3
Screening for Skin Cancer

Today, most Americans are probably familiar with advice to limit sun exposure—actually, exposure to sunlight and other sources of ultraviolet radiation such as sunlamps—to reduce their risk of skin cancer. They are also likely to have heard or seen messages sponsored by the American Cancer Society (ACS), the American Academy of Dermatology (AAD), or other groups explaining how to check their skin for warning signs of skin cancer, especially melanoma.1

From a public health perspective, the advice to limit sun exposure—especially during the first two decades of life—is a form of primary prevention, which includes counseling and educational interventions that aim to keep people from developing health problems in the first place. Another primary prevention strategy that has been widely advised, sunscreen use, was recently reported to have helped prevent one type of skin cancer in a controlled clinical trial (Green et al., 1999).

The advice about skin self-examination is a form of secondary prevention, which promotes early identification of risk factors or subclinical disease in people who have not developed symptoms. Skin self-examination for suspicious moles or other skin features is a form of secondary prevention. Clinical screening—examination by a physician or other trained individual of the skin of an

1  

For melanoma, warning symptoms include new or changing pigmented spots that are asymmetrical, larger than 6 mm, irregular in border, or varied in color including blue, black, or gray. For nonmelanoma skin cancer, warning symptoms include the appearance of a new spot, bump, or sore that does not heal or go away in about two weeks, and may or may not bleed or crust. See AAD (1994b) and ACS (1999b).



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 38
Extending Medicare Coverage for Preventive and Other Services 3 Screening for Skin Cancer Today, most Americans are probably familiar with advice to limit sun exposure—actually, exposure to sunlight and other sources of ultraviolet radiation such as sunlamps—to reduce their risk of skin cancer. They are also likely to have heard or seen messages sponsored by the American Cancer Society (ACS), the American Academy of Dermatology (AAD), or other groups explaining how to check their skin for warning signs of skin cancer, especially melanoma.1 From a public health perspective, the advice to limit sun exposure—especially during the first two decades of life—is a form of primary prevention, which includes counseling and educational interventions that aim to keep people from developing health problems in the first place. Another primary prevention strategy that has been widely advised, sunscreen use, was recently reported to have helped prevent one type of skin cancer in a controlled clinical trial (Green et al., 1999). The advice about skin self-examination is a form of secondary prevention, which promotes early identification of risk factors or subclinical disease in people who have not developed symptoms. Skin self-examination for suspicious moles or other skin features is a form of secondary prevention. Clinical screening—examination by a physician or other trained individual of the skin of an 1   For melanoma, warning symptoms include new or changing pigmented spots that are asymmetrical, larger than 6 mm, irregular in border, or varied in color including blue, black, or gray. For nonmelanoma skin cancer, warning symptoms include the appearance of a new spot, bump, or sore that does not heal or go away in about two weeks, and may or may not bleed or crust. See AAD (1994b) and ACS (1999b).

