9
Medical Screening

In this chapter and the next, the committee turns to issues of screening in the context of the Radiation Exposure Screening and Education Program (RESEP), which is administered by the Health Resources and Services Administration (HRSA) in the Department of Health and Human Services. The broader framework is the role that RESEP plays relative to the Radiation Exposure Compensation Act (RECA) and its amendments, as administered by the Department of Justice (DOJ).

HISTORICAL CONTEXT OF SCREENING IN RESEP

The original 1990 RECA legislation does not mention screening for the diseases it covers. The omission may have arisen because RECA’s primary purpose is compensation. The inaccessibility to services for determining eligibility for compensation led to a change in the original RECA statute, but Sec. 2, “Findings” of the RECA amendments of 2000, still does not explicitly mention screening. However, screening could be construed as implied in Congress’ sixth finding (Public Law 106-245 [S. 1515] July 10, 2000 Radiation Exposure Compensation Act Amendments of 2000 106 PL 245; 114 Stat. 501) that

it should be the responsibility of Federal Government in partnership with State and local governments and appropriate healthcare organizations, to initiate and support programs designed for the early detection, prevention and education on radiogenic diseases in approved States to aid the thousands of individuals adversely affected by the mining of uranium and the testing of nuclear weapons for the Nation’s weapons arsenal. (emphasis added)



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program 9 Medical Screening In this chapter and the next, the committee turns to issues of screening in the context of the Radiation Exposure Screening and Education Program (RESEP), which is administered by the Health Resources and Services Administration (HRSA) in the Department of Health and Human Services. The broader framework is the role that RESEP plays relative to the Radiation Exposure Compensation Act (RECA) and its amendments, as administered by the Department of Justice (DOJ). HISTORICAL CONTEXT OF SCREENING IN RESEP The original 1990 RECA legislation does not mention screening for the diseases it covers. The omission may have arisen because RECA’s primary purpose is compensation. The inaccessibility to services for determining eligibility for compensation led to a change in the original RECA statute, but Sec. 2, “Findings” of the RECA amendments of 2000, still does not explicitly mention screening. However, screening could be construed as implied in Congress’ sixth finding (Public Law 106-245 [S. 1515] July 10, 2000 Radiation Exposure Compensation Act Amendments of 2000 106 PL 245; 114 Stat. 501) that it should be the responsibility of Federal Government in partnership with State and local governments and appropriate healthcare organizations, to initiate and support programs designed for the early detection, prevention and education on radiogenic diseases in approved States to aid the thousands of individuals adversely affected by the mining of uranium and the testing of nuclear weapons for the Nation’s weapons arsenal. (emphasis added)

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program Sec. 4 of the RECA amendments gives an explicit requirement for screening. The HRSA administrator must make competitive grants available “for the purpose of carrying out programs to screen individuals described under section 4(a)(1)(A)(i) or 5(a)(1)(A) of the Radiation Exposure Compensation Act (42 U.S.C. 2210 note) for cancer as a preventative health measure” (PL 106-245 Sec. 417C(b)(1)) (emphasis added). In other words, RECA, as amended, does require cancer screening as a preventive health intervention for the purpose of early detection (CFR, Vol. 67, No. 83, page 21256). To carry out that mandate, HRSA created RESEP and announced the first fiscal-year competitive application cycle on April 30, 2000, in the Federal Register. In its announcement, HRSA described the legislation as providing the authority for competitive grants for “individual cancer screening” (CFR, Vol. 67, No. 83, page 21257). Although the amended legislation might be read as linking screening to cancer, this CFR section provides for the “availability of $3.0 million to eligible entities for the purpose of carrying out programs to screen eligible individuals for cancer and other radiogenic diseases” (ibid) (emphasis added). Because RECA legislation covers both nonmalignant and malignant radiogenic diseases, RESEP is consistent with the overall intent of the amended RECA in not limiting screening to cancer. HRSA charges its Bureau of Primary Health Care with administering the RESEP grant program. On May 2, 2002, and June 5, 2003, HRSA published a program information notice providing guidance to potential grantees about the RESEP competitive process. This guidance includes eligibility requirements, program expectations, review criteria, and award factors for the grants. RESEP grantees are expected to pursue multiple core activities: education, screening and early detection, referrals for medical treatment, eligibility assistance, quality assurance, staffing, data collection, finance, performance reports, and outreach. We examine only the screening and early detection component here. Screening and Early Detection in RESEP HRSA treats screening and early detection as a single activity. The notice did distinguish between nonmalignant and malignant radiogenic diseases. RESEP identifies basic screening protocols and differentiates basic screening and an array of steps collectively considered referrals for medical treatment (advanced testing, diagnosis, evaluation, and treatment), as follows: Screening and early detection: Basic screening protocols. Referrals for medical treatment: Advanced testing. Diagnosis.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program Evaluation. Treatment. Table 9.1 lists the HRSA requirements for a screening and early detection program with the basic screening protocols for both nonmalignant and malignant diseases as the agency categorized them. Not all the tests listed (in the table), however, detect clinical diseases and processes that have been associated scientifically with exposure to radiation. Referrals for Medical Treatment If basic screening protocols reveal abnormalities in patients, then the providers must refer the patients to a hospital, a special clinic or an imaging center, or doctor’s office for advanced testing. Table 9.2 elucidates what HRSA means by advanced testing THE NATURE OF SCREENING For our purposes, screening is of at least two different types: (1) screening for medical reasons, with the underlying purpose of improving health outcomes; and (2) screening to identify persons who may be eligible for compensation under RECA. This committee distinguishes these two concepts with respect to the underlying purposes of RESEP, using medical screening for the former and compensational screening for the latter. This distinction is not traditional but is required because RESEP specifies screening for some diseases not traditionally recommended for medical screening to improve health outcomes. Because a term is required to describe this RESEP concept, we coin the term compensational screening. This chapter is concerned largely with epidemiologic and statistical factors related to medical screening, the limitations of medical screening, and the implications for screening in RESEP. We comment as well on future research, and we introduce issues that extend beyond the purely technical questions of screening itself. Chapter 10 discusses these findings in the context of RECA and RESEP, which are oriented more toward compensation than toward medical interventions. Screening Definitions Screening has numerous meanings. One general meaning is sifting or filtering objects to separate what is wanted or desirable from what is not, as when a miner during the Alaskan gold rush screened for gold. In medical contexts, screening is commonly directed at groups of individuals who are asymptomatic, using relatively speedy, inexpensive tests and procedures. Such screening is

