifts in stage may also herald improved survival and decreased mortality, though stage shift alone does not establish benefit.
Biologic cellular characteristics of cancer, such as grade, hormone sensitivity, and gene overexpression are recognized as important predictors of cancer behaviors. For example, high-grade cancer may be fast growing and quick to metastasize regardless of stage at the time of diagnosis. Therefore, detection of these cancers when they are small may not affect outcome. Randomized controlled trials are most definitive in determining screening benefits.Observed Patterns of Cancer Stage at Diagnosis
The Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute gathers cancer incidence data from 11 geographic areas, covering approximately 14% of the U.S. population. These population-based data of long duration (1973–present) are a unique and important resource in monitoring stage-related survival.Interpreting Changes in Relative Survival Over Time
Increases in survival over time, however, even when based on data from tumor registries, such as SEER that include all cases in a given population, are difficult to interpret. They may reflect the benefits of early detection or improved treatment or both, but they may also result from lead-time bias and overdiagnosis, both of which occur commonly with screening.
Lead-time bias will result in longer estimated survival of people with cancers that have been identified through screening because the time preceding the clinical diagnosis of the cancer is included in the calculation of survival.
Overdiagnosis may result from finding cancers that would never have become manifest clinically. By definition, these cancers have a good prognosis. For example, autopsy series have shown a high percentage of occult early prostate carcinomas in elderly men who died of causes unrelated to prostate cancer. The discovery of these cancers through screening could increase the number of cases and give the appearance of stage shift, and of increases in survival or cure rates, without necessarily reducing mortality. An analysis of data reported by the SEER program between 1950 and 1996 found that changes over time in 5-year relative survival rates for 20 major cancers were essentially unrelated to trends in mortality rates for those cancers over the same period. The authors suggest that changes in 5-year survival rates are largely due to earlier diagnosis and to detection of subclinical cases that might never have surfaced clinically. They conclude that inferences about the effectiveness of early diagnosis or treatment should not be drawn from temporal changes in 5-year survival rates, but rather should be based on changes in mortality rates. Thus, changes in 5-year survival rates or stage shifts are not appropriate measures of the effectiveness of screening for early disease. Reductions in incidence rates for late-stage tumors represent a better measure of progress due to screening than 5-year survival trends, although such evidence is less compelling than reductions in mortality.Study Designs
Varying study designs may be available to support a given summary. The strongest design would be obtained from a randomized controlled trial. It is, however, not always practical to conduct such a trial to address every question surrounding the field of screening. For each summary of evidence statement, the associated strength of study designs are listed. There are five study designs that are generally used in judging the evidence. In order of strength of design, the five levels are as follows:
- Evidence obtained from randomized controlled trials.
- Evidence obtained from nonrandomized controlled trials.
- Evidence obtained from cohort or case-control studies.
- Evidence obtained from ecologic and descriptive studies (e.g., international patterns studies, time series).
- Opinions of respected authorities based on clinical experience, descriptive studies, or reports of expert committees.
Experimental trials are designed to correct for or eliminate selection, lead-time, length, healthy volunteer, and other biases when prospectively testing a detection procedure to determine its effect on health outcome. The highest level of evidence and greatest benefit from screening is mortality reduction in a randomized controlled trial. For most sites, such evidence is not available. Theoretically it is possible to conduct randomized trials for most interventions, but the sample size that is needed, the expense, and the duration of such trials for most cancers, frequently make this approach impractical. Therefore, evidence obtained by other methods is often used.
In certain cases, a preliminary alternative to using mortality reduction to evaluate a new screening modality could be a relatively short-term (e.g., several years) comparison of interval cancer rates observed in a randomized trial comparing the new test and the “standard” screening modality. If the new screening test has the potential to improve disease-specific mortality, repeated applications over a discrete period of time should result in a lower proportion of patients in the intervention arm presenting with symptomatic cancer (of the type screened for) between negative screens. That is, through increased early detection and resulting treatment, the new screening test prevents a higher percentage of clinically important asymptomatic lesions from progressing to overt cancer. Unlike cross-sectional sensitivity comparisons in which study participants receive both new and older screening modalities, this trial design allows for an estimation of the degree of overdiagnosis generated by a screening test. This comparison should take place within the context of a randomized controlled trial.
Case-control and cohort studies provide indirect evidence for the effectiveness of screening, but it is difficult to eliminate the contribution of selection bias and healthy volunteer biases evident in these studies.
