Enthusiasm for Cancer Screening in the United States FREE

There is a growing recognition among medical professionals that cancer screening is a double-edged sword. While some individuals may benefit from early detection, others may only be diagnosed and treated for cancer unnecessarily.^1- 2 In recent years, the public has been exposed to expert debate about many of the most basic assumptions of screening: some scientists have challenged the utility of mammography for women younger than age 50 years^3- 5 or even for women at any age^6- 7; questions have been raised about how often to be screened for cervical cancer^8- 9; and whether to be screened at all for prostate^8- 9 or lung cancer.¹⁰ Emerging from these debates is a growing consensus that to make good decisions about screening, the public needs access to balanced information about its potential benefits and harms.^11- 12

But the public has long received a different message. Public health officials, physicians, and disease advocacy groups have worked hard over a number of years to persuade individuals living in the United States about the importance of cancer screening. It is practically impossible to read a major newspaper or popular magazine, watch television, ride public transportation, visit the beauty parlor,¹³ or even lick a stamp¹⁴ without seeing a public service announcement promoting some form of cancer screening. Most recently, aggressive direct to consumer advertising is bringing a variety of new advanced, but unproven screening tests including brain magnetic resonance imaging, lung and total-body computed tomographic (CT) screening,¹⁵ and genetic testing for "cancer genes" to the public.¹⁶

To understand the issues facing those who hope to bring balanced information to the public, we conducted a national telephone survey during 2001 and 2002 to learn about adults' experience with a broad array of screening tests. Specifically, we explored general beliefs about early detection, personal commitment to screening, screening as an obligation, attitudes toward false-positive results, and desire for total-body CT scanning, a new and potentially comprehensive screening method.

Our goal was to interview a nationally representative sample of adults for whom screening for cancer was relevant. Because our focus was on screening, we specifically targeted a screen-eligible population. Thus, we excluded individuals with a history of cancer because a prior cancer diagnosis may change how one thinks about screening and early detection. We further restricted our sample to women aged 40 years or older and to men aged 50 years or older because it is at these ages that most cancer screening is recommended (a notable exception being Papanicolaou testing). This project was approved by the institutional review boards at Dartmouth Medical School, Hanover, NH, and at the University of Massachusetts, Boston.

We used random digit dialing to obtain a national probability sample of households in the continental United States with telephone service. Figure 1 details the steps of our sampling procedure. The process began with the generation of a random list of 4000 US telephone numbers for the goal of 500 completed interviews. From this list, 1702 working residential telephone numbers were identified. Interviewers successfully completed a 3-minute "screening" interview with an English-speaking adult at 1208 of the residences to identify individuals meeting study criteria. At least one eligible adult resided in 697 of the households (if ≥1 eligible adult was identified, a computer selected a respondent so that each eligible person had an equal chance of being selected). A total of 500 individuals completed the interview.

There is some debate in the survey research literature about how best to calculate a response rate in this setting. The specific question is whether to account for residences that could not be screened. To maximize transparency of this issue, the American Association of Public Opinion Research¹⁷ suggests calculating 2 response rates using 2 different denominators. The simplest approach (commonly seen in the literature) is to ignore unscreened residences and use known eligible households as the denominator. In our case, 500 responses were obtained from 697 eligible households—a response rate of 72% among individuals known to be eligible. However, there are almost certainly some eligible households among those not screened. The second approach attempts to account for this by increasing the denominator to include the estimated number of eligible households among unscreened households. This estimate takes the proportion eligible among those households screened (in our case, 697/1208 = 0.58) as the best estimate of the proportion eligible among those households not screened. Thus in our case, among the 494 households not screened, 287 (0.58 × 494) would be expected to be eligible. Using this approach, we obtained 500 responses from an estimated 984 (697 + 287) eligible households—providing responses from 51% of those estimated to be eligible (N = 500).