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program TABLE 9.1 Screening Protocols in RESEP Screening Protocols Compensable Nonmalignant Radiogenic Diseases Compensable Malignant Radiogenic Diseases Medical and occupational history Must include date of exposure, place, duration of employment, and tobacco use Must include date of exposure, place, duration of employment; special attention to symptoms of and risk factors for primary cancers or other diseases covered by RECA; should also include tobacco, alcohol, and caffeine use Physical examination Emphasis on pulmonary, cardiovascular, and renal systems Complete examination to include all cancers covered by RECA Chest radiography Standard posterior-anterior view chest radiograph for presence of radiologic fibrosis, silicosis, or pneumoconiosis As indicated by physician Pulmonary-function testing As needed, can include spirometry, lung volumes, arterial blood gases, and DLCOa See physical examination Routine testing Other routine laboratory work and electrocardiography as required Routine laboratory work as indicated by physician Other Program Requirements Included in Screening and Early Detection     Followup Patient contact via telephone; report results to patient, primary care physician, or both; periodic re-evaluation See case management below Case management Not required Extensive followup to ensure that care was received; documented monitoring of patient progress; all operative, consultative, procedural, and pathology reports and physician, hospital, and healthcare facility discharge summaries maintained aDLCO, the diffusing capacity of the lung for carbon monoxide (CO), i.e., measurement of carbon monoxide transfer from inspired gas to pulmonary capillary blood (American Thoracic Society, 1995).

