Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Using mammography for cancer control: An unrealized potential

Using mammography for cancer control: An unrealized potential Benefits of Mammography The most convincing evidence that screenDr. Howard is ProgramDirector for Studiesin Social Epidemiology, Health Promotion Sci encesBranch, in the Division of CancerPreven tion andControl of theNational CancerInstitute in Bethesda,Maryland. VOL. 37, NO 1 JANUARY/FEBRUARY 1987 ing for breast cancer is an effective means of secondary prevention comes from the Health Insurance Plan of Greater New York Screening Project (HIP), a randomized clinical trial conducted between 1963 and 1970.8_Il This study of members tested the benefit of four annual examinations involv ing mammography and breast palpation. Half of the 62,000 women in the study were randomly assignedto the intervention group, and the other half, the controls, their usual care. Those in the intervention group were offered the opportunity to participate in the screening program, and 65 percent ap peared for screening at least once. Those who refused to be screened at all were still considered to be part of the intervention group, so that selective participation would not bias the results. Within the control group, mammography was used increasingly for differential diagnosis but not as part of a program for early detection among asymp tomatic women 8l0 Many published reports have emanated from the HIP study,8'5 and follow-up of the original cohort is continuing. Basically, the investigators have concluded that the screening program substantially reduced breast cancer mortality among women age 50 and older at the time of entry.9M Data suggest that women younger than 50 may also have benefited, but this possibility is open to debate.8-9- In those 50 and older, 16.17 the intervention group showed a reduction in mortality of approximately 50 percent at five years and 33 percent at 10 years. com pared with equivalent women in the control group.9―° Potential biasing factors such as lead time, length bias, and overdiagnosis were taken into account in the analysis. Although the HIP study involved two screening modalities, their relative contri bution to the observed reductions in mor tality is a matter of speculation. For the intervention group as a whole, more screening-detected cancers were discov ered by physical examination than by mam mography.―Among women age 50 to 59 at diagnosis. however, 42 percent of the cancers were detected by mammography alone, and another 18 percent were discov ered by both modalities.―There appears to be some consensus that one third of the reduction in mortality in the total group offered screening was due to the impact of mammography.―'8-'9 as it functioned in the !960s. Further evidence that mammography per se facilitates the early detection of breast cancer comes from the Breast Cancer De tection Demonstration Project (BCDDP), in which approximately 280,000 American women were enrolled in the mid-!970s.202' Because the BCDDP was a demonstration project rather than a clinical trial, no con trol group was included. The data strongly suggest. however, that annual mammo graphic examinations played a key role in detecting ‘¿minimal cancers― in situ car ( cinomas and invasive cancers smaller than one cm), for which the prognosis would appear to be particularly favorable. Among the 1,153 minimal cancersidentified through the screening program, 57 percent were detected by mammography alone, com pared with six percent detected by physical examination alone 2oMoreover, mammog raphy alone was also responsible for de tecting a much higher proportion of infil trating cancers greater than or equal to one cm. than was physical examination alone (34 percent versus nine percent).2°Thus, 42 percent of all program-detected cancers in the BCDDP were discovered through mammography alone, compared with only nine percent from breast palpation alone.20 In contrast to the HIP study. in the BCDDP mammography seemsto have been effective in the 40-to-49-year age group as well as in older women. Mammography alone was responsible for 35 percent of cancer detections in the younger group. and it was positive along with physical exami nation in another 50 percent of thesecases 20 In the HIP study, the equivalent figures were 19 percent and 19 percent respec tively.―2° When mammography was re moved from the BCDDP as a routine screening modality for women younger than 50. the detection of minimal cancers de creased in this age group.2° The American Cancer Society (ACS) has concluded, therefore, that “¿there evidence that is The data suggest that only about 15 to 20 percent of American women age 50 and over have ever had .a mammogram and that a much smaller proportion are being examined with systematic regularity. screening with mammography can detect very small, localized breast cancers in women 35 to 49 years old, which in turn suggests the possibility of better survival rates in this age group. “¿22 Recent research results from Sweden and Holland lend additional credence to the proposition that mammography alone can significantly reduce mortality from breast cancer—atleast among older women. The Swedish study, which began in 1977, is of particular interest because it is a random ized controlled trial of single-view mam mography without breast palpation for the early detection of breast cancer.23 As was true in the HIP study, the control group has received usual care, which for 13 percent has included mammographic examina tions.23 Although the screening interval in the Swedish trial is almost three years for women age 50 or older, the 40 percent reduction in breast cancer mortality in the 50-to-74 age group (p = 0.003) is similar to CA-A CANCER JOURNAL FOR CLINICIANS that for the 50-to-64 age group in the HIP study. where annual screening was proto col.-3 Moreover, as in the HIP study, women age 40 to 49 in the Swedish trial are yet to show a significant mortality benefit from the screening effort.3 For all age groups. further follow-up and more extensive anal yses are anticipated. One of the screening studies from Hol land also examined the effect of single view mammography alone on breast cancer mortality. Preliminary results from the Nijmegen project suggest that the mortality in women age 35 and over can be reduced by about 50 percent through regular mam mographic screening of all eligible women. A variety of studies suggest that physicians regard mammography as a relatively unimportant prevention strategy, even for women age 50 and older. This was a case-control design. however. rather than a randomized trial, and various sources of possible bias are still being eval uated. Furthermore, the investigators ac knowledged that they cannot yet study the effect of screening on women younger than 55 because of the small number of cases and the higher attendance rates in the youngerage groups.Data from theother Dutchcase-control (the study DOM project in Utrecht) indicate that screening with the combination mammography and physi of cal examination gives protection against dying from breast cancer, especially among elderly women and those screenedtwice. 25.26 Again, however, the researchers are con sidering the possibility that important bias ing factors may have been overlooked, and they need a longer follow-up period to eval uate mortality trends. f@rsomewhat in their screening guidelines for breast cancer. The National Cancer In stitute (NC h has adopted a more conserva tive position regarding mammography than the ACS has. For asvmptomatic women 50 and older. both groups recommend annual or rout inc screening with mammography and breast palpation. 2) For younger women. however, the NCI would restrict mammographic screening to those at high risk—that is. women with a personal or family history of breast cancer. 30,31 The ACS. on the other hand. recommends a baseline mammogram for all asymptomatic women age 35 to 40 and screenings every one to two years for asymptomatic women age 40 to 4929 The American College of Radiology makes similar recommenda tions for the various age groups.32 If the NCI and ACS guidelines had been adopted by the medical and lay communi ties. one would expect routine utilization of mammography in the 50-plus age cate gory and lesser usage among younger women. The literature indicates, however, that only a small proportion of American women age 50 and older regularly use mammography for screening purposes. It is also unclear whether the infrequent use among younger women is consistent with any of the guidelines. Patterns of Use It is difficult to determine the extent to which mammographic examinations have been performed for screening rather than diagnostic purposes. There is, therefore, an element of uncertainty in estimating the proportion of women who have been screened with mammography at some time in their lives and the proportion who are routinely screened. The weight of the data in Table 1does, however, suggest that only about 15 to 20 percent of American women age 50 and older have ever had a mammo gram and that a much smaller proportion are being examined with systematic regu larity. Furthermore, both of these estimates are undoubtedly inflated by the use of di agnostic mammography. since most mam mograms are performed for diagnostic pur poses.37 Current Use ofMammography Guidelines for Use The two major organizations concernedwith cancer prevention in the United States dif VOL 37, NO 1 JANUARY/FEBRUARY 1987 An obvious exception to this picture are women inmetropolitan reported areas higher the findings of the Gallup Organization mammography use(24percent)hanthose t shown in Table 1 According to the esti in the suburbs (21 percent) or in nonmetro mates in that survey. 