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services asymptomatic person—is also a form of secondary prevention, and it is the focus of this chapter. Because primary prevention of skin cancer emphasizes actions to be taken by children and young adults, secondary prevention is the main issue for those over 65. For people who already have a medical problem, usual clinical management may include measures to prevent additional problems or complications. These measures, sometimes described as tertiary prevention, include such steps as identification and elimination of oral infections before organ transplants and treatment with immunosuppressive drugs afterwards. Medicare coverage for diabetes outpatient self-management training and supplies, which was approved by Congress in 1997 as a preventive service, is another example of tertiary prevention. The primary, secondary, and tertiary labels for preventive services are not rigidly applied. For example, much tertiary prevention is viewed as treatment and thus not subject to Medicare’s preventive services exclusion. The exclusion for outpatient drugs would, however, still apply to most of the medications used for tertiary prevention. The premise underlying both self-examination and clinical screening programs is that detecting a disease earlier than would happen in usual health care will result in earlier treatment that saves lives and reduces the physical and emotional burden of illness. In addition, screening is often promoted as a way of reducing the overall costs associated with treating disease, especially late-stage disease. Nonetheless, when claims about the benefits of particular screening programs are subjected to systematic evaluation, the evidence supports some but is negative, mixed, limited, or otherwise inadequate to support others (see, e.g., Eddy, 1991; Russell, 1994; USPSTF, 1996).2 As controversies over assessments of breast cancer screening for women ages 40 to 49 demonstrate, conclusions that the evidence does not clearly support screening for a particular disease can generate considerable controversy, given the understandable hopes that screening will prevent or reduce the mortality, disability, and other suffering caused by the disease (see, e.g., Eddy, 1997; Ransohoff and Harris, 1997; Taubes, 1997a,b). Medicare does not cover screening for skin cancer in asymptomatic people. It does, however, cover a physician visit initiated by a concerned patient who has noticed, for example, a change in the color of a mole (clinically described as a pigmented nevus or, more generally, skin lesion), or a new skin growth. Simi- 2   Claims about other prevention strategies become similarly complex when rigorously investigated. Proponents of evidence-based medicine will note that the results of the controlled trial of sunscreen use that was mentioned on the previous page supported—for squamous cell carcinoma—a prevention strategy that has long been promoted on the basis of biological plausibility without direct evidence (Hill, 1999). For basal cell carcinoma, the trial did not find a significant effect of sunscreen use. The trial also tested the use of beta-carotene supplements to prevent skin cancer. This strategy received some attention based on animal studies but was not confirmed in this trial.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services larly, if a physician notices such a suspicious sign during a visit for another purpose and extends the visit to investigate further, Medicare may pay more for the visit if it meets certain criteria for a higher level “evaluation and management service” (which is Medicare payment terminology for a physician visit). In either situation, if the patient is referred to a dermatologist for further assessment, that referral visit is also covered. Appendix B, which was prepared by researchers at the Oregon Health Sciences University Evidence-Based Practice Center presents the review of the scientific literature on skin cancer screening that was commissioned for this study. Consistent with provisions of the legislation authorizing this study, the review was developed in conjunction with both the IOM committee and the U.S. Preventive Services Task Force (USPSTF). It also follows the general strategy used by an earlier IOM committee evaluating thyroid cancer screening in people exposed to radioactive iodine from atomic weapons testing (IOM/NRC, 1999).3 The rest of this chapter discusses the committee’s analytic approach; the burden of illness associated with different forms of skin cancer; the procedures used for screening and evidence about their effectiveness; estimated five-year costs to Medicare of three alternative screening approaches; statements on skin cancer screening from other groups; and the committee findings and conclusions. ASSESSMENT APPROACH: INTERVENTION, POPULATION, AND OUTCOMES Following the general approach set forth in Chapter 2, the committee began by defining the specific procedures or activities that constitute skin cancer screening. As described in Appendix B, skin cancer screening may rely on a case-finding strategy, when a person seeing a health care professional for another reason is offered a total skin examination (or a partial skin examination with referral to a specialist depending on the findings). Such screening may focus on all people or only those identified as high-risk. Another strategy involves mass screening in which people are invited and self-select to undergo a total skin examination by a health care professional and then are referred to their primary care physician or a specialist for follow-up. The committee considered two case-finding approaches to screening for skin cancer in addition to the mass screening approach just described. The first case-finding approach involves a visual examination of the entire skin (including scalp and nails as well as the mouth) and a patient history to identify risk factors such as family history of skin cancer, level and frequency of sun exposure, or recent change in a mole. The other case-finding strategy separates the process 3   Reflecting differences in IOM and Task Force objectives and presentation needs, the Appendix in this report will differ slightly from the Task Force document expected early in 2000.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services into two phases, so that a total skin examination is offered only for those identified by their history as being at high risk of skin cancer. The committee also considered a mass screening approach as a third approach. In all approaches, those with lesions identified as suspicious for skin cancer are offered biopsies. Because skin cancer screening searches for disease in people who have not noticed symptoms of the disease, the committee defined it as excluding several services already covered by Medicare. These covered services include a skin examination and history undertaken by a physician (a) in response to a patient’s concern about a skin abnormality, including both new and changing skin features, (b) after the incidental discovery of a suspicious skin lesion during an examination for another purpose, or (c) during a subsequent referral visit to a specialist. Also excluded from the definition of screening are Medicare-covered follow-up visits and skin examinations for patients previously diagnosed and treated for skin cancer or other conditions that put them at higher risk for skin cancer. In contrast to a new program of clinical skin cancer screening, all these services are considered part of usual patient care. In this analysis the population of primary interest is Medicare beneficiaries age 65 and over. Nonetheless, evidence related to all age groups was reviewed.4 The major health benefits sought from skin cancer screening are reduced mortality and associated suffering, lower rates of recurrence and subsequent treatment, and less treatment-related discomfort or disfigurement. Another potential benefit of screening is improved patient knowledge, self-examination skills, and risk reduction behaviors. Harms from skin cancer screening are also possible, for example, when false positive or inconclusive results lead to unnecessary treatment, pain, scarring, and anxiety, although any pain or scarring is likely to be modest for such false positives. In addition, false negatives may induce complacency and disregard for subsequent skin changes possibly indicative of cancer. Most people have some moles, differently pigmented areas of skin, and other skin features, and the number of skin changes and growths tends to increase with age. Although most of these skin features will not be or become malignant, some may prompt suspicion and further testing.5 For younger people, possible harms of 4   The benefit would also, presumably, apply to Medicare beneficiaries who are under age 65, for instance those who qualify as disabled. The younger group of beneficiaries is a relatively small part of Medicare, and skin cancer is associated with age, so this discussion emphasizes older beneficiaries. 5   For example, actinic keratoses (scaly or crusty patches that may be flat or elevated and are associated with cumulative sun exposure) are relatively common and do develop into squamous cell carcinoma in a small percentage of cases. Because they can cause itching and other discomfort, some patients would likely seek evaluation and treatment independent of a screening program. In addition, some actinic keratoses would—like cancerous lesions—be identified incidentally during the course of a physician visit for another purpose.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services FIGURE 3–1 Evidence pyramid for assessing a screening intervention. SOURCE: Adapted from IOM/NRC 1999, p. 89. screening and false positives include labeling and denial of health coverage by insurance companies, employers, or others. The committee viewed these possible harms of skin cancer screening as relatively minor. To guide this assessment of a screening intervention, the committee used Figure 3–1, which is adapted from the simpler, three-tier figure introduced in Chapter 2. Thus, the committee sought evidence that skin cancer is important in terms of prevalence, incidence, and mortality or morbidity (disease burden) for the population 65 and over; amenable to effective treatment; detectable by a screening test that is reliable, accurate, and safe and that detects the disease at an earlier stage than in usual care; and more effectively treated when detected at an earlier stage with overall benefits of treatment outweighing any harms. Appendix B describes the evidence search strategy more specifically. A test that met all the criteria in Figure 3–1 would clearly have benefits compared to usual care, but the extent of the benefit relative to the cost to Medicare, a health plan, or society generally would still need to be considered. As explained earlier, this report provides only estimates of costs to Medicare and not others, and it does not include formal assessment of the cost-effectiveness of skin cancer screening compared to other interventions (e.g., promotion of skin self-examination). In addition, this chapter covers only a subset of feasibility issues that might arise in implementing a new program of skin cancer screening. Although not considered in depth here, certain ethical considerations relevant to screening recommendations should also be noted (see, e.g., Malm, 1999). Screening—even when targeted to higher-risk groups—typically involves examination or testing of a large number of healthy people who will turn out not to have the disease or who may have subclinical disease that will never progress to cause problems. For this reason, those making recommendations about screening interventions may want the margin of benefits over harms to be