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program TABLE 9.2 Advanced Testing Protocols in RESEP Advanced Testing for Compensable Nonmalignant Radiogenic Diseases Advanced Testing for Compensable Malignant Radiogenic Diseases Measurement of lung volumes and diffusion capacity (if not performed earlier) Resting and exercise arterial blood gases if not medically contraindicated High resolution computed tomography and computed tomography-scans; chest tomography, bronchoscopy, ventilation and perfusion lung scanning, pulmonary angiography, and thoracentesis, Pleural biopsy, magnetic resonance imaging, and positron emission tomography scans if required 24-hour urine studies and supplementary blood tests if not previously performed Renal ultrasonography, radionuclide scanning, magnetic resonance imaging, and renal biopsy if required Endoscopy Tissue biopsies and fine-needle aspiration Imaging studies, including computed tomography scans, magnetic resonance imaging, mammography and other breast imaging techniques, radionuclide imaging, ultrasonography, regular x-rays and contrast studies generally intended to distinguish people with some undiagnosed and even unsuspected ailment or pathologic condition from those who are well (at least with respect to that ailment or condition). Screening may also be used to detect risk factors for disease that may give individuals a higher than average probability of illness; it may also be directed at individuals who are known or suspected to have risk factors. Typically, medical screening is directed at large numbers of apparently well individuals (for example, the public at large) with the aim of finding those who have undiagnosed problems, genetic traits, or other characteristics that may benefit from medical intervention. The underlying assumption is that early detection can improve outcomes through early treatment or changes in lifestyle. Screening is used to facilitate appropriate care early in a disease, before serious signs, symptoms, or complications develop. An Institute of Medicine (IOM) committee stated that such screening for cancer “refers to the early detection of cancer or premalignant disease in persons without signs or symptoms suggestive of the target condition (the type of cancer that the test seeks to detect)” (IOM, 2003b, p. 156). Positive results from screening tests are usually not conclusive, so confirmatory tests are needed before a diagnosis is established; if such a diagnosis is made, then followup, referral, and treatment would typically be expected to ensue. Other terms are sometimes used in conjunction with or as synonyms for medical screening. Among them are case-finding and early detection, although these terms have not been used with total consistency.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program The distinction rests on whether such testing is done in the context of an established provider-patient relationship. We use patient screening to refer to testing within such a relationship and case-finding if the person screened and the testing program or provider have no established relationship. Thus, this distinction refers to the context of the screening program. Much of the discussion in this chapter and the rest of this report applies to both contexts. When what we say refers properly to both contexts, we simply use the term screening. PRINCIPLES OF SCREENING Basic Tenets Several principles should guide an analysis of screening and case-finding, whether for medical purposes, for compensation, or both. First, screening asymptomatic individuals presumes that efficacious and effective therapies exist for treating the disease either in its early (preclinical) stages or at later stages upon clinical detection (that is, upon diagnosis in patients with signs or symptoms). If that is not the case, then either screening is needless or treatment should await clinical detection. Second, screening always carries some risk of harm. These risks arise from the tests themselves, from diagnostic labeling (whether accurate or not), and from false-positive results that lead unnecessarily to further tests and possibly therapeutic interventions. These risks or harms must be balanced against the assumed benefits of screening. This balancing is especially important because medical tests almost always carry a risk of physical, social, or psychological harm; in the RESEP context, these harms must be weighed against the possibility of compensation. At low probabilities of disease, screening engenders substantial risk of false-positive results and their consequences. Screening for more than one disease amplifies that problem. Screening for one disease using multiple tests is also problematic. Third, particularly for compensational purposes, medical screening tests are appropriate only after the individuals are fully informed of the risks that these tests pose (see Chapter 8). Fourth, screening may provide useful information about related diseases. One example is screening for diabetes among people who have high blood pressure or high blood lipid concentrations but not symptoms or signs of diabetes (Harris et al., 2003); positive results of diabetes screening (especially impaired glucose tolerance or impaired fasting glucose) may call for changes in the management of the other conditions (perhaps more than for the possible diabetes itself). Fifth, if publicly supported screening programs are to be effective, ethical, and equitable, some means must be available for screened populations to gain access to appropriate followup, diagnosis, and therapy.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program Sixth, once a person is screened, he or she should, in general, be informed of any clinical information that the test revealed. Screening, even for compensational purposes, may provide information that should be linked to actions that might improve the medical situation for the person screened. Once screening for compensational purposes has been done, the patient has already been exposed to the risks of the test and of a possibly false-positive result. Having undertaken this risk, the screened individual should have the opportunity to receive whatever medical benefit such information might provide. Stated another way, in general once a person is screened, he or she should be informed of any clinical information that the test revealed. In most medical circumstances, withholding incidentally discovered information would be unethical (see Chapter 8). This principle, however, must also reflect some sensitivity to the screened individual’s culture, background, and preferences for receiving information (especially bad information). Individuals of different cultural backgrounds evince different preferences for information (Blackhall et al., 1995). For example, traditional Navajo culture takes a worldview that language shapes reality and that revealing negative information (truth telling) may “cause” those bad outcomes to occur. Moreover, an important element of traditional Navajo culture is to think and speak in a positive way and to avoid thinking or speaking in a negative way (Carrese and Thodes, 1995; Gostin, 1995). Seventh, screening programs must use reliable and valid tests and procedures. Of particular importance are sensitivity, specificity, and positive and negative predictive values. Increasingly, policy makers, clinicians, and others concerned with screening issues are called on to deal with odds ratios, likelihood ratios, and receiver operating characteristic (ROC) curves. Eighth, in specific populations, judgments about the feasibility of screening must consider access to health care and health insurance. To summarize, for general screening to be useful on a population basis, certain circumstances must exist (Frame and Carlson, 1975; Eddy, 2004): The condition must be present in the population and have an important effect on the quality and length of life. The incidence of the condition must be sufficient to justify the risks of the screening. Acceptable methods of treatment or prevention in the preclinical or early stage of disease must be available. The condition must have an asymptomatic period during which detection and treatment would substantially improve clinical outcomes, such as reducing morbidity or mortality. A screening protocol or test that is sufficiently accurate and acceptable to patients and that has reasonable costs must be available to detect the condition in the asymptomatic period. The benefits of screening must exceed its harms.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program The benefits of a screening program must justify its costs (including its induced costs) and its use of resources. The health care system must be able to provide referral services that address followup (further evaluation and continuing care) of patients identified by screening. Patients who choose to participate must be fully informed of the risks posed by screening and the benefits of screening and freely consent to be screened. Applications to RESEP The committee believes that in the RECA-RESEP legislation screening aims principally not at diagnosing and treating patients but rather at facilitating compensation for the harms caused by uranium mining and related activities and by the aboveground nuclear tests of the 1950s (discussed in Chapters 2 and 7). Although the legislation speaks to screening for the purpose of improving medical outcomes, early diagnosis of the majority of diseases identified for RESEP screening (such as lung cancer, pancreatic cancer, and ovarian cancer) improves health outcomes little or not at all. Furthermore, the current RESEP screening protocols include tests, such as pulse oximetry and spirometry, that have not been shown to improve health outcomes in asymptomatic populations. Most tests used for screening are also those used to establish a medical diagnosis. When a patient’s signs or symptoms are the reasons, or part of the reason, for using a given test, the procedure is more properly called medical diagnosis than screening. The hallmark of a screening program is identification of preclinical (that is, undiagnosed) disease. Patients or populations that are screened do not have clinical evidence of disease or, if they do, then neither the patients nor their health care providers have yet appreciated that evidence. The target for which the screening is undertaken can be a propensity or risk factor (presumably modifiable) for developing a disease, a disease itself, or a complication of a disease that the patient is known to have. For convenience and readability, we use the term disease to refer to all three entities. In many screening scenarios, the likelihood of the disease that the screening program seeks to identify is relatively low (low pretest probabilities). Moreover, most tests and screening protocols are imperfect in that their specificity is less than 100% (meaning that their ability to rule out disease is not infallible). Because of those two factors, screening can produce substantial numbers of false-positive results; depending on the aftermath of screening when false-positive results occur, the potential benefits of screening may be decreased (or may even not outweigh the risks of error). Because screening tests commonly have sensitivity below 100%, false-negative results will also occur, but these errors may be far less problematic in the screening setting (see discussion later in this chapter). Nonetheless, a negative screening test does not

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program guarantee that the patient does not have the disease, risk factor, or condition being sought. As discussed later in this chapter, a tradeoff often exists between false-positive and false-negative results. Typically, for a given test, altering the test or its interpretation to decrease one of these errors will increase the other. Thus, establishing the criterion for or definition of a positive test result requires well-reasoned balancing of the consequences of each kind of potential error and the frequency of the disease in the population to be screened. CONTEMPORARY SCREENING PROTOCOLS General Issues Many groups have explored traditional medical screening, especially screening for various forms of cancer. Among the authoritative sources are the US Preventive Services Task Force (USPSTF) (Harris et al., 2001; http://www.ahrq.gov/clinic/uspstfix.htm) and (until mid-2004) the Canadian Task Force for Preventive Health Care. Appendix E presents current information on cancer screening tests of interest to this committee; some are for RECA-designated diagnoses and others for nonradiogenic cancers; as a more complete context for considering medical screening in primary care, we also present there a fuller account of the types of screening tests the USPSTF recommends or strongly recommends for general populations other than RECA populations. As seen in Appendix E, relatively few screening tests for cancer have been recommended for widespread use. The limited number or nature of such recommendations turns on several issues. In many instances, no effective treatment for preclinical disease exists or a treatment appears more effective than it truly is. A common issue that arises here is lead time bias, which refers to an apparent improvement in survival if survival is measured from the detection of disease and not from the onset of disease. Some authors argue that such early identification can be beneficial even if overall survival is unchanged, because a patient and his or her family might be able to plan better for the patient’s limited lifetime. A common counterargument is that such early labeling of a patient as having cancer (or another serious ailment) may diminish quality of life during the lead time. The benefit or the loss is surely very patient-specific. Thus, screening should be predicated on extensive counseling and informed consent for both practical and ethical reasons. The effectiveness of a treatment can be overestimated because of length bias, which refers to the fact that disease that develops more slowly (such as a slow-growing or relatively benign cancer) may be more likely to be detected by a screening test, whereas a rapidly progressive disease is more likely to present clinically. These patterns mean that a substantial proportion of disease that is found on screening (as contrasted with disease that presents clinically with signs