41 percent of women politan areas (12 percent): Jewish women in the 50+ agegroup have received a “¿breast showed a much higher usage rate (38 per x-ray―at some point in their lives, and 15 cent) than any other religious group: women percent have such examinations every year. with household incomes above $15,000 a Even among those in the youngest age group year had higher usage patterns than those (20 to 39), 33 percent reported having re with lesser incomes: involvement in other ceived a breast x-ray. It is possible that health activities (including breast self-ex some of these respondents confused chest amination) was positively associated with x-rays with mammography. mammography use: usage increased with increasing knowledge of breast cancer de In spite of the age guidelines of the NC! and ACS, the national surveys show no tection and treatment procedures and in consistent relationship between age and the creasing self-estimates of the likelihood of utilization of mammography—except that getting breast cancer. Such factors as geo graphic region. concern about the side ef in each study the youngest age group re ports the smallest proportion of mammog fects of breast cancer, and beliefs about the raphy use (Table 1). Two of the surveys progress of breast cancer treatment were (Lieberman33 and Gallup35) show almost not meaningfully associated with mam no difference in utilization between women mography usage. The analysis by Lane and Fine of pa in the 50+ age group and those 30 to 49 or 40 to 49. The NCI's Office of Cancer Com tient compliance with mammography re munications (0CC) survey,owever, h shows ferrals is also relevant.39 They found no an inverse relationship between age and significant association between a host of mammography use after age 35,34The pro sociodemographic factors (such asage. race, portion who reported ever having had a maritaltatus, s education, income)and and the compliance behavior of their respon mammogram declined from 25 percent for the 35-to-49-year age group to 16 percent dents. Furthermore, there was no relation for those age 65 and older. ship between a personal or family history of breast cancer and patient compliance. With respect to other predictors of mammography use, Lieberman shows a sizable positive relationship between edu cation and mammography experience.33 Referral Patterns Among women who attended college. 18 A variety of studies over the last 15 years percent reported having had a mammo gram. compared with 12 percent for those suggest that physicians regard mammogra phy as a relatively unimportant prevention with a high-school education and nine per strategy, even for women age 50 and older. cent for the grade-school group. The 0CC In 1970. Bates and Mulinare studied the survey found only a slight positive effect (four percentage points over the entire screening practices of internists and gen range34),while the Gallup Organization did eral practitioners in upstate New York.4° The investigators found that only one per not assessthe impact of education.35 cent of those surveyed reported that they The 0CC survey also evaluated rela tionships between mammography use and usually prescribed mammography for other demographic factors, and it measured women age 40 to 65. In the mid-1970s. an audit of internist records in North Carolina the effects of various perceptions. atti showed no mammography referrals for tudes, and behavior patterns among Amer women age 50 and overwho visitedheir t ican women. 34.38 brief, the findings were In physicians for general checkups.41-42 In as follows: There was no difference in 1980, a record review at a general hospital mammography usage between whites and in Cleveland showed that only four percent blacks, but the “¿other― category race of eligible women age 50 to 60 received a showed considerably less experience; CA-A CANCER JOURNAL FOR CLINICIANS [I D t) a, 0@ a,, ‘¿a, a,. C. @C. ‘¿a, ,., Co DNCN C@1 e@ - .C@) N, a,.@. II. II II II zzz —¿. ,,@. a, a, .0) ,—.000 I .1' + CD -@‘T a0' ,,a,@ ‘¿â€˜a, ‘¿,,,@ ‘¿a, 1, cr ,. ‘¿@ a―. c_;,' @c VOL 37. NO 1 JANUARY/FEBRUARY1987 mammogram during their usual care.43 In another study. from 1978 to 1980, only two percent of women age 50 to 70 at a general medicine clinic in Indianapolis received mammography during their usual care.'@' Somewhat more favorable findings come from the study by Dietrich and Goldberg of primary-care physicians in Northern Cali fornia.45 In the early 1980s, these investi gators reviewed the charts of generalists and'subspecialists involved in group prac tice to determine their compliance with pre ventive medicine guidelines. Less compli ance was shown for mammography than for any other preventive procedure except flu immunization, and mammography was also regarded as the least important proce dure. However, approximately 15 percent of eligible women age 50 to 59 actually receive mammography over an annual in terval. Among physicians who rated the procedure high in importance, patient cov erage increased to 38 percent. compared with 11 percent for the physicians who rated it low in importance. Unlike the other preventive techniques (such as Pap smears and breast palpation), mammography was less likely to be pre scribed in the context of a complete physi cal examination than during other medical encounters,45 suggesting its use as a diag nostic tool in response to detected symp toms. This finding by Dietrich and Gold berg differs somewhat from the 1983 Gallup survey of patients themselves. Among women age 50 and older who reported ever having had a breast x-ray, 62 percent said it was part of a regular checkup. The com parable figure for medical palpation of the breasts, however, was 84 percent, indicat ing that mammography had the higher di agnostic/prevention ratio.35 In the California study, mammography usage did not differ significantly by physi cian specialty. In Battista's study of pri mary-care physicians in Quebec, however, mammography was prescribed much more frequently in community health centers and family medicine teaching centers than in fee-for-service practices.46-47At that time (1981 and 1982), the Canadian Task Force on the Periodic Health Examination rec ommended annual mammography for women 50 to 59,46 Yet, only eight percent of the entire sample of physicians said that they complied with this recommendation even if the physical breast examination was normal.46 Moreover, the figure of eight percent was based on interview data, and there are indications from billing audits that these reports may have been somewhat ex aggerated.46 As was true in California, physicians' useof mammography was much less prevalent than their commitment to other preventive procedures, such as breast palpation (99 percent) or Pap smears (91 percent). 46 Two recent surveys in the US show a wide discrepancyetweenbeliefs phy b by Two recent US surveys show a wide discrepancy between beliefs by physicians about the effectiveness of mammography and their actual utilization practices. sicians about the effectiveness of mam mography and their actual utilization prac tices. In 1982, Cummings et al mailed questionnaires to a random sample of fam ily physicians in New York State and re ceived a 60 percent response.48 The great majority of respondents believed that mam mography is effective for detecting breast cancer in its early stages. Again, however, only eight percent recommended annual mammography forasymptomaticwomen older than 50. Twenty-nine percent did not recommend mammography at all for this age group, and another 16 percent recom mended it only for symptomatic patients. A nationwide telephone survey of pri mary-care physicians was conducted for the ACS in 1984.29 Because the response rate reached 91 percent, the investigators contend that its results are applicable to all primary-care physicians in the continental US. Forty-one percent of the respondents CA-A CANCERJOURNAL FORCLINICIANS agreed “¿completely― the ACS guide with lines concerning mammography. but only Il percent said that they observed these guidelines with all relevant patients. In comparison. 80 percent of the physicians universally followed the guidelines for breast physical examination, and 75 percent ad hered to the recommendations for the Pap test. Approximately one fifth of the re spondents said that over the last five years they had increased their use of mammog raphy for early detection. Less than half of the entire group. however, had ever used the procedure for asymptomatic women with no history of cancer. Obstetricians and gynecologists in the radiation. “¿unnecessary― biopsies. over diagnosis. and financial costs. In addition, physicians and patients have developed at titudinal barriers to mammography that are not reflections of the objective realities. Radiation Risks The radiation hazards of mammography. as the technique was applied in the 1960s and early 1970s, were publicized by Bailar in a widely quoted article.'9 Based on cer The recognized deterrents to the use of mammography include the risks of radiation, “¿unnecessary― biopsies, overdiagnosis, and financial costs. ACS survey were more committed to mam mography than were internists and general practitioners. Only two thirds of the OB GYN group. however, had ever ordered a mammogram for asymptomatic women, and only 17 percent universally followed the ACS guidelines for mammography. In light of these findings, the ACS has concluded that: “¿Too physicians are withhold many ing mammography in the absence of symp toms, at a time when the main purpose of mammography is to find an asvPnptomatic cancer.―49 Deterrents to the Use of Mammography Although mammography has proved to be effective in detecting early cancers, the his tory of its usage provides evidence for cau tion in its application as a screening tech nology. The recognized deterrents to the use of mammography include the risks of VOL 37, NO 1 JANUARY/FEBRUARY 1987 tain assumptions about the dose-response relationship, the latent period, and so on, Bailar “¿regretfully― concluded that “¿there seems to be a possibility that the routine use of mammography in screening asymp tomatic women may eventually take almost as many lives as it saves.― acknowl He edged that mammography can contribute to the reduction of breast cancer mortality but that screening by medical history and phys ical examination alone “¿will probably pro vide much or most of the same benefit las a three-way screenj without risk from irra diation, at least in women under some fairly high age limit.― In essence. Bailar argued that the promotion of mammography as a general public health measure was prema ture. Concerns about the hazards of radiation in the BCDDP led to a set of Working Group investigations of these potential risks.@°On basis of estimated risk-bene the fit ratios, the experts concluded that annual mammography should continue in the BCDDP for women age 50 and older but not for younger women, unless they had a personal or family history of breast can cer.5° he review panel also recommended T that the BCDDP should continue its moni toring and control measures for radiation dosageA retrospective analysis of dosage lev els actually administered in the BCDDP showed substantial declines from the radia tion levels in the HIP study.8-'9'2'5' and there were reductions over time in the BCDDP itself.21-50-52 During the HIP study. the average skin dosage from a mammo graphic examination was 7.7 rad.8―9 By the end of the BCDDP. however, the aver age surface exposure was 0.47 R for film screen units and 1.2 R for xeromammog raphy.5' Similarly, estimates of midline dosage levels in the two studies show large differences in favor of the BCDDP. 