Development. To learn how the public thinks about screening, we conducted 2 focus groups with adults aged 40 years or older to discuss cancer screening tests in general, experiences with specific tests, and perceptions of the pros and cons of such testing. The focus group and all subsequent survey development was done in collaboration with experts at the Center for Survey Research, a professional survey research firm affiliated with the University of Massachusetts. A draft survey instrument was developed based on the results of the focus groups. Experienced interviewers then conducted 10 cognitive interviews to ensure that the questions were understood and that the answers were meaningful. After revising the draft based on this feedback, 17 eligible adults identified by random digit dialing completed the survey; these interviews were audiotaped and then coded to identify questions that were difficult for interviewers to read or for respondents to answer. Final revisions were made to the survey based on the pretest results.

Instrument. The survey consisted of a general screening module (general questions about the value of early detection, worry about cancer, and interest in total-body CT screening) and 4 modules about common screening tests (mammography, Papanicolaou smear, prostate-specific antigen [PSA] testing, and sigmoidoscopy or colonoscopy). The screening test modules contained a parallel series of questions about the value of the test, testing preferences (eg, frequency, starting and stopping age, current behavior), experience with abnormal test results, understanding of screening controversies or uncertainties and risk perceptions about the corresponding cancer. All respondents received the general screening module and the sigmoidoscopy or colonoscopy module; women also received the mammography and Papanicolaou test modules, while men received the PSA module.

Administration. From December 2001 through July 2002, interviews were conducted by professional interviewers from the University of Massachusetts Center for Survey Research. All interviewers received special training on the purposes and procedures of this particular study; all underwent routine monitoring for quality control and feedback from a supervisor. The interviews took an average of 20 minutes (range, 10-54 minutes). Answers to the questions were directly entered into the computer-assisted telephone interviewing system by the interviewer.

Analysis. We created weights to account for differential probability of selection into our sample. An individual's probability of selection was a function of the number of residential voice telephone lines (ie, more phone lines, higher probability of selection) and the number of eligible adults at the residence (ie, more eligible adults, lower probability of selection).

Survey researchers sometimes create a second set of weights to force the sample proportions for selected demographic characteristics to match those in the population. Theoretically such poststratification weights reduce bias resulting from differences in response rates among demographic subgroups. The technique is controversial, however, because it requires a substantial assumption that nonrespondents would answer questions similarly to respondents. Thus, if Native American males who did not graduate from high school were underrepresented in the sample, the responses of these few individuals would be weighted upward to represent the US population proportion of this subgroup. Ironically, the more this kind of weighting has the potential to influence the results (ie, when the sample looks least like the target population), the more heroic the assumption.

Because the distribution of most demographic characteristics in our sample closely approximated those in the 2000 US Census (Table 1), poststratification weighting is unlikely to influence our results. Nevertheless, because the lowest education and oldest age groups were underrepresented, we created poststratification weights to match the US Census distribution on age, sex, race, Hispanic origin, educational attainment, and region.¹⁸ Analyses using these weights yielded results nearly identical (ie, ± 1%-2%) to those using only the probability weights. For simplicity, and to avoid the assumptions inherent in poststratification weighting, we present results using only the probability weights. All analyses were performed using STATA statistical software (Version 7; College Station, Tex).

Most adults (87%) living in the United States believe routine cancer screening is "almost always a good idea." Seventy-four percent believe that finding cancer early saves lives "most" or "all of the time" (Table 2). Fifty-three percent believe screening usually reduces the amount of treatment needed when cancer is found. But enthusiasm for screening also reflects a desire to know about the presence of cancer—regardless of its implications. Two thirds of individuals would want to be tested for a cancer even if nothing could be done. Fifty-six percent would want to be tested for what is sometimes called pseudodisease, cancers so slow growing that even untreated would never cause problems during the person's lifetime. Thirty-five percent believed they had had too few cancer screening tests in the past (64% thought they had had "about the right number"). Virtually no one interviewed (2%) thought they had had "too many" cancer screening tests.

We found that most adults say they have had cancer screening tests: 99% of US women aged 40 years or older reported having a Papanicolaou test and having 89% mammography; 71% of men aged 50 years or older had a PSA test; and 46% of the men and women in these age groups had a sigmoidoscopy or colonoscopy (our findings closely reflect those from the 2001 Behavioral Risk Factor Surveillance Survey¹⁹ whose corresponding numbers were 96%, 88%, 75%, and 48%, respectively). We also found that most women who had been screened with a Papanicolaou test or mammography and most men who had been screened with a PSA plan to undergo at least annual testing (Table 3).