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program or symptoms) will be, on average, less aggressive. Sometimes, screening might detect disease that might never become clinically evident. This problem is well recognized when screening for cancers of the prostate, lung, thyroid, and brain; in fact, the term “incidentaloma” is sometimes used to describe the detection of such disease. The patients detected by screening have an excellent prognosis, of course. Such patients inflate survival rates for patients with disease identified by screening but overall unnecessarily consume society’s resources and generate unreasonable anxiety and potential complications for the patient without substantial benefit. Another common reason that cancer-screening programs receive unfavorable recommendations rests on the occurrence of false-positive results (or, said another way, on the imperfect specificity of almost all medical tests). Because clinicians must inform patients about each positive result and because they often recommend additional diagnostic studies, such false-positive results carry substantial risks. They include the risk of physical harm during the additional studies (the workup, which not infrequently can involve invasive diagnostic testing). Other issues include psychologic harm (from anxiety and from labeling, insofar as patients consider themselves as “damaged goods” and societal harm and discrimination (in employability and insurability). The same psychologic and societal harms may affect family members as well. In addition, identifying disease that benefits little from treatment (lung or pancreatic cancer, for example) can lead to futile surgery, which can engender major costs, pain, disability (from reduced lung volume in the case of pulmonary neoplasms), and surgical mortality. Although most screening tests are themselves of low risk and noninvasive, some screening tests expose individuals to ionizing radiation. Such exposure itself can be oncogenic (cancer-causing). Recently, investigators have calculated that the oncogenic risks posed by some radiographic studies outweigh potential benefits from screening (Brenner and Elliston, 2004). The third common reason for unfavorable recommendations about medical screening has rested (openly or covertly) on economic arguments. Our world of limited resources is increasingly constrained with respect to health care and faces many competing demands. Thus, committees that produce clinical practice guidelines, insurers that fund medical care, and even providers that treat individual patients may consider the relatively high cost of some proposed screening programs in light of other uses to which the resources might be applied (the opportunity costs of health care services forgone). Whether these issues are dealt with formally or informally is a matter for the groups in question, but all decision-makers need to consider how best to deal with such tradeoffs in resource-constrained circumstances. The economic cost incurred as a result of false-positive tests is important. One recent article documents that a large proportion of persons screened for cancer (43%, higher among men than among women) had at least one false-

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program positive result, and more than four of five of these patients had followup care (Lafata et al., 2004). Medical costs in the year after the screening were statistically significantly higher among those with false-positive tests than among those with negative tests (on average, about $1,000 higher). The economic costs of a screening program thus extend well beyond the screening test itself. They include the required infrastructure for providing the test and the followup for abnormal results, the counseling that must be undertaken before the screening test is performed and when informing patients about abnormal results, and the choices they then must make. (Note that we use the term patients here for both patient screening and for case-finding because, even in the latter context, once clinicians inform individuals about an abnormal result, they become patients and require a provider-patient relationship.) Economic costs of the screening program must also include the costs of the therapeutic or preventive intervention to be undertaken in patients whose positive results are confirmed. In settings in which the benefit of treating a patient with a disease that has been detected early is small, therapeutic costs may not be economically reasonable for society to undertake. Prevalence and Benefit In assessing the therapeutic benefit of a screening program, one must consider the perspective of the population that is screened, not the subpopulation that is identified as having disease and then treated because of the program. In the population to be screened, disease prevalence can markedly limit benefit. Consider a screening program for a single cancer for which early detection provides a huge therapeutic benefit of 10 years of added life expectancy, as illustrated in Table 9.3. Assume that a screening test has a high sensitivity of 90% and that the prevalence of disease in the population to be screened is 2 cases per 1,000 individuals. Further assume (theoretically) that the screening test has perfect specificity (100%—a positive result is pathognomonic of disease). In a population of 10,000 patients screened, 20 will have the cancer and 18 will be identified and treated early. These 18 patients will gain 180 years overall (10 years of life per patient detected). The 20 patients who have cancer will, on average, gain only 9 years of life (180/20). The 10,000 patients who TABLE 9.3 Effect of Perspective on Benefit of Good Treatment Group Years of Life Saved Number of Individuals Gain per Individual Patients identified 180 18 10 years Individuals with disease 180 20 9 years Screened individuals 180 10,000 0.018 years = 6.6 days