0.2I By the fifth year of that program, the average midline dose was 0.035 rad for film screen units and 0.36 rad for xeromammogra phy .@‘ Continuing technological improve ments have led the ACS to conclude that: “¿Modernechnology has reduced the ra t diation exposure of low-dose mammogra phy to the point of negligible risk, if risk exists at all, and has increased its diagnos tic capabilities at the same time.―49 Unnecessary Biopsies The purpose of screening with mammog raphy is to detect breast cancer at an earlier stage than it would be detected by clinical palpation, breast self-examination, or ac cidental discovery. Mammography can also help in the interpretation of the findings from clinical examinations. Physicians who believe in aggressive early detection with mammography willingly recommend sur gical biopsies on the basis of mammo graphic suspicions alone.53 Other physi cians are reluctant to biopsy lesions that they cannot feel.54 Even in the BCDDP, the proportion of recommended biopsies varied greatly from clinic to clinic. In the most aggressive center, the percentage of screening examinations resulting in surgi cal recommendations (including biopsies, aspirations, and consultations) was eight times higher than in the least aggressive center.55 It is reasonable to suppose that the difference in biopsies recommended was largely a function of the difference in atti tudes toward mammographic findings. More than eight percent of all partici pants in the BCDDP had biopsies per formed.2o@56 he great majority resulted from T project screening examinations at an aver age rate of one biopsy for every 48 screen ings;20'56 the comparable rate for the HIP study was one biopsy for every 101 screen ings.'° Compliance with biopsy recom mendations in the BCDDP decreased ap preciably over time as physicians adjusted to high benign-malignancy ratios in biopsy experiences.55 These ratios were inversely related to age, varying from 16.4 for women 35 to 39, to 2.7 for those 70 to 7420.21 Among all agegroupscombined,theratio of benign to malignant biopsies performed because of recommendations from screen ing in the BCDDP was 5420 Generally speaking, this is higher than the ratios for settings in which no special screening ef forts were undertaken,576° although there, too, the ratios vary—apparently by age, physician orientation, and type of hospital. Concerns about false-positive findings in mammographic screening are not limited totheUS. Investigators associated withthe NijmegenprojectnHollandobserved i that “¿the positive predictive value of mammog raphy is 35 percent; thus, for each screen detected case, two women had a referral and one of them had a biopsy that they would not have had if there had been no screening. “¿24 Canada, physician skep In ticism about the specificity of mammogra phy has deterred compliance with biopsy recommendations from the National Breast Screening Study.54 To counter this skepti cism, directors of the study have used the medical media to promote the value of mammography in detecting early breast cancer. Basically, they argue that “¿where there is no increase in biopsies, there can be no impact of screening.―54 Overdiagnosis The major attractiveness of mammography as a screening technology is its capacity to detect minimal cancers that portend favor able prognoses. But as Eddy has observed, this benefit contains the seeds of potential problems: “¿Unfortunately there is no sharp boundary between nonmalignant and ma lignant cells, and it is quite possible to overdiagnose as a very early cancer a lesion that is not cancer and would never become cancer. As well as increasing the number of ‘¿cancers' etected, this can inflate the d number of cancers detected in the earliest stage, and since these lesions would never become clinically significant cancers, it can inflate survival statistics.' ‘¿58 CA-A CANCER JOURNAL FOR CLINICIANS The magnitude of overdiagnosis is not easily estimated. Inferences can be drawn from statistical modeling. mirror-biopsies. and differences in diagnosed casesbetween intervention and control groups in screen ing trials. Minimal cancers in the HIP study group constituted only 10 to 16 percent of all the detected breast cancers.6' and over time the totality of cancers in the study group equalled that of the controls.8 This suggests that the screening effort did not result in overdiagnosis.6' but merely in ear lier detection of cancers that would ulti mately have surfaced on their own. The BCDDP presents a more questionable pic ture. however, becauseof the large propor tion of minimal cancers and the absence of a control group. It has been argued, there fore, that some of the cancers detected by mammography alone might never have caused morbidity or mortality. Lesions can qualify as cancers from a histologic per spective without qualifying on the basis of their biologic behavior2',6Two of the European studies acknowl edge links between mammography and overdiagnosis. The Swedish investigators suggest that some of the excess Stage I cancersdetected in the mammography group might never have surfaced clinically.23 The Nijmegen investigators contend that the diagnosis and treatment of “¿clinically irrel evant non-invasive cancers―may be an important side effect of screening.24 Simi larly, Baines of the Canadian National Breast Screening Study accepts the possibility of overdiagnosis as a reality one must live with.54 Where screening with mammography is aggressive, concerns about overdi agnosis may go beyond the issue of clinical surfacing. For example, in the BCDDP it was alleged by the mass media that some women were unnecessarily losing their breasts because of incorrect interpretations of biopsy specimens. In a small proportion of cases, lesions initially defined as mini mal cancers were later judged to be be nign.5° he Working Group that evaluated T the pathology of minimal cancers could not procure all the relevant slides, and they found differences of opinion in the original pathology reviews and in the subsequent reviews by experts.5° Recognizing that minimal cancers can present special diag nostic problems. the Working Group con cluded that treating physicians must take into account clinical as well as pathologic information.50 According to Baines, the uncertainties associated with aggressive mammographic screening are not uncommon in medical care. and the risk of overdiagnosis is en twined with the quest for early diagnosis:54 “¿There is unanimity with respect to rarely diagnosis or treatment of any condition. There rarely is a right answer. Perhapsbreast cancer in situ is harmless. Perhaps ‘¿early detection' will not prevent deaths. Perhaps those who repudiate mammography are right. But until we know, perhaps women deserve the right to the earliest diagnosis possible. Financial Costs The economic cost of mammography var ies greatly from area to area and from phy sician to physician. Recent data from Illi nois show a median statewide price of $105. but the charge can surpass $200. including physician fees.2' Part of this cost may re flect the underutilization of existing mam mographic equipment. In screening clinics with high-volume participation. the aver age costs decrease. Attempts by the ACS to lower the cost of mammography have resulted in lower charges in some areas, ranging from $36 to $77,63_6s Where screening programs are subsidized, the ac tual costs to the patient may be further reduced or be entirely free.M At the present time, few insurance companies cover the cost of mammography for asymptomatic women. Similarly, mammography is not authorized by Medi care for routine screening purposes,@ but only as a diagnostic test under certain specified conditions. Physicians, there fore, may have to “¿bootleg― or disguise the routine use of mammography by attach ing the more acceptable label of differential diagnosis. VOL. 37, NO 1 JANUARY/FEBRUARY 1987 Estimates have been made of the poten tial costs of national screening programs using mammography. These estimates vary according to eligibility criteria, screening intervals, number of views, and so on. In 1980, Eddy speculated that the yearly costs of annual mammography for the 31 million American women over age 50 would be approximately 1.5 billion dollars, at $50 an examination.58 This total did not include the add-on costs of follow-up examinations and biopsies for women with positive or suspicious mammographic findings,67 nor did it take into account financial savings that might result from earlier diagnosis and treatment, and consequent reductions in morbidity and mortality. Attitudinal Barriers Survey data indicate that the deterrents to mammography discussed in scientific jour nals and the mass media have influenced the beliefs of physicians and patients. In each of the physician surveys, the impor tant reasons for negative attitudes toward mammography (Table 2) included the risks of radiation, financial costs, and concerns about the effectiveness, reliability, or ne cessity of the procedure.29'46-48 Physicians in the US were also deterred by the small likelihood of detecting breast cancer in any particular screening examination.29'48 Studies of lay attitudes toward mam mography suggest that the recommenda tion of physicians is a critical intervening variable that influences the beliefs and be havior of patients.68 In a 1977 Los Angeles survey, 93 percent of respondents reported that they would be at least somewhat likely to have a mammogram if their physician recommended it.36 In the same study, 97 percent of women who reported being ad vised to have a mammogram said that they actually had it. Furthermore, a recent sur vey by the Saturday Evening Post indicates that the absence of physician recommen dations is the greatest deterrent to the use of mammography.69 Among the respon dents over age 50 who were not having annual mammograms, 64 percent gave as a reason that it was not suggested by their physicians. Yet, other data show that patients fre quently reject mammography referrals. In Cummings' survey, nearly two thirds of the physicians who reported using mammog raphy said that patients often or sometimes refuse the procedure when it is recom mended.48 This is documented by a Long Island study of referral com?liance in a breast cancer screening project.'9 From 1978 to 1980, only 45 percent of the participants referred for