To further gauge how personally committed individuals are to screening, we asked those who were currently being screened how they would respond if their physician told them to be screened less often. Fifty-eight percent of women said they would overrule their physician if he or she suggested less frequent Papanicolaou tests. Seventy-seven percent of men would continue to undergo prostate screening and 74% of men and women would continue with colon cancer screening (colonoscopy or sigmoidoscopy) even if their physician recommended against testing. Few individuals thought there would ever be a time when they would stop having routine screening tests; ranging from a high of 35% saying they would ever stop having Papanicoloau tests to a low of 21% for stopping mammography. In addition, if cost was not a concern, some would like to be screened as frequently as every 6 months: 4% for colonoscopy or sigmoidoscopy, 13% for Papanicoloau test, 16% for mammography, and 19% for PSA test.

To learn whether the public views screening as an obligation (eg, the right thing to do, or something individuals owe to their loved ones), we asked respondents to judge whether a person in average health would be "irresponsible" if he or she did not have screening (Figure 2). When asked about a 55-year-old person in average health, responses ranged from 79% (rating forgoing Papanicoloau tests as irresponsible) to 54% (for colonoscopy). A substantial proportion also believed that an 80-year-old who chose not to be tested was irresponsible: ranging from 41% for mammography to 32% for colonoscopy.

Overall, 38% of men and women in our sample had had at least 1 false-positive screening result that required further testing (11% for PSA, 30% for Papanicoloau, and 35% for mammography). Many of these individuals underwent invasive follow-up procedures (Table 4). While most individuals found out they did not have cancer within 2 weeks, 25% of women with abnormal Papanicoloaou test results, 13% with abnormal mammograms, and 25% of men with false-positive PSA test results waited more than 1 month for this information. Many individuals characterized this time as either "very scary" or the "scariest time" of their lives (43% for Papanicoloau test, 37% for mammography, and 58% for PSA test). Yet, looking back, 98% were glad they had had the initial screening test.

Finally, to gauge general enthusiasm for screening, we examined the public's interest in total-body CT scanning, a relatively new technology now aggressively marketed to consumers. We first described a total-body CT as a "3-D look inside your body using a CT scanner. A CT scan gives a very detailed picture of your lungs, liver, heart, and other internal organs, as well as bones and arteries. A total body scan can find many diseases like cancer before they can be found by routine check-ups. The body scan is quick and painless." After hearing this description, 86% said they wanted to have a free total-body CT. To learn about the strength of desire for CT, we asked those individuals who chose a free CT whether they would prefer a total-body CT scan or receiving $1000 in cash. Eighty-five percent would choose the total-body CT scan (ie, 73% of the entire sample). Only 27% thought there might be any downside to having a total-body scan (mostly discomfort during the procedure or anxiety); 14% mentioned concern for false-positive results, and 3% mentioned the downside of unnecessary subsequent testing.

Most people in the United States are firmly committed to cancer screening. Most individuals would overrule a physician who recommended against cancer screening and could not imagine a time they would stop being tested. This enthusiasm is not dampened by false-positive test results. While nearly half the respondents who had experienced a false-positive result described the episode as extremely scary, virtually all were glad they had had the original screening test.

Because it is a new and highly visible technology,¹⁵ we specifically asked respondents about total-body CT screening. There are no data to support the benefit (or even safety) of total-body CT screening, and it is not endorsed by any professional medical organization.²⁰ In fact, total-body CT screening is actually discouraged by the American College of Radiology and the American Association of Physicists in Medicine.^21- 22 Nevertheless, total-body CT scans are directly and aggressively marketed to consumers. While there have been a number of media reports about total-body CT,^23- 25 our study is the first to systematically document the substantial public interest: almost three quarters would choose such testing instead of receiving $1000 in cash.