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program FIGURE 9.6 Multiple tests: AND versus OR. up) together require that both tests be positive—a criterion often called the AND criterion—then the joint sensitivity of the testing pair is less than either individual sensitivity, and the joint specificity is higher than either individual specificity. If the tests are statistically independent, then the joint sensitivity will be Sensitivity1 × Sensitivity2, and the joint specificity will be 1 − [1 − Specificity1) × (1 − Specificity2)]. Of course, many tests are not statistically independent of one another; instead, they may depend on the stage of disease. In such cases, the joint specificity of two or more tests interpreted by the AND criterion may not be as high as this expression might predict. As a special case, consider the circumstance in which the second test is a repeat of the first, perhaps even on the same tissue or blood sample. In that case, the two results would be very dependent, and one would not TABLE 9.5 Interpretation of Two Diagnostic Tests Parameter Individual Joint Test 1 Test 2 AND OR Sensitivity 70% 80% 56% 94% Specificity 90% 95% 99.5% 85.5% Positive Predictive Value Disease Prevalence 1% 6.6% 13.8% 52.8% 6.1% Disease Prevalence 20% 58.6% 76.2% 95.7% 56.5%

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program expect the joint specificity to increase at all, save perhaps to account for truly stochastic laboratory variation. In general, if the two tests are correlated, then their joint specificity must be adjusted to account for the degree of positive or negative correlation. For example, if the two sensitivities are 70% and 80% and the two specificities are 90% and 95%, then the joint sensitivity will be 56% and the joint specificity 99.5%. In a screening setting with low prevalence of disease, a higher specificity will probably produce an improved posterior (posttest) probability—that is, a higher positive predictive value. Such improved specificity with the AND criterion of combination might be useful if the prevalence of disease were low or if the burden of false-positive results were high, that is, when excluding the presence of a disease may have a high priority. If those tests were applied in a population in which the prevalence of disease is 1%, the positive predictive value if test 1 alone were positive would be 6.6%, the positive predictive value if test 2 alone were positive would be 13.8%, but the positive predictive value if both were positive would be 53%. In contrast, if the tests were to be applied in a population in which the prevalence of disease is 20%, the positive predictive value if test 1 alone were positive would be 59%, the positive predictive value if test 2 alone were positive would be 76%, but the positive predictive value if both were positive would be 96%. Conversely, if denoting a screening protocol as positive and proceeding to the next step requires only that either test be positive—a criterion often called the OR criterion—then the joint sensitivity of the testing pair is higher than either individual sensitivity and the joint specificity is lower than either individual specificity. If the tests are independent, then the joint sensitivity will be 1 − [1 − Sensitivity1) × (1 − Sensitivity2)], and the joint specificity will be Specificity1 × Specificity2. Using the same example as above, the joint sensitivity will be 94% and the joint specificity will be 85.5%. In a screening setting with low prevalence, a lower specificity will probably produce a lower positive predictive value. Again, in a population in which the prevalence of disease is 1%, the positive predictive value after either test was positive would be only 6.1%. In contrast, in a population in which the prevalence of disease is 20%, the positive predictive value after either test was positive would be 57%. The improved sensitivity of the OR criterion of combination might be useful if the burden of false-positive results were low and if finding a disease had high priority. As seen in Table 9.5, the difference between the AND criterion and the OR criterion is proportionally greater in the low-prevalence setting that is typical of screening. Multiple Diseases and Conditional Probabilities For simplicity, we have been dealing up to this point with a single disease and a cutoff criterion that classifies all test results into one of four categories: true

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program positive, false positive, true negative, and false negative. That allows us to think in terms of the summary measures sensitivity and specificity. Unfortunately, the real world does not always support this simplification. Consider a circumstance with three diagnostic possibilities: disease 1, disease 2, and no disease. Assume that we perform a test to establish the presence of disease 1 but that the likelihood of a positive test is different among patients with disease 2 than among patients with no disease. In such circumstances, the “specificity” of the test cannot be defined, because it depends on the relative likelihood of disease 2 and of no disease among patients who do not have disease 1. These factors make the mathematics a bit more complicated because one needs to think in terms of conditional probabilities (for example, the conditional probability of a positive test given the presence of disease 2). One can, however, derive unexpectedly useful diagnostic information by doing so. A negative test (presumably for disease 1) can actually change the relative likelihoods of disease 2 and no disease (Gorry et al., 1978). Imaging Studies The situation becomes substantially more complex if one considers imaging studies, such as chest x rays, computed tomography scanning (CT scans), magnetic resonance imaging (MRI), or nuclear scanning. For each of those studies, one does not have a single parameter or axis of results, such as an ejection fraction, a PSA concentration, or an oxygen saturation. Rather, the person interpreting the image can identify any of a large number of abnormalities; these might be a pulmonary nodule, a diffuse pattern of fibrosis, a pulmonary infiltrate, or an enlarged cardiac silhouette. For most imaging studies, interpreters do not limit their reports to a single criterion of positivity, such as a pulmonary nodule of at least 5 mm in diameter or a renal mass. Instead, they typically report one or more of a substantial number of incidental findings if for no other reason than to protect themselves from an accusation of missing a cancer, one of the more common reasons for malpractice actions. For adrenal, pituitary, and thyroid images, these tiny lesions have sometimes been called incidentalomas. Thus, estimating the specificity (true negative rate) of an imaging study is difficult. The literature can provide information about specific findings, for instance, how often a solitary pulmonary nodule larger than 5 mm in diameter is present in the absence of cancer. It does not tell us how often a radiograph of a healthy individual is reported as completely normal. In part, these decisions depend on how hard the imager looks. Because many incidental findings will generate further workup and anxiety among individuals screened with imaging techniques, the false-positive cascade and its burden will increase considerably if a screening protocol includes several imaging studies.