mammography complied with these referrals. Compliance was signifi cantly affected by the presence or absence of self-reported breast symptoms, the pres ence or absence of abnormal findings on physical examination, and the source of payment for mammography. Compliance, therefore, ranged from a low of eight per cent for asymptomatic women with normal examination findings who had to pay for mammography to 93 percent for sympto matic women with abnormal findings who had mammography free of charge. Almost two thirds of the noncompliers with project-paid mammography were con cerned about radiation hazards or other risks.39 According to the Gallup poll, only 38 percent of women age 50 and older feel that mammography is a safe procedure that they would not hesitate to have;35 among younger women, the figure is considerably smaller. Similarly, in Lieberman's survey more than half the respondents felt that mammography involves risks, but the ma jority of the sample also felt that the bene fits of the procedure outweighed its risks.33 Increasing the Use of Mammography Few attempts have been made to alter the attitudes and practices of physicians and patients regarding mammography, and most of these efforts have been limited to resi dency programs in the setting of a single hospital or clinic. Because mammography is used so infrequently in the community at large, small increases in utilization have a greater likelihood of being statistically sig nificant without indicating a meaningful change. For example, McDonald et al de veloped a system of computer reminders to CA-A CANCERJOURNAL FORCLINICIANS breast 42% Ineffectiveness:.,@ 51% of procedure ,: ,@.. -‘ .28% .71. ,‘‘!:‘@.‘-‘‘ , .. ‘¿@ ç ‘¿ . .‘, ‘¿ . ‘¿ six ‘¿percent the sample were omitted of with .@ not .1 ‘¿:, ‘¿ .‘ mammography. guidelines. recommend VOL 37, NO 1 JANUARY/FEBRUARY 1987 increase preventive care at the Indiana Uni versity School of Medicine.@' Two years after the intervention began, residents in the reminder group had prescribed mam mography for eight percent of eligible pa tients age 50 to 70, compared with two percent in the control group (p = .0005). Other early detection tests showed a much greater absolute effect. For example, stud ies of occult blood test usage showed 55 percent usage for the reminder group ver sus 22 percent for the controls. More impressive results for mammog raphy were obtained by Cohen and col leagues at Cleveland Metropolitan General Hospital.43 Their intervention consisted of a checklist and radiology requests attached to the patient's chart by a trained assistant. Over a four-month period, house-officer compliance with mammography guidelines for women age 50 to 60 reached 32 percent in the intervention group. compared with four percent for the controls (p <.001). The longer-term effects of the intervention remain to be determined. Two studies by Fox et al address the issue of durability.70-7' They tested the ef fectiveness of educational interventions on the screening behavior of residents and es tablished family practitioners at the Uni versity of Michigan. Patient eligibility was expanded to include women age 35 to 49. In the first study. a single seminar coupled with reinforcement cues increased the use of mammography for eligible patients from four percent at baseline to 11 percent over a three-month period.70 During the next three months, usage remained essentially stable without further reinforcement. In contrast, mammography use by control group physicians never surpassed three percent of eligible patients. The results of the second study showed a similar pattern, even though the target population was lim ited to women 35 to 49 years of age.7' Consistent with other research,46 Fox et al found that staff self-reports of mammogra phy usage overestimated the actual refer rals by an average of 50 percent.7' A much more comprehensive effort to increase the use of mammography was launched in 1982 by the Illinois Division of the ACS.65 Unlike the interventions for res idents, this effort focuses on patients as well as physicians as potential instruments of change. The multifaceted promotional program has included public awareness campaigns, professional education, quality control for mammography equipment, and pressure to reduce the costs of mammogra phy. Since the program began, there have been substantial increases in the number of women receiving mammograms, the pro portion of mammograms ordered for screening purposes, the number of facili ties conforming to radiation guidelines, and the proportion of women with favorable attitudes toward routine mammography. The average cost of mammography throughout Over time, total cancers detected in the HIP study group equalled those of the controls—suggesting that screening did not result in overdiagnosis. the state, however, has remained stable over time, although certain areas have substan tially reduced costs in response to ACS pressure. Despite the promotional program in Il linois, the vast majority of eligible women are not receiving mammography on a rou tine basis,65'72 and in the survey of attitu dinal change, only 17 percent of respon dents expressed a strong inclination to have such examinations.65 The most favorable result of the Illinois program was the large increase in conformity to ACS guidelines for radiation. With the support of the Illi nois Radiological Society, conformity in creased from 69 percent of the relevant facilities to 97 percent over a two-year pe rind.65 The public, therefore, could be more assured of the minimal risks of radiation exposure. The ACS is continuing to moni tor mammography facilities in Illinois in terms of dosage levels, image quality, and the capacity to serve the screening needs of eligible women. CA-A CANCER JOURNAL FOR CLINICIANS Prospects for Change By any preventive criterion, the level of mammography usage in the US leaves much room for improvement. In response to this reality, the ACS has launched the Breast Cancer Detection Awareness Program to increase the use of mammography and pro mote awareness among health profession als about the importance of early detection of breast cancer in general. In addition, the NCI has requested grant applications aimed at increasing and sustaining the use of state of-science mammography and breast pal pation in geographically defined popula tions age 50 and older.73 Applicants are The HIP experience suggests that proof of the benefit of mammography is not in itself sufficient to convince physicians and patients to use it. encouraged to promote mammography among the medical profession and the women themselves. Prospects for change in the use of mam mography will depend on the success or failure of consensus policymaking, mar keting strategies, cost reductions, risk re ductions, and technological improvements in sensitivity and specificity. There may also be a positive association between the credibility of lumpectomy as a treatment procedure and the credibility of mammog raphy as an early detection procedure. If the treatment for small, minimal, and bor derline cancers detected by mammography were less disfiguring than standard mastec tomies, fear of diagnosis and overdiagnosis might be reduced. Analogously, the attrac tiveness of mammography should increase with increased proof of its effectiveness in reducing breast cancer mortality.74 As fur ther data accumulate, policymakers may be better able to determine the most efficient interval for mammography screening. If that interval can be lengthened beyond the current yearly recommendation, costs should decline along with fears of radiation. VOL 37, NO 1 JANUARY/FEBRUARY 1987 The author assumes that there are ra tional as well as irrational barriers to the use of mammography and that both have consequences that are real in practical terms. Moreover, the distinction between rational and irrational concerns is not always clear cut. In certain communities, for example, use of equipment that yields excessive lev els of radiation and poor image quality may justify a negative attitude toward routine mammography. Where necessary, existing facilities should be upgraded for the protec tion of current patients and those who may respond to promotional programs. Diagnostic radiologists will also argue that the interpretation of screening mam mograms requires special training and ex perience,53-75 and that a successful national campaign to increase the use of mammog raphy would reveal a shortage of skilled personnel37—not only radiologists but spe cially trained surgeons capable of assessing “¿definiteand equivocal abnormalities identified on screening mammograms.―53 In addition, it would ultimately be neces sary to expand available facilities to ac commodate increased demand.65 The HIP experience suggests that proof of the benefit of mammography is not in itself sufficient to convince physicians and patients to use it. However, one could al ternatively argue that the HIP study did not establish the benefit of mammography per se and that this verdict awaits the comple tion of ongoing studies in Europe23-24-76and Canada.7779 Meanwhile, those who are convinced of the value of mammography face the task of persuading others. To im prove the probability of success, promo tional efforts need scientific evaluation. In terventions that involve publicity in the mass media, professional education, organiza tional endorsements, mass mailings, and personal contact must be evaluated in terms of outcome and process measures.8° Just as it is necessary to test the effectiveness of mammography through experimental and quasi-experimental designs. it is also nec essary to test promotional strategies through systematic research. For example. the use of control groups can reduce the possibility of biased interpretations of program re sults. If data supporting the value of mam mography continue to accumulate, usage might gradually increase without large-scale promotional campaigns. The thrust of re cent advocacy makes this a plausible hy pothesis.74 In 1984 and 1985, Lancet gave editorial encouragement to mammographic screening,lS8i and in 1985 Cairns in Sci entific American singled out screening for breast cancer as one of the few effective strategies to reduce cancer mortality.82 Whether these indicators portend greater acceptance of mammography by the medi cal profession remains to be seen. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png CA: A Cancer Journal for Clinicians Wiley

Using mammography for cancer control: An unrealized potential

CA: A Cancer Journal for Clinicians , Volume 37 (1) – Jan 1, 1987

Loading next page...