Our findings should be interpreted in light of several potential limitations. First, since we used the telephone to select our respondents, the 5% of adults living in households without telephone service²⁶ are not represented. Next, while our response rate was good—72% among individuals known to be eligible and 51% among those estimated to be eligible—systematic bias between respondents and nonrespondents is still possible. This concern is lessened by the fact that our sample's demographics and screening rates closely approximated official US Census statistics. Although the elderly and persons with less formal education were underrepresented in our sample, stratified analyses demonstrated that the beliefs examined did not differ importantly from these characteristics. In addition, our findings about false-positive PSA test results are based on only 10 men and should be interpreted cautiously.

Our findings are also limited because they cannot tell us exactly why people living in the United States are so enthusiastic about screening. Our data do, however, argue against one important possibility—that enthusiasm stems from an exaggerated sense of cancer risk. We asked a variety of questions to understand cancer risk perceptions and found no evidence of a widespread overestimation of cancer risk. Instead we found that most individuals believe they are at average or lower risk for each of the cancers considered, report that they worry "a little" or "not at all" about getting cancer, and recognize that cancer is typically not a rapidly fatal disease (data not shown).

Our work suggests that screening is not seen as a choice but as an obligation. The idea that "you owe it to your children to be screened" and that it "would be selfish" to forgo screening was expressed repeatedly in our focus groups. This sentiment was borne out in the survey: two thirds, for example, believed that a 55-year-old person who did not have routine screening was "irresponsible." While the sense of obligation may, in part, stem from the intuitive appeal of early detection (ie, everyone knows that "an ounce of prevention is worth a pound of cure"), other factors are at work.

Health care marketing has created an environment in which screening is portrayed as the safest course of action. Screening has long been marketed as the preeminent weapon in the war on cancer.²⁷ In the name of improved population health, many well-meaning public health agencies and disease advocacy groups use powerful messages to persuade individuals to undergo screening with slogans like "take the test not the chance" or "don't be a victim" or by the use of fear- and guilt-inducing images (eg, a picture of the young children who lost an unscreened parent to cancer). Other efforts to promote screening may be more self-interested. For example, an increasing number of independent total-body CT scan centers market themselves directly to the public with slogans like "new technology that could save your life"; these advertisements are often accompanied by personal anecdotes from individuals who believe their lives were saved by the early diagnosis of an unsuspected tumor. Regardless of the source of the message, the net effect is the same: screening is always seen as the right thing to do.

Ironically, even what physicians see as the important harms of screening—false-positive results and detection of pseudodisease—reinforce the apparent case for screening. Virtually everyone who had a false-positive test result in our survey was glad they had been tested and intended to be tested again. Apparently, the relief experienced when the confirmatory test result comes back negative overwhelms the substantial fear that came before. Even pseudodisease—often taken to be the most important harm of screening because it results in the unnecessary diagnosis and treatment of cancer—may reinforce enthusiasm for screening: more people appear to have cancer (prevalence increases because "occult" cancers are detected) and prognosis improves (due to lead time and overdiagnosis biases).²⁸

Some clinicians will see our results as welcome evidence of the success of public health campaigns for widely recommended cancer screening tests. Others will have quite a different take. They will see disturbing evidence that these same campaigns have communicated a misleadingly simple and 1-sided message—a message that discourages meaningful discussions about the use of these tests in settings when the recommendations are less clear (eg, screening at younger ages, at advanced age, or for individuals with multiple comorbidities).²⁹

However, we would hope that everyone can agree that these messages have an undesirable adverse effect: a public that is primed to believe there is value in having any test that is marketed as being able to find early cancer. In the case of unproven tests such as total-body CT, excessive enthusiasm makes it extremely easy for exaggerated marketing tactics to succeed. Consequently, some have suggested that the government replicate what it has done with pharmaceutical advertising and regulate the direct-to-consumer advertising of unproven tests,³⁰ requiring that the advertisements are factually true and provide balanced information about benefit and harm.³¹

But there are limits to what physicians should expect from governmental intervention. The public's enthusiasm for cancer screening and resistance to do less stems in large part from messages the medical establishment itself has promulgated. Unfortunately, these messages have helped create an environment that hinders discourse on the prudent use of existing tests and is ripe for the premature diffusion of new ones. The challenge now is to balance messages and reduce the public's risk for overtesting and overtreatment.