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program Repeated Studies Our discussion of screening has implicitly considered a screening test or battery of tests at a single time. Many RECA diseases, however, evolve, either developing a new manifestation or reaching the threshold of detectability. That threshold may vary from test to test for a specific disease (see the discussion above of lead time bias and the inherent biases of using more sensitive screening tests). Thus, we can ask, if an initial screening test for a disease is negative, should the test be repeated in the future? If so, when? The many groups that have recommended only a limited number of screening tests have not uniformly analyzed the complex question of retesting. They often address it informally or in terms of expert opinion. For the initial screening test, the appropriate measure of pretest probability is the prevalence of disease in the population. If a screening test is to be repeated, then the pretest probability of disease is usually the interval incidence of disease, a number almost always lower than the prevalence. The interval incidence of disease refers to individuals known not to have the disease at the beginning of a period (because of a prior negative evaluation at that time) who develop the disease by the end of that period. If the initial screening test is negative, one may well be selecting a subset of the population with a lower-than-average propensity for developing the disease. In such repeat screening scenarios, the pretest likelihood is often far lower for the second and later screens than for the first. Arguably, the evidence for repeating a test must be even stronger than it was for the initial screen. Some experts pose counterarguments, however. On the one hand, for instance, one might say that disease identified on a repeated screen is more likely to be rapidly developing, although whether more aggressive disease is more or less amenable to the benefit of early treatment is not clear. On the other hand, one might argue that disease identified on the repeated screen may be more likely to have existed in an earlier stage and thus be more amenable to early treatment. Answers must lie in scientific evidence and in studies that compare outcomes of various frequencies of repetition of screening. Few such studies have been reported. Conclusions of Other Groups As mentioned above, various advisory bodies have considered screening options for many diseases, and relatively few have been recommended for routine screening. We have presented a fairly extensive listing in Appendix E to convey where the current view of medical screening for primary care; materials there are taken from recommendations of the second or third Guide to Clinical Preventive Services of the USPSTF or, where appropriate, recommendations of the equivalent Canadian task force. We have included all “recommendation and

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program rationale” statements concerning neoplastic diseases, whether they are specifically RECA conditions or not; we have also developed a listing of other “A” and “B” recommendations (respectively, strongly recommend and recommend) from the USPSTF for a variety of other conditions. The intent is to convey the extent of coverage of clinical preventive services as they apply to screening, to underscore the point that very few diseases are at present amenable to screening (in asymptomatic populations) with reliable and valid tests in which the benefits of screening likely outweigh the risks or harms of such screening. We note, for example, that screening for cancers of the cervix, breast, and colon is recommended; screening for cancer of the prostate receives mixed reviews, and extensive discussion with the patient is suggested before it is undertaken. Screening for cancers of the lung, pancreas, ovaries, and thyroid is not recommended, in part because prevalence is low (ovary and thyroid) and in part because the therapeutic benefit is small (lung, pancreas, thyroid). To modify those recommendations in RECA populations, one would need to believe that the prevalence of those diseases that offer some benefit of early detection and treatment is substantially higher than it is in other populations. As explained in other chapters, that proposition probably does not hold. First, we are now many years after exposure, so that the excess relative risk of the cancers has declined. Second, some downwinders had had relatively low exposures to the radiation particles that would be expected to induce most of the cancers. Of course, for many cancers and geographies the data about exposure include substantial uncertainties. Third, as the exposed population ages and comorbidities develop, the medical benefit of early detection in improving life expectancy wanes. Thus, for downwinders and onsite participants, little rationale can be advanced for expanding medical screening recommendations in the RECA populations beyond contemporary recommendations for the general population (as in Appendix E) that are based on an assumption of achieving improved health outcomes through early detection and on an acceptable balance of benefits and harms. For miners, millers, and transporters who were exposed to dust and silica, the same issues arise as in mining for other materials, for example, exposure to silica. In general, silicosis and coal workers’ pneumoconiosis (CWP) are progressive diseases managed by minimizing exposure.3 Few data suggest that early identification and treatment can modify prognosis, save once the fibrotic lung disease is established, save to decrease further exposure to silica (Sharma et al., 1991; Bates et al., 1992; Banks et al., 1993; Banks, 2005). Small trials of systemic or bronchoalveolar lavage with steroids have shown modest statistical but not clinical improvement (Sharma et al., 1991). Thus, early identification of silicosis in 3   Black lung is a common term for coal worker’s pneumoconiosis caused by excessive exposure to coal mine dust. Silicosis is related to the excessive quartz dust exposure (http://www.thoracic.org/news/atsnews/news0197.html#ats9, accessed for this purpose 10/2/2004).

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program asymptomatic individuals who are no longer actively employed as miners would be of limited medical benefit. Silicosis is, however, associated with an increased prevalence of mycobacterial disease. Performing a purified protein derivative (PPD) or tuberculin test to screen former miners, millers, and ore transporters for tuberculosis is a rational approach. Such skin testing need not be predicated on demonstrating silicosis on a chest x ray. If the skin test were positive, following up with a chest x ray, and perhaps with other appropriate tests to confirm the presence of active disease (if the x ray were suspicious) would be the current standard of care. Offering prophylactic antituberculosis therapy to patients with positive skin reactions who do not have active disease might be valuable as well. The demonstration of silicosis on chest x ray might improve compliance with early treatment for latent mycobacterial infections. Some data suggest that lung cancer to have a somewhat increased frequency in patients with silicosis, but the data are inconsistent; moreover, early detection does not appear to confer much, if any benefit. Several groups, including the National Institutes of Health and the National Kidney Foundation, suggest screening for early evidence of chronic renal disease in populations at risk for renal failure, such as patients with diabetes and hypertension. Recently, those recommendations have been extended to include a far broader population (Levey et al., 2003; National Kidney Foundation, 2002). If chronic renal dysfunction is identified early, various interventions such as angiotensin-converting enzyme [ACE] inhibitor drugs, diet modification, and lipid-lowering therapies, can slow the progression of renal disease with substantial benefit. Recent data suggest that the “MDRD” glomerular-filtration-rate (GFR) calculator, developed as a part of the modification of diet in renal disease study (Levey et al., 2003; http://www.nephron.com/mdrd/default.html, accessed 3/13/ 2005), is probably the most accurate means of identifying such patients. Because of the putative relationship between exposure to soluble uranium salts and renal failure (perhaps more acute than chronic) and because chronic renal disease is a RECA-covered disease for some uranium workers, screening millers and ore transporters for renal dysfunction is reasonable. The committee finds insufficient supporting evidence to support additional medical screening or medical case-finding in the RECA populations beyond the level and type of screening advised for general populations and persons with occupational exposures similar to those of in miners in general. Specifically the committee recommends that the Health Resources and Services Administration base RESEP medical screening efforts in asymptomatic individuals on robust scientific evidence that such screening improves health outcomes and that its benefits outweigh its risks. The committee believes that the current legislation and regulations that promulgate screening for the RECA-compensable diseases for the purpose of preventing disease and improving health arise because the distinction between medical screening and compensational screening, as explicated in this chapter, has not