 
/lp/wiley/using-mammography-for-cancer-control-an-unrealized-potential-0St0v3gk8K

References (55)

Publisher
Wiley
Copyright
Copyright © 1987 American Cancer Society
ISSN
0007-9235
eISSN
1542-4863
DOI
10.3322/canjclin.37.1.33
Publisher site
See Article on Publisher Site

Abstract

Benefits of Mammography The most convincing evidence that screenDr. Howard is ProgramDirector for Studiesin Social Epidemiology, Health Promotion Sci encesBranch, in the Division of CancerPreven tion andControl of theNational CancerInstitute in Bethesda,Maryland. VOL. 37, NO 1 JANUARY/FEBRUARY 1987 ing for breast cancer is an effective means of secondary prevention comes from the Health Insurance Plan of Greater New York Screening Project (HIP), a randomized clinical trial conducted between 1963 and 1970.8_Il This study of members tested the benefit of four annual examinations involv ing mammography and breast palpation. Half of the 62,000 women in the study were randomly assignedto the intervention group, and the other half, the controls, their usual care. Those in the intervention group were offered the opportunity to participate in the screening program, and 65 percent ap peared for screening at least once. Those who refused to be screened at all were still considered to be part of the intervention group, so that selective participation would not bias the results. Within the control group, mammography was used increasingly for differential diagnosis but not as part of a program for early detection among asymp tomatic women 8l0 Many published reports have emanated from the HIP study,8'5 and follow-up of the original cohort is continuing. Basically, the investigators have concluded that the screening program substantially reduced breast cancer mortality among women age 50 and older at the time of entry.9M Data suggest that women younger than 50 may also have benefited, but this possibility is open to debate.8-9- In those 50 and older, 16.17 the intervention group showed a reduction in mortality of approximately 50 percent at five years and 33 percent at 10 years. com pared with equivalent women in the control group.9―° Potential biasing factors such as lead time, length bias, and overdiagnosis were taken into account in the analysis. Although the HIP study involved two screening modalities, their relative contri bution to the observed reductions in mor tality is a matter of speculation. For the intervention group as a whole, more screening-detected cancers were discov ered by physical examination than by mam mography.―Among women age 50 to 59 at diagnosis. however, 42 percent of the cancers were detected by mammography alone, and another 18 percent were discov ered by both modalities.―There appears to be some consensus that one third of the reduction in mortality in the total group offered screening was due to the impact of mammography.―'8-'9 as it functioned in the !960s. Further evidence that mammography per se facilitates the early detection of breast cancer comes from the Breast Cancer De tection Demonstration Project (BCDDP), in which approximately 280,000 American women were enrolled in the mid-!970s.202' Because the BCDDP was a demonstration project rather than a clinical trial, no con trol group was included. The data strongly suggest. however, that annual mammo graphic examinations played a key role in detecting ‘¿minimal cancers― in situ car ( cinomas and invasive cancers smaller than one cm), for which the prognosis would appear to be particularly favorable. Among the 1,153 minimal cancersidentified through the screening program, 57 percent were detected by mammography alone, com pared with six percent detected by physical examination alone 2oMoreover, mammog raphy alone was also responsible for de tecting a much higher proportion of infil trating cancers greater than or equal to one cm. than was physical examination alone (34 percent versus nine percent).2°Thus, 42 percent of all program-detected cancers in the BCDDP were discovered through mammography alone, compared with only nine percent from breast palpation alone.20 In contrast to the HIP study. in the BCDDP mammography seemsto have been effective in the 40-to-49-year age group as well as in older women. Mammography alone was responsible for 35 percent of cancer detections in the younger group. and it was positive along with physical exami nation in another 50 percent of thesecases 20 In the HIP study, the equivalent figures were 19 percent and 19 percent respec tively.―2° When mammography was re moved from the BCDDP as a routine screening modality for women younger than 50. the detection of minimal cancers de creased in this age group.2° The American Cancer Society (ACS) has concluded, therefore, that “¿there evidence that is The data suggest that only about 15 to 20 percent of American women age 50 and over have ever had .a mammogram and that a much smaller proportion are being examined with systematic regularity. screening with mammography can detect very small, localized breast cancers in women 35 to 49 years old, which in turn suggests the possibility of better survival rates in this age group. “¿22 Recent research results from Sweden and Holland lend additional credence to the proposition that mammography alone can significantly reduce mortality from breast cancer—atleast among older women. The Swedish study, which began in 1977, is of particular interest because it is a random ized controlled trial of single-view mam mography without breast palpation for the early detection of breast cancer.23 As was true in the HIP study, the control group has received usual care, which for 13 percent has included mammographic examina tions.23 Although the screening interval in the Swedish trial is almost three years for women age 50 or older, the 40 percent reduction in breast cancer mortality in the 50-to-74 age group (p = 0.003) is similar to CA-A CANCER JOURNAL FOR CLINICIANS that for the 50-to-64 age group in the HIP study. where annual screening was proto col.-3 Moreover, as in the HIP study, women age 40 to 49 in the Swedish trial are yet to show a significant mortality benefit from the screening effort.3 For all age groups. further follow-up and more extensive anal yses are anticipated. One of the screening studies from Hol land also examined the effect of single view mammography alone on breast cancer mortality. Preliminary results from the Nijmegen project suggest that the mortality in women age 35 and over can be reduced by about 50 percent through regular mam mographic screening of all eligible women. A variety of studies suggest that physicians regard mammography as a relatively unimportant prevention strategy, even for women age 50 and older. This was a case-control design. however. rather than a randomized trial, and various sources of possible bias are still being eval uated. Furthermore, the investigators ac knowledged that they cannot yet study the effect of screening on women younger than 55 because of the small number of cases and the higher attendance rates in the youngerage groups.Data from theother Dutchcase-control (the study DOM project in Utrecht) indicate that screening with the combination mammography and physi of cal examination gives protection against dying from breast cancer, especially among elderly women and those screenedtwice. 25.26 Again, however, the researchers are con sidering the possibility that important bias ing factors may have been overlooked, and they need a longer follow-up period to eval uate mortality trends. f@rsomewhat in their screening guidelines for breast cancer. The National Cancer In stitute (NC h has adopted a more conserva tive position regarding mammography than the ACS has. For asvmptomatic women 50 and older. both groups recommend annual or rout inc screening with mammography and breast palpation. 2) For younger women. however, the NCI would restrict mammographic screening to those at high risk—that is. women with a personal or family history of breast cancer. 30,31 The ACS. on the other hand. recommends a baseline mammogram for all asymptomatic women age 35 to 40 and screenings every one to two years for asymptomatic women age 40 to 4929 The American College of Radiology makes similar recommenda tions for the various age groups.32 If the NCI and ACS guidelines had been adopted by the medical and lay communi ties. one would expect routine utilization of mammography in the 50-plus age cate gory and lesser usage among younger women. The literature indicates, however, that only a small proportion of American women age 50 and older regularly use mammography for screening purposes. It is also unclear whether the infrequent use among younger women is consistent with any of the guidelines. Patterns of Use It is difficult to determine the extent to which mammographic examinations have been performed for screening rather than diagnostic purposes. There is, therefore, an element of uncertainty in estimating the proportion of women who have been screened with mammography at some time in their lives and the proportion who are routinely screened. The weight of the data in Table 1does, however, suggest that only about 15 to 20 percent of American women age 50 and older have ever had a mammo gram and that a much smaller proportion are being examined with systematic regu larity. Furthermore, both of these estimates are undoubtedly inflated by the use of di agnostic mammography. since most mam mograms are performed for diagnostic pur poses.37 Current Use ofMammography Guidelines for Use The two major organizations concernedwith cancer prevention in the United States dif VOL 37, NO 1 JANUARY/FEBRUARY 1987 An obvious exception to this picture are women inmetropolitan reported areas higher the findings of the Gallup Organization mammography use(24percent)hanthose t shown in Table 1 According to the esti in the suburbs (21 percent) or in nonmetro mates in that survey. 