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services greater or clearer than they would for interventions intended to cure or improve the well-being of those who are ill. The general enthusiasm for screening may lead to screening promotions that do not adequately inform the public and the professions about the potential for harm (e.g., see Woloshin and Schwartz, 1999b). POPULATION BURDEN OF DISEASE The skin cancers considered here are melanoma, basal cell carcinoma, and squamous cell carcinoma.6 The latter two are often lumped together as nonmelanoma skin cancers, which also include other, far less common skin malignancies such Kaposi’s sarcoma and cutaneous T-cell lymphoma. These uncommon conditions are not the focus of general skin cancer screening programs. Melanomas originate in cells that can produce melanin, a pigment found in the skin, hair, eyes, and sometimes elsewhere. Basal cell carcinoma and squamous cell carcinoma originate in the epidermis, the outermost layer of the skin. Melanoma Melanoma is the least common but most deadly of the three skin cancers considered here. It accounts for less than 5 percent of reported cases but about 80 percent of skin cancer deaths (ACS, 1999a). Although melanoma occurs most commonly in the skin, it may be found in the eye and, rarely, elsewhere. It may arise either de novo or from an existing mole and can spread through the lymph system or blood. Risk factors for melanoma include light skin color, older age, large numbers of common moles, atypical moles, history of severe sunburns (especially at younger ages), and family history of the disease. Unlike basal cell and squamous cell carcinomas, a majority of melanomas occur on skin surfaces not normally exposed to the sun.7 Melanoma of the skin is primarily a disease of fair-skinned people. It is also more common in white men than in white women with an incidence in 1996 of 19.3 per 100,000 for the former and 13.2 per 100,000 for the latter (SEER, 1999).8 (The corresponding rates for black men and women are 1.3 per 100,000 and 0.6 per 100,000.) The disease is also much more common in older age 6   This discussion draws on AAD, 1994a; ACS, 1999b,c; Albert et al., 1990; NCI, 1999b–e; USPSTF, 1996; Whited and Grichnik, 1998). 7   There are four types of melanoma: superficial spreading melanoma (the most common form of melanoma), acral lentiginous melanoma, nodular melanoma and lentigo maligna. Lentigo maligna melanoma is most often found among the elderly, and usually on areas with high cumulative sun exposure. 8   Unless otherwise indicated, data come from the Surveillence, Epidemiology and End Results (SEER) program of the National Cancer Institute.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services groups; the 1992–1996 age-adjusted incidence rate for those age 65 and over was 72.3 per 100,000 for white men and 31.8 per 100,000 for white women. In 1999, approximately 44,000 cases of melanoma are expected to be diagnosed (ACS, 1999a). Based on 1994–1996 data, the lifetime risk of being diagnosed with invasive melanoma is 1.84 percent in white men and 1.34 percent in white women, which ranks it 6th among common cancers for white men and 11th for white women. The age-adjusted incidence of the disease has been increasing, from 5.7 per 100,000 population in the 1973 to 13.8 per 100,000 in 1996 with higher rates of increase occurring during the earlier part of this period (SEER, 1999). As noted in Appendix B, changes in record-keeping procedures, diagnostic criteria, and preventive practices may be contributing to increased detection and reporting. Some suggest that more melanomas of a relatively nonaggressive and clinically unimportant kind are being detected, but no studies have established this (Burton and Armstrong, 1995; Swerlick and Chen, 1996, 1997). The age-adjusted mortality from melanoma has, however, also been increasing, rising from 1.6 per 100,000 in 1973 to 2.3 per 100,000 in 1996. This trend suggests that the increasing incidence of the disease is not just an artifact of increased detection or changing criteria for diagnosis. Again, older white men are most at risk. For 1994–1996, their age-adjusted mortality rate was 17.4 per 100,000 compared to 7.5 for older white women. Most of the increase in overall mortality since 1973 has been among white men. White men have a 0.38 percent lifetime risk of dying from melanoma, which ranks it 13th among major kinds of cancers; for white women the comparable risk of dying is 0.28 and 18th in rank. In comparison, the lifetime risks for white men of dying from lung or prostate cancer are 6.94 percent and 3.09 percent, respectively; for white women, the lifetime risks of dying from lung cancer or invasive breast cancer are 4.77 percent and 3.47 percent, respectively. In 1999, about 7,300 deaths are expected from melanoma (ACS, 1999a). When melanomas of the skin are diagnosed in their early stages while thin and localized (i.e., no spread beyond the development site), prospects for long-term survival are very good. Data from the Surveillance, Epidemiology and End Results (SEER) program of the National Cancer Institute show a five-year relative survival rate (1989–1995) for localized disease of over 95 percent; for regional disease that involves nearby lymph nodes, the rate drops to 58 percent; and for distant metastatic disease, it is only 13 percent. SEER data also show that more than 80 percent of melanomas are diagnosed while still local. Nonmelanoma Skin Cancers Basal cell carcinoma and squamous cell carcinoma are the most common skin cancers, accounting for about one million new cases a year and about 1,900 deaths (ACS, 1999a). Risk factors include older age, sun exposure (primarily