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program been clear. With that distinction in mind, the committee recommends that HRSA not extend its medical screening beyond the generally accepted screening protocols that apply to the US population at large. However, the committee further recommends that uranium miners, millers and ore transporters also be screened for diseases generally recommended for screening in other mining populations and that uranium millers and ore transporters be screened for chronic renal disease. In particular, the committee observes that some screening protocols that HRSA grantees now use exceed these guidelines. In our judgment, HRSA should cease funding any medical screening efforts that do not conform to the recommendation. Correspondingly, the committee recommends that, once an individual has been shown to be administratively eligible for compensation under RECA (including employment, residence, or a calculated PC/AS at or above some established cutoff criterion), the individual be offered medical screening recommended in generally accepted protocols that apply to the population at large (see Appendix E-2). FUTURE RESEARCH The committee has referred earlier to authoritative screening recommendations from the USPSTF and other groups (see Appendix E). The dozens of systematic reviews prepared for the USPSTF and the related USPSTF statements of recommendations and rationales typically have extensive comments on future research issues, including the reviews of cancer screening technologies. These can be found on the USPSTF website (www.ahrq.gov/clinic/prevenix.htm) and in the peer-reviewed literature. We briefly raised the question of the costs of screening in general and of the RESEP program in particular. Not much is known about the cost-effectiveness of screening, in part because of the unproven effectiveness of many screening modalities (such that examining cost-effectiveness has little meaning). Very little information is available on the cost-effectiveness of alternative screening strategies; the chief exceptions are screening for colorectal cancer and type 2 diabetes. Insofar as HRSA and RESEP engage in appropriate medical screening in the future, data from strong cost-effectiveness analyses would help HRSA to specify RESEP screening strategies that would optimize use of federal resources. We note in particular the need for more head-to-head comparisons of screening technologies and for collection of cost data for cost-effectiveness analyses. The committee encourages Congress to expand financial support for rigorous studies of diagnostic and screening tests, including cost-effectiveness analyses. As documented elsewhere in this report, the six current HRSA grantees operate screening programs with quite different orientations. Some do screening similar to a broad-based annual physical examination, with attention to both

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program RECA-compensable disorders and to non-RECA ailments typical of the age and sex of patients that they see; their orientation might thus be characterized as more medical than compensational. Others (notably the Indian Health Service RESEP program in Shiprock, New Mexico) apparently direct their focus nearly exclusively to RECA conditions and do not expand their screening efforts beyond that; their orientation might be characterized as more compensational than medical. The committee believes that such variation poses problems of inequity that HRSA should investigate in more detail; one aim of such a study is to determine whether some individuals are receiving more services than others for equivalent expenditures of public dollars in RESEP. On the other hand, the best practices among current grantees are not clear. It would not be unreasonable for HRSA to undertake a trial to establish which protocols are most effective. Of course, participation in such a trial should also be predicated on obtaining informed consent for the trial from each participant. The RECA populations being screened by HRSA grantees with RESEP funding are an interesting group to study insofar as they comprise individuals with presumably greater than average risks of disease. The sensitivity, specificity, and positive and negative predictive values (or reliability and validity) of a variety of screening tests may well differ in those populations from the values in populations of either average or low risk. Epidemiologic and clinical studies might well be undertaken to clarify the yield from various screening tests; such work would shed light on the utility of the individual tests and combinations of tests under those circumstances. HEALTH-CARE ISSUES BEYOND SCREENING Evolution of Technology Apart from the issues we have raised above about medical and compensation screening, we note that some testing, particularly diagnostic testing, must and will continue. For uranium miners, for example, the increased risk of lung cancer and restrictive lung disease can be addressed through traditional services, such as pulmonary function tests and chest x-rays. Those methods have been applied to studies of underground miners (for example, of coal workers’ pneumoconiosis in the United States and uranium miners in many countries) and can be said to be appropriate current practice as carried out by the current HRSA grantees. However, as new imaging and sputum-cytology methods emerge, the HRSA grantees should be encouraged to consider new and better tests (especially those with test properties better than those of traditional screening or diagnostic tests) and abandon tests that appear to be less effective. Such evolution should, however, be based on evidence. The committee would generalize that point by saying that both HRSA and the medical professions more broadly must be sensitive to