41 percent of women politan areas (12 percent): Jewish women in the 50+ agegroup have received a “¿breast showed a much higher usage rate (38 per x-ray―at some point in their lives, and 15 cent) than any other religious group: women percent have such examinations every year. with household incomes above $15,000 a Even among those in the youngest age group year had higher usage patterns than those (20 to 39), 33 percent reported having re with lesser incomes: involvement in other ceived a breast x-ray. It is possible that health activities (including breast self-ex some of these respondents confused chest amination) was positively associated with x-rays with mammography. mammography use: usage increased with increasing knowledge of breast cancer de In spite of the age guidelines of the NC! and ACS, the national surveys show no tection and treatment procedures and in consistent relationship between age and the creasing self-estimates of the likelihood of utilization of mammography—except that getting breast cancer. Such factors as geo graphic region. concern about the side ef in each study the youngest age group re ports the smallest proportion of mammog fects of breast cancer, and beliefs about the raphy use (Table 1). Two of the surveys progress of breast cancer treatment were (Lieberman33 and Gallup35) show almost not meaningfully associated with mam no difference in utilization between women mography usage. The analysis by Lane and Fine of pa in the 50+ age group and those 30 to 49 or 40 to 49. The NCI's Office of Cancer Com tient compliance with mammography re munications (0CC) survey,owever, h shows ferrals is also relevant.39 They found no an inverse relationship between age and significant association between a host of mammography use after age 35,34The pro sociodemographic factors (such asage. race, portion who reported ever having had a maritaltatus, s education, income)and and the compliance behavior of their respon mammogram declined from 25 percent for the 35-to-49-year age group to 16 percent dents. Furthermore, there was no relation for those age 65 and older. ship between a personal or family history of breast cancer and patient compliance. With respect to other predictors of mammography use, Lieberman shows a sizable positive relationship between edu cation and mammography experience.33 Referral Patterns Among women who attended college. 18 A variety of studies over the last 15 years percent reported having had a mammo gram. compared with 12 percent for those suggest that physicians regard mammogra phy as a relatively unimportant prevention with a high-school education and nine per strategy, even for women age 50 and older. cent for the grade-school group. The 0CC In 1970. Bates and Mulinare studied the survey found only a slight positive effect (four percentage points over the entire screening practices of internists and gen range34),while the Gallup Organization did eral practitioners in upstate New York.4° The investigators found that only one per not assessthe impact of education.35 cent of those surveyed reported that they The 0CC survey also evaluated rela tionships between mammography use and usually prescribed mammography for other demographic factors, and it measured women age 40 to 65. In the mid-1970s. an audit of internist records in North Carolina the effects of various perceptions. atti showed no mammography referrals for tudes, and behavior patterns among Amer women age 50 and overwho visitedheir t ican women. 34.38 brief, the findings were In physicians for general checkups.41-42 In as follows: There was no difference in 1980, a record review at a general hospital mammography usage between whites and in Cleveland showed that only four percent blacks, but the “¿other― category race of eligible women age 50 to 60 received a showed considerably less experience; CA-A CANCER JOURNAL FOR CLINICIANS [I D t) a, 0@ a,, ‘¿a, a,. C. @C. ‘¿a, ,., Co DNCN C@1 e@ - .C@) N, a,.@. II. II II II zzz —¿. ,,@. a, a, .0) ,—.000 I .1' + CD -@‘T a0' ,,a,@ ‘¿â€˜a, ‘¿,,,@ ‘¿a, 1, cr ,. ‘¿@ a―. c_;,' @c VOL 37. NO 1 JANUARY/FEBRUARY1987 mammogram during their usual care.43 In another study. from 1978 to 1980, only two percent of women age 50 to 70 at a general medicine clinic in Indianapolis received mammography during their usual care.'@' Somewhat more favorable findings come from the study by Dietrich and Goldberg of primary-care physicians in Northern Cali fornia.45 In the early 1980s, these investi gators reviewed the charts of generalists and'subspecialists involved in group prac tice to determine their compliance with pre ventive medicine guidelines. Less compli ance was shown for mammography than for any other preventive procedure except flu immunization, and mammography was also regarded as the least important proce dure. However, approximately 15 percent of eligible women age 50 to 59 actually receive mammography over an annual in terval. Among physicians who rated the procedure high in importance, patient cov erage increased to 38 percent. compared with 11 percent for the physicians who rated it low in importance. Unlike the other preventive techniques (such as Pap smears and breast palpation), mammography was less likely to be pre scribed in the context of a complete physi cal examination than during other medical encounters,45 suggesting its use as a diag nostic tool in response to detected symp toms. This finding by Dietrich and Gold berg differs somewhat from the 1983 Gallup survey of patients themselves. Among women age 50 and older who reported ever having had a breast x-ray, 62 percent said it was part of a regular checkup. The com parable figure for medical palpation of the breasts, however, was 84 percent, indicat ing that mammography had the higher di agnostic/prevention ratio.35 In the California study, mammography usage did not differ significantly by physi cian specialty. In Battista's study of pri mary-care physicians in Quebec, however, mammography was prescribed much more frequently in community health centers and family medicine teaching centers than in fee-for-service practices.46-47At that time (1981 and 1982), the Canadian Task Force on the Periodic Health Examination rec ommended annual mammography for women 50 to 59,46 Yet, only eight percent of the entire sample of physicians said that they complied with this recommendation even if the physical breast examination was normal.46 Moreover, the figure of eight percent was based on interview data, and there are indications from billing audits that these reports may have been somewhat ex aggerated.46 As was true in California, physicians' useof mammography was much less prevalent than their commitment to other preventive procedures, such as breast palpation (99 percent) or Pap smears (91 percent). 46 Two recent surveys in the US show a wide discrepancyetweenbeliefs phy b by Two recent US surveys show a wide discrepancy between beliefs by physicians about the effectiveness of mammography and their actual utilization practices. sicians about the effectiveness of mam mography and their actual utilization prac tices. In 1982, Cummings et al mailed questionnaires to a random sample of fam ily physicians in New York State and re ceived a 60 percent response.48 The great majority of respondents believed that mam mography is effective for detecting breast cancer in its early stages. Again, however, only eight percent recommended annual mammography forasymptomaticwomen older than 50. Twenty-nine percent did not recommend mammography at all for this age group, and another 16 percent recom mended it only for symptomatic patients. A nationwide telephone survey of pri mary-care physicians was conducted for the ACS in 1984.29 Because the response rate reached 91 percent, the investigators contend that its results are applicable to all primary-care physicians in the continental US. Forty-one percent of the respondents CA-A CANCERJOURNAL FORCLINICIANS agreed “¿completely― the ACS guide with lines concerning mammography. but only Il percent said that they observed these guidelines with all relevant patients. In comparison. 80 percent of the physicians universally followed the guidelines for breast physical examination, and 75 percent ad hered to the recommendations for the Pap test. Approximately one fifth of the re spondents said that over the last five years they had increased their use of mammog raphy for early detection. Less than half of the entire group. however, had ever used the procedure for asymptomatic women with no history of cancer. Obstetricians and gynecologists in the radiation. “¿unnecessary― biopsies. over diagnosis. and financial costs. In addition, physicians and patients have developed at titudinal barriers to mammography that are not reflections of the objective realities. Radiation Risks The radiation hazards of mammography. as the technique was applied in the 1960s and early 1970s, were publicized by Bailar in a widely quoted article.'9 Based on cer The recognized deterrents to the use of mammography include the risks of radiation, “¿unnecessary― biopsies, overdiagnosis, and financial costs. ACS survey were more committed to mam mography than were internists and general practitioners. Only two thirds of the OB GYN group. however, had ever ordered a mammogram for asymptomatic women, and only 17 percent universally followed the ACS guidelines for mammography. In light of these findings, the ACS has concluded that: “¿Too physicians are withhold many ing mammography in the absence of symp toms, at a time when the main purpose of mammography is to find an asvPnptomatic cancer.―49 Deterrents to the Use of Mammography Although mammography has proved to be effective in detecting early cancers, the his tory of its usage provides evidence for cau tion in its application as a screening tech nology. The recognized deterrents to the use of mammography include the risks of VOL 37, NO 1 JANUARY/FEBRUARY 1987 tain assumptions about the dose-response relationship, the latent period, and so on, Bailar “¿regretfully― concluded that “¿there seems to be a possibility that the routine use of mammography in screening asymp tomatic women may eventually take almost as many lives as it saves.― acknowl He edged that mammography can contribute to the reduction of breast cancer mortality but that screening by medical history and phys ical examination alone “¿will probably pro vide much or most of the same benefit las a three-way screenj without risk from irra diation, at least in women under some fairly high age limit.― In essence. Bailar argued that the promotion of mammography as a general public health measure was prema ture. Concerns about the hazards of radiation in the BCDDP led to a set of Working Group investigations of these potential risks.@°On basis of estimated risk-bene the fit ratios, the experts concluded that annual mammography should continue in the BCDDP for women age 50 and older but not for younger women, unless they had a personal or family history of breast can cer.5° he review panel also recommended T that the BCDDP should continue its moni toring and control measures for radiation dosageA retrospective analysis of dosage lev els actually administered in the BCDDP showed substantial declines from the radia tion levels in the HIP study.8-'9'2'5' and there were reductions over time in the BCDDP itself.21-50-52 During the HIP study. the average skin dosage from a mammo graphic examination was 7.7 rad.8―9 By the end of the BCDDP. however, the aver age surface exposure was 0.47 R for film screen units and 1.2 R for xeromammog raphy.5' Similarly, estimates of midline dosage levels in the two studies show large differences in favor of the BCDDP. 