Ecological studies can demonstrate association between the use of screening and a stage shift in cancer that can provide indirect evidence of the value of screening. Such evidence is particularly compelling for the effectiveness of screening for cervical cancer. Ecological correlation of mortality and intensity of screening has been used in this context. Such studies do not prove a mortality-reduction effect, and the potential for bias to invalidate inferences from nonexperimental studies or to give misleading results can be substantial.[17-22]
Descriptive uncontrolled studies based on the experience of individual physicians, hospitals, and nonpopulation-based registries may yield some information about screening. The performance characteristics of various detection tests, such as sensitivity, specificity, and PPVs are generally first reported in such descriptive studies. The first evidence that screening may be successful is an increase in the incidence of early cancers and a decreased incidence of late-stage metastatic cancers (stage shift); later, a reduction in deaths may occur. These descriptive studies do not establish efficacy because of the absence of an appropriate control group.
A more detailed description of how the overall evidence regarding benefits and harms of screening tests is graded by the PDQ Screening and Prevention Editorial Board can be found in the PDQ summary on Levels of Evidence for Cancer Screening and Prevention Studies.Disease-Specific and All-Cause Mortality Endpoints
Disease-specific mortality has been the most widely accepted endpoint in randomized clinical trials of cancer screening; however, the validity of this endpoint rests on the fundamental assumptions that the cause of death can be accurately determined and that the screening and subsequent treatment have negligible effects on other causes of death. Recent reviews of randomized clinical trials of cancer screening suggest that misclassification in cause of death has been a major problem and that misclassification has led to an overestimation of the effectiveness (or an underestimation of the harms) of screening.[23-25] In contrast to disease-specific mortality, all-cause mortality depends only on an accurate ascertainment of deaths and when they occur and therefore is not affected by misclassification in cause of death. One major limitation of the all-cause mortality endpoint; however, is that it is unlikely to reveal a statistically significant effect of cancer screening because this intervention is usually targeted to a disease that causes only a small proportion of all deaths. Nevertheless, all-cause mortality should be considered in conjunction with disease-specific mortality to reduce the possibility that a major harm (or benefit) from screening is hidden by misclassification in cause of death.Measures of Risk
Several measures of risk are used in cancer research. Absolute risk or absolute rate measures the actual cancer risk or rate in a population or subgroup (e.g., U.S. population, or whites or African Americans). For example, the SEER Program reports risk and rate of cancer in specific geographic areas of the United States.
Rates are often adjusted (e.g., age-adjusted rates) to allow a more accurate comparison of rates over time or among groups. The purpose of the adjustment is to make the groups more alike with respect to important characteristics that may affect the conclusions. For example, when the SEER Program compares cancer rates over time in the United States, the rates are adjusted to one age distribution. If this were not done, cancer rates would seem to increase over time simply because the U.S. population is getting older and the risk of cancer is higher in older age groups.
RR compares the risk of developing cancer among those who have a particular characteristic or exposure with those who do not. RR is expressed as a ratio of risks or rates; it ranges from infinity to the inverse of infinity (i.e., zero). If the RR is greater than one, the exposure or characteristic is associated with a higher cancer risk; if the RR is one, the exposure and cancer are not associated with one another; if the RR is less than one, the exposure is associated with a lower cancer risk (i.e., the exposure is protective). RR is often used in clinical trials of cancer prevention and screening to estimate the reduction in cancer risk or risk of death, respectively.
An odds ratio (OR) is often used as an estimate of the RR. It, too, indicates whether there is an association between an exposure or characteristic and cancer. It compares the odds of an exposure or characteristic among cancer cases with the odds among a comparison group without cancer. Although not as intuitively understood as rates or risk, OR is used because it is statistically more valid in some settings when other measures of risk are not valid. For relatively uncommon events/diseases such as a cancer diagnosis, it can be interpreted like a RR is interpreted; however, it becomes a progressively inaccurate estimate of the RR as the underlying absolute risk of disease in the population under study rises above 10%. ORs are typically used in case-control studies to identify potential risk factors or protective factors for cancer.
Risk or rate difference (or excess risk) compares the actual cancer risk or rate among at least two groups of people, based on an important characteristic or exposure, by subtracting the risks or rates from one another (e.g., subtracting lung cancer rates among nonsmokers from that of cigarette smokers estimates the excess risk of lung cancer due to smoking). This can be used in public health to estimate the number of cancer cases that could be avoided if an exposure were reduced or eliminated in the population.
Population-attributable risk measures the proportion of cancers that can be attributed to a particular exposure or characteristic. It combines information about the RR of cancer associated with a particular exposure and the prevalence of that exposure in the population, and estimates the proportion of cancer cases in a population that could be avoided if an exposure were reduced or eliminated.
Number needed to screen estimates the number of people that must participate in a screening program for one death to be prevented over a defined time interval.
Average life-years saved estimates the number of years that an intervention saves, on average, for an individual who receives the intervention. This reflects mortality reduction and life extension (or avoidance of premature deaths).References
- American Cancer Society.: Cancer Facts and Figures 2014. Atlanta, Ga: American Cancer Society, 2014. Available online. Last accessed March 26, 2014.