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services childhood exposure for basal cell carcinoma and both early and recent exposure for squamous cell carcinoma), and fair hair, eyes, or skin. For squamous cell carcinoma, immunosuppression and cigarette smoking are an additional risk factors. European and U.S. data suggest that people who have been diagnosed with these cancers are at higher risk of subsequent skin cancers (including melanoma) and other noncutaneous cancers (Kahn et al., 1998; Karagas, 1994; Karagas et al., 1992; 1998). Because of their relatively low lethality, neither cancer is included in the major reports issued by the SEER program. Basal cell carcinoma accounts for about three-quarters of all skin cancers. This cancer is slow growing and usually does not spread to other parts of the body. It is highly curable by surgical removal but may, if neglected, cause death, functional impairment, or disfigurement. Because basal carcinomas often appear on the face, scarring or coloration changes associated with even minor surgical treatment may cause distress. Squamous cell carcinoma occurs primarily in the skin but may also occur elsewhere, including the mouth and genitals. Another skin condition, actinic keratosis, is a concern because a small proportion may develop into squamous cell carcinomas (Mittelbronn et al., 1998; NCI, 1999e; Schwartz, 1997). Like basal cell carcinoma, squamous cell carcinoma usually is slow growing, but it is more likely to be lethal. It is usually curable if detected early but may cause death, functional impairment, or severe disfigurement if neglected. Squamous cell carcinoma of the skin accounts for about one-fifth of all skin cancers and most of the deaths from nonmelanoma skin cancer (ACS, 1999a). AVAILABILITY OF EFFECTIVE TREATMENT Melanoma Surgical treatment is widely accepted as effective in achieving long-term survival for patients with localized melanoma (Holmstrom, 1992; Kelly et al., 1984; NCI, 1999b,c; NIH, 1992). Recent research has focused on identifying the effect on survival of more limited surgical excisions that remove smaller margins of skin surrounding the tumor. Results suggest that smaller margins are acceptable for early-stage disease and reduce the need for skin grafts (Karakousis et al., 1996). Other research has examined the effectiveness of routine lymph node resection when there is no clinical indication of spread to the lymph nodes, but trials to date have not established its value for most patients (Hochwald and Coit, 1998). A more focused approach for identifying and treating people with subclinical spread of the disease, the sentinel lymph node biopsy (which targets the lymph nodes that first drain a primary tumor site), is being tested (Gershenwald et al., 1998; Glass et al., 1998).

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services The thinner the melanoma, the higher the survival rate (Halpern and Schuchter, 1997; Sabin et al., 1997; Straume and Akslen, 1996). For patients with thicker melanomas or melanomas that have spread, effective treatments are still being sought. For example, the NCI recommends that physicians discuss with such patients the possible benefits and harms of enrolling in clinical trials that are testing various interventions such as chemotherapy, biological therapy, or therapy with immunologically active agents such as interferons (NCI, 1999b). If cancer has spread to regional lymph nodes, surgery to remove the tumor and affected lymph nodes can be successful, but survival rates are lower than for less advanced disease (Karakousis et al., 1998). For melanoma that has spread beyond the skin and lymph nodes, treatment is primarily palliative, aimed at relieving pain and other symptoms rather than at improving long-term survival. Nonmelanoma Skin Cancers Most nonmelanoma skin cancers can be successfully treated with one of several kinds of surgical procedures including surgical excision, cryosurgery (which uses liquid nitrogen to kill cancer cells), and electrodessication and curettage (drying the lesion with electric current and then scraping the debris away) (NCI, 1999d; Preston and Stern, 1992). These procedures are also commonly used to remove premalignant actinic keratoses. Radiation therapy, chemotherapy, and other nonsurgical approaches are also used in certain situations. The choice of procedure depends on the type of tumor, its location, and patient history (e.g., past nonmelanoma skin cancers). For small excisions, surrounding skin may be stretched over the wound and stitched; large excisions may require grafts of skin from elsewhere on the body. Microsurgical techniques have been developed to check tissue as it is removed to minimize the removal of healthy tissue and subsequent scarring. In general, the smaller the lesion is, the easier the procedure (Thomas and Amonette, 1988). An important benefit of early detection and treatment of these cancers is reduced scarring and a better cosmetic appearance. SCREENING AND DIAGNOSTIC PROCEDURES Screening As described earlier, the committee reviewed evidence for a mass screening strategy and two case-finding approaches to screening. One case finding approach used a history and total skin examination for all patients, and the second used the patient history to identify high-risk people to undergo total skin examination. This examination takes approximately three to five minutes for an

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services experienced clinician if nothing suspicious is detected that requires further investigation.9 In examining the skin for possible melanoma, American clinicians are advised to remember the “ABCDs” (and sometimes E) of pigmented lesions suggestive of melanoma: Asymmetry (a lesion not regularly round or oval), Border irregularity (poorly defined edges, scalloping, notching), Color variegation, and Diameter greater than 6 millimeters (Friedman, et al., 1985; Harris et al., 1999). The fifth item sometimes suggested is Elevation of the lesion (Thomas et al., 1998). Public awareness and education programs promoting skin self-examination describe the same or similar warning signs (AAD, 1994b), and about half of all melanomas are initially found by patients (Koh et al., 1992). Physicians, however, are more likely to detect thin, early-stage melanomas (Epstein et al., 1999). If nothing suspicious is found during the skin examination, the result of the screening examination is described as negative. If a clinician identifies a mole or other lesion suspicious for melanoma or nonmelanoma skin cancer, the examination is considered positive. The next step—and the “gold standard” for identification of any skin cancer—is a skin biopsy, usually a simple and easily tolerated outpatient procedure. Diagnosis If examination of the skin detects something suspicious for melanoma or nonmelanoma skin cancer, a diagnostic biopsy and microscopic examination of the removed tissue is used to confirm or rule out the presence of skin cancer. If melanoma or other cancerous tumor is identified, the tissue analysis also identifies characteristics useful for “staging” the cancer (judging how advanced it is) and assessing a patient’s prognosis. An excisional biopsy removes the suspicious lesion with a narrow margin of normal-appearing skin as well as a portion of underlying subcutaneous fat (Arndt et al., 1995; Geisse, 1994). Such an excisional biopsy may also serve as effective treatment for early-stage lesions. For large lesions, an incisional biopsy may be used to remove enough tissue for a diagnosis to guide subsequent treatment decisions. Either kind of biopsy usually requires local anesthetic; scarring is likely but the extent will depend on the type, size, and location of the biopsy. As part of the diagnosis, cancer staging systems are intended to classify characteristics of the disease related to prognosis. The most common staging system for melanoma uses information about the tumor’s size; the extent to which it has invaded the skin or nearby tissue; involvement of the lymph nodes; and spread or metastasis to more distant sites (NCI, 1999b). Appendix B pro- 9   Dermatologists may also use in vivo epiluminescence microscopy, a noninvasive test that is still being assessed for accuracy (Kittler et al., 1999; USPSTF, 1996).