0.2I By the fifth year of that program, the average midline dose was 0.035 rad for film screen units and 0.36 rad for xeromammogra phy .@‘ Continuing technological improve ments have led the ACS to conclude that: “¿Modernechnology has reduced the ra t diation exposure of low-dose mammogra phy to the point of negligible risk, if risk exists at all, and has increased its diagnos tic capabilities at the same time.―49 Unnecessary Biopsies The purpose of screening with mammog raphy is to detect breast cancer at an earlier stage than it would be detected by clinical palpation, breast self-examination, or ac cidental discovery. Mammography can also help in the interpretation of the findings from clinical examinations. Physicians who believe in aggressive early detection with mammography willingly recommend sur gical biopsies on the basis of mammo graphic suspicions alone.53 Other physi cians are reluctant to biopsy lesions that they cannot feel.54 Even in the BCDDP, the proportion of recommended biopsies varied greatly from clinic to clinic. In the most aggressive center, the percentage of screening examinations resulting in surgi cal recommendations (including biopsies, aspirations, and consultations) was eight times higher than in the least aggressive center.55 It is reasonable to suppose that the difference in biopsies recommended was largely a function of the difference in atti tudes toward mammographic findings. More than eight percent of all partici pants in the BCDDP had biopsies per formed.2o@56 he great majority resulted from T project screening examinations at an aver age rate of one biopsy for every 48 screen ings;20'56 the comparable rate for the HIP study was one biopsy for every 101 screen ings.'° Compliance with biopsy recom mendations in the BCDDP decreased ap preciably over time as physicians adjusted to high benign-malignancy ratios in biopsy experiences.55 These ratios were inversely related to age, varying from 16.4 for women 35 to 39, to 2.7 for those 70 to 7420.21 Among all agegroupscombined,theratio of benign to malignant biopsies performed because of recommendations from screen ing in the BCDDP was 5420 Generally speaking, this is higher than the ratios for settings in which no special screening ef forts were undertaken,576° although there, too, the ratios vary—apparently by age, physician orientation, and type of hospital. Concerns about false-positive findings in mammographic screening are not limited totheUS. Investigators associated withthe NijmegenprojectnHollandobserved i that “¿the positive predictive value of mammog raphy is 35 percent; thus, for each screen detected case, two women had a referral and one of them had a biopsy that they would not have had if there had been no screening. “¿24 Canada, physician skep In ticism about the specificity of mammogra phy has deterred compliance with biopsy recommendations from the National Breast Screening Study.54 To counter this skepti cism, directors of the study have used the medical media to promote the value of mammography in detecting early breast cancer. Basically, they argue that “¿where there is no increase in biopsies, there can be no impact of screening.―54 Overdiagnosis The major attractiveness of mammography as a screening technology is its capacity to detect minimal cancers that portend favor able prognoses. But as Eddy has observed, this benefit contains the seeds of potential problems: “¿Unfortunately there is no sharp boundary between nonmalignant and ma lignant cells, and it is quite possible to overdiagnose as a very early cancer a lesion that is not cancer and would never become cancer. As well as increasing the number of ‘¿cancers' etected, this can inflate the d number of cancers detected in the earliest stage, and since these lesions would never become clinically significant cancers, it can inflate survival statistics.' ‘¿58 CA-A CANCER JOURNAL FOR CLINICIANS The magnitude of overdiagnosis is not easily estimated. Inferences can be drawn from statistical modeling. mirror-biopsies. and differences in diagnosed casesbetween intervention and control groups in screen ing trials. Minimal cancers in the HIP study group constituted only 10 to 16 percent of all the detected breast cancers.6' and over time the totality of cancers in the study group equalled that of the controls.8 This suggests that the screening effort did not result in overdiagnosis.6' but merely in ear lier detection of cancers that would ulti mately have surfaced on their own. The BCDDP presents a more questionable pic ture. however, becauseof the large propor tion of minimal cancers and the absence of a control group. It has been argued, there fore, that some of the cancers detected by mammography alone might never have caused morbidity or mortality. Lesions can qualify as cancers from a histologic per spective without qualifying on the basis of their biologic behavior2',6Two of the European studies acknowl edge links between mammography and overdiagnosis. The Swedish investigators suggest that some of the excess Stage I cancersdetected in the mammography group might never have surfaced clinically.23 The Nijmegen investigators contend that the diagnosis and treatment of “¿clinically irrel evant non-invasive cancers―may be an important side effect of screening.24 Simi larly, Baines of the Canadian National Breast Screening Study accepts the possibility of overdiagnosis as a reality one must live with.54 Where screening with mammography is aggressive, concerns about overdi agnosis may go beyond the issue of clinical surfacing. For example, in the BCDDP it was alleged by the mass media that some women were unnecessarily losing their breasts because of incorrect interpretations of biopsy specimens. In a small proportion of cases, lesions initially defined as mini mal cancers were later judged to be be nign.5° he Working Group that evaluated T the pathology of minimal cancers could not procure all the relevant slides, and they found differences of opinion in the original pathology reviews and in the subsequent reviews by experts.5° Recognizing that minimal cancers can present special diag nostic problems. the Working Group con cluded that treating physicians must take into account clinical as well as pathologic information.50 According to Baines, the uncertainties associated with aggressive mammographic screening are not uncommon in medical care. and the risk of overdiagnosis is en twined with the quest for early diagnosis:54 “¿There is unanimity with respect to rarely diagnosis or treatment of any condition. There rarely is a right answer. Perhapsbreast cancer in situ is harmless. Perhaps ‘¿early detection' will not prevent deaths. Perhaps those who repudiate mammography are right. But until we know, perhaps women deserve the right to the earliest diagnosis possible. Financial Costs The economic cost of mammography var ies greatly from area to area and from phy sician to physician. Recent data from Illi nois show a median statewide price of $105. but the charge can surpass $200. including physician fees.2' Part of this cost may re flect the underutilization of existing mam mographic equipment. In screening clinics with high-volume participation. the aver age costs decrease. Attempts by the ACS to lower the cost of mammography have resulted in lower charges in some areas, ranging from $36 to $77,63_6s Where screening programs are subsidized, the ac tual costs to the patient may be further reduced or be entirely free.M At the present time, few insurance companies cover the cost of mammography for asymptomatic women. Similarly, mammography is not authorized by Medi care for routine screening purposes,@ but only as a diagnostic test under certain specified conditions. Physicians, there fore, may have to “¿bootleg― or disguise the routine use of mammography by attach ing the more acceptable label of differential diagnosis. VOL. 37, NO 1 JANUARY/FEBRUARY 1987 Estimates have been made of the poten tial costs of national screening programs using mammography. These estimates vary according to eligibility criteria, screening intervals, number of views, and so on. In 1980, Eddy speculated that the yearly costs of annual mammography for the 31 million American women over age 50 would be approximately 1.5 billion dollars, at $50 an examination.58 This total did not include the add-on costs of follow-up examinations and biopsies for women with positive or suspicious mammographic findings,67 nor did it take into account financial savings that might result from earlier diagnosis and treatment, and consequent reductions in morbidity and mortality. Attitudinal Barriers Survey data indicate that the deterrents to mammography discussed in scientific jour nals and the mass media have influenced the beliefs of physicians and patients. In each of the physician surveys, the impor tant reasons for negative attitudes toward mammography (Table 2) included the risks of radiation, financial costs, and concerns about the effectiveness, reliability, or ne cessity of the procedure.29'46-48 Physicians in the US were also deterred by the small likelihood of detecting breast cancer in any particular screening examination.29'48 Studies of lay attitudes toward mam mography suggest that the recommenda tion of physicians is a critical intervening variable that influences the beliefs and be havior of patients.68 In a 1977 Los Angeles survey, 93 percent of respondents reported that they would be at least somewhat likely to have a mammogram if their physician recommended it.36 In the same study, 97 percent of women who reported being ad vised to have a mammogram said that they actually had it. Furthermore, a recent sur vey by the Saturday Evening Post indicates that the absence of physician recommen dations is the greatest deterrent to the use of mammography.69 Among the respon dents over age 50 who were not having annual mammograms, 64 percent gave as a reason that it was not suggested by their physicians. Yet, other data show that patients fre quently reject mammography referrals. In Cummings' survey, nearly two thirds of the physicians who reported using mammog raphy said that patients often or sometimes refuse the procedure when it is recom mended.48 This is documented by a Long Island study of