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program the evolution of useful technologies, adopting those of improved performance and discontinuing use of those approaching obsolescence. Screening Recommendations from Other Groups As reflected in this chapter and Appendix E-1, the committee used the USPSTF guidelines to help develop recommendations on medical or compensational screening of RECA populations. As the clinical and research communities develop new or improved screening methods and put them into practice, the USPSTF and other professional or clinical bodies will review such guidelines and, as appropriate, revise them. The Institute of Medicine instituted a modern-day understanding of clinical practice guidelines as the behest of the then Agency for Health Care Policy and Research (IOM, 1990a; IOM, 1992; Lohr, 1998, 1999; Lohr et al., 1998). This evolution in clinical-practice guidelines since that time for both primary and specialty or referral care, may have implications for both RECA populations and the RESEP program. Updates to clinical practice guidelines of all sorts can be found through the National Guidelines Clearinghouse supported by the Agency for Healthcare Research and Quality (AHRQ) (www.guideline.gov) and, specifically for the USPSTF, through the AHRQ Preventive Services Web site (www.ahrq.gov/clinic/prevenix.htm) and the Put Prevention into Practice program (www.ahrq.gov/clinic/ppipix.htm). The committee recommends that HRSA regularly monitor and follow screening guidelines developed by the US Preventive Services Task Force and published by the Agency for Healthcare Research and Quality Psychologic Issues in Exposed Populations Of concern to the committee are psychologic disorders potentially present among downwinders who at various times have not been fully informed about the risks, although small, of exposure to fallout from the NTS tests. Major depression is a particular concern because of the already high prevalence of depression in the general population, its serious potential consequences (such as suicide), and its responsiveness to treatment if it is identified. Generalized anxiety disorder, posttraumatic stress disorder, and a more chronic concern among these populations about their view of living in a “contaminated environment” (Kuletz, 1998, p336), has been called “chronic environmental stress disorder” (IOM, 1999), and it may be associated with NTS activity and its aftermath over the past 50 years. In the end, the committee could not amass sufficient scientific data or develop a plausible chain of reasoning to include these conditions as compensable under RECA. In the setting of a nuclear accident or other catastrophic event, rapidly pro-

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program viding the public with accurate information can minimize the mental health effects (IOM, 1999, p106). For a variety of reasons, that was not always the case for downwind RECA populations. Some people who testified at the committee’s information-gathering meetings described experiences and impressions that may place them at risk for some of these disorders. We heard a good deal about chronic emotional pain and sorrow related to the cancer deaths of many family members and friends. We also heard many of those who testified express anger and frustration at the government for (supposedly) misleading them and their predecessors about the dangers of uranium mining, the effects of fallout, and the meaning of the “dust” found on many crops, homes, and other locations. They presumed that this dust was a consequence of the nuclear-weapons tests at the NTS. Finally, many spoke of their strong sense of “not being heard” by either HRSA or DOJ. Although downwinders were not exposed to a catastrophic threat (such as a nuclear accident), the paucity of accurate information provided to them for many years may have contributed to their psychologic burden. However, even here the committee did not find any scientific studies to validate the contributory effect of this paucity of information. In addition, as noted in Chapter 7, the committee has been unable to identify any data that evaluate the psychologic effects of chronic environmental stressors during either the testing and mining periods or, more recently, among RECA downwinders or other populations with similar radiation-exposure experiences. Hence, the committee does not believe that compensation, as outlined in Chapter 8, is appropriate for downwinders with psychologic disorders. Nevertheless, the possibility exists that depression and generalized anxiety disorder are more prevalent in RECA populations than in the general population and might have been ameliorated if more complete information had been provided to these populations in the past. Of course, such information must be provided in a culturally sensitive an appropriate manner. The committee notes three important factors: (1) rapid, simple, low-cost screening tests for depression are available and have been validated (Pignone et al., 2002); (2) depression can have serious consequences that can be ameliorated by timely treatment; and (3) the USPSTF recommends screening for depression in the general adult population within the context of the delivery of health care services by institutions or providers with appropriate quality assessment programs in place (http://www.ahrq.gov/clinic/3rduspstf/depressrr.htm). In sum, screening for depression constitutes good medical practice by providers with good quality-of-care procedures; the link between quality of care and ethics (Lohr, 1995) strengthens this conclusion and is implicit in reports from the Institute of Medicine spanning more than a decade (IOM, 1990b; IOM, 2001). Given the above, HRSA may want to consider screening for depression in their

OCR for page 219
Assessment of the Scientific Information for the Radiation Exposure Screening and Education Program medical screening protocols. The low-cost, potentially high-benefit strategy of screening for depression, with appropriate referrals and follow through, may ameliorate some psychologic burdens in the RECA population.4,5 CONCLUSION This chapter has reviewed in depth the conceptual and statistical underpinnings of screening for medical purposes. We have discussed the limitations and risks of screening for diseases for which early detection offers little benefit. We have also shown how screening in low-risk populations and screening for multiple diseases will produce a substantial burden of false-positive results. Finally, we have offered several recommendations to reduce potential harms and mistakes from the use of multiple screening tests among individuals for whom the medical benefits cannot outweigh the harms and the likelihood of compensation may be low. Some recommendations are directed at agents that would need to amend RECA (or RESEP) legislation; others are directed at steps that HRSA can consider without statutory changes. The next chapter picks up on the new issue in RESEP: screening for compensable disease. 4   For the reasons provided in detail in this chapter concerning the potential risks of depression developing in exposed populations (about which we heard testimony from Drs. Robert Ursano and Evelyn Bromet), the seriousness of the depression in terms of morbidity and mortality from suicide, and the treatability of depression, committee members Stephen G. Pauker and Catherine Borbas find it inconsistent with good medical or public health practice merely to state that “HRSA may want to consider screening for depression” in RECA populations. It is their opinion that such screening should be “recommended,” as it is for adults in the general population (see USPSTF). Further, it is noteworthy that few, if any, current HRSA grantees have listed screening for depression in their current protocols, emphasizing the need for this specific recommendation. 5   Committee member Kathleen N. Lohr wishes to support a recommendation that HRSA expand its screening activities to include mental and emotional disorders (particularly major depression, generalized anxiety disorder, and post-traumatic stress disorder), following through as needed with appropriate referrals to medical and psychiatric care relevant to the diagnoses in question. The report still contains much evidence, both from the published literature and more anecdotally from the numerous presentations at the committee’s information-gathering meetings, that major and/or minor depression may well have a prevalence in these RECA populations higher than that for the general adult population. Screening for depression in adults, with the qualifications noted in the report as to the medical infrastructure needed for high-quality care, is a formal recommendation of the US Preventive Services Task Force. Numerous easy-to-use screening methods exist, and the practice of such screening is spreading (Santora and Carey, 2005).