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services vides more information on diagnosis and staging of melanoma and nonmelanoma skin cancers. Accuracy of Screening Tests The literature review in Appendix B identified four recent studies of screening by total skin examination for both melanoma and nonmelanoma skin cancers (De Rooij et al., 1995, 1997; Jonna et al., 1998; Limpert, 1995; Rampen et al., 1995).10 As summarized in Table B-1, these studies found suspicious lesions in from 4 to 28 percent of those screened. Between 30 and 58 percent of those found to have a suspicious lesion who then followed up and had a biopsy were diagnosed with some form of skin cancer, mostly basal cell carcinomas.11 It is important to distinguish between the number who were referred for biopsy on the basis of a positive screen and the usually smaller number who actually followed up and had a biopsy. When analyses of the outcome of the screening program also included those who were referred for a biopsy but did not have one, the percentages of those with a positive screening result who were diagnosed with skin cancer dropped considerably for two studies. This latter kind of analysis—including all those referred for further testing—better reflects the reality that people do not always follow up as advised. In a study focused on the much less common melanoma, one very large, free mass-screening project (over 280,000 participants) reported 0.3 percent (or 763) of the participants had lesions suspicious for melanoma (Koh et al., 1996). An additional 1.3 percent had lesions for which “rule-out melanoma” biopsies were recommended. The diagnosis of melanoma was confirmed by biopsy in 130 of the 763 patients with a suspected melanoma. Thus, of those with a lesion suspicious for melanoma, about 19 percent of those who had a biopsy were diagnosed with the disease. Not all of the participants (especially those in the “rule-out” melanoma category) followed up by having a biopsy as advised. Among those in the rule-out category who had a biopsy, an additional 234 cases of melanoma were identified, but finding those cases required biopsies for additional 2,316 disease-free individuals. Two commonly used measures of screening test accuracy—specificity and sensitivity—cannot usually be computed for skin cancer screening studies because only the suspicious lesions that are detected are biopsied, and people with negative screening result are not followed. Sensitivity (the proportion of people who have the disease and who also have positive test results) is relevant in assessing screening tests because the lower the sensitivity, the more likely a test is 10   Earlier studies reviewed in USPSTF, 1996, and Elwood, 1994, and 1996, were consistent with these studies. 11   As discussed further in Appendix B, the proportion of people with a positive test result who are found to have the disease is referred to as the positive predictive value of a screening test.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services screened or not screened (i.e., to continue “usual” care) and then followed both groups long enough for differential outcomes (e.g., five-year mortality rates) to be evident. Because cancer screening trials typically require very large study populations, long follow-up periods, and significant administrative complexities, they usually have been undertaken only for cancers that affect many people and cause major mortality and morbidity. The committee identified no randomized trials of clinical skin cancer screening. Discussions during the committee’s June 1999 workshop indicated that such a clinical trial would require from one-half million to one million participants and would not likely be funded in the United States. The committee heard that a trial is underway in Australia (where skin cancer rates are much higher than in the United States) involving at least 500,000 people in 60 communities that have been randomly assigned to have a screening program established or to continue with current care. Current care includes intensive education and awareness campaigns for both the general population and the health professions. The study was reported to be in its first year of a 10-year follow-up period. The committee also found no case-control studies of the effectiveness of clinical skin cancer screening in reducing mortality or morbidity. Lacking direct evidence of a link between skin cancer screening and better outcomes, the committee searched for indirect evidence. Because the underlying assumption is that screening will lead to earlier detection of disease, the committee considered evidence for this link in the pathway from screening to better outcomes (see arc 3 in the diagram in Figure B-3 of Appendix B). Appendix B summarizes eight screening studies (none of which included an unscreened comparison group) that measured thickness of detected melanomas. Four found no melanomas over 1.0 mm, and another study found only 8 percent in this category. Three studies reported 67 to 87 percent of melanomas were 1.5 mm or less. Data from SEER indicate that more melanomas are being discovered at a thinner stage than in the past but are still diagnosed more often at thicker stages than reported in the screening studies (Dennis, 1999). One recently reported French study suggested that delay in diagnosis of melanomas was less important to prognosis than aggressive tumor growth, a finding consistent with other retrospective studies trying to identify reasons for delay in diagnosis and its impact (Richard et al., 1999). In general, however, the evidence suggests that screening can identify melanomas at a thinner stage than usual care.14 14   The committee notes that “usual” care is not precisely defined (compared to a studied intervention) and that such care may change over time. For example, on the one hand, participation in or competition from managed care may prompt shorter office visits and less direct patient access to specialists; on the other hand, HMOs may encourage more patients to have a periodic preventive care visit. The committee found no specific evidence to document changes in usual primary or specialist care related to identification