referral com?liance in a breast cancer screening project.'9 From 1978 to 1980, only 45 percent of the participants referred for mammography complied with these referrals. Compliance was signifi cantly affected by the presence or absence of self-reported breast symptoms, the pres ence or absence of abnormal findings on physical examination, and the source of payment for mammography. Compliance, therefore, ranged from a low of eight per cent for asymptomatic women with normal examination findings who had to pay for mammography to 93 percent for sympto matic women with abnormal findings who had mammography free of charge. Almost two thirds of the noncompliers with project-paid mammography were con cerned about radiation hazards or other risks.39 According to the Gallup poll, only 38 percent of women age 50 and older feel that mammography is a safe procedure that they would not hesitate to have;35 among younger women, the figure is considerably smaller. Similarly, in Lieberman's survey more than half the respondents felt that mammography involves risks, but the ma jority of the sample also felt that the bene fits of the procedure outweighed its risks.33 Increasing the Use of Mammography Few attempts have been made to alter the attitudes and practices of physicians and patients regarding mammography, and most of these efforts have been limited to resi dency programs in the setting of a single hospital or clinic. Because mammography is used so infrequently in the community at large, small increases in utilization have a greater likelihood of being statistically sig nificant without indicating a meaningful change. For example, McDonald et al de veloped a system of computer reminders to CA-A CANCERJOURNAL FORCLINICIANS breast 42% Ineffectiveness:.,@ 51% of procedure ,: ,@.. -‘ .28% .71. ,‘‘!:‘@.‘-‘‘ , .. ‘¿@ ç ‘¿ . .‘, ‘¿ . ‘¿ six ‘¿percent the sample were omitted of with .@ not .1 ‘¿:, ‘¿ .‘ mammography. guidelines. recommend VOL 37, NO 1 JANUARY/FEBRUARY 1987 increase preventive care at the Indiana Uni versity School of Medicine.@' Two years after the intervention began, residents in the reminder group had prescribed mam mography for eight percent of eligible pa tients age 50 to 70, compared with two percent in the control group (p = .0005). Other early detection tests showed a much greater absolute effect. For example, stud ies of occult blood test usage showed 55 percent usage for the reminder group ver sus 22 percent for the controls. More impressive results for mammog raphy were obtained by Cohen and col leagues at Cleveland Metropolitan General Hospital.43 Their intervention consisted of a checklist and radiology requests attached to the patient's chart by a trained assistant. Over a four-month period, house-officer compliance with mammography guidelines for women age 50 to 60 reached 32 percent in the intervention group. compared with four percent for the controls (p <.001). The longer-term effects of the intervention remain to be determined. Two studies by Fox et al address the issue of durability.70-7' They tested the ef fectiveness of educational interventions on the screening behavior of residents and es tablished family practitioners at the Uni versity of Michigan. Patient eligibility was expanded to include women age 35 to 49. In the first study. a single seminar coupled with reinforcement cues increased the use of mammography for eligible patients from four percent at baseline to 11 percent over a three-month period.70 During the next three months, usage remained essentially stable without further reinforcement. In contrast, mammography use by control group physicians never surpassed three percent of eligible patients. The results of the second study showed a similar pattern, even though the target population was lim ited to women 35 to 49 years of age.7' Consistent with other research,46 Fox et al found that staff self-reports of mammogra phy usage overestimated the actual refer rals by an average of 50 percent.7' A much more comprehensive effort to increase the use of mammography was launched in 1982 by the Illinois Division of the ACS.65 Unlike the interventions for res idents, this effort focuses on patients as well as physicians as potential instruments of change. The multifaceted promotional program has included public awareness campaigns, professional education, quality control for mammography equipment, and pressure to reduce the costs of mammogra phy. Since the program began, there have been substantial increases in the number of women receiving mammograms, the pro portion of mammograms ordered for screening purposes, the number of facili ties conforming to radiation guidelines, and the proportion of women with favorable attitudes toward routine mammography. The average cost of mammography throughout Over time, total cancers detected in the HIP study group equalled those of the controls—suggesting that screening did not result in overdiagnosis. the state, however, has remained stable over time, although certain areas have substan tially reduced costs in response to ACS pressure. Despite the promotional program in Il linois, the vast majority of eligible women are not receiving mammography on a rou tine basis,65'72 and in the survey of attitu dinal change, only 17 percent of respon dents expressed a strong inclination to have such examinations.65 The most favorable result of the Illinois program was the large increase in conformity to ACS guidelines for radiation. With the support of the Illi nois Radiological Society, conformity in creased from 69 percent of the relevant facilities to 97 percent over a two-year pe rind.65 The public, therefore, could be more assured of the minimal risks of radiation exposure. The ACS is continuing to moni tor mammography facilities in Illinois in terms of dosage levels, image quality, and the capacity to serve the screening needs of eligible women. CA-A CANCER JOURNAL FOR CLINICIANS Prospects for Change By any preventive criterion, the level of mammography usage in the US leaves much room for improvement. In response to this reality, the ACS has launched the Breast Cancer Detection Awareness Program to increase the use of mammography and pro mote awareness among health profession als about the importance of early detection of breast cancer in general. In addition, the NCI has requested grant applications aimed at increasing and sustaining the use of state of-science mammography and breast pal pation in geographically defined popula tions age 50 and older.73 Applicants are The HIP experience suggests that proof of the benefit of mammography is not in itself sufficient to convince physicians and patients to use it. encouraged to promote mammography among the medical profession and the women themselves. Prospects for change in the use of mam mography will depend on the success or failure of consensus policymaking, mar keting strategies, cost reductions, risk re ductions, and technological improvements in sensitivity and specificity. There may also be a positive association between the credibility of lumpectomy as a treatment procedure and the credibility of mammog raphy as an early detection procedure. If the treatment for small, minimal, and bor derline cancers detected by mammography were less disfiguring than standard mastec tomies, fear of diagnosis and overdiagnosis might be reduced. Analogously, the attrac tiveness of mammography should increase with increased proof of its effectiveness in reducing breast cancer mortality.74 As fur ther data accumulate, policymakers may be better able to determine the most efficient interval for mammography screening. If that interval can be lengthened beyond the current yearly recommendation, costs should decline along with fears of radiation. VOL 37, NO 1 JANUARY/FEBRUARY 1987 The author assumes that there are ra tional as well as irrational barriers to the use of mammography and that both have consequences that are real in practical terms. Moreover, the distinction between rational and irrational concerns is not always clear cut. In certain communities, for example, use of equipment that yields excessive lev els of radiation and poor image quality may justify a negative attitude toward routine mammography. Where necessary, existing facilities should be upgraded for the protec tion of current patients and those who may respond to promotional programs. Diagnostic radiologists will also argue that the interpretation of screening mam mograms requires special training and ex perience,53-75 and that a successful national campaign to increase the use of mammog raphy would reveal a shortage of skilled personnel37—not only radiologists but spe cially trained surgeons capable of assessing “¿definiteand equivocal abnormalities identified on screening mammograms.―53 In addition, it would ultimately be neces sary to expand available facilities to ac commodate increased demand.65 The HIP experience suggests that proof of the benefit of mammography is not in itself sufficient to convince physicians and patients to use it. However, one could al ternatively argue that the HIP study did not establish the benefit of mammography per se and that this verdict awaits the comple tion of ongoing studies in Europe23-24-76and Canada.7779 Meanwhile, those who are convinced of the value of mammography face the task of persuading others. To im prove the probability of success, promo tional efforts need scientific evaluation. In terventions that involve publicity in the mass media, professional education, organiza tional endorsements, mass mailings, and personal contact must be evaluated in terms of outcome and process measures.8° Just as it is necessary to test the effectiveness of mammography through experimental and quasi-experimental designs. it is also nec essary to test promotional strategies through systematic research. For example. the use of control groups can reduce the possibility of biased interpretations of program re sults. If data supporting the value of mam mography continue to accumulate, usage might gradually increase without large-scale promotional campaigns. The thrust of re cent advocacy makes this a plausible hy pothesis.74 In 1984 and 1985, Lancet gave editorial encouragement to mammographic screening,lS8i and in 1985 Cairns in Sci entific American singled out screening for breast cancer as one of the few effective strategies to reduce cancer mortality.82 Whether these indicators portend greater acceptance of mammography by the medi cal profession remains to be seen.

Journal

CA: A Cancer Journal for CliniciansWiley

Published: Jan 1, 1987

There are no references for this article.