- Kramer BS: The science of early detection. Urol Oncol 22 (4): 344-7, 2004 Jul-Aug. [PUBMED Abstract]
- Nishizawa S, Kojima S, Teramukai S, et al.: Prospective evaluation of whole-body cancer screening with multiple modalities including [18F]fluorodeoxyglucose positron emission tomography in a healthy population: a preliminary report. J Clin Oncol 27 (11): 1767-73, 2009. [PUBMED Abstract]
- Croswell JM, Kramer BS, Kreimer AR, et al.: Cumulative incidence of false-positive results in repeated, multimodal cancer screening. Ann Fam Med 7 (3): 212-22, 2009 May-Jun. [PUBMED Abstract]
- Sima CS, Panageas KS, Schrag D: Cancer screening among patients with advanced cancer. JAMA 304 (14): 1584-91, 2010. [PUBMED Abstract]
- Lu D, Fall K, Sparén P, et al.: Suicide and suicide attempt after a cancer diagnosis among young individuals. Ann Oncol 24 (12): 3112-7, 2013. [PUBMED Abstract]
- Baade PD, Fritschi L, Eakin EG: Non-cancer mortality among people diagnosed with cancer (Australia). Cancer Causes Control 17 (3): 287-97, 2006. [PUBMED Abstract]
- Burda BU, Norris SL, Holmer HK, et al.: Quality varies across clinical practice guidelines for mammography screening in women aged 40-49 years as assessed by AGREE and AMSTAR instruments. J Clin Epidemiol 64 (9): 968-76, 2011. [PUBMED Abstract]
- Yamamoto K, Hayashi Y, Hanada R, et al.: Mass screening and age-specific incidence of neuroblastoma in Saitama Prefecture, Japan. J Clin Oncol 13 (8): 2033-8, 1995. [PUBMED Abstract]
- Bessho F: Effects of mass screening on age-specific incidence of neuroblastoma. Int J Cancer 67 (4): 520-2, 1996. [PUBMED Abstract]
- Woods WG, Tuchman M, Robison LL, et al.: A population-based study of the usefulness of screening for neuroblastoma. Lancet 348 (9043): 1682-7, 1996 Dec 21-28. [PUBMED Abstract]
- Soderstrom L, Woods WG, Bernstein M, et al.: Health and economic benefits of well-designed evaluations: some lessons from evaluating neuroblastoma screening. J Natl Cancer Inst 97 (15): 1118-24, 2005. [PUBMED Abstract]
- Woolf SH: Screening for prostate cancer with prostate-specific antigen. An examination of the evidence. N Engl J Med 333 (21): 1401-5, 1995. [PUBMED Abstract]
- Welch HG, Schwartz LM, Woloshin S: Are increasing 5-year survival rates evidence of success against cancer? JAMA 283 (22): 2975-8, 2000. [PUBMED Abstract]
- Irwig L, Houssami N, Armstrong B, et al.: Evaluating new screening tests for breast cancer. BMJ 332 (7543): 678-9, 2006. [PUBMED Abstract]
- Hakama M, Miller AB, Day NE, eds.: Screening for cancer of the uterine cervix. Lyon, France: International Agency for Research on Cancer, 1986.
- Connor RJ, Prorok PC, Weed DL: The case-control design and the assessment of the efficacy of cancer screening. J Clin Epidemiol 44 (11): 1215-21, 1991. [PUBMED Abstract]
- Friedman DR, Dubin N: Case-control evaluation of breast cancer screening efficacy. Am J Epidemiol 133 (10): 974-84, 1991. [PUBMED Abstract]
- Janzon L, Andersson I: The Malmo mammographic screening trial. In: Miller AB, Chamberlain J, Day NE, et al., eds.: Cancer Screening. Cambridge: Cambridge University Press, 1991, pp 37-44.
- Moss SM: Case-control studies of screening. Int J Epidemiol 20 (1): 1-6, 1991. [PUBMED Abstract]
- Weiss NS, Lazovich D: Case-control studies of screening efficacy: the use of persons newly diagnosed with cancer who later sustain an unfavorable outcome. Am J Epidemiol 143 (4): 319-22, 1996. [PUBMED Abstract]
- Suzuki KJ, Nakaji S, Tokunaga S, et al.: Confounding by dietary factors in case-control studies on the efficacy of cancer screening in Japan. Eur J Epidemiol 20 (1): 73-8, 2005. [PUBMED Abstract]
- Black WC: Overdiagnosis: An underrecognized cause of confusion and harm in cancer screening. J Natl Cancer Inst 92 (16): 1280-2, 2000. [PUBMED Abstract]
- Olsen O, Gøtzsche PC: Screening for breast cancer with mammography. Cochrane Database Syst Rev (4): CD001877, 2001. [PUBMED Abstract]
- Black WC, Haggstrom DA, Welch HG: All-cause mortality in randomized trials of cancer screening. J Natl Cancer Inst 94 (3): 167-73, 2002. [PUBMED Abstract]