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services Studies of survival following excision indicate that greater thickness is associated with poorer survival. These data, although suggestive, are not definitive evidence that detecting thinner melanomas through screening, as opposed to detection through usual care by alert health care professionals, will improve survival. One reason is that screening of asymptomatic people might just lengthen the time between diagnosis and death (lead-time bias). A second possibility is that screening may mostly discover more nonaggressive tumors that exist for long periods of time while missing many faster growing, more lethal tumors (length bias) that arise between screenings. Randomized clinical trials are the best strategy for assessing the effects of screening programs on mortality and morbidity. The committee identified one case-control study of skin self-examination, which reported suggestive evidence that such examination might reduce the risk of lethal melanoma (Berwick et al., 1996). However, the study also found that older men—those most at risk of melanoma—were less likely than women to examine their own skin, although they were much more likely to have a melanoma identified by a spouse. It is reasonable to expect that patients who examine their own skin for skin cancer will do so more frequently than a physician would for any class of patients, so the generalizability of this study to clinical screening is not clear. Conversely, the extent to which effective self-examination could occur in the absence of initial education by a clinician is also uncertain. If Congress was persuaded to extend coverage for a new program of skin cancer screening in asymptomatic people, the statute or implementing regulations would have to address the question of screening frequency. No evidence is available to guide such a policy decision. The most limited option would be to pay for a single screening examination in combination with education about skin self-examination. Whatever the frequency of screening, skin cancer screening could be incorporated in a periodic preventive services visit that also included other recommended preventive services. Again, regardless of how skin cancers or suspicious lesions are identified, current Medicare policy covers follow-up services. ESTIMATED COSTS TO MEDICARE OF EXTENDING COVERAGE As discussed in Chapter 2, the cost estimation approach used by the committee follows the generic practices (e.g., not discounting estimates to present value) employed by the Congressional Budget Office (CBO) in making estimates for Congress. A more detailed presentation of the committee’s cost estimates appears in Appendix E, which was prepared by the Lewin Group in con-     of skin cancers. One analysis indicated that patients in managed care plans were less likely to have skin care provided by a dermatologist than those in fee-for-service plans (Feldman et al., 1996). Most Medicare beneficiaries are not enrolled in HMOs, but their care may still be influenced by managed care depending on their community and the composition of an individual physician’s practice.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services sultation with the committee and background paper authors. To illustrate how Medicare costs would be affected by different skin cancer screening strategies and behaviors, the committee developed estimates for the three models of screening described earlier. As summarized in Box 3-1 and explained below, for the five-year period 2000 to 2004, net estimated costs to Medicare range from about $150 million for the most limited screening scenario to about $900 million dollars for the most expansive. The committee’s estimates of Medicare costs are based on a series of assumptions, some of which have supporting evidence or data but others of which BOX 3–1 Summary of Estimated Costs to Medicare for Covering a New Program of Screening Asymptomatic Beneficiaries for Skin Cancer Screening Strategy Assumptions Case Finding Approach 1. 30% of beneficiaries are screened each year by total skin examination during a physician visit that includes other covered services; those with skin lesions identified as positive (suspicious for melanoma or nonmelanoma skin cancers) are referred to a dermatologist, who may order a biopsy. Estimate assumes 5% of those screened would be referred to and visit a dermatologist; 50% of this group would have a biopsy. Case Finding Approach 2. Same as above except that beneficiaries would first be assessed for skin cancer risk; only those identified as high risk would be examined. Estimate assumes that physicians would identify 10% of beneficiaries as high risk and then examine them; 20% of those examined would be referred to and visit a dermatologist; 50% of this group would have a biopsy. Mass Screening Campaign. Public information campaign encourages beneficiaries to assess their own risk for skin cancer and then see a dermatologist if they conclude they are at high risk. Estimate assumes that 10% of beneficiaries are at high risk and that 20% of that group visit a dermatologist, who biopsies 20% of this group. Cost Estimate Assumptions for Years 2000 to 2004 (see also Appendix E) In 2000, Medicare pays $20 for screening added to a physician visit, $50 for a separate visit, and $90 for a biopsy. Medicare payments increase at 2% per year. Costs are not discounted to present value. Costs are subject to 20% cost sharing.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services Cost Estimate Assumptions for Years 2000 to 2004 No cost offsets for program savings are attributed to screening. Net costs are total estimated gross costs offset by 25% for premium increase borne by Medicare beneficiaries. Data Sources (see Appendix E for specifics) Beneficiary population data from HCFA Office of the Actuary. Assumptions about use of screening visits, referrals, and biopsies and cost of biopsy based on advice of IOM committee and its consultants. Payment per visit data (before adjustments) from 1998 Medicare Physician Fee Schedule. Estimated Costs (in millions) to Medicare Summed Over 2000–2004   Screening Model   Case Finding 1 Case Finding 2 Mass Screening Gross 5-Year Cost to Medicare $1,199.4 $510.8 $199.5 25% Medicare Premium Offset 299.9 127.7 49.9 Net 5-Year Cost to Medicare 899.5 383.1 149.6 are best guesses based on committee judgment in the absence of such information. The estimates are intended to suggest the order of magnitude of the costs to Medicare of extending coverage, but they could be considerably higher or lower than what Medicare might actually spend were coverage policies changed. The tables in Appendix E allow readers to vary some of the committee’s assumptions and calculate alternative estimates.15 The unit cost of skin cancer screening services (physician visits and diagnostic biopsies) is not high, which means that the cost estimates are driven primarily by the number of individuals who would be screened. Some 39 million beneficiaries are expected to be enrolled in Medicare in 2000. A majority visit a primary care physician each year, so the number of persons who could be offered screening is quite large. Even if all beneficiaries were entitled to screening, however, not all would seek or accept screening nor would all physicians advise or conduct screening for their patients. A recent survey of primary care physi- 15   The CBO and HCFA actuaries would undoubtedly be asked to develop their own estimates for any specific proposals under active consideration by Congress.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services cians in Dade County, Florida, and New Haven, Connecticut, reported that about 30 percent said they routinely performed a full-body skin examination on all their patients, and about 30 percent of the remainder say they did so for high-risk patients (Kirsner et al., 1999). A retrospective analysis of patient records at two Veterans Affairs Medical Centers found skin cancer screening documented for only 28 percent of 200 patients and only 18 percent when those without a skin-related complaint were excluded (Federman et al., 1997). Reports on use of other preventive services likewise show considerably less than complete adherance to recommendations by both physicians and consumers (HCFA, 1998a, and USPSTF, 1996; see also Chapter 6).16 For the single-step case-finding approach, the estimates below assume that of 30 percent of Medicare beneficiaries, 5 percent would then be referred to a dermatologist who would do biopsies on half of those referred. For the two-step strategy involving an initial risk assessment, the estimates assume that primary care physicians would identify 10 percent of beneficiaries as high risk and refer them to a dermatologist for further evaluation. Both estimates assume that other beneficiaries would have skin examinations (and referrals and biopsies) prompted by a physician’s incidental discovery of a skin abnormality or patient concern about a skin abnormality. These services are already covered by Medicare. Consistent with the CBO approach, estimates are not discounted to present value, and total direct costs for screening are offset by 25 percent to reflect the increase in the Medicare premium that would be paid by beneficiaries based on the projected increase in expenditures resulting from extended coverage. The screening models assume that the costs listed for visits and biopsies are actual Medicare reimbursements with nothing subtracted for cost sharing by patients. Absent relevant evidence, the models assume no offsetting savings to Medicare during the period. The committee did not attempt to assess the cost-effectiveness of skin cancer screening. During its workshop, however, the committee was presented with an analysis that modeled the cost-effectiveness of a single occurrence of skin cancer screening among high-risk individuals, compared to no screening (see Freedberg et al., in press). This analysis assumed that the screening was performed by a dermatologist. More important, the analysis also assumed (1) that the screening would result in melanomas being detected at earlier, thinner stages; (2) that these screening-detected melanomas would be more effectively 16   Skin cancer screening is less demanding than some other screening services because it is not invasive, nor does it involve expensive, specialized equipment that is not available in the office of a primary care physician (thus requiring the patient to go elsewhere). Still, the patient or physician would have to be motivated to request or recommend the skin examination. For a physician, discussing skin cancer screening and doing a risk assessment or skin examination could add several minutes to office visits that are often scheduled for less than 15 minutes.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services treated than if they had been detected in the course of usual care; and (3) that the result would be reduced Medicare costs for treating skin cancer. Analysts then estimated the cost of screening a population of one million and the quantity of life saved (mortality avoided) due to the screening. The result was an estimated $29,170 expended per year of life saved, a figure not greatly different from several other cancer screening strategies. It is important to note that the assumptions of the analysis are significantly different than those used here, namely, that screening would in fact detect melanomas at earlier stages leading to earlier and less costly treatment. The analysis did not assess the costs or benefits of screening for nonmelanoma skin cancers. In developing its estimates, the committee took into account past experience with cost estimates that have assumed that more people would take advantage of screening benefits than actually do. Underuse and underprovision of effective screening services are important public health issues. Chapter 6 discusses the challenge of “putting prevention into practice,” that is, turning clinical recommendations and coverage into actual delivery of preventive services to those likely to benefit from them. STATEMENTS OF OTHERS ABOUT SKIN CANCER SCREENING A number of organizations have made statements and recommendations about clinical screening for skin cancer. The organizations vary in the extent to which they explicitly link their conclusions to systematic assessments of the evidence. The U.S. Preventive Services Task Force (USPSTF, 1996) stated that “there is insufficient evidence for or against routine screening for skin cancer by primary care physicians using total body skin examination…. Clinicians should remain alert for skin lesions with malignant features…when examining patients for other reasons” especially those with established risk factors (p. 148). “A recommendation to consider referring [patients with melanocytic precursor or marker lesions] to skin cancer specialists for evaluation and surveillance may be made on the grounds of patient preference or anxiety…although evidence of benefit from such referral is lacking.” The USPSTF is again reviewing the evidence related to skin cancer screening and could reaffirm or change its 1996 recommendation, but an announcement is not expected before the release of this report. Since 1985, the American Academy of Dermatologists has sponsored free skin cancer examinations as part of a public education program (Koh et al., 1996). The Academy’s materials for the public stress sun avoidance and skin self-examination rather than routine clinical skin examinations for asymptomatic individuals. Statements from the National Institutes of Health (NIH) are not fully consistent. The National Cancer Institute’s on-line PDQ information system for

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services physicians states “there is insufficient evidence to establish whether a decrease in mortality occurs from routine examination of the skin” (NCI, 1999f). In contrast, a 1992 NIH consensus conference stated that “there is sufficient evidence to warrant screening programs for melanoma in the United States…. The public should be encouraged to ask their primary care physicians and nurses for periodic skin examinations when seeing them for other purposes, for example, a physical examination…[and] should be made aware of (1) the increased risk of melanoma related to excessive sun exposure, particularly in childhood; (2) the clinical appearance of early melanoma; (3) the excellent prognosis associated with detection and treatment of early melanoma; and (4) the need for regular skin examinations by themselves and by their health professionals.” A specific screening interval is not cited. The consensus statement does not link its recommendations to specific citations of the literature or evaluate the strength or quality of the evidence but does note that no randomized clinical trials are available. The American College of Preventive Medicine (ACPM) recommends “periodic total cutaneous examinations be performed, targeting populations at high risk for malignant melanoma.” The ACPM, however, finds insufficient evidence to set a screening interval more precisely and recommends well-conducted observational or case-controlled studies or randomized clinical trials to better identify the screening interval and the risk-benefit ratio for different groups. The strongest and most comprehensive screening recommendation comes from the American Cancer Society. It recommends a cancer-related screening—including a skin examination—every year for those over 40 (ACS, 1999b). COMMITTEE FINDINGS AND CONCLUSIONS In developing its findings and recommendations, the committee recognized that the pathway from adoption of a new program of skin cancer screening to improved health outcomes for Medicare beneficiaries would have many uncertainties. Figure 3–2 illustrates a simple pathway and indicates some of the uncertainties associated with each element of the pathway. This figure does not include every possible step or uncertainty but rather summarizes some major variables that would likely affect the success of a screening program. These are the sort of potential issues to be weighed in the formulation of any final policy. Most of these uncertainties would affect the cost of covering screening. Findings After reviewing the literature, considering the discussion at its workshop, and drawing on its members’ judgment, the committee reached several findings relevant to decisions about coverage of a new program for skin cancer screening for Medicare beneficiaries. The first findings listed below relate to the assessment criteria depicted in Figure 3–1. The last relate to directions for further research.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services FIGURE 3–2 Causal Pathway: Skin cancer screening, with examples of uncertainty that could affect outcome at several key points. Note: Events are in bold and in the main path; examples of variables that increase uncertainty of outcomes are offset to the right. Disease Burden. Basal cell and squamous cell carcinoma are relatively common among older people. Squamous cell carcinoma is sometimes lethal and both can cause disfigurement or functional impairment. Melanoma is much less common but more often lethal. Older white males appear to be at particular risk of developing and dying from melanoma.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services Treatment Effectiveness. Basal and squamous cell carcinomas are effectively treated by excision or other therapies. Earlier diagnosis and treatment is likely to result in less scarring. Excision is also effective for early-stage melanoma, but effective treatments have not been found for late-stage melanoma. Accuracy of Screening Tests. Although studies are limited and show somewhat mixed results for total skin examination, the test appears acceptably accurate when performed by a dermatologist. Diagnostic biopsy is also not perfectly accurate but has good results in making the basic distinction between benign and malignant lesions. Effect of Screening on Outcomes. Because basal cell carcinoma and squamous cell carcinoma are highly treatable and rarely lethal, it is unlikely that a new program of screening asymptomatic people could appreciably improve survival rates. Direct evidence is not available on the effect of screening on morbidity and disfigurement from these conditions. Direct evidence is not available to support conclusions about the effect of clinical screening on mortality or other health outcomes related to melanoma. Physicians identify a substantial proportion of the melanomas and tend to detect them at a thinner stage than do patients, and thinner melanomas have a better prognosis. This indirect evidence is only suggestive about the possible benefits of a new program of skin cancer screening for those without symptoms, in part because of inadequate knowledge about the rate of growth of melanomas, especially thin melanomas in older people. Another uncertainty involves the degree to which beneficiaries, particularly those at greatest risk, would avail themselves of a screening benefit and pursue recommended follow-up care. Benefits Versus Harms. No controlled studies provide direct evidence about the benefits or harms of skin cancer screening. Patient perspectives on possible harms have not been explicitly assessed. Unlike breast, prostate, and certain other cancers, unnecessary surgery for a lesion misdiagnosed as skin cancer is unlikely to be life- or function-altering, disfiguring, or very painful. Scarring and anxiety may be expected from unnecessary diagnostic and treatment services if screening falsely identifies someone as having skin cancer, but such scarring usually will be minimal. Possible Directions for Future Research The committee identified several areas where further research would be helpful, although it did not attempt to set priorities. As noted above, a randomized controlled trial of skin cancer screening is underway in Australia, although substantial results are years away. Also, the much higher incidence of skin can-

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services cer in Australia would have to be taken into account in assessing the study’s relevance to this country. In the absence of evidence from controlled trials, it would be useful to have more research on the early stages of cancerous and precancerous lesions and their progression to more advanced states. Such research would help answer questions about how quickly different kinds of melanomas progress in different risk groups and about how likely it is that earlier detection of disease through a new program of clinical screening would make a difference in outcomes. More research would also be useful to understand how frequently and how quickly actinic keratoses develop into squamous cell carcinoma and what factors predict such progression. Key issues in arguments against screening involve the effect of the intervention on healthy participants, which both exposes many individuals to unnecessary harm including inconvenience, discomfort, and anxiety, and drives up the cost of the program relative to any benefit. These concerns would be reduced if those most likely to benefit from the program could be accurately and efficiently identified, for example, during a preventive care office visit. To this end, further research on the effect of training primary care physicians in the accurate assessment of skin features should examine change in actual clinical practice (not just identification of photographs). Accurately identifying those most at risk of skin cancer is part of the problem, but reaching the members of that group in the community is a different problem. One question is what kinds of communication strategies will encourage people at higher risk to limit sun exposure, to be alert to the warning signs of cancer (especially melanoma), to visit a physician when something suspicious is found, and to follow up on referrals for further assessment or treatment. Except for communication of sun exposure, a particular focus should be identifying ways of communicating more effectively with older white males. Continued work to develop and assess educational programs and skin self-examination initiatives makes sense. For example, although outcome data are limited, research suggests that women are more likely to self-identify melanoma than men and that men are more likely than women to have a melanoma identified by a family member. This may suggest investigating whether self-examination education programs might also emphasize the role of family members and close friends in being alert for, and telling one another about, abnormal-appearing areas of skin, which should then be professionally evaluated. More generally, in addition to research that could clarify the benefits, harms, and cost-effectiveness of clinical skin cancer screening and primary prevention programs, other interesting lines of investigation exist in the area of treatment. These include more effective chemotherapy for nonlocalized melanoma and vaccination or immunotherapy for melanoma.

OCR for page 38
Extending Medicare Coverage for Preventive and Other Services Conclusions In summary, the committee concluded that evidence for the effectiveness of skin cancer screening is insufficient to support positive or negative conclusions about the adoption of a new program of clinical screening of asymptomatic Medicare beneficiaries. Direct evidence that detection of skin cancer through clinical screening leads to better health outcomes is lacking rather than negative, inconsistent, or ambiguous. The indirect evidence for screening is suggestive but not compelling. The committee is aware that Medicare coverage has been extended for other services (e.g., prostate-specific antigen testing and bone densitometry testing) for which direct and indirect evidence of benefit is inconclusive or disputed. Those precedents are not sufficient grounds for covering a new program of clinical screening for skin cancer. Because evidence does support benefits of early detection and treatment as part of usual medical care, clinicians and patients should continue to be alert to the common signs of skin cancer—with a particular emphasis on older white males and on melanoma—and should investigate suspicious signs further. Medicare already covers skin examination and testing by primary care physicians and dermatologists prompted by patient concern about a skin abnormality or by incidental physician discovery of an abnormality during a visit for other purposes. Further, dermatological and other organizations should continue skin cancer educational programs for people of all ages, including programs that encourage people to limit sun exposure and inform themselves about skin cancer risk factors and warning signs, especially those for melanoma. Perhaps the major challenge related to the Medicare population is identifying and implementing better ways of reaching the group at highest risk of death from skin cancer—